overfit_and_underfit.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "fTFj8ft5dlbS"
   },
   "source": [
    "##### Copyright 2018 The TensorFlow Authors."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "cellView": "form",
    "colab": {},
    "colab_type": "code",
    "id": "lzyBOpYMdp3F"
   },
   "outputs": [],
   "source": [
    "#@title Licensed under the Apache License, Version 2.0 (the \"License\");\n",
    "# you may not use this file except in compliance with the License.\n",
    "# You may obtain a copy of the License at\n",
    "#\n",
    "# https://www.apache.org/licenses/LICENSE-2.0\n",
    "#\n",
    "# Unless required by applicable law or agreed to in writing, software\n",
    "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
    "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
    "# See the License for the specific language governing permissions and\n",
    "# limitations under the License."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "cellView": "form",
    "colab": {},
    "colab_type": "code",
    "id": "m_x4KfSJ7Vt7"
   },
   "outputs": [],
   "source": [
    "#@title MIT License\n",
    "#\n",
    "# Copyright (c) 2017 François Chollet\n",
    "#\n",
    "# Permission is hereby granted, free of charge, to any person obtaining a\n",
    "# copy of this software and associated documentation files (the \"Software\"),\n",
    "# to deal in the Software without restriction, including without limitation\n",
    "# the rights to use, copy, modify, merge, publish, distribute, sublicense,\n",
    "# and/or sell copies of the Software, and to permit persons to whom the\n",
    "# Software is furnished to do so, subject to the following conditions:\n",
    "#\n",
    "# The above copyright notice and this permission notice shall be included in\n",
    "# all copies or substantial portions of the Software.\n",
    "#\n",
    "# THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\n",
    "# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\n",
    "# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL\n",
    "# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\n",
    "# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING\n",
    "# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER\n",
    "# DEALINGS IN THE SOFTWARE."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "C9HmC2T4ld5B"
   },
   "source": [
    "# Explore overfitting and underfitting"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "kRTxFhXAlnl1"
   },
   "source": [
    "<table class=\"tfo-notebook-buttons\" align=\"left\">\n",
    "  <td>\n",
    "    <a target=\"_blank\" href=\"https://www.tensorflow.org/beta/tutorials/keras/overfit_and_underfit\"><img src=\"https://www.tensorflow.org/images/tf_logo_32px.png\" />View on TensorFlow.org</a>\n",
    "  </td>\n",
    "  <td>\n",
    "    <a target=\"_blank\" href=\"https://colab.research.google.com/github/tensorflow/docs/blob/master/site/en/r2/tutorials/keras/overfit_and_underfit.ipynb\"><img src=\"https://www.tensorflow.org/images/colab_logo_32px.png\" />Run in Google Colab</a>\n",
    "  </td>\n",
    "  <td>\n",
    "    <a target=\"_blank\" href=\"https://github.com/tensorflow/docs/blob/master/site/en/r2/tutorials/keras/overfit_and_underfit.ipynb\"><img src=\"https://www.tensorflow.org/images/GitHub-Mark-32px.png\" />View source on GitHub</a>\n",
    "  </td>\n",
    "  <td>\n",
    "    <a href=\"https://storage.googleapis.com/tensorflow_docs/docs/site/en/r2/tutorials/keras/overfit_and_underfit.ipynb\"><img src=\"https://www.tensorflow.org/images/download_logo_32px.png\" />Download notebook</a>\n",
    "  </td>\n",
    "</table>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "19rPukKZsPG6"
   },
   "source": [
    "As always, the code in this example will use the `tf.keras` API, which you can learn more about in the TensorFlow [Keras guide](https://www.tensorflow.org/guide/keras).\n",
    "\n",
    "In both of the previous examples—classifying movie reviews and predicting fuel efficiency—we saw that the accuracy of our model on the validation data would peak after training for a number of epochs, and would then start decreasing.\n",
    "\n",
    "In other words, our model would *overfit* to the training data. Learning how to deal with overfitting is important. Although it's often possible to achieve high accuracy on the *training set*, what we really want is to develop models that generalize well to a *testing set* (or data they haven't seen before).\n",
    "\n",
    "The opposite of overfitting is *underfitting*. Underfitting occurs when there is still room for improvement on the test data. This can happen for a number of reasons: If the model is not powerful enough, is over-regularized, or has simply not been trained long enough. This means the network has not learned the relevant patterns in the training data.\n",
    "\n",
    "If you train for too long though, the model will start to overfit and learn patterns from the training data that don't generalize to the test data. We need to strike a balance. Understanding how to train for an appropriate number of epochs as we'll explore below is a useful skill.\n",
    "\n",
    "To prevent overfitting, the best solution is to use more training data. A model trained on more data will naturally generalize better. When that is no longer possible, the next best solution is to use techniques like regularization. These place constraints on the quantity and type of information your model can store.  If a network can only afford to memorize a small number of patterns, the optimization process will force it to focus on the most prominent patterns, which have a better chance of generalizing well.\n",
    "\n",
    "In this notebook, we'll explore two common regularization techniques—weight regularization and dropout—and use them to improve our IMDB movie review classification notebook."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Tell TF not to consume all GPU memory so you can run more than one notebook at once\n",
    "import os\n",
    "os.environ['TF_FORCE_GPU_ALLOW_GROWTH'] = 'true'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m\u001b[37mcgpu03\u001b[m  Sat Jul 13 16:18:11 2019\n",
      "\u001b[36m[0]\u001b[m \u001b[34mTesla V100-SXM2-16GB\u001b[m |\u001b[31m 35'C\u001b[m, \u001b[32m  0 %\u001b[m | \u001b[36m\u001b[1m\u001b[33m    0\u001b[m / \u001b[33m16130\u001b[m MB |\n"
     ]
    }
   ],
   "source": [
    "!gpustat"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "5pZ8A2liqvgk"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.0.0-beta1\n"
     ]
    }
   ],
   "source": [
    "from __future__ import absolute_import, division, print_function, unicode_literals\n",
    "\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "print(tf.__version__)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "1cweoTiruj8O"
   },
   "source": [
    "## Download the IMDB dataset\n",
    "\n",
    "Rather than using an embedding as in the previous notebook, here we will multi-hot encode the sentences. This model will quickly overfit to the training set. It will be used to demonstrate when overfitting occurs, and how to fight it.\n",
    "\n",
    "Multi-hot-encoding our lists means turning them into vectors of 0s and 1s. Concretely, this would mean for instance turning the sequence  `[3, 5]` into a 10,000-dimensional vector that would be all-zeros except for indices 3 and 5, which would be ones."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "QpzE4iqZtJly"
   },
   "outputs": [],
   "source": [
    "NUM_WORDS = 10000\n",
    "\n",
    "(train_data, train_labels), (test_data, test_labels) = keras.datasets.imdb.load_data(num_words=NUM_WORDS)\n",
    "\n",
    "def multi_hot_sequences(sequences, dimension):\n",
    "    # Create an all-zero matrix of shape (len(sequences), dimension)\n",
    "    results = np.zeros((len(sequences), dimension))\n",
    "    for i, word_indices in enumerate(sequences):\n",
    "        results[i, word_indices] = 1.0  # set specific indices of results[i] to 1s\n",
    "    return results\n",
    "\n",
    "\n",
    "train_data = multi_hot_sequences(train_data, dimension=NUM_WORDS)\n",
    "test_data = multi_hot_sequences(test_data, dimension=NUM_WORDS)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "MzWVeXe3NBTn"
   },
   "source": [
    "Let's look at one of the resulting multi-hot vectors. The word indices are sorted by frequency, so it is expected that there are more 1-values near index zero, as we can see in this plot:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "71kr5rG4LkGM"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x2aab2bc72080>]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(train_data[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "lglk41MwvU5o"
   },
   "source": [
    "## Demonstrate overfitting\n",
    "\n",
    "The simplest way to prevent overfitting is to reduce the size of the model, i.e. the number of learnable parameters in the model (which is determined by the number of layers and the number of units per layer). In deep learning, the number of learnable parameters in a model is often referred to as the model's \"capacity\". Intuitively, a model with more parameters will have more \"memorization capacity\" and therefore will be able to easily learn a perfect dictionary-like mapping between training samples and their targets, a mapping without any generalization power, but this would be useless when making predictions on previously unseen data.\n",
    "\n",
    "Always keep this in mind: deep learning models tend to be good at fitting to the training data, but the real challenge is generalization, not fitting.\n",
    "\n",
    "On the other hand, if the network has limited memorization resources, it will not be able to learn the mapping as easily. To minimize its loss, it will have to learn compressed representations that have more predictive power. At the same time, if you make your model too small, it will have difficulty fitting to the training data. There is a balance between \"too much capacity\" and \"not enough capacity\".\n",
    "\n",
    "Unfortunately, there is no magical formula to determine the right size or architecture of your model (in terms of the number of layers, or the right size for each layer). You will have to experiment using a series of different architectures.\n",
    "\n",
    "To find an appropriate model size, it's best to start with relatively few layers and parameters, then begin increasing the size of the layers or adding new layers until you see diminishing returns on the validation loss. Let's try this on our movie review classification network.\n",
    "\n",
    "We'll create a simple model using only ```Dense``` layers as a baseline, then create smaller and larger versions, and compare them."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "_ReKHdC2EgVu"
   },
   "source": [
    "### Create a baseline model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "QKgdXPx9usBa"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense (Dense)                (None, 16)                160016    \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 16)                272       \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 1)                 17        \n",
      "=================================================================\n",
      "Total params: 160,305\n",
      "Trainable params: 160,305\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "baseline_model = keras.Sequential([\n",
    "    # `input_shape` is only required here so that `.summary` works.\n",
    "    keras.layers.Dense(16, activation='relu', input_shape=(NUM_WORDS,)),\n",
    "    keras.layers.Dense(16, activation='relu'),\n",
    "    keras.layers.Dense(1, activation='sigmoid')\n",
    "])\n",
    "\n",
    "baseline_model.compile(optimizer='adam',\n",
    "                       loss='binary_crossentropy',\n",
    "                       metrics=['accuracy', 'binary_crossentropy'])\n",
    "\n",
    "baseline_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "LqG3MXF5xSjR"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING: Logging before flag parsing goes to stderr.\n",
      "W0713 16:18:20.482950 46912496740800 deprecation.py:323] From /usr/common/software/tensorflow/gpu-tensorflow/2.0.0-beta-py36/lib/python3.6/site-packages/tensorflow/python/ops/math_grad.py:1250: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.where in 2.0, which has the same broadcast rule as np.where\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 25000 samples, validate on 25000 samples\n",
      "Epoch 1/20\n",
      "25000/25000 - 3s - loss: 0.4948 - accuracy: 0.7950 - binary_crossentropy: 0.4948 - val_loss: 0.3382 - val_accuracy: 0.8756 - val_binary_crossentropy: 0.3382\n",
      "Epoch 2/20\n",
      "25000/25000 - 2s - loss: 0.2525 - accuracy: 0.9099 - binary_crossentropy: 0.2525 - val_loss: 0.2846 - val_accuracy: 0.8880 - val_binary_crossentropy: 0.2846\n",
      "Epoch 3/20\n",
      "25000/25000 - 2s - loss: 0.1833 - accuracy: 0.9350 - binary_crossentropy: 0.1833 - val_loss: 0.2900 - val_accuracy: 0.8850 - val_binary_crossentropy: 0.2900\n",
      "Epoch 4/20\n",
      "25000/25000 - 2s - loss: 0.1469 - accuracy: 0.9490 - binary_crossentropy: 0.1469 - val_loss: 0.3217 - val_accuracy: 0.8750 - val_binary_crossentropy: 0.3217\n",
      "Epoch 5/20\n",
      "25000/25000 - 2s - loss: 0.1191 - accuracy: 0.9604 - binary_crossentropy: 0.1191 - val_loss: 0.3386 - val_accuracy: 0.8749 - val_binary_crossentropy: 0.3386\n",
      "Epoch 6/20\n",
      "25000/25000 - 2s - loss: 0.0964 - accuracy: 0.9703 - binary_crossentropy: 0.0964 - val_loss: 0.3672 - val_accuracy: 0.8718 - val_binary_crossentropy: 0.3672\n",
      "Epoch 7/20\n",
      "25000/25000 - 2s - loss: 0.0784 - accuracy: 0.9773 - binary_crossentropy: 0.0784 - val_loss: 0.4062 - val_accuracy: 0.8686 - val_binary_crossentropy: 0.4062\n",
      "Epoch 8/20\n",
      "25000/25000 - 2s - loss: 0.0648 - accuracy: 0.9831 - binary_crossentropy: 0.0648 - val_loss: 0.4460 - val_accuracy: 0.8632 - val_binary_crossentropy: 0.4460\n",
      "Epoch 9/20\n",
      "25000/25000 - 2s - loss: 0.0506 - accuracy: 0.9884 - binary_crossentropy: 0.0506 - val_loss: 0.4827 - val_accuracy: 0.8625 - val_binary_crossentropy: 0.4827\n",
      "Epoch 10/20\n",
      "25000/25000 - 2s - loss: 0.0417 - accuracy: 0.9914 - binary_crossentropy: 0.0417 - val_loss: 0.5294 - val_accuracy: 0.8586 - val_binary_crossentropy: 0.5294\n",
      "Epoch 11/20\n",
      "25000/25000 - 2s - loss: 0.0317 - accuracy: 0.9949 - binary_crossentropy: 0.0317 - val_loss: 0.5697 - val_accuracy: 0.8558 - val_binary_crossentropy: 0.5697\n",
      "Epoch 12/20\n",
      "25000/25000 - 2s - loss: 0.0244 - accuracy: 0.9966 - binary_crossentropy: 0.0244 - val_loss: 0.6076 - val_accuracy: 0.8561 - val_binary_crossentropy: 0.6076\n",
      "Epoch 13/20\n",
      "25000/25000 - 2s - loss: 0.0192 - accuracy: 0.9979 - binary_crossentropy: 0.0192 - val_loss: 0.6516 - val_accuracy: 0.8548 - val_binary_crossentropy: 0.6516\n",
      "Epoch 14/20\n",
      "25000/25000 - 2s - loss: 0.0146 - accuracy: 0.9989 - binary_crossentropy: 0.0146 - val_loss: 0.6890 - val_accuracy: 0.8530 - val_binary_crossentropy: 0.6890\n",
      "Epoch 15/20\n",
      "25000/25000 - 2s - loss: 0.0115 - accuracy: 0.9996 - binary_crossentropy: 0.0115 - val_loss: 0.7157 - val_accuracy: 0.8535 - val_binary_crossentropy: 0.7157\n",
      "Epoch 16/20\n",
      "25000/25000 - 2s - loss: 0.0090 - accuracy: 0.9997 - binary_crossentropy: 0.0090 - val_loss: 0.7477 - val_accuracy: 0.8527 - val_binary_crossentropy: 0.7477\n",
      "Epoch 17/20\n",
      "25000/25000 - 2s - loss: 0.0073 - accuracy: 0.9998 - binary_crossentropy: 0.0073 - val_loss: 0.7747 - val_accuracy: 0.8528 - val_binary_crossentropy: 0.7747\n",
      "Epoch 18/20\n",
      "25000/25000 - 2s - loss: 0.0059 - accuracy: 0.9999 - binary_crossentropy: 0.0059 - val_loss: 0.8034 - val_accuracy: 0.8514 - val_binary_crossentropy: 0.8034\n",
      "Epoch 19/20\n",
      "25000/25000 - 2s - loss: 0.0048 - accuracy: 0.9999 - binary_crossentropy: 0.0048 - val_loss: 0.8291 - val_accuracy: 0.8518 - val_binary_crossentropy: 0.8291\n",
      "Epoch 20/20\n",
      "25000/25000 - 2s - loss: 0.0041 - accuracy: 1.0000 - binary_crossentropy: 0.0041 - val_loss: 0.8523 - val_accuracy: 0.8511 - val_binary_crossentropy: 0.8523\n"
     ]
    }
   ],
   "source": [
    "baseline_history = baseline_model.fit(train_data,\n",
    "                                      train_labels,\n",
    "                                      epochs=20,\n",
    "                                      batch_size=512,\n",
    "                                      validation_data=(test_data, test_labels),\n",
    "                                      verbose=2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "L-DGRBbGxI6G"
   },
   "source": [
    "### Create a smaller model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "SrfoVQheYSO5"
   },
   "source": [
    "Let's create a model with less hidden units to compare against the baseline model that we just created:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "jksi-XtaxDAh"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_3 (Dense)              (None, 4)                 40004     \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 4)                 20        \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 1)                 5         \n",
      "=================================================================\n",
      "Total params: 40,029\n",
      "Trainable params: 40,029\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "smaller_model = keras.Sequential([\n",
    "    keras.layers.Dense(4, activation='relu', input_shape=(NUM_WORDS,)),\n",
    "    keras.layers.Dense(4, activation='relu'),\n",
    "    keras.layers.Dense(1, activation='sigmoid')\n",
    "])\n",
    "\n",
    "smaller_model.compile(optimizer='adam',\n",
    "                      loss='binary_crossentropy',\n",
    "                      metrics=['accuracy', 'binary_crossentropy'])\n",
    "\n",
    "smaller_model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "jbngCZliYdma"
   },
   "source": [
    "And train the model using the same data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Ofn1AwDhx-Fe"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 25000 samples, validate on 25000 samples\n",
      "Epoch 1/20\n",
      "25000/25000 - 3s - loss: 0.5865 - accuracy: 0.7443 - binary_crossentropy: 0.5865 - val_loss: 0.4649 - val_accuracy: 0.8457 - val_binary_crossentropy: 0.4649\n",
      "Epoch 2/20\n",
      "25000/25000 - 2s - loss: 0.3667 - accuracy: 0.8840 - binary_crossentropy: 0.3667 - val_loss: 0.3459 - val_accuracy: 0.8756 - val_binary_crossentropy: 0.3459\n",
      "Epoch 3/20\n",
      "25000/25000 - 2s - loss: 0.2710 - accuracy: 0.9092 - binary_crossentropy: 0.2710 - val_loss: 0.3046 - val_accuracy: 0.8835 - val_binary_crossentropy: 0.3046\n",
      "Epoch 4/20\n",
      "25000/25000 - 2s - loss: 0.2223 - accuracy: 0.9245 - binary_crossentropy: 0.2223 - val_loss: 0.2878 - val_accuracy: 0.8868 - val_binary_crossentropy: 0.2878\n",
      "Epoch 5/20\n",
      "25000/25000 - 2s - loss: 0.1907 - accuracy: 0.9361 - binary_crossentropy: 0.1907 - val_loss: 0.2827 - val_accuracy: 0.8879 - val_binary_crossentropy: 0.2827\n",
      "Epoch 6/20\n",
      "25000/25000 - 2s - loss: 0.1687 - accuracy: 0.9440 - binary_crossentropy: 0.1687 - val_loss: 0.2848 - val_accuracy: 0.8866 - val_binary_crossentropy: 0.2848\n",
      "Epoch 7/20\n",
      "25000/25000 - 2s - loss: 0.1503 - accuracy: 0.9507 - binary_crossentropy: 0.1503 - val_loss: 0.2916 - val_accuracy: 0.8843 - val_binary_crossentropy: 0.2916\n",
      "Epoch 8/20\n",
      "25000/25000 - 2s - loss: 0.1355 - accuracy: 0.9568 - binary_crossentropy: 0.1355 - val_loss: 0.3009 - val_accuracy: 0.8826 - val_binary_crossentropy: 0.3009\n",
      "Epoch 9/20\n",
      "25000/25000 - 2s - loss: 0.1228 - accuracy: 0.9612 - binary_crossentropy: 0.1228 - val_loss: 0.3188 - val_accuracy: 0.8768 - val_binary_crossentropy: 0.3188\n",
      "Epoch 10/20\n",
      "25000/25000 - 2s - loss: 0.1124 - accuracy: 0.9650 - binary_crossentropy: 0.1124 - val_loss: 0.3272 - val_accuracy: 0.8772 - val_binary_crossentropy: 0.3272\n",
      "Epoch 11/20\n",
      "25000/25000 - 2s - loss: 0.1021 - accuracy: 0.9686 - binary_crossentropy: 0.1021 - val_loss: 0.3409 - val_accuracy: 0.8754 - val_binary_crossentropy: 0.3409\n",
      "Epoch 12/20\n",
      "25000/25000 - 2s - loss: 0.0935 - accuracy: 0.9722 - binary_crossentropy: 0.0935 - val_loss: 0.3574 - val_accuracy: 0.8731 - val_binary_crossentropy: 0.3574\n",
      "Epoch 13/20\n",
      "25000/25000 - 2s - loss: 0.0855 - accuracy: 0.9756 - binary_crossentropy: 0.0855 - val_loss: 0.3736 - val_accuracy: 0.8712 - val_binary_crossentropy: 0.3736\n",
      "Epoch 14/20\n",
      "25000/25000 - 2s - loss: 0.0782 - accuracy: 0.9785 - binary_crossentropy: 0.0782 - val_loss: 0.3919 - val_accuracy: 0.8694 - val_binary_crossentropy: 0.3919\n",
      "Epoch 15/20\n",
      "25000/25000 - 2s - loss: 0.0720 - accuracy: 0.9818 - binary_crossentropy: 0.0720 - val_loss: 0.4097 - val_accuracy: 0.8668 - val_binary_crossentropy: 0.4097\n",
      "Epoch 16/20\n",
      "25000/25000 - 2s - loss: 0.0663 - accuracy: 0.9835 - binary_crossentropy: 0.0663 - val_loss: 0.4266 - val_accuracy: 0.8659 - val_binary_crossentropy: 0.4266\n",
      "Epoch 17/20\n",
      "25000/25000 - 2s - loss: 0.0597 - accuracy: 0.9861 - binary_crossentropy: 0.0597 - val_loss: 0.4454 - val_accuracy: 0.8646 - val_binary_crossentropy: 0.4454\n",
      "Epoch 18/20\n",
      "25000/25000 - 2s - loss: 0.0544 - accuracy: 0.9890 - binary_crossentropy: 0.0544 - val_loss: 0.4674 - val_accuracy: 0.8620 - val_binary_crossentropy: 0.4674\n",
      "Epoch 19/20\n",
      "25000/25000 - 2s - loss: 0.0497 - accuracy: 0.9902 - binary_crossentropy: 0.0497 - val_loss: 0.4844 - val_accuracy: 0.8624 - val_binary_crossentropy: 0.4844\n",
      "Epoch 20/20\n",
      "25000/25000 - 2s - loss: 0.0451 - accuracy: 0.9920 - binary_crossentropy: 0.0451 - val_loss: 0.5069 - val_accuracy: 0.8598 - val_binary_crossentropy: 0.5069\n"
     ]
    }
   ],
   "source": [
    "smaller_history = smaller_model.fit(train_data,\n",
    "                                    train_labels,\n",
    "                                    epochs=20,\n",
    "                                    batch_size=512,\n",
    "                                    validation_data=(test_data, test_labels),\n",
    "                                    verbose=2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "vIPuf23FFaVn"
   },
   "source": [
    "### Create a bigger model\n",
    "\n",
    "As an exercise, you can create an even larger model, and see how quickly it begins overfitting.  Next, let's add to this benchmark a network that has much more capacity, far more than the problem would warrant:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "ghQwwqwqvQM9"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_2\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_6 (Dense)              (None, 512)               5120512   \n",
      "_________________________________________________________________\n",
      "dense_7 (Dense)              (None, 512)               262656    \n",
      "_________________________________________________________________\n",
      "dense_8 (Dense)              (None, 1)                 513       \n",
      "=================================================================\n",
      "Total params: 5,383,681\n",
      "Trainable params: 5,383,681\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "bigger_model = keras.models.Sequential([\n",
    "    keras.layers.Dense(512, activation='relu', input_shape=(NUM_WORDS,)),\n",
    "    keras.layers.Dense(512, activation='relu'),\n",
    "    keras.layers.Dense(1, activation='sigmoid')\n",
    "])\n",
    "\n",
    "bigger_model.compile(optimizer='adam',\n",
    "                     loss='binary_crossentropy',\n",
    "                     metrics=['accuracy','binary_crossentropy'])\n",
    "\n",
    "bigger_model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "D-d-i5DaYmr7"
   },
   "source": [
    "And, again, train the model using the same data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "U1A99dhqvepf"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 25000 samples, validate on 25000 samples\n",
      "Epoch 1/20\n",
      "25000/25000 - 3s - loss: 0.3422 - accuracy: 0.8509 - binary_crossentropy: 0.3422 - val_loss: 0.2925 - val_accuracy: 0.8819 - val_binary_crossentropy: 0.2925\n",
      "Epoch 2/20\n",
      "25000/25000 - 2s - loss: 0.1398 - accuracy: 0.9500 - binary_crossentropy: 0.1398 - val_loss: 0.3352 - val_accuracy: 0.8709 - val_binary_crossentropy: 0.3352\n",
      "Epoch 3/20\n",
      "25000/25000 - 2s - loss: 0.0422 - accuracy: 0.9880 - binary_crossentropy: 0.0422 - val_loss: 0.4701 - val_accuracy: 0.8668 - val_binary_crossentropy: 0.4701\n",
      "Epoch 4/20\n",
      "25000/25000 - 2s - loss: 0.0064 - accuracy: 0.9991 - binary_crossentropy: 0.0064 - val_loss: 0.5978 - val_accuracy: 0.8670 - val_binary_crossentropy: 0.5978\n",
      "Epoch 5/20\n",
      "25000/25000 - 2s - loss: 8.1483e-04 - accuracy: 1.0000 - binary_crossentropy: 8.1483e-04 - val_loss: 0.6799 - val_accuracy: 0.8696 - val_binary_crossentropy: 0.6799\n",
      "Epoch 6/20\n",
      "25000/25000 - 2s - loss: 2.1655e-04 - accuracy: 1.0000 - binary_crossentropy: 2.1655e-04 - val_loss: 0.7154 - val_accuracy: 0.8709 - val_binary_crossentropy: 0.7154\n",
      "Epoch 7/20\n",
      "25000/25000 - 2s - loss: 1.3062e-04 - accuracy: 1.0000 - binary_crossentropy: 1.3062e-04 - val_loss: 0.7406 - val_accuracy: 0.8710 - val_binary_crossentropy: 0.7406\n",
      "Epoch 8/20\n",
      "25000/25000 - 2s - loss: 9.4787e-05 - accuracy: 1.0000 - binary_crossentropy: 9.4787e-05 - val_loss: 0.7593 - val_accuracy: 0.8707 - val_binary_crossentropy: 0.7593\n",
      "Epoch 9/20\n",
      "25000/25000 - 2s - loss: 7.3431e-05 - accuracy: 1.0000 - binary_crossentropy: 7.3431e-05 - val_loss: 0.7738 - val_accuracy: 0.8708 - val_binary_crossentropy: 0.7738\n",
      "Epoch 10/20\n",
      "25000/25000 - 2s - loss: 5.9207e-05 - accuracy: 1.0000 - binary_crossentropy: 5.9207e-05 - val_loss: 0.7865 - val_accuracy: 0.8708 - val_binary_crossentropy: 0.7865\n",
      "Epoch 11/20\n",
      "25000/25000 - 2s - loss: 4.8800e-05 - accuracy: 1.0000 - binary_crossentropy: 4.8800e-05 - val_loss: 0.7976 - val_accuracy: 0.8710 - val_binary_crossentropy: 0.7976\n",
      "Epoch 12/20\n",
      "25000/25000 - 2s - loss: 4.1160e-05 - accuracy: 1.0000 - binary_crossentropy: 4.1160e-05 - val_loss: 0.8072 - val_accuracy: 0.8710 - val_binary_crossentropy: 0.8072\n",
      "Epoch 13/20\n",
      "25000/25000 - 2s - loss: 3.5078e-05 - accuracy: 1.0000 - binary_crossentropy: 3.5078e-05 - val_loss: 0.8158 - val_accuracy: 0.8710 - val_binary_crossentropy: 0.8158\n",
      "Epoch 14/20\n",
      "25000/25000 - 2s - loss: 3.0349e-05 - accuracy: 1.0000 - binary_crossentropy: 3.0349e-05 - val_loss: 0.8239 - val_accuracy: 0.8710 - val_binary_crossentropy: 0.8239\n",
      "Epoch 15/20\n",
      "25000/25000 - 2s - loss: 2.6475e-05 - accuracy: 1.0000 - binary_crossentropy: 2.6475e-05 - val_loss: 0.8311 - val_accuracy: 0.8708 - val_binary_crossentropy: 0.8311\n",
      "Epoch 16/20\n",
      "25000/25000 - 2s - loss: 2.3330e-05 - accuracy: 1.0000 - binary_crossentropy: 2.3330e-05 - val_loss: 0.8383 - val_accuracy: 0.8708 - val_binary_crossentropy: 0.8383\n",
      "Epoch 17/20\n",
      "25000/25000 - 2s - loss: 2.0671e-05 - accuracy: 1.0000 - binary_crossentropy: 2.0671e-05 - val_loss: 0.8445 - val_accuracy: 0.8708 - val_binary_crossentropy: 0.8445\n",
      "Epoch 18/20\n",
      "25000/25000 - 2s - loss: 1.8432e-05 - accuracy: 1.0000 - binary_crossentropy: 1.8432e-05 - val_loss: 0.8507 - val_accuracy: 0.8709 - val_binary_crossentropy: 0.8507\n",
      "Epoch 19/20\n",
      "25000/25000 - 2s - loss: 1.6526e-05 - accuracy: 1.0000 - binary_crossentropy: 1.6526e-05 - val_loss: 0.8561 - val_accuracy: 0.8708 - val_binary_crossentropy: 0.8561\n",
      "Epoch 20/20\n",
      "25000/25000 - 2s - loss: 1.4902e-05 - accuracy: 1.0000 - binary_crossentropy: 1.4902e-05 - val_loss: 0.8616 - val_accuracy: 0.8709 - val_binary_crossentropy: 0.8616\n"
     ]
    }
   ],
   "source": [
    "bigger_history = bigger_model.fit(train_data, train_labels,\n",
    "                                  epochs=20,\n",
    "                                  batch_size=512,\n",
    "                                  validation_data=(test_data, test_labels),\n",
    "                                  verbose=2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Fy3CMUZpzH3d"
   },
   "source": [
    "### Plot the training and validation loss\n",
    "\n",
    "<!--TODO(markdaoust): This should be a one-liner with tensorboard -->"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "HSlo1F4xHuuM"
   },
   "source": [
    "The solid lines show the training loss, and the dashed lines show the validation loss (remember: a lower validation loss indicates a better model). Here, the smaller network begins overfitting later than the baseline model (after 6 epochs rather than 4) and its performance degrades much more slowly once it starts overfitting."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "0XmKDtOWzOpk"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot_history(histories, key='binary_crossentropy'):\n",
    "    plt.figure(figsize=(16,10))\n",
    "\n",
    "    for name, history in histories:\n",
    "        val = plt.plot(history.epoch, history.history['val_'+key],\n",
    "                       '--', label=name.title()+' Val')\n",
    "        plt.plot(history.epoch, history.history[key], color=val[0].get_color(),\n",
    "                 label=name.title()+' Train')\n",
    "\n",
    "    plt.xlabel('Epochs')\n",
    "    plt.ylabel(key.replace('_',' ').title())\n",
    "    plt.legend()\n",
    "\n",
    "    plt.xlim([0,max(history.epoch)])\n",
    "\n",
    "\n",
    "plot_history([('baseline', baseline_history),\n",
    "              ('smaller', smaller_history),\n",
    "              ('bigger', bigger_history)])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Bi6hBhdnSfjA"
   },
   "source": [
    "Notice that the larger network begins overfitting almost right away, after just one epoch, and overfits much more severely. The more capacity the network has, the quicker it will be able to model the training data (resulting in a low training loss), but the more susceptible it is to overfitting (resulting in a large difference between the training and validation loss)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ASdv7nsgEFhx"
   },
   "source": [
    "## Strategies to prevent overfitting"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "4rHoVWcswFLa"
   },
   "source": [
    "### Add weight regularization\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "kRxWepNawbBK"
   },
   "source": [
    "You may be familiar with Occam's Razor principle: given two explanations for something, the explanation most likely to be correct is the \"simplest\" one, the one that makes the least amount of assumptions. This also applies to the models learned by neural networks: given some training data and a network architecture, there are multiple sets of weights values (multiple models) that could explain the data, and simpler models are less likely to overfit than complex ones.\n",
    "\n",
    "A \"simple model\" in this context is a model where the distribution of parameter values has less entropy (or a model with fewer parameters altogether, as we saw in the section above). Thus a common way to mitigate overfitting is to put constraints on the complexity of a network by forcing its weights only to take small values, which makes the distribution of weight values more \"regular\". This is called \"weight regularization\", and it is done by adding to the loss function of the network a cost associated with having large weights. This cost comes in two flavors:\n",
    "\n",
    "* [L1 regularization](https://developers.google.com/machine-learning/glossary/#L1_regularization), where the cost added is proportional to the absolute value of the weights coefficients (i.e. to what is called the \"L1 norm\" of the weights).\n",
    "\n",
    "* [L2 regularization](https://developers.google.com/machine-learning/glossary/#L2_regularization), where the cost added is proportional to the square of the value of the weights coefficients (i.e. to what is called the squared \"L2 norm\" of the weights). L2 regularization is also called weight decay in the context of neural networks. Don't let the different name confuse you: weight decay is mathematically the exact same as L2 regularization.\n",
    "\n",
    "L1 regularization introduces sparsity to make some of your weight parameters zero. L2 regularization will penalize the weights parameters without making them sparse—one reason why L2 is more common.\n",
    "\n",
    "In `tf.keras`, weight regularization is added by passing weight regularizer instances to layers as keyword arguments. Let's add L2 weight regularization now."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "HFGmcwduwVyQ"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 25000 samples, validate on 25000 samples\n",
      "Epoch 1/20\n",
      "25000/25000 - 3s - loss: 0.5059 - accuracy: 0.8188 - binary_crossentropy: 0.4651 - val_loss: 0.3718 - val_accuracy: 0.8768 - val_binary_crossentropy: 0.3298\n",
      "Epoch 2/20\n",
      "25000/25000 - 2s - loss: 0.2969 - accuracy: 0.9101 - binary_crossentropy: 0.2516 - val_loss: 0.3349 - val_accuracy: 0.8853 - val_binary_crossentropy: 0.2871\n",
      "Epoch 3/20\n",
      "25000/25000 - 2s - loss: 0.2470 - accuracy: 0.9310 - binary_crossentropy: 0.1969 - val_loss: 0.3362 - val_accuracy: 0.8858 - val_binary_crossentropy: 0.2846\n",
      "Epoch 4/20\n",
      "25000/25000 - 2s - loss: 0.2243 - accuracy: 0.9412 - binary_crossentropy: 0.1710 - val_loss: 0.3519 - val_accuracy: 0.8810 - val_binary_crossentropy: 0.2973\n",
      "Epoch 5/20\n",
      "25000/25000 - 2s - loss: 0.2087 - accuracy: 0.9486 - binary_crossentropy: 0.1530 - val_loss: 0.3644 - val_accuracy: 0.8772 - val_binary_crossentropy: 0.3077\n",
      "Epoch 6/20\n",
      "25000/25000 - 2s - loss: 0.1971 - accuracy: 0.9542 - binary_crossentropy: 0.1393 - val_loss: 0.3763 - val_accuracy: 0.8756 - val_binary_crossentropy: 0.3179\n",
      "Epoch 7/20\n",
      "25000/25000 - 2s - loss: 0.1893 - accuracy: 0.9570 - binary_crossentropy: 0.1302 - val_loss: 0.3993 - val_accuracy: 0.8712 - val_binary_crossentropy: 0.3394\n",
      "Epoch 8/20\n",
      "25000/25000 - 2s - loss: 0.1825 - accuracy: 0.9592 - binary_crossentropy: 0.1219 - val_loss: 0.4091 - val_accuracy: 0.8707 - val_binary_crossentropy: 0.3479\n",
      "Epoch 9/20\n",
      "25000/25000 - 2s - loss: 0.1762 - accuracy: 0.9612 - binary_crossentropy: 0.1143 - val_loss: 0.4298 - val_accuracy: 0.8660 - val_binary_crossentropy: 0.3674\n",
      "Epoch 10/20\n",
      "25000/25000 - 2s - loss: 0.1712 - accuracy: 0.9643 - binary_crossentropy: 0.1084 - val_loss: 0.4383 - val_accuracy: 0.8677 - val_binary_crossentropy: 0.3749\n",
      "Epoch 11/20\n",
      "25000/25000 - 2s - loss: 0.1663 - accuracy: 0.9662 - binary_crossentropy: 0.1024 - val_loss: 0.4518 - val_accuracy: 0.8638 - val_binary_crossentropy: 0.3874\n",
      "Epoch 12/20\n",
      "25000/25000 - 2s - loss: 0.1623 - accuracy: 0.9673 - binary_crossentropy: 0.0974 - val_loss: 0.4722 - val_accuracy: 0.8624 - val_binary_crossentropy: 0.4071\n",
      "Epoch 13/20\n",
      "25000/25000 - 2s - loss: 0.1570 - accuracy: 0.9711 - binary_crossentropy: 0.0915 - val_loss: 0.4828 - val_accuracy: 0.8621 - val_binary_crossentropy: 0.4171\n",
      "Epoch 14/20\n",
      "25000/25000 - 2s - loss: 0.1540 - accuracy: 0.9710 - binary_crossentropy: 0.0879 - val_loss: 0.4997 - val_accuracy: 0.8586 - val_binary_crossentropy: 0.4333\n",
      "Epoch 15/20\n",
      "25000/25000 - 2s - loss: 0.1526 - accuracy: 0.9716 - binary_crossentropy: 0.0860 - val_loss: 0.5117 - val_accuracy: 0.8598 - val_binary_crossentropy: 0.4446\n",
      "Epoch 16/20\n",
      "25000/25000 - 2s - loss: 0.1504 - accuracy: 0.9723 - binary_crossentropy: 0.0825 - val_loss: 0.5261 - val_accuracy: 0.8561 - val_binary_crossentropy: 0.4579\n",
      "Epoch 17/20\n",
      "25000/25000 - 2s - loss: 0.1490 - accuracy: 0.9732 - binary_crossentropy: 0.0808 - val_loss: 0.5323 - val_accuracy: 0.8565 - val_binary_crossentropy: 0.4638\n",
      "Epoch 18/20\n",
      "25000/25000 - 2s - loss: 0.1467 - accuracy: 0.9744 - binary_crossentropy: 0.0777 - val_loss: 0.5604 - val_accuracy: 0.8532 - val_binary_crossentropy: 0.4911\n",
      "Epoch 19/20\n",
      "25000/25000 - 2s - loss: 0.1446 - accuracy: 0.9752 - binary_crossentropy: 0.0751 - val_loss: 0.5625 - val_accuracy: 0.8542 - val_binary_crossentropy: 0.4926\n",
      "Epoch 20/20\n",
      "25000/25000 - 2s - loss: 0.1371 - accuracy: 0.9802 - binary_crossentropy: 0.0673 - val_loss: 0.5669 - val_accuracy: 0.8539 - val_binary_crossentropy: 0.4974\n"
     ]
    }
   ],
   "source": [
    "l2_model = keras.models.Sequential([\n",
    "    keras.layers.Dense(16, kernel_regularizer=keras.regularizers.l2(0.001),\n",
    "                       activation='relu', input_shape=(NUM_WORDS,)),\n",
    "    keras.layers.Dense(16, kernel_regularizer=keras.regularizers.l2(0.001),\n",
    "                       activation='relu'),\n",
    "    keras.layers.Dense(1, activation='sigmoid')\n",
    "])\n",
    "\n",
    "l2_model.compile(optimizer='adam',\n",
    "                 loss='binary_crossentropy',\n",
    "                 metrics=['accuracy', 'binary_crossentropy'])\n",
    "\n",
    "l2_model_history = l2_model.fit(train_data, train_labels,\n",
    "                                epochs=20,\n",
    "                                batch_size=512,\n",
    "                                validation_data=(test_data, test_labels),\n",
    "                                verbose=2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "bUUHoXb7w-_C"
   },
   "source": [
    "```l2(0.001)``` means that every coefficient in the weight matrix of the layer will add ```0.001 * weight_coefficient_value**2``` to the total loss of the network. Note that because this penalty is only added at training time, the loss for this network will be much higher at training than at test time.\n",
    "\n",
    "Here's the impact of our L2 regularization penalty:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "7wkfLyxBZdh_"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA7AAAAJNCAYAAAAIxpmxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAAgAElEQVR4nOzdeXhV9YH/8fe5NysJAUISlrCEfSdhlX1VQAXXWqVatdaxnbq07mvbqVPHzkzHabX9aa11aW1dasellUq1FVDBIqsgiOyyhB0SliRkOb8/IoiCGJZws7xfz8OT3HPOPfdzE56QD9/vOd8gDEMkSZIkSarpIrEOIEmSJElSVVhgJUmSJEm1ggVWkiRJklQrWGAlSZIkSbWCBVaSJEmSVCtYYCVJkiRJtUJcrAMcq4yMjDAnJyfWMSRJkiRJ1WDu3LnbwjDMPNK+Wldgc3JymDNnTqxjSJIkSZKqQRAEa79on1OIJUmSJEm1ggVWkiRJklQrWGAlSZIkSbVCrbsG9khKS0tZv349xcXFsY6io0hKSqJVq1bEx8fHOookSZKkWqhOFNj169fTsGFDcnJyCIIg1nF0BGEYsn37dtavX0+7du1iHUeSJElSLVQnphAXFxfTtGlTy2sNFgQBTZs2dZRckiRJ0nGrEwUWsLzWAn6PJEmSJJ2IOlNgYy0ajZKXl0dubi59+/Zl5syZJ/X8V155JS+88AIAV199NUuWLDmh8+3du5emTZtSUFDwme3nnXcezz///Bc+b9q0aUycOPGEXluSJEmSjocF9iRJTk5mwYIFLFy4kPvvv58777yz2l7rscceo3v37id0jpSUFMaNG8dLL710cFtBQQFvv/22BVWSJElSjWSBrQaFhYU0adIEgD179jB27Fj69u1Lr169ePnll4HKEdCzzz6b3NxcevbsyXPPPQfA3LlzGTlyJP369WP8+PHk5+cfdv5Ro0YxZ84cAFJTU7n77rvJzc1l0KBBbN68GYCtW7dy4YUXMmDAAAYMGMA777xz2HkmT57Ms88+e/Dxiy++yIQJE2jQoAGzZ89myJAh9OnThyFDhrBs2bKT+0WSJEmSpGNkgT1JioqKyMvLo2vXrlx99dV8//vfByqXjnnxxReZN28eb775JjfffDNhGPLaa6/RsmVLFi5cyOLFi5kwYQKlpaVcf/31vPDCC8ydO5errrqKu++++6ivu3fvXgYNGsTChQsZMWIEv/71rwH47ne/y4033sh7773Hn/70J66++urDnjthwgTmzp3L9u3bAXj22WeZPHkyAF27dmXGjBnMnz+fe++9l7vuuutkfrkkSZIk6ZjViWV0Pu/iX806bNvE3i34+uAcivaXc+UTsw/b/5V+rbiof2t27N3Pvz499zP7nvvW4C99zQNTiAFmzZrF5ZdfzuLFiwnDkLvuuosZM2YQiUTYsGEDmzdvplevXtxyyy3cfvvtTJw4keHDh7N48WIWL17MGWecAUB5eTktWrQ46usmJCQcnPLbr18/Xn/9dQDeeOONz1wnW1hYyO7du2nYsOFnnnvOOefwwgsvcOGFF7JgwQLGjRsHVE4nvuKKK1i+fDlBEFBaWvqlXwNJkiRJqk51ssDG2uDBg9m2bRtbt25lypQpbN26lblz5xIfH09OTg7FxcV07tyZuXPnMmXKFO68807GjRvH+eefT48ePZg16/AC/kXi4+MP3t03Go1SVlYGQEVFBbNmzSI5Ofmoz588eTI//vGPCcOQc889l/j4eAC+//3vM3r0aF588UXWrFnDqFGjju+LIUmSJEknSZ0ssEcbMU1OiB51f3pKQpVGXI/mww8/pLy8/OBdfrOysoiPj+fNN99k7dq1AGzcuJH09HQuu+wyUlNTefLJJ7njjjvYunUrs2bNYvDgwZSWlvLRRx/Ro0ePY84wbtw4fvGLX3DrrbcCsGDBAvLy8g47bvTo0VxxxRX88pe/5KGHHjq4vaCggOzsbACefPLJ4/gqSJIkSdLJVScLbCwcuAYWIAxDnnrqKaLRKJdeeimTJk2if//+B6+RBVi0aBG33norkUiE+Ph4Hn74YRISEnjhhRe44YYbKCgooKysjO9973vHVWAffPBBrr32Wnr37k1ZWRkjRozgkUceOey4SCTChRdeyB//+EdGjBhxcPttt93GFVdcwQMPPMCYMWOO86siSZIkSSdPEIZhrDMck/79+4cH7sB7wNKlS+nWrVuMEulY+L2SJEmSdDRBEMwNw7D/kfZ5F2JJkiRJUq1ggZUkSZIk1QoWWEmSJElSrWCBlSRJkiTVChZYSZIkSVKtYIGVJEmSJNUKFtiTJBqNkpeXR25uLn379mXmzJkn9fxXXnklL7zwAgBXX301S5YsOaHzTZ06lby8PPLy8khNTaVLly7k5eVx+eWXV/kc5eXlDB8+/IRySJIkSarfwjBkT0lZlY6Nq+Ys9UZycjILFiwAKsvhnXfeyfTp06vltR577LETPsf48eMZP348AKNGjeKnP/0p/fsfvtRSWVkZcXFH/msSjUZ56623TjiLJEmSpPqlsLiUt5dv480PtzDto62c3i2L+y/o/aXPcwS2GhQWFtKkSRMA9uzZw9ixY+nbty+9evXi5ZdfBmDv3r2cffbZ5Obm0rNnT5577jkA5s6dy8iRI+nXrx/jx48nPz//sPOPGjWKOXPmAJCamsrdd99Nbm4ugwYNYvPmzQBs3bqVCy+8kAEDBjBgwADeeeedKud/7LHHuOSSS5g4cSJnnnkmhYWFjBkzhr59+9K7d2/+8pe/AJXltnHjxgC88cYbjB07lgsuuIAuXboc00iuJEmSpPrjhmfm0+fe1/nO7+cx9YNNnNYunZGds6r0XEdgT5KioiLy8vIoLi4mPz+ff/zjHwAkJSXx4osvkpaWxrZt2xg0aBDnnHMOr732Gi1btuTVV18FoKCggNLSUq6//npefvllMjMzee6557j77rt5/PHHv/B19+7dy6BBg7jvvvu47bbb+PWvf80999zDd7/7XW688UaGDRvGxx9/zPjx41m6dGmV38+sWbNYsGABTZo0obS0lJdffpmGDRuyZcsWhg4dysSJEw97zrx581iyZAlZWVkMGjSId999l0GDBh3jV1KSJElSXbCnpIx3Vmxj2rItLNlYyEvXDiUIAro0b0jr9GRGdcmiT+vGxEWrPq5a5wrsj/78AUs2Fp7Uc3ZvmcYPJ/U46jGHTiGeNWsWl19+OYsXLyYMQ+666y5mzJhBJBJhw4YNbN68mV69enHLLbdw++23M3HiRIYPH87ixYtZvHgxZ5xxBlB5jWmLFi2O+roJCQkHy2S/fv14/fXXgcoR0UOvky0sLGT37t00bNiwSu953LhxB0eRwzDk9ttv5+233yYSibBu3Tq2bdt2cPT1gEGDBh3Mm5eXx5o1ayywkiRJUj3z9vJt/L9pK3hvzQ5Ky0NSE+MY3imDwuIyGiXHc+3ojsd97jpXYGuCwYMHs23bNrZu3cqUKVPYunUrc+fOJT4+npycHIqLi+ncuTNz585lypQp3HnnnYwbN47zzz+fHj16MGvWrCq/Vnx8PEEQAJXXpJaVVV78XFFRwaxZs0hOTj6u95CSknLw89/+9rcUFBQwb9484uLiaNWqFcXFxYc9JzEx8eDnh2aRJEmSVDftLSlj5srtTFu2hUtPa0v3lmmUlJWzfc9+rhrWjtFdsujXtgnxxzDKejR1rsB+2UjpqfDhhx9SXl5O06ZNKSgoICsri/j4eN58803Wrl0LwMaNG0lPT+eyyy4jNTWVJ598kjvuuIOtW7cya9YsBg8eTGlpKR999BE9ehz7exo3bhy/+MUvuPXWWwFYsGABeXl5x/V+DryHuLg4Xn/9dTZs2HBc55EkSZJU++0tKePZ99YxbdkW/rlqB/vLK0hJiDKwXTrdW6YxpmsWY7s1q5bXrnMFNlYOXAMLlVNun3rqKaLRKJdeeimTJk2if//+5OXl0bVrVwAWLVrErbfeSiQSIT4+nocffpiEhAReeOEFbrjhBgoKCigrK+N73/vecRXYBx98kGuvvZbevXtTVlbGiBEjeOSRR47rvX39618/+B769u1Lp06djus8kiRJkmqfov3lvLtqOyEhY7o2IxoJ+O+pH9KqSQOuGNKW0V2y6J+TTkJc5SjrgRmi1SEIw7DaTl4d+vfvHx64A+8BS5cupVu3bjFKpGPh90qSJEmq+dZu33twiZtZK7dTUlZBv7ZN+NO/DgFg+54SmqYmfslZjk8QBHPDMDx8jU8cgZUkSZKkeq+4tJz31xcwsF06AP/+l6W8sXQz7TNS+NppbRjdJevgPqDayuuXscBKkiRJUj20bsc+pi3bwpvLtjJz5TaKSyt4986xNG+UxK3ju3DP2d3IyUj58hOdQhZYSZIkSaoHSsrKCUNIio/yysKN3PDMfADapDfg4v6tGdU1iyYp8QB0aV615TdPNQusJEmSJNVRG3cVMW3ZVt5ctoWZK7Zx19nduPS0tgxql873J3ZndJdM2mWkVOuNl04mC6wkSZIk1THFpeWc98t3+HDTbgCyGydzft9serRsBEBWWhLfHNYulhGPiwVWkiRJkmqxkrJyZq7YzmuLN1ERhvz3RbkkxUfp17YJF/ZtxagumXTMSq01o6xHE4l1gLoiNTX1sG0PPPAA3bt3p3fv3owdO5a1a9cedsyoUaOYOnXqZ7b97Gc/4zvf+c4xv54kSZKk+uPt5du4/pn59Pv3N/jGk+/x6qJ8AA4slXrf+b34lxHt6dSsYZ0or2CBrVZ9+vRhzpw5vP/++3zlK1/htttuO+yYyZMn8+yzz35m27PPPsvkyZNPVUxJkiRJtcCOvft5/r11FO0vB2Du2p3MXLGNSbkteOIbA5j7/dP574ty60xZPRKnEFej0aNHH/x80KBBPP3004cd85WvfIV77rmHkpISEhMTWbNmDRs3bmTYsGHs2bOHc889l507d1JaWsqPf/xjzj333FP5FiRJkiTF0MZdRUz9YBNTP9jE7NU7qAihaWoCY7s145oR7bluTEeikbpbWD/PAnuK/OY3v+HMM888bHvTpk0ZOHAgr732Gueeey7PPvssF198MUEQkJSUxIsvvkhaWhrbtm1j0KBBnHPOOXX6f1QkSZKk+q6krJzEuCgrtuzm9AdmANApK5VrR3dkfI/m9GiZBkByQjSWMWOibhbYJ84+fFuP82Dgv8D+ffD7iw7fn/c16HMp7N0Oz1/+2X3fePWE4jz99NPMmTOH6dOnH3H/gWnEBwrs448/DlTOXb/rrruYMWMGkUiEDRs2sHnzZpo3b35CeSRJkiTVHGEYsmhDwScjrZsZkNOE+y/oTYfMVH44qTsjO2fSPtN74EBdLbA1yBtvvMF9993H9OnTSUxMPOIx5513HjfddBPz5s2jqKiIvn37AvD73/+erVu3MnfuXOLj48nJyaG4uPhUxpckSZJUjR76+3KefW8dG3YVEY0EDGqfzoCcdACCIOAbQ2vfUjfVqW4W2KONmCY0OPr+lKYnPOJ6wPz58/nWt77Fa6+9RlZW1hcel5qayqhRo7jqqqs+c/OmgoICsrKyiI+P58033zziXYwlSZIk1Q7FpeXMXLmNt5Zv4/tndycSCdhVVEq3FmnceEZnxnbNoklKQqxj1mh1s8DGwL59+2jVqtXBxzfddBNTpkxhz549XHRR5ZTlNm3a8Morrxzx+ZMnT+aCCy74zB2JL730UiZNmkT//v3Jy8uja9eu1fsmJEmSJJ1Ue0rKePPDLUz9YBNvfriFvfvLaZgYx5VDcmjbNIXvT+we64i1igX2JKmoqDhs20033VTl559//vkH12s6ICMjg1mzZh3x+D179hxbQEmSJEmnxPY9JQA0TU1k9urtXP/MfDJSEzgnryXjezRnSIcMEuJc0fR4WGAlSZIk6QRt2FXE1MWVy928t2YH143uyE3jujC0YwbPf2sw/do2qVfL3VQXC6wkSZIkHacwDLn40XeZvXoHAJ2bpXLd6I5MzG0JQGJclIHt0mMZsU6xwEqSJElSFRxY7ua1xZtYu30fv7y0L0EQMCCnCaO7ZDG+RzOXu6lm1VpggyCYAPwciAKPhWH4k8/tbwM8BTT+5Jg7wjCcUp2ZJEmSJOlYLNlYyPNz1vG3DzaxsaD44HI3xaXlJMVHuXW8N1s9VaqtwAZBEAV+CZwBrAfeC4LglTAMlxxy2D3A82EYPhwEQXdgCpBTXZkkSZIk6cuUllfw9vJt9GrViIzURBZt2MUzsz9meKdMbhrXxeVuYqg6R2AHAivCMFwFEATBs8C5wKEFNgTSPvm8EbCxGvNIkiRJ0hGFYci8j3fx8oIN/OX9fHbs3c+/n9eTrw9qyzm52Uzs3ZKURK/AjLXq/A5kA+sOebweOO1zx/wb8LcgCK4HUoDTqzFPtUpNTT1saZsHHniAxx57jLi4ODIzM3n88cdp27btwf3bt29n7NixAGzatIloNEpmZiYAs2fPJiGhav+r841vfIM77riDLl26nKR3I0mSJNUfRfvLmfDzGazdvo/EuAind2/GeXnZjOicAUByQjTGCXVAdRbYI90jOvzc48nAk2EY/k8QBIOB3wVB0DMMw88sqhoEwTXANQBt2rSplrDVoU+fPsyZM4cGDRrw8MMPc9ttt/Hcc88d3N+0aVMWLFgAwL/927+RmprKLbfccth5wjAkDEMikSOvFfXEE09UzxuQJEmS6qAthcW8snAjW3aXcNdZ3UhOiDKuezO6NE9jfI9mNEyKj3VEfYHqXD13PdD6kMetOHyK8DeB5wHCMJwFJAEZnz9RGIaPhmHYPwzD/gdGKGuD0aNH06BBAwAGDRrE+vXrq/zcFStW0LNnT7797W/Tt29f8vPzueaaa+jfvz89evTg3nvvPXjssGHDWLBgAWVlZTRu3Jg77riD3NxcBg8ezJYtW076+5IkSZJqm93Fpbwwdz2XPfZPBt3/d3786lLmrNlBeUXlGNvdZ3fnK/1aWV5ruOossO8BnYIgaBcEQQJwCfDK5475GBgLEARBNyoL7NZqzBQzv/nNbzjzzDOP6TlLlizhm9/8JvPnzyc7O5uf/OQnzJkzh4ULF/L666+zZMmSw55TUFDAyJEjWbhwIYMHD+bxxx8/WW9BkiRJqlX2l1VQWl45ufOpmWu45Y8LWbtjL9eO7sgbN43k/74zlGjkSBNHVVNV2xTiMAzLgiC4DphK5RI5j4dh+EEQBPcCc8IwfAW4Gfh1EAQ3Ujm9+MowDD8/zfjY/PUO2LToBNN/TvNecOZPvvy4L/D0008zZ84cpk+ffkzP69ChAwMGDDj4+JlnnuE3v/kNZWVlbNy4kSVLltC9e/fPPCc5OflgUe7Xrx9vvfXWceeWJEmSapuKipC5H+/kxfkbmLIon/84vxdn9WrBV/u3ZnCHDPq2aUwQWFprq2q9jdYna7pO+dy2Hxzy+RJgaHVmiLU33niD++67j+nTp5OYmHhMz01JSTn4+fLly/n5z3/O7Nmzady4MZdddhnFxcWHPefQGz9Fo1HKysqOP7wkSZJUSxSXlvPg35fz8oKNbNhVRFJ8hHHdm5PdOBmArLQkstKSYpxSJ6ru3Qf6BEZKT7b58+fzrW99i9dee42srKwTOldhYSENGzYkLS2N/Px8pk6dyoQJE05SUkmSJKn22VRQzEebdzOicyaJcRGmfrCJDlmp3DK+M+O6N3fZmzrI7+hJsm/fPlq1anXw8U033cSUKVPYs2cPF110EVB5B+VXXvn8ZcBV07dvX7p3707Pnj1p3749Q4fW6YFrSZIk6YgKikp5bXE+L83fyLurt5OWFM+ce04nPhrhr98dQUJcdd7mR7EWnOglp6da//79wzlz5nxm29KlS+nWrVuMEulY+L2SJEnS8Xp29sf84JUP2F9WQU7TBpybl815fbJpl5Hy5U9WrREEwdwwDPsfaZ8jsJIkSZJqnIqKkH+u3sHLCzZw8YDW9GnThO4t0/jawDac1yeb3FaNvBlTPWSBlSRJklRjLM0v5KUFG/jzgo1sLCimQUKUfm2b0KdNE3q3akzvVo1jHVExZIGVJEmSFFNF+8tJTohSWl7BJY++y56SMkZ2zuT2M7tyRvdmNEiwtqhSnfmbEIahUwhquNp2vbUkSZKqz659+5myaBMvLdjA1t0l/OPmkcRHIzxyWT86N0ulaeqxLUGp+qFOFNikpCS2b99O06ZNLbE1VBiGbN++naQk196SJEmqz+au3cGvpq/izWVbKC0PaZ+Zwvl9stlfXkFiXJTBHZrGOqJqsDpRYFu1asX69evZunVrrKPoKJKSkj6z1JAkSZLqvvKKkHdXbad9ZgotGiWzpbCE+et2cfngHM7Ly6ZndpqDUKqyOlFg4+PjadeuXaxjSJIkSaJy9t38dbuY8n4+f35/I5sLS7j5jM5cP7YTZ3RvxrgezYlGLK06dnWiwEqSJEmqGcrKKzjjf2ewette4qMBIztn8f2JLTm9WzMA4qKRGCdUbWaBlSRJknRcwjDk/fUFvLoon/yCYh6a3Ie4aITz+2ST3TiZ07s3o1FyfKxjqg6xwEqSJEk6Jss37+aFuet5dVE+63cWER8NGNEpk9LyCuKjEW4Y2ynWEVVHWWAlSZIkHVUYhizeUEibpg1olBzPu6u285u3VzOsUwbfHduJcd2b06iBI62qfhZYSZIkSYcJw5APNhbyl/fzmbIon4937OM/L+zFxQPacH7fVkzKbUnjBgmxjql6xgIrSZIk6TP2lJQx8cG3WLN9H3GRgCEdM7hudEfG9ai8EVNqojVCseHfPEmSJKkeC8OQJfmFvPp+PsWlFfxgUndSE+MY2TmTf22ZxrjuzWmS4kiragYLrCRJklQPLd+8m5cWbGDKok2s3raXaCRgTNcswjAkCAJ+dG7PWEeUDmOBlSRJkuqBMAz5cNNu2memkBgX5c8LN/LI9FUMbt+Ua0a0Z3yP5qQ70qoazgIrSZIk1VEHSuuURfm8+n4+q7bt5fEr+zOmazOuGJLDFUNyaJqaGOuYUpVZYCVJkqQ6aHNhMZN//S6rtu4lEsCg9k355vB25LVuAmBxVa1kgZUkSZJquTAM+WjzHl5dlE9iXIRrR3ckMzWRbi3SuGpoO8b3aE5mQwuraj8LrCRJklRLrdiymz8vzOfVRfms2LKHSABn9WoBQCQS8Muv9Y1xQunkssBKkiRJtciKLXvokJlCEAT85u3VPPveOk5rl84VQ3oywZFW1XEWWEmSJKmGW7FlN6++v4lXF23ko817eOW6ofRu1ZjrxnTixjM6k9UwKdYRpVPCAitJkiTVUKu37eXa389jSX4hQQADctL50Tk9aN2kAQDZjZNjnFA6tSywkiRJUg0RhiHzPt5JYXEZo7tk0aJREo2S4/nhpO6c1asFzdIcaVX9ZoGVJEmSYmzXvv28OH8Dz8z+mI8276Fndhqju2SRFB/lmWsGxTqeVGNYYCVJkqQY+vWMVfz0b8soKasgt1UjfnJBLybltox1LKlGssBKkiRJp9COvfv5v3nrObt3C1o0SqZ9Zgpf7d+aSwa2pkfLRrGOJ9VoFlhJkiSpmoVhyKxV23lm9jqmLt7E/vIKUhLjmDywDWO7NWNst2axjijVChZYSZIkqRrtL6vgrAffYsWWPaQlxfG109oweWAbujRvGOtoUq1jgZUkSZJOooqKkJkrtzP/451cP7YTCXERzuzZnHYZKZzVqwVJ8dFYR5RqLQusJEmSdBJs2V3MH+es57n31vHxjn2kpyRwxdAc0pLiuXlcl1jHk+oEC6wkSZJ0gv72wSa+8/t5lFWEDGqfzs3jOjO+R3NHW6WTzAIrSZIkHaNNBcU8P2cdXZo3ZHyP5vTPSeeqYe24eEBrOmSmxjqeVGdZYCVJkqQqKK8Imf7RFv7wz3X848PNVIRw9bB2jO/RnPSUBO46q1usI0p1ngVWkiRJqoJvPvUe05ZtJSM1kW+N7MAlA1rTtmlKrGNJ9YoFVpIkSfqcsvIK/vHhFv40bz0/vSiXhknxXD64LZcMaM3Ybs2Ij0ZiHVGqlyywkiRJ0ifW7djHc++t449z17G5sISshoms2rqX3NaNGdO1WazjSfWeBVaSJEmisryO+O83CYBRXbL493NbM6ZrFnGOtko1hgVWkiRJ9dLa7Xt59r11FJeW88NJPWid3oAfn9eTUV2yyG6cHOt4ko7AAitJkqR6Y39ZBX9bsolnZ6/j7RXbiARwZs8WhGFIEARcelrbWEeUdBQWWEmSJNV5BwrqQ/9YzkP/WEF242RuPqMzF/VvTfNGSbGOJ6mKLLCSJEmqk5Zv3s2ri/KZsiif2yd0ZWy3Zlw8oDX92jZheKdMopEg1hElHSMLrCRJkuqM4tJyHp62kimL8lm+ZQ9BAP3bNiEhrvJGTK2aNKBVkwYxTinpeFlgJUmSVGuFYchHm/ewYdc+xnRtRkI0wgtz15PdJJkfndODCT2b0yzNKcJSXWGBlSRJUq0ShiEfbtrNlE+mB6/cupdmaYm8e2cWkUjA328eSVJ8NNYxJVUDC6wkSZJqvDAMAQiCgP+euoz/N20lkQBOa9eUK4e2Y0KP5gRB5TWtllep7rLASpIkqUYKw5APNhYyZVE+f128iQcv6UOvVo0Y36M5LRsnM6FnczJSE2MdU9IpZIGVJElSjVKwr5RHZqzkr4vyWbN9H9FIwOD2TSmrqAAgt3Vjcls3jnFKSbFggZUkSVJMhWHIog0F7C4uY2jHDBLjI/z+3bXktm7Mt0d2YFyP5qSnJMQ6pqQawAIrSZKkUy4MQ95fX1B5I6bF+azbUUSPlmm8esNwkuKjzL77dK9llXQYC6wkSZJOiTAMD95o6fY/vc/zc9YTHw0Y2jGD68d0Ylz3ZgePtbxKOhILrCRJkqpNRUXI/HW7mLIon9cWb+KP3x5My8bJnJeXzYCcdMZ1b06jBvGxjimplrDASpIk6aTbXFjMr6av4q+L88kvKCY+GjC8UyZ7S8oAGNIxI8YJJdVGFlhJkiSdsIqKkHkf7yQIoF/bdCJBwB9mr2VYxwxuHd+Fsd2a0SjZkVZJJ8YCK0mSpONSUTDzYi4AACAASURBVBEyZ+3OT9ZpzWdzYQmjumTy5DcGktkwkfnfH0dygteySjp5LLCSJEmqskNvxPTNp97jzWVbSYiLMKpzJmf3bsGYrlkHj7W8SjrZLLCSJEn6Uht2FfHE26uZumQTf/3uCFIT45g8sA3n9clmbLdmpCb6a6Wk6udPGkmSJH2hDzcV8uj0VbyycCMAp3drRkFRKamJcYzr0TzG6STVNxZYSZIkHdGabXuZ8LO3aJAQ5YohOVw1rB3ZjZNjHUtSPWaBlSRJElB5U6a/LdnMyq17uHZ0R3IyUvifi3IZ2y2Lxg0SYh1PkiywkiRJ9V1xaTkvzd/AozNWsWrbXjpkpvAvw9uTEBfhwn6tYh1Pkg6ywEqSJNVjby/fxo3PL2Dr7hJ6ZTfil1/ry4SezYlGglhHk6TDWGAlSZLqmU0FxezdX0aHzFRyMhrQvUUa11zcniEdmh5cIkeSaiILrCRJUj2xYstufjV9FS8t2MDQjhk8+Y2BtGrSgKeuGhjraJJUJRZYSZKkOm7+xzv55ZsreWPpZpLiI3xtYBuuHt4+1rEk6ZhZYCVJkuqgioqQEIhGAmau3M6ctTv47thOXDEkh/QU7ygsqXaywEqSJNUh+8sqeGlB5R2FbxjbiXNyW/KNoTl8Y2gODRL81U9S7eZPMUmSpDpgd3Epz8z+mN+8vZrNhSV0a5FGWlLlr3oWV0l1hT/NJEmS6oDLH5/N/I93Mbh9U/7rK7mM6JThHYUl1TkWWEmSpFpo1dY9PDVzDbdN6EpKYhy3jutCSmIcua0bxzqaJFUbC6wkSVItMv/jnfxq+iqmLtlEfDTCuB7NGdoxgyEdM2IdTZKqnQVWkiSpFthbUsZVT77HP1fvIC0pjmtHdeSKITlkNkyMdTRJOmUssJIkSTVUaXkFizYU0LdNE1IS42iWlsQ9Z3fjkoFtSE301zhJ9Y8/+SRJkmqYvSVlPPveOn7z1iq27dnP23eMJqthEg9O7hPraJIUUxZYSZKkGmLn3v08/s5qfjtrLQVFpQxsl86Pz+9JZqrThCUJLLCSJEkxV14REo0EFBSV8vC0lYztlsW3Rnagb5smsY4mSTWKBVaSJClGFq0v4JHpKyktr+DRy/uTk5HCzDvGkJWWFOtoklQjWWAlSZJOoTAMeWv5Nn41YyXvrNhOw8Q4LhvcloqKkEgksLxK0lFYYCVJkk6hJ95Zw71/WUJWw0TuPLMrk09rQ1pSfKxjSVKtYIGVJEmqZh9uKqS4tIK81o25sF8rUhPjOLdPSxLjorGOJkm1igVWkiSpmmzcVcQDr3/En+atZ2BOOs99azCNkuP56oDWsY4mSbWSBVaSJOkkKygq5ZHpK3n87dWEIVw9rB3Xju4Y61iSVOtZYCVJkk6yV9/P5+FpKzm/TzY3ndGZ1ukNYh1JkuoEC6wkSdIJqqgI+cuifAJgUm5LLurfirzWjeneMi3W0SSpTrHASpIknYCZK7dx/5QPWbShgBGdM5mU25L4aMTyKknVwAIrSZJ0HJZv3s19U5YybdlWWjZK4oGv5nJeXnasY0lSnWaBlSRJOg7rdxUxb+1O7jqrK5cPziEp3iVxJKm6WWAlSZKq4MCdhROiEW48ozOjOmfy9h1jSEuKj3U0Sao3LLCSJElHUVJWztPvfsxD/1hOQVEplwxoTRiGBEFgeZWkU8wCK0mS9AXeXbWdW19YyLodRQzvlMHtE7rSM7tRrGNJUr1lgZUkSfqckrJyEuOipKck0Dg5gfuu6sWIzpmxjiVJ9Z4FVpIk6RMfbirkJ3/9kNTEOH7xtb50btaQV64bShAEsY4mScICK0mSRH5BEQ/87SNemLeeholxXDem48HrXC2vklRzWGAlSVK99saSzVz7h3mEIVw9rB3Xju5I4wYJsY4lSToCC6wkSap3SsrK2b5nPy0bJ5PXpjHn5rXk+jGdaJ3eINbRJElHYYGVJEn1RkVFyJ/f38h/T11GRmoiL35nCBmpifzXV3JjHU2SVAUWWEmSVC/MXLGN//jrUhZvKKRbizRuOqOz17dKUi1jgZUkSXXeq+/nc+0f5pHdOJn/vTiXc3OziUQsr5JU21hgJUlSnbRxVxEbdhUxICedsd2y+NE5Pbh4QGuS4qOxjiZJOk4WWEmSVKcUFJXy8LSVPPHOalo2TubvN40kKT7KFUNyYh1NknSCLLCSJKlOKCkr53ez1vKLN1dQUFTKeXnZ3HRGZ6cKS1IdYoGVJEl1woyPtvHjV5cyvFMGt0/oSs/sRrGOJEk6ySywkiSp1pq5YhsbdhVxUf/WnN4tixe+PZj+OemxjiVJqiYWWEmSVOss2VjIf039kGnLttK5WSoX9G1FNBJYXiWpjrPASpKkWmPZpt3852sf8o8Pt5CWFMddZ3Xl8sE5RL3OVZLqhWotsEEQTAB+DkSBx8Iw/MkRjvkq8G9ACCwMw/Br1ZlJkiTVLmEYsm9/OSmJcewvq2DBul3cfEZnLh+cQ6MG8bGOJ0k6haqtwAZBEAV+CZwBrAfeC4LglTAMlxxyTCfgTmBoGIY7gyDIqq48kiSpdikrr+DVRfk8PG0lPVo24n++mkuvVo2YdecYEuNcy1WS6qPqHIEdCKwIw3AVQBAEzwLnAksOOeZfgF+GYbgTIAzDLdWYR5Ik1QLFpeX8ce56Hp2xknU7iuiYlcrwThkH91teJan+qs4Cmw2sO+TxeuC0zx3TGSAIgneonGb8b2EYvlaNmSRJUg33szeW88j0leS1bsz3z+7O6d2auZarJAmo3gJ7pH9pwiO8fidgFNAKeCsIgp5hGO76zImC4BrgGoA2bdqc/KSSJClmtu4u4fF3VjO6SxYD26VzxZC2jOicweD2TQkCi6sk6VPVWWDXA60PedwK2HiEY94Nw7AUWB0EwTIqC+17hx4UhuGjwKMA/fv3/3wJliRJtdC6Hft4dMYqnp+zjv3lFTRMimNgu3RaNEqmRaPkWMeTJNVA1Vlg3wM6BUHQDtgAXAJ8/g7DLwGTgSeDIMigckrxqmrMJEmSaoB7/7yEp2atIRLAhX1bcc2I9rTPTI11LElSDVdtBTYMw7IgCK4DplJ5fevjYRh+EATBvcCcMAxf+WTfuCAIlgDlwK1hGG6vrkySJCl25n28k97ZjYiLRmjVJJmrhubwzWHtad4oKdbRJEm1RBCGtWtGbv/+/cM5c+bEOoYkSaqCMAyZtmwr/2/aCt5bs5OHJvdhUm7LWMeSJNVgQRDMDcOw/5H2VecUYkmSVE+VV4QH13Bdml9Iy0ZJ/HBSd8Z2c8l3SdLxs8BKkqSTJgxDgiAgAH72xkdEgoCfXpTLObktSYiLxDqeJKmWs8BKkqQTtru4lKff/Zj/m7eel64dSkpiHE9/8zSapyW5hqsk6aSxwEqSpOO2bU8JT7yzmt/OWsvu4jKGd8pgx979pCTG0bKxS+FIkk4uC6wkSTou63fuY+z/TGd/eQVn9mzOt0d2oHerxrGOJUmqwyywkiSpypZt2s3C9bv4av/WtGrSgO+d3plxPZrRwTVcJUmngAVWkiR9qblrd/LwtBW8sXQLjRvEM6l3S5ITovzrqA6xjiZJqkcssJIk6QstzS/kh698wOzVO2jSIJ4bT+/M5YPbkpwQjXU0SVI9ZIGVJEmfUVZeQWFxGekpCSTHR9mws4gfTOzOJQNb0yDBXx0kSbHjv0KSJAmA4tJy/jRvPY/OWEWnrIY8dkV/cjJSeOu20S6FI0mqESywkiTVc7uLS/nDPz/msbdXs3V3Cb1bNeIr/Vod3G95lSTVFBZYSZLquSfeWcMDr3/EsI4Z/OziPIZ0aEoQWFolSTWPBVaSpHpm254SHp2xikHt0xnTtRlfH9SWkZ0zyW3tGq6SpJrNAitJUj1xoLj+btZaSsrKaZAQZUzXZjRJSaBJSkKs40mS9KUssJIk1QOPzljJ/76+nJKycs7Ny+a6MR3pkJka61iSJB0TC6wkSXXUtj0lpCbGkRQfpVFyPBN6Nre4SpJqtUisA0iSpJNr6+4S7nt1CcP+8x88P2cdABcPaMP/XpxneZUk1WqOwEqSVEds3V3CozNW8rt317K/rILz8rIZ1jEj1rEkSTppLLCSJNUR1z8zj9mrd3DeJ9e4tne0VZJUx1hgJUmqpbbuLuGxt1ZxzYj2NE1N5J6zu9MgIWpxlSTVWRZYSZJqmS27i3l0+iqe/mflVOFerRoxsXdLemY3inU0SZKqlQVWkqRaoqIi5P6/Lj14jev5fVpx3ZiOtMtIiXU0SZJOCQusJEk13J6SMlIT44hEAjbuKubsXi0trpKkeskCK0lSDbVldzG/mr6KZ2Z/zCvXDaNjVioPTe5DJBLEOpokSTFhgZUkqYY5UFyffnctZRUh5+Vlk5wQBbC8SpLqNQusJEk1SNH+csb97wx2F5dxXl4214/pSI5ThSVJAiywkiTF3Jbdxbz6fj5XDskhOSHKj87pQW6rxhZXSZI+xwIrSVKMfH6q8PBOmXTMSuXcvOxYR5MkqUaywEqSdIoVFJXy4N+XHyyu5/fJ5rrRThWWJOnLWGAlSTpFyitCopGA+GjAX97fyKTclhZXSZKOgQVWkqRqtqWwmIenr2TWyu385fphNEiIY9otow/eWViSJFWNBVaSpGpyoLj+4Z8fU1YRckGfbPbuL6dRcsTyKknScbDASpJUDRZvKODCh2ceLK7XjelI26ZOFZYk6URYYCVJOkm2FBbz4abdjOicSbcWaVw9vB0X929Dm6YNYh1NkqQ6wQIrSdIJOnSqcMOkOGbeMZaEuAi3ju8a62iSJNUpFlhJko5TQVEpD/19Ob/7ZDmcC/tmc93oTiTERWIdTZKkOskCK0nScVq2aTePv7OaC/q24oYxnZwqLElSNbPASpJ0DP65ajuLNxbyzWHtGNgunem3jqZ1usVVkqRTwQIrSVIVrN+5j/v/+iGvvp9Pm/QGXHpaG5Lio5ZXSZJOIQusJElHUbS/nIenr+RX01cSBHDj6Z25ZkR7kuJdx1WSpFPNAitJ0lFs2V3MI9NWMr5nc+48systGyfHOpIkSfWWBVaSpM9ZvKGAqR9s4uZxXWjbNIVpt46yuEqSVANYYCVJ+sS2PSX8dOoynpuzjvQGCXx9cFuyGiZZXiVJqiEssJKkem9/WQVPzVzDg39fTlFpOVcPa8f1YzuRlhQf62iSJOkQFlhJUr134EZN/XOacM/E7nTITI11JEmSdARfWmCDIEgPw3DHqQgjSdKpsmLLHn47aw0/mNidRg3imXLDcJo3Sop1LEmSdBSRKhzzzyAI/hgEwVlBEATVnkiSpGpUUFTKvX9ewoSfzeDFeRtYvmUPgOVVkqRaoCpTiDsDpwNXAQ8FQfAc8GQYhh9VazJJkk6i8oqQ595bx0//toyd+/ZzyYDW3DyuCxmpibGOJkmSquhLC2wYhiHwOvB6EASjgaeB7wRBsBC4IwzDWdWcUZKkE1YRhjz+zmo6Zqbyg0nd6ZndKNaRJEnSMarKNbBNgcuArwObgeuBV4A84I9Au+oMKEnS8Vq/cx+/fHMFd53VjYZJ8TzzL4PISE3AK2IkSaqdqjKFeBbwO+C8MAzXH7J9ThAEj1RPLEmSjt+Buwr/avpKggDO6tWC4Z0yyWzodGFJkmqzqhTYLmEYhkEQpAVB0DAMw90HdoRh+J/VmE2SpGMShiF/fj+f+6csJb+gmIm9W3DnWd3Ibpwc62iSJOkkqEqB7RcEwRNAQyAIgmAXcFUYhnOrN5okScfuj3PWkZ6SwM8v6cPAdumxjiNJkk6iqhTYx4HvhGH4FkAQBMOAJ4De1RlMkqSq2LanhAde/4h/HdmB1ukNePCSPqQlxxONeJ2rJEm1ShhCWclRD6lKgd19oLxWnjN8OwiC3Ud7giRJ1W1/WQVPzVzDg39fTlFpOQNymtA6vQFNUhJiHU2SpPonDKF0HxQXQnkJNMmp3L5qGuxcU7m9uABKCqFBBoy6vXL/81fAxvmV24sLoXHro75MVQrs7CAIfgU8A4TAxcC0IAj6VuYM5x3XGzxOW3YfvZFLkuq+N5dt4d//vIRV2/Yyuksm90zsTofM1FjHkiSpdgpD2L/30xK5fy+06le5b9U0yF9Yuf3A/iCA8z+5n+8rN8DSVyq3h+WV2xq3he+9X/n52/9beQ6AIAKJDaFln09fu0kOxCVCYhokpUFqc+BbXxi1KgU275OPP/zc9iFUFtoxVTjHSbN1dwnFpeUkxUdP5ctKkmqQN5ZshgCeuHIAo7tmxTqOJEknXxhWFsXyUti/B8rLoKK08nFFGTRqVVn89myBnWs/u6+iDNqNhPgkWDsTVvz90/J54ONlL0B8MrzxI3jn55+WTwAC+MEOiETggxdh7pOfls+kRpByyL+9LfMgmlBZPg+U0JTMT/ef+8vK8yWlQUJq5Xs61Bk/OsKb/+ICG4RheBxfzdhJbNEp/PMbbzGuR/NYR5EknSIFRaU89PflnNmrBf3aNmFPSRkJ0QgJcZFYR5Mk1UWlRZUlMLEh7N0O83/3SUE8pET2OA+y+8GOVfDW/xxeMId+F9oMgg1z4bW7KvdVlH163KQHoe1g+GgqvPSvhz//qqnQ5jSY/3t4+TuHZ/z2O9C8J/zzV/DX2w7ff8MCSG9XOQL693s/LZeJjSo/XvIHSG5c+frr/nnI/rTKktphDESin464Hql8VpMgCOaGYdj/SPu+dAQ2CIJGVI6+jvhk03Tg3jAMC05exKqLRgKmLMq3wEpSPVBeEfL8nHX8dOoyduzbT2bDRPq1bUJqYlUmEEmSdAzKSipHKj/4P1j2Vxh+Mwy/CfZtgzcOmYwaRCEaD1ndKgtsyW5Y8Q+IxkEkvnJfJL5yGu6B4+MSIJICkbhP9sdBQoPK/Q1bQPfzPt1+4GNai8r92f1g/P2f2x8PaS0r93eeAOntP7svGld5XoDB18PQ731x+ew8vvLPF0lKO/6vaTX40hHYIAj+BCwGnvpk09eB3DAML6jmbEfUvEOPMOOyB5hzz+lOI5akOuy9NTv44csfsCS/kIE56fxgUnd6ZjeKdSxJUl0ThvDK9bDkFSgpgOQm0O0c6Hs5tOoPFRVQVvxpgTxFo5D12QmNwAIdwjC88JDHPwqCYMHJiXbsGiXHs6ekjLeWb+OM7s1iFUOSVM3mrNnJrn37eWhyHyb2bkHgLwySpJOhvAxWT4dN78OwGysLaek+6DYJep5fee1oNP7T4yORT0dLFXNVKbBFQRAMC8PwbYAgCIYCRdUb64ulJsWR3CCeV9/faIGVpDqkaH85v5qxks7NGnJWrxZcNSyHK4fkkJzgbBtJ0gmqKIc1b1dOD176Z9i3vfI6zwFXV17n+pXHY51QVVSVAvtt4LefXAsLsBO4ovoiHV0AjOvejCmLNnk3YkmqA0rLK/jzwo38dOoyNhYUc9XQdpzVqwWJcf58lySdgIqKyjvrRuNhzuMw5RaIT4EuZ0KP86Hj6ZV36VWtctQCGwRBBOgShmFuEARpAGEYFp6SZEdxVq8WPD9nvdOIJamWe2XhRu6fspT8gmJ6tEzjZ5f0YWC79FjHkiTVVhUVsP69yqVflrwEY38AeV+rvKY1JRM6jXM6cC131AIbhmFFEATXAc/XhOJ6wNCOGTRKjmfKonwLrCTVMvkFRTRIiKNRcjwBkNM0hf84vxcjO2cSiXidqyTpOJSXVd4p+IOXoHA9RBOh0xnQuE3l/obNKpe9Ua1XlSnErwdBcAvwHLD3wMYwDHdUW6ovER+NML5HM/7qNGJJqjWW5hfy67dW8cqCjXzv9E5cN6YTE3u3YFJuy1hHkyTVNmEI+Qth20fQ+6uVy8Z8PKtyXdSxP6icJlzDln/RyVGVAnvVJx+vPWRbCLQ/+XGqzmnEklQ7vLNiG7+asYoZH22lQUKUrw9uy7l52QDeWViSVHVhCJs/qJwe/MH/wY5VkNiocg3VuAT45usQcWCrrqtKge0WhmHxoRuCIIj51c5OI5akmquiIjw4Hfixt1axNL+QW8d34bLT2tKoQfyXPFuSpEOEYeVSNzMfhNd/AEEE2o2Aod+tvLY1LqHyOMtrvVCVAjsT6FuFbadUfDTCuO7NeG3xJkrKyr1bpSTVALuLS3l29jqemrWGZ/5lEK3TG/CTC3vTuEG8P6clSVW3fSUs/r/KkdZx/155x+DOZ0JCCnQ7F1IzY51QMfKFBTYIguZANpAcBEEfKlewAUgDasStu87q3YI/zl3PWx9t43RHYSUpZvILinjynTX84Z8fs7ukjMHtm7J3fxkAzdJiPmlHklQblBbBuw9XltZNiyq3tR4EkU8qS2bnyj+q1442AjseuBJoBTxwyPbdwF3VmKnKhnb4dBqxBVaSYmNPSRlj/2c6JWUVnNWrBf8yvB29WzWOdSxJUm2wax3sXF05JTiaALMfhbRsGP8flde2NsqOdULVMF9YYMMwfAp4KgiCC8Mw/NMpzFRlCXFOI5akUy0MQ95ZsZ23V2zjjjO7kpoYx/0X9KJvmya0Tq8RE3QkSTVZ4UZY8nLlFOH1s6FhC7hxSeU1rNfO9u7BOqqqXAP7lyAIvgbkHHp8GIb3Vleoo9q/5zMPnUYsSadGaXkFf3l/I4/OWM3S/EIyGyZyzYj2pKckHLyrsCRJRzXjp/CPHwMhNOtVueRN9/MgEqncb3nVl6hKgX0ZKADmAiXVG6cKdq6FioqDf8mHdsggLSnOacSSVI0WrS/gmt/NIb+gmI5ZqfzXhb05t09LZ75IkioduFNwyW7YuACKdkDRTti7DVZNgwn3Q/Ne0HYojLoTel4AGZ1inVq1UFUKbKswDCdUe5KqKt8Pa96C9iOBT6YR92jOVKcRS9JJlV9QxJbCEnJbNyYnowHdWqRx3/k9GdU56+ASOZKkOqaiAvZtryyfRTtg3ydFtFkPaJkHe7bAlFs+2b7r02PG/xgGXA07VsNTEz97zqYdK4ssQNvBlX+k41SlZXSCIOgVhuGiak9TFZEoLPj9wQILcHbvFrwwdz1vL9/G2G6OwkrSiViaX8ivZ6zilYUb6dSsIVNuGEbDpHgev3JArKNJkqqqooL/z959x1dZnn8c/9zZIWQnhJWEkcEmyN5DFBAFVKoo7lWto9baaqu/ttqtttXWOmrV1ipVK25RxIUyZO8hOyEQRggBkpB57t8fd0ICMoLk5CTh+369zusk53nOyXVagXzPfT3XTcmB6gAa1BxadHIrpV/8vvLxymNFedBlIgy9B0oPwWMp3369YT9xAdb4w571EBoNUYnQqof7OqGbOy+2I1z7nnssNMbdB2k+gtSd2gTYIcB1xpituBZiA1hrbQ+vVnYiodGw9l244LEjPfJVbcQfrMxRgBUR+Y6WZO7n8U828NXGXJoF+XP1wGRuGNweY7TaKiLiU+UlrjU3LM59/82HcHAHFO2vDqFxqTD0x+7433pD3hawnurX6P49uPSfrs13/lNuUahZjAuZYfHua4DgCBj3aOWx6Or7sBbueFgs3LHwxLUGhbmJwiJeUpsAO87rVZyO0Fgo3wFr3oLe1wJqIxYR+a7KKjxUeCwhgf5k5RWyftchfjImnan9k4hqFuTr8kREzk7ffAjbF8Leb2DverfNTJs+cNMsd/zTh2HPWvd1UHMXQv1r/J3dbTLYiuoV0GYxEN2u+vj9WdVDk45lDPS/xStvS6QuGGvtqU8yZgiQaq190RgTDzS31m71enXH0adPH7v4Oj8IjYIbPz7y+Ofr93D9vxbx/LV9tAorInIKh4rLeHXhdl6Yu5VrB7Xj1uEdKavw4LFWHwKKiHhbSQHkflMdUPesdzttXD/DHZ82BTbNcteOxqdDfCd3DWqXie74/kwICHG/DwcE++59iHiJMWaJtbbP8Y6dcgXWGPNLoA+QDrwIBAIvA4PrssjTknElfPJLyN14ZHrZ4JTKNuJVaiMWETmRnAOH+dfcbUxbkMWhknIGdIihR5tIAAL9T/BpvIiIfDeH8yF3gwupe7+B0Q+BfwDM+j9Y/II7xz8IYlPd9alVO21M/Lu7VM4/8PivG51cf+9BpIGpTQvxxUAvYCmAtXanMSbcq1WdSs8prnVi+TQY/UugRhvxGrURi4icyP3TV/HVxr2M79Gam4e2p0fbKF+XJCLS+BXluZDasjsEh8PK12HWL+BQTvU5ASHQ/1Y3+KjX1dDxXLeyGt3OhdqawmLrtXyRxqQ2AbbUWmuNMRbAGBPm5ZpOLbwlpIyGFa/CqAfdRejA+O6aRiwiUsVay9xN+3h+zhZ+e3F3WkeF8sD4zoQGdiMxRhMhRUROW9UKae5GWPBs9cpq4R53/Nr33ACj8FbQYUR1+298J4hKOvI7K23OcTcROW21CbCvG2OeBaKMMTcDNwDPebesWsi4Ev53LWz+HFJHA2ojFhEBN5jpg5U5/OPLLazNOUh8eDDbcgtpHRVKWoJvG2hERBqFsmLY/nX1NapV92N+Dz0vh5KDsPI1F1DTzq8OqS0rN+loP9TdRKTOnTLAWmsfM8acBxzEXQf7C2vtLK9Xdirp49xUteUvHwmwQQF+nNelJR+vVRuxiJydissqOP8vX5KVV0Rqi+Y8MrkHEzNa6+9DEZFjWeu2oqkZUNsNhR6XuW1pXqocmBQSBS06Q+eLqq89bdXLTfLVNmMi9a42Q5zCgM+stbOMMelAujEm0Fpb5v3yTiIgGLpfBktedH/JhEYDML5HS6YvzWbuplxGddIqrIg0fQcOl/HFN3uYmNGGkEB/pvRLpHPLCIanxePnp1+uROQsV17ipvbmbYHAENfa66mAR1PcHqpVQmMgMsl9Hd4SrnnXrao2b/HtoHqiLWhExOtq00L8JTDUGBMNfAIsBi4HpnqzsFrpNRUWPgur3oB+NwMwJCWe8JAAPli5SwFWRJq0/KJSXpi7jRfnbuVQcTkZiVEkx4bxgxEpvi5NcPOJyAAAIABJREFURKR+lRa5vVJLCyGxn3vszVsgcx4cyAYqt41sP8wFWD9/GPADaBZd3f4bFlf9esZAh+H1/CZEpDZqE2CNtbbIGHMj8Ddr7SPGmGXeLqxWWvaAhG6w/JUjATYowI/zK9uIS8u7ExSgT8hEpGk5WFzGs7M38+95mRSUlDO2a0vuPDeF5Fjfz9gTEfGa4oNwcKfbbgZgzuOw8WO3slo17Tc2Be5c4r4OiYKkgRDTofoW27H69Yb/pH7rF5E6UasAa4wZiFtxvfE0nud9xkDGVJj5M9i9FhK6ANVtxHM27dUqrIg0GdZajDF4PJaX5mcyPD2eO0el0KllhK9LExGpG0V50CzGfb3uPVj3vguoeVugKBcCm8HPd7rfAQv3gvVAx1EQ074yoNboQLngEd+8BxHxqtoE0buBnwFvWWvXGGM6AJ97t6zT0OMytxn08ldgzG8BtRGLSNOy91AJz321heVZ+bz2/QFENQviq5+OJKpZkK9LExE5PbayldcY2LkM1s+oDqh5W6A4H3661YXYXath2xwXTjtdUL2K6qlw+6ZW/t4nImeX2kwhng3MBjDG+AG51tq7vF1YrYXFQdpYN8p89K/AP7ByGnECs9RGLCKN2O6DxTw7ewvTFmZSWu5hYkYbikorCAsOUHgVkYbvwA7Y/OnRATVvK1z/IbTqATkr4KvH3P6oMR2g+2R3byp/bxtxP4z8mW/fg4g0OLWZQjwNuBWoAJYAkcaYP1trH/V2cbWWMRXWvw8bZ7lP6IDx3Vvx5tIdzN2Uy8hOLXxcoIjI6VmSmccVzy2gwmO5uFcbbh+ZQvs4XeMqIg1ISQFkL3LBdP9WF07ztsD5v4aU0bBnHbx7J/gFQHQ7F06TB0NQ5d9lPS6HnldCwAk+kNMWNSJyHLVpIe5irT1ojJkKzADuwwXZhhNgU8+DsHjXRlwZYIekxrk24lU5CrAi0ijsyD9M1r4iBnaMpXubKK4ZkMzVA5M1nElEfKtgL+Qsh53L3X3Xi91q6YHt8J9J7pyAEIiuvA41INQ9ljwQfrgCItq6lt9jBYbW33sQkSajNgE20BgTCEwCnrTWlhljrJfrOj3+ge5TvAXPQGEuhMURHODPeV0S+HjNLkovVhuxiDRc2/OKeOqLzbyxZDstI0OYfe9IggL8ePDCLr4uTUTONoW5UFrgVkxLC+HJfnAwu/p4bIrbVxVcWL32fXcf3urbe6MGhVWvtoqI1JHaBNhngW3ACuBLY0wycNCbRX0nGVNh/pOw8nUY+ANAbcQi0rBtzyviyc82MX1pNn7GcHnfRG4bkYKfn9rmRKSebPnCtQHvXO6uST2wHTpdCFNeceGz0wXuGtXWvdz2hSE1pp4HBEP7oT4rXUTOTrUZ4vRX4K81Hso0xoz0XknfUUIX95fr8mlHAuyQ1DjCg9VGLCINS9V2OJv2FPDW8h1M7Z/ErSM60ipS7XQi4iVFeW7q785lbmV19C/d458+DDuWQExHSOwH/W5xe6dWuaDhXDEmIgK1G+IUCfwSGFb50GzgYeCAF+v6bjKmwox73SeIrXqqjVhEGpTNewt48rNNtIkK5d4x6YxIj2fufaOIDw/2dWki0pQczofQKPf1nMdh8fOQn1V9PKEbnPsLNyTpkuegWWz1+SIiDVxtWohfAFYDl1V+fzXwInCJt4r6zrpPhpk/h2WvQKueAIzv0Yo3l+1g7uZcRqZrFVZE6t/G3Yf422ebeG/lToID/LhlWEcAjDEKryJyZooPuBXUqgFLO5dDfibctw1CoyG4ObTpDX1vglYZ7vejmmE1tqPPShcR+S5qE2A7WmsvrfH9Q8aY5d4q6IyERkOn8bDqdTfCPSC4uo14ZY4CrIjUu+fnbOU3H6wlNNCf7w/ryE1D2xPXXKFVRL6Dw/urg2q3yRCVCKvfhPfvdsej27nLqfrcUP2cvje5m4hIE1GbAHvYGDPEWjsHwBgzGDjs3bLOQMZVsOYt2PARdJmoNmIRqXdrdx4kPCSAxJhmDOwQyw9GdOTGIR2ICTvBXociIsey1rX47tvsrlPNWQ77t1Ufj+ngAmzaWLj6bbey2izGZ+WKiNSX2gTYW4GXKq+FBdgPXOu9ks5Qx5FulPuyV6DLRAAu6K42YhHxvtU7DvDEpxuZtXY3l/VpyyOTe9KldQRdWkec+skicvY6nO/md+xcVt0G3PcmGHSH2yt151K3snrOte6+ZliNaOVuIiJniZMGWGOMH5Bure1pjIkAsNY2vC10avLzh55TYO4TcGgXhLdkaJprI56hNmIR8YKV2fk88clGPl2/h4iQAH40Oo3rBrfzdVki0tCUHYYD2bA/EwJDoN0Qt6fqoyngKXPnRCZB6wyIae++j2gNd6/yXc0iIg3MSQOstdZjjLkDeL3BB9eaMq6COX+Bla/B4B9WtxGv3c1vyz1qIxaROvXm0h0sydrPveencc2gdkSEBPq6JBHxhaqAmp8JHg+kne8ef/0ayPoaCnZXn9tuKFz3vttLdfxjEJnohiyFxfqmdhGRRqI2LcSzjDH3Aq8BhVUPWmvzvFbVmYpLgcT+ro140F1gjNqIRaTOLN6WxxOfbuS2ER0Z1DGOu0encu+YdJoH1+avVBFptGoG1OID0K1yxuX798D6948OqHFp1QE2MhFSIyAqGaKS3C0utfrc3tfV21sQEWnsavPbVtUou9trPGaBDnVfTh3KmArv3eVGy7ftozZiETljC7bs44lPNzJv8z5iw4LIKywFIKqZhjOJNAllxdUB9eAOOOca9/jnv4Ml/zo6oAaGQddL3KClqCRIPf/ogBqdXH3umN/W69sQEWnKThlgrbXt66OQOtf1YvjwPlj2MrTtQ3CAP6Mr24h/V+Eh0F9txCJSe7dPW8oHK3OIDw/mwfGdmdo/mdAgf1+XJSKno2ZAzc+EHpdDUBgsfA6+fAwKdh19fucJbs/UqORvB9SopOrzhtxdv+9DROQsdsIAa4y5CjDW2v8c8/jNQKG1dpq3izsjIRHQZYLbH23s7yEwlAu6t+KtZTuYuymXEVqFFZGTsNYyf8s++rWLIcDfj6EpcfRJjuaKfkmEBCq4ijRIRwXULEgfB+EtYc3b7kPtYwNq237Qsltli+/obwfU4MoJ4r2mupuIiPjcyVZgfwwMO87jrwGfAw07wIJrI175Gqz/ALpPZmhqZRvxqhwFWBE5Lmstszfs5a+fbmRpVj5PTMlgYkYbpvRLOvWTRaR+5G+HzHnQ5hx3LWnmPPjf9d8OqBGtXYCNbHv8gBpeuf1M+lh3ExGRBu9kAdbfWnvo2AettQeNMbUasWmMGQs8AfgD/7TW/uEE500G/gf0tdYurs1r10q7oW4c/bKXoftkQgJdG/HMNbv57cVqIxaRatZaPlm3hyc/28iK7AO0iQrlN5O6MbZbS1+XJiKlhbB6uguq2+bCgSz3+Pg/uwAb3vLkAbVtH3cTEZFG72QBNtAYE2atLaz5oDEmHDjlxBJjjD/wd+A8IBtYZIx511q79jivdxew4HSLPyU/P8i4AmY/4j6tjUpUG7GIHJe18MhH6ykur+D3l3Tn0nPaasstEV+wFvZ+A5lzICweukwE64H37obQaEgeBIPucPcturjnxHSAiX/3bd0iIlIvThZgnwfeMMbcZq3dBmCMaYcLpc/X4rX7AZustVsqn/sqMBFYe8x5vwYeAe49ncJrLeNKmP1HWPkqDPsJQ1PjaK42YpGzXlmFh3eW7+SVBZm8dEM/wkMCeeG6vrSKDCFA3Rki9W/xi7DpE8iaD0X73GOdJ7gAGxwOdy11K6zG+LZOERHxqRMGWGvtY8aYAmC2MaY5buucQuAP1tqna/HabYDtNb7PBvrXPMEY0wtItNa+X7nXbN2LbudaiZdPg6H3ujbizi34eO1ufqtpxCJnneKyCt5Yks0zszeTvf8wnVtFsPtgMeEhgSTGNPN1eSJNX0UZ5KyAzLlwcCeM+6N7fO07kLcF0sa61dXkQRBdYyOE6HY+KVdERBqWk26jY619BnimMsCa410TexLH+4jUHjlojB/wF+C6U76QMbcAtwAkJX2HQSoZV8Lbt7lPdZMHcUH3Vry9fCfzNu9jeFr86b+eiDRKeYWljHviS3YfLKFXUhQPT+zKyPQWGK3oiHjf2ndh8fOwfSGUFbnH4tKhvBQCgmDKNAjSh0giInJyp9wHFsBaW/AdXjsbSKzxfVtgZ43vw4FuwBeVvzy2BN41xkw4dpCTtfYfwD8A+vTpYzldXSbCjJ/A8lcgeRDD0uJdG/HKHAVYkSbuYHEZSzL3MzK9BTFhQUzq1YbhqfEM7Bir4CriDaWFLqRmznOrrJc8B5FtoGA3FOZCr6sgebBbYW1e41IehVcREamFWgXY72gRkGqMaQ/sAKYAV1YdtNYeAOKqvjfGfAHcW6dTiKsEhUHXSW4fuHGPEBIUxujOLZi5dhe/qeimNmKRJmh/YSkvzt3Ki/O2UVLm4eufn0tMWBA/G9fZ16WJNE05K+CDH8POZeApB+MHrXpCUa4LsH1vgn43+7pKERFp5LwWYK215caYO4CZuG10XrDWrjHGPAwstta+662ffVwZV7ntdNa+AxlXqo1YpInKKyzl2dmb+c/XmRSVVjCmawJ3jEwlJuyUw9NFpDaK8qpXVzPnQu/roM8NEBoDxh8G3eVWWBP7QUhE9fPU8SAiInXglAHWGLMYeBGYZq3dfzovbq2dAcw45rFfnODcEafz2qctaYAbs7/sFci4Um3EIk2MtRZjDAXF5bw4dxvjurfkByNSSG8Z7uvSRBq38hIICHbXqj43Enavdo8HhEDbvi64AkQlwo0zfVeniIicFWqzAjsFuB63j2tVmP3YWnv616L6kjFumNNnv4G8rYTEtOdctRGLNHqZ+wp5+ovN7C8q5dmr+5AU24z5PxtFbPNgX5cm0jjlb69eXd02103/vfpNN2gpsR90vRjaDYHWvVywFRERqUenTG3W2k3W2geANGAa8AKQZYx5yBgT4+0C61TPKwADK/4LwPjurcgvKmP+5n2+rUtETtvG3Ye4+9VljHzsC95ctoOWESFUeNznagqvIrVkLRzIrv7+f9fB493gre+7S27i0qDTBdXHL/wLDLvXdTUpvIqIiA/U6hpYY0wP3CrsBcB04BVgCPAZkOG16upaZFvoMAKW/xeG33+kjfiDlTkMUxuxSKPx3oqd3PnfZTQL8uemoR24aUh7WkSE+LoskYYvZyVs+gT2bYLcjZC7AUoL4L5MCG4OnSdA0kA3IbhFV/BTd5KIiDQstbkGdgmQDzwP3G+tLak8tMAYM9ibxXlFr6tg+o2w7StCOgxXG7FII7EkMw+Phb7tYhiWGs8Pz03lukHtiNZwJpFqJYdg95rqcJq7EfZthKn/c3Mgts2BTx+C5i0hLhW6XQItu1c/v9slvqtdRESkFk4aYI0xfsB0a+3vjnfcWtv4/qXrNB6CI92esB2Gc0H3VryzfCfzN+/TKqxIA2OtZf7mffzts03M37KPISlxvHxTfyKbBfKj89J8XZ6Ib1SUw/5tLphWhdR+N7stazZ/Bq9f487zD4KYjtCiC3gq3GO9roJeUyEk0mfli4iInImTBlhrrccYMxY4boBtlAJDofulro34gkcZnhZPWJA/M1apjVikIZm3OZdHZ37Dsqx8WoQH8+D4zlzZP8nXZYnUn8P7IXeTC6ktu7mAumsV/GMkeMqqzwtrAenj3PGkQXDl6251NSoZ/PyPfs2a29qIiIg0QrW5BnaWMeZe4DWgsOpBa22e16rytoypsPgFWPMWIb2vY3SXBGau2cWvJ6mNWMSXPB5LhbUE+vuRta+IPQdL+M2kbkzu3ZaQQP9Tv4BIY1NRDvmZblJ+TAcoPgjTLnerq4V7q88b9hMXUKOSYdCdLqDGpUFsCoRGVZ/XPB7SxtT/+xAREakn5lS74Rhjth7nYWut7eCdkk6uT58+dvHixWf2ItbC3/u7f/Rv/JiZa3bx/f8s4aUb+mkVVsQHyis8vL8yh79/vomp/ZO4bnB7yio8APpQSZqG8lK3DQ3A57+HPZXXqe7b7FZTM6bCpKfcv0//mQRRSRBbGVKrVlP9azV3UUREpNEzxiyx1vY53rFT/mtorW1f9yX5mDHuGqBZv4DcjQxP66A2YhEfKC338ObSbJ76YjNZeUWkJ4STGNMMUHCVRmzbHMhZUTlIqfI61Rad4dp33fF170FFqQunaWPcfavKgf7GwDXv+K52ERGRBq622+h0A7oAR/apsNa+5K2i6kWPy+GTh2D5K4SM/hXndlYbsUh9++Gry/hw9S56tI3kwfG9Gd05AT8/4+uyRL6trBgCK/8JzFoAu1dBwR4o2O3uAa5we4zz5aOw5QsIjXbhNPV8aNu7+rVunaPtaURERL6j2myj80tgBC7AzgDGAXOAxh1gw1tCymhY8SqM+j8u6N6Kd1fs5Ost+xiaqlVYEW8oKCnnla8zubR3W+KaB3PT0PZM6ZfEsNQ4jFFwlXpWUeauMy3YDQV7IeVcN/Ro9XRY++7RAbWiBB7c41ZIl70Ey14GDITFQfMEiG7n2n+NgYuegKBwCIs9/s9VeBUREfnOarMCOxnoCSyz1l5vjEkA/undsupJr6nw+kzY/Dkj0kceaSNWgBWpWweKyvjXvG28MHcrBw6XEREayBX9kuidHOPr0qQpKimA/Kzq8FmwGwr3wJB7oFkMLPonfP47KNp39PPu3eSGIOVnub1Umye4wUnNE6B5C/CUg38gjPo/d2sWd/zrUqPb1cvbFBERORvVJsAertxOp9wYEwHsAXwywKnOpY2D0BhY/jIhqaM5t3MCH63excMT1UYsUhestTz28Tf8e14mBSXljO6cwB2jUshIjDr1k0VqKikAvwDXxrs/EzZ8VBlQq0LqHpj4JLTsDmvegnfvOPr5ASHQ80oXYGM6QJeJ1cG0eYK7Ve2NOuRH7nYi4S299z5FRETkpGoTYBcbY6KA54AlQAGw0KtV1ZeAIOj+PVjyIhTlqY1YpI7kFZYSExaEMYZt+4oYkR7P7SNT6NxKe1DKcZQdhuzFENkWYtq7vU8//dXRLbxlRXD5y9D5Iti3CT78KRh/CIuvDqFVU/XbD4PJL1QH0+YtIDjCtfcCdBzlbiIiItLo1GYK8Q8qv3zGGPMREGGtXendsupRr6mw8FlYPZ0RvW5QG7HId5RfVMqMVbt4e/kOFm/LY+bdw0hNCOevU3rhr8FMUlN5KWydDZlzIXM+7FjitpIZ9wj0/z5g3fTe5i2gbd/qEBrfyT0/ebBr920W465ZPVZ0sruJiIhIk1PbKcRtgOSq840xw6y1X3qzsHrTqickdHfTiPvdzKjOCcxcs5tfT/QQoDZikVPanlfEw++v5Ytv9lBWYekQH8bdo9OICA0EUHgVKMyFzHmujTftfBdW/zvFHWvdCwb+wIXSxH7usbhUuH3BiV8vMKR6IrCIiIicVWozhfiPwOXAWqCi8mELNI0AC24V9qP7YfdaxndvyXsrdvL1ljyGpMb5ujKRBqe8wsP8LfvwM4bBKXFEhAayftdBrh3Yjkm92tC1dYQmCgusfQc2f+aCa+4G91j74S7ABoXBjR+7FdWgMN/WKSIiIo1KbVZgJwHp1toSbxfjM92/Bx8/CMtfYcSoh2kW5M8Hq3YqwIpUstayMvsAby/fwXsrcsgtKGFwSiyDU+KIDA3ky5+MVGg9W1kL+za7duD9W2H0r9zjS1+C7QshaQBkXOlWWFtlVD+vTe/jvZqIiIjISdUmwG4BAoGmG2DD4iBtLKx8jZDRv+JctRGLHOXu15bzzvKdBPn7MbJTPJMy2jCyU4sjxxVez0KbPnEhNXOe20sVIKwFDL8PAkPh4mchNPr416iKiIiIfEe1CbBFwHJjzKfUCLHW2ru8VpUv9LoK1r8PG2cxvnsvtRHLWWvPoWLeX5HDB6ty+Oc1fYgOC2JCz9YM7BDLuG6tiGwW6OsSpT5VlMHO5ZUDl+bB+McgKgnytsKOpW6ab/IgSBrkrl2t+jAjTH93ioiISN2rTYB9t/LWtKWMdtsxLH+FEZeOqWwjzlGAlbNCUWk5M1bt4p3lO5i7KRePhS6tIsg5UEx0WBDndk7wdYlS3/asgw/vg+xFbgsbgNhUOLTbBdg+N0C/m31bo4iIiJx1arONzr/roxCf8w+EHpfDgmcIKd3PqE4tmLlmF7+e2FVtxNIklZZ7yD9cSovwEPYeKuHe/60gMSaUH4xIYWJGa1ITwn1dotSH4gOQtcCtsGbNd38P9r0RQiLhcB6cc03lCutAt5VNFbUGi4iIiA+cMMAaY1631l5mjFmFmzp8FGttD69W5gu9roL5T8LK17mwx8W8vzKHBVvzGJyiVVhpGjwey6Jteby9fCczVuUwsEMsz1zdm+TYMD64awhdWmmCcJNXXgoBQVBRDs+Pdu3BWPALdFvaBFd+cBHRGm6d49NSRURERI51shXYH1beX1gfhTQILTpD63NcG/GNt9AsyJ/3V+YowEqT8MKcrfzzqy3sPFBMaKA/Y7omcMk5bY8c79o60ofVidfkZ0Hm/OprWCPbwDXvgH+AC6xpY90Ka5s+ENTM19WKiIiInNQJA6y1NqfyPrPqMWNMHLDPWvutFdkmI+NKmHEvIbmr1UYsjdqO/MN8sHIn1w5qR3CAP4Ul5aS3DOe+cZ04r0sCzYJqcwm8NCrWusAaney+f+MGWD3dfR0cCckD3dClKhf+pf5rFBERETkDJ2shHgD8AcgDfg38B4gD/Iwx11hrP6qfEutZ98kw8wFYPo3x3X+kNmJpVPKLSvlgVQ7vLNvJwm15AHRvE8XAjrHceW6qj6uTM1JeCody4NAuaJ0BAcGwYSasfN09fnCnO1ZRAvdnuVbgzhMgcYALri266LpVERERafROtgTzJPBzIBL4DBhnrf3aGNMJ+C/QNANsaDR0Gg+rXmfEiF8RGuimESvASkO3ac8hxj3xFWUVlo7xYfz4vDQmZrQhKVZtoQ2atXB4f2UIzYFDOyFtHDSPh28+hM9/6x4vyq1+zu2LID4NDmyHHYshvDW0OQfCW0FcGlB5HXPXST55SyIiIiLecrIAG2Ct/RjAGPOwtfZrAGvt+iY/5CVjKqx5k9CtH3Nu5yRmrt7FwxPURiwNR3mFh3mb9/H28h20CA/h/nGd6BjfnNtHpjC6cwJdW2sYU4NQUV65alq1Qlp53+NyaNkNNn0Cr06F8uKjn3f129B8JAQ2c6G09TluqFJ4K3cf0cqd1/cmdxMRERE5S5wswHpqfH34mGNN9xpYgI4j3YrG8mmM7/GE2oilwVi94wBvLMnm/ZU55BaUEB4SwJX9kgAwxnD36DQfV3iWKTvshiMdPCaknnON6+TYtQKeG3X0c/yD3fCklt0gur0LoDXDadU9QIfh7iYiIiIiwMkDbE9jzEFcL1po5ddUfh/i9cp8yc8fek6BuY8zYpxHbcTiU9tyC0mObYYxhlcWZDJ9yQ5GdWrBpF6tGZHegpBAXddYpzwVUFbkriEtL4XFz9dYPa1s8T3nGhjyIyg+CC9fWv3cZrHuw6/SIvd9bApc9Nejg2loNFStjsd2hDG/rf/3KCIiItJImcY2UPic3r3t0iVLvP+DcjfBk71h9EPcnjWMrzfvY8HPz1UbsXhdflEpK7IPsGJ7Pp+s283K7AO8fftgMhKj2H2wmJBAfyJDA31dZtNhLezbBFu+gM2fw7avoMtEmPgkeDzwmxZg/CC8ZXUQ7TIBul7sjm9f4Fp6w1u5wUoiIiIickaMMUustX2Od6zR7aPxza5D/OHD9dw9OtW7K09xKZDYH5a/wvhhl/HByhwWbs1jkFZhpQ4Vl1WwZucB4poHkxwbxtKs/Vzy1Lwjx7u1ieCBCzqTGB0KQEJE025+qDelhRAU5r5+/nzIXui+jkpywbTzBPe9nx/cu+HoVdOa/PzchF8RERERqReNLsBGhATyzOzNzFq7i0e/15NzkqK998MypsJ7dzEqfPuRNmIFWDkTpeUe3lyazYrsfFZsP8A3uw9R4bHcMTKFe8ekk54Qzk/HppPRNopubSOJCNFKa50oLYTM+bDlc7fSWrQP7lnnQmnPyyHjCugwEmLaf/u5zWLqvVwREREROb5G10Lcp08f++dpH3L/9JXsOljMTUM78NMx6d5p7S0+CI+lQc8p3H7oGrURS61Za8nKKzrSCpwQEcwtwzri8Vh6PvSxy02JUfRsG0XPxCh6JUUR11ztp3Wmoty1/fr5wfy/w6xfgqcM/IMgaQB0GAED71DLr4iIiEgD1KRaiAGGpcUz80fD+N2M9WzNLcTfz0vbhYREuGvdVr/JReN+qDZiOaGCknKaB7s/TvdPX8lHa3aRX1QGQHCAHxf1dFNl/fwMs+4ZTkJEsLa5qUs1r2Pd8gVs/QquedvtjdqyBwy4zU0XTxwAQdoXV0RERKSxanwBtrwYrCU8JJDfX9Kd8goPxhgy9xUybWEWPxqdVrfXxmZMhZWvMcouIjQwQm3EQkFJOauyD7AiO5+Vla3AZRUeFj4wGoCYsCDGdGlJz8QoerSNJL1lOIE1Vu1bRuo61jphrWsB3rMOXp4MB7Pd41FJ0HWS20MVoP1QdxMRERGRRq/xBdi9G+AfI+DcX0DHUUfaeT9bv4dnZ29h1prdPDK5B33a1dF1a+2GQlQSQaumMarTL5i5ZhcPT+zmvVVfaVBKyz2s33WQFdvzubxvEkEBfjw28xv+NW8bAEkxzeiVFEVGYhTlFR4C/P346dhOvi26qTr2Otb0C2DUAxCVDIl9of2PXWtwdPvjD1wSERERkUav8QXYyLZQlAcvX+LC5ehfQds+XD+4PWkJ4fz0jZV879n53DC4Pfeen05o0Bmuxvr5Qc8rYfYfuWTs//HBqlIWbN3HoI5ahW2q1u86yLRhFpqfAAAgAElEQVQFWazIPsC6nQcprfAA0Cspmm5tIrm8byLD0+Pp2TaKmLAgH1d7lph2OWz69OjrWGNT3LGgZvC9f/m0PBERERGpH40vwDaLgTvnwuIX4ctH4Z+j4a5lENOewSlxzPzRMP7w4Tqen7OVsOAA7jkv7cx/ZsYVMPsPDC2aRWhgBh+szFGAbeSsteQcKGZldj7Ltx9gZXY+d4xMYVBKHHsPlTB9STbd2kRy/eB2R1qB20S5rWw6t4qgc6sIH7+DJsha2Le5eoX18H64foY7FtMRBqS5FdakgbqOVUREROQs1SinEC9evNh9U3LIrcp0neS+X/qS+wU3Kon5m/fRo20kYcEBbMstJCEi5MxWY/91IRzI5vbY51mwLY8FPx+tNuJG5EBRGSXlFbSICCF7fxEXPzWPvYdKAAj0N3RqGcE956UxslOLI9dV6//ferToefjqz9XXsUYmuaFLF/4F/Ly437OIiIiINDhNbgrxEcHh1eG1cB98eB94yqHPjQwc+mMIDqDCY7n5pcWUVnh4dHJP+rX/jtfGZkyFt29lavedfLDaT23EjcDqHQd4bdF25mzKZWtuIdcMTObhid1oGRHC8LR4ureJpEfbSDq3ijhq8Je2SfKio65jnQ1TXoHoZPdnuc05MPQe9yFUTAddxyoiIiIi39K4V2CPdSAbZv8Rlr3sJpAOvB0G3cm87BLum76S7P2HuXZgO346Np1mQaeZ3UsL4bE0yjtNoNuyCUzu3ZbfTOp+5m9IvOKaFxby5Ya9BAf4MSwtnl5JUQzuGEfPxChfl3Z22rMOZvwEti+AitLq61jH/A5a6s+RiIiIiFQ72Qps0wqwVfZugM9/AxtnwZ1LIaIVhSXlPPLRev49P5OkmGa8clN/EmNO8zq6d26H1W/xo6TX+SrzsNqIGwhrLUuz8pm5Zhf3j+2En5/h+TlbCfAzTMpoQ2SzQF+X2PRZC4W5kJ/pbvszYccSSBsL51wNh3bDK5dCh5G6jlVERERETqrpthCfSHwaXPaS+6U5PAGsJez923io/VDGdR3Dywt30KpyL05rLaa2rYoZV8Gyl7k2aiVvrUlWG7GP7S8s5c1lO3htURYbdhcQFuTP5X0T6RjfnBuHtPd1eU1PUV5lQM1yATU/C+LSoP8tYD3w585uSnCVqCQXVsH9Obx1ji+qFhEREZEmpGkG2CrhCe6++ADkbYZVrzMgNpUBox4Ev17kFZZy078X8ZMxnRjYMfbUr5c0AGI60H3v+4QE3smMVZpG7CsrtufzvWfnU1ruoWdiFH+4pDsX9mxN8+Cm/Z+0VxUfqA6mVUE1NBpG3O+O/2O4e6xKSCR0m+y+9vOHi55wU8KjkiEq0V3XKiIiIiJSh86O3/ZDo+CmT2H9+/Dpr+F/10LrXuwf/lf2FZZyxXNfc83AZO4b24mwkwUgYyDjSvw/+w3f63gbH67ezUMTuqmNuB7sOVTM9CU7CA8J4KoByXRpHcH1g9sxKaONtrSprdLCbwdUTzmM+6M7Pm0KZM2rPj+oObQfXv39eQ+DX6BbWY1Kcn+uauo11fvvQURERETOak3zGtiT8VTAildh8Qtw7bscJoTHZyzhHwv20jY6lD9e2uPkq6oHsuEv3djQ6TbOXz6E/948oHart3LaKjyWLzfs5dVFWXy6bg/lHsuEnq356xW9fF1aw1R2uDKcZsH+be6+YDdc/Kz78GX6TbDqf9XnB4RCi85wy+fu+w0fQ1mRmwoclexWXzUJWERERETq2dl3DezJ+Pm7laLK1aLQijJ+lnUr309N5e6943lxbsTJA2xkW+g4kpSc9wgNHMKMVTkKsF7ywFureHXRdmLDgrhhSPsj17eetcpL3AcoVeG0aiX1wschJAK++APMfbz6fP8gt1JaVgRBYdD7ejdUKbqdezws/uiAmnZ+fb8jEREREZHTcvYF2GN5KqDH5cTM+xv/LvuY0maXQ34CmRUx7Nh/mEEpxwmzGVPxm34jtybt5D+rg/nVhK5qIz5DZRUePl23h9cWZfGLi7rSPi6My/smMjQ1nvO6JBAUcBbszVpRDgd3VE/xrbof+XOIaQ9LX4IZ91af7xfgPlAp2ucCbNeLIaFbdYtv8wTwq/G/W7vB9f+eRERERETqkAJsYAiMuA/63oSZ82eCFz4H66YzPenv/HVdGFP7J/GzCzofPRyo03gIjuRSv9n8paA1C7fmaRX2O9qWW8hri7fzxpJs9h4qISEimO15RbSPC6NXUjS9knxdYR2yFgr21Aio29x9n+uhTW/Y+DG8ekX1+cYPItpC4V4XYDuMhEnPuHAanQzhrVxHQZXWGe4mIiIiItJEKcBWCYuFMb+FAbfB0pe4bdDFFM3ayPJ5H3Hx+ix+Obk/Q1IrV2MDQ6H7pbRZ/l/iAi9WG/F3VFBSzpjHv6SswsOoTi2Y0jeJEenxBPg34tXWw/nfXkFNHwcp58KulfDssKPPD2sBqee5ANvmHLjor9XXoEa2Bf8ae9jGpbibiIiIiMhZ6uwb4nQ6Sosof6wTBaWWv5ZN5JxLf8yF51TuL5q9BP45ipfj7+Hx/YNZ8PNz1UZ8Cht2H+LVhdvZtq+QF67rC8BHq3eRkRhFy8p9eRu8qkFJRwLqNhc8u13q9kl95Jj9Z4MjYdQD0P/7UFIAy6dVB9SoJAhq5pO3ISIiIiLSUGmI03cV1IyAa98i/JOH+MXW/+D54nMwP6Ow02TC2pwDcelcUPEZDxb0YdG2PAZ00CrssYpKy3l/ZQ6vLsxiaVY+gf6G87u0pKS8guAAf8Z2a+nrEo9WUQ4Hs49eQY1KhN7XuRbgRzq4oUhVAkLA3OICbGg0jPmdWzmNSnZBNTS6+tzg5tD/lnp/SyIiIiIiTYUC7Km06Y3/te/Cli/w++QheOd2fjqrkPCOA/lV9ynEfP4QnQJ2MWNVjgJsDR6Pxc/P8NayHTzw1mo6xIfxwAWdueScNsQ2D/ZdYda6rWVqBtTAEBh0pzv+9EDI3VB9vvGHLhNdgDUGzv81BEdUB9SwFtWDkoyBgbfX+1sSERERETlbKMDWVocRcPNwSrbOo+03MTz35RZSQtdxo/Hj7rhFPLgqmV9edHZPIz5YXMY7y3fy6sIspvRN5OqB7ZjQszWpLcLp2y4aU197ino8bhV17wYXRksOwoj73bGXL4XNnx59ftu+1QF28N1gK6oDakSbo69D7XtT/bwHERERERH5FgXY02EMwR0G87MOMK5TNC1eugdjPZyb/wbPlHRi0bZeZ+Uq7JLMPP67cDsfrMzhcFkFnVtFHFllDQ8JpF/7GO/84PIS2LcZ9m2EzhPcCuinD8PXTx/d5ts8AYbf546fc7UbqlQVUKOS3FCuKpX7A4uIiIiISMOjIU5noPhwIRuev5keuR8AsCZqBF2vrZwi28QdLq0gNMht4XLp0/NYn3OQCRmtmdI3iR5tI+t2tfXwfghq7lZCN86CRf90K6v7t4H1uHN+vAHCE2D1dDdgKy4V4tMhLg3CjrOXr4iIiIiINEgnG+KkAHumykuxf0ojpzyCZmX7eLHri9xw4UgiQwLcil8T4vFYvt6yj/8u2s7n6/fw5U9HEhMWxNbcQlqEBxMWXAcL+vnbYf37LqBWtQAX7oFbvoDWvWDVG/DVn11AjUurDKmp0KLL0a2+IiIiIiLSKGkKsTcFBGF6XE7CoucZWPwXcpeW8drGL3mvxdPEJ6bDkHvcHrONTIXHUuGxBAX4UVRazotzt/H64u1k7isiIiSAyb3bUu5xq5/t48JO44XLXctvzYCa+w2MfBDSzoe8zfDR/RAS5cJp2vkQl+6GJQF0n+xuIiIiIiJy1lGArQsZV+K/4BnGBy5hd6er2bo7n8+2lfG97KcwS/6FGXQXDPwBBIfXa1nWWowxWGtZteMABw6XHbnlF5XRrU0kw9PiKSgp55aXFh91vKCknHtGp3HnuankF5Xx6Mxv6N8+hrtHpzKuWytCAv1P/sOLDxwdUJOHuDC6fys8NaD6vMhEt5JatXqaOADu3eTafpvYCraIiIiIiJwZBdi60KonJHTn2vx5fG/bhcz+yQie+rwtY2d/yZ8i3qP7F7+Dhf+A7892e4RmL4bMuW4P0YDgyvsQSBvjBgrlb3dtswGhR47vLPSQW9GcA8XlR0Jmi/AQzuuSAMA9ry9n14Hio0LoBd1a8cfJPQC45Kl5lHuObhe/blA7hqfFExzgR0m5h4SIENISwokMDSQitHr4Unx4MF/cO4J2x660WgsHd7qAGhgGSf3dYKXHe0DBrurz/ALdNaxp50N0e7jkORda41Ih6JjXDAxxNxERERERkWMowNaVXlNp99H9RJdsYlV2L+4dk86Yri1JiLgaDq6h+OOHyN59gKJDzQlf/CHtlz/2rZf4Sfs32VEaxoR9LzCl+LWjjrUGRhb/ixKCeCDgZa7w/wyPfzA0bw4BwfzwoOWe2KdIiAjh5oAZdA5cQXhec3gzGhMQzGc9Atk94AEiQwNpsWMWYUXZBAbthKVLCQwIYfroKEg9z/2wvRug/DAE5ML+QwQGhNAuGKAybH5wL+xYDLkbobTAPZY+HpKmucDd7VJo3qJ6iFJUMvhX/qfmHwA9LvPK/wUiIiIiItK0aYhTXSnch/1TOi+Wn09m75/z0MRuRx2+8rmvmbd5HwB+eAimlPTYQN7+fh8oL+bht5YyJz+G8GYhpAXspoPJITnSn/PToqC8mG927CWr41QimwXTZtdnROUuJphSAjwlbtXTemDyC+6Hff57+GaGe7y82N0HNYO7lrnjr051g5JqikqGu1e6r1+aCFu+OPp4q57w/S/d169dBSUFlUOU0tw1qvHpLrSKiIiIiIicAU0hri+vTuXgxjmMMc8y5+dj8Pervobzm12HWL3jAJGhgUQ2CyQiJJDoZoG0iPBBu2x5aXWwLT9cGYCtC6PgWpwLdtc4pxhCo6HrxfVfq4iIiIiInFU0hbi+9LqKiPXv07V0IYu39aV/h+rpw+ktw0lvWb9DnE4oIMjdTqTtcf9bERERERER8Sk/XxfQpKSchyesBZcHfMmMVTm+rkZERERERKRJUYCtS/4B+PW8nFF+S/l61Td4PI2rPVtERERERKQhU4CtaxlT8aeCwYc/Z3Hmfl9XIyIiIiIi0mQowNa1Fp2paNWLywJmM2PlTl9XIyIiIiIi0mQowHqB/zlX0clksXnVPLURi4iIiIiI1BEFWG/odikVfkGcWzxLbcQiIiIiIiJ1RAHWG0Kj8aSPZ6L/PGauyPR1NSIiIiIiIk2CAqyXBPa+imhTQNHq99VGLCIiIiIiUgcUYL2lw0gOhyRwXsknLMlSG7GIiIiIiMiZUoD1Fj9//HtdwXC/FXy5eKWvqxEREREREWn0FGC9KKj31fgbS9C6N9RGLCIiIiIicoYUYL0pLoV9MecwruxTlmTm+boaERERERGRRk0B1svC+l9Dit9OVnz9ia9LERERERERadQUYL0spOellJhgYjaqjVhERERERORMKMB6W0gEu9uMYXTFHJZtyfF1NSIiIiIiIo2WAmw9iBt6PRGmiG1zXvN1KSIiIiIiIo2WAmw9aJY6gtyAlnTd8jzvfTHf1+WIiIiIiIg0Sgqw9cHPj7AJj5Dsn8vIzyfx4Ut/xFPh8XVVIiIiIiIijYoCbD0J7TGRwDu+Znd4F8Zt+R3r/jSW4n3bfV2WiIiIiIhIo6EAW48CYtvR4Z5PmJt2Hx0Kl1H+5AAOLngZrKYTi4iIiIiInIoCbD0zfv4MvvLnLBr7Hhs9rYj48HYKXpoCBXt9XZqIiIiIiEiDpgDrI8MGDsDvhpk84Xc1QVs/oexv/WDN274uS0REREREpMFSgPWhnsmxTL7rUe5o/jjrD0fC/66FN26EojxflyYiIiIiItLgKMD6WJuoUP50xxT+lPR3/lQ2mYo1b2GfGggbZvq6NBERERERkQZFAbYBCA8J5J/XD2Rfn7uZUPxrdpaGwrTL4J3bofiAr8sTERERERFpEBRgG4gAfz9+O6kbk8aNY9ShX/FG6GXY5dPgqUGw+XNflyciIiIiIuJzCrANiDGGm4d14ImpA3iw4BK+H/R7Sv2C4T+T4IMfQ2mhr0sUERERERHxGQXYBmhst5a8dstAlnlSGJT/EDs63QCLnoenB0PmfF+XJyIiIiIi4hMKsA1Uz8Qo3r59MLGRkQxfeR6fDXgBrAdeHAczH4Cyw74uUUREREREpF4pwDZgbaJCeeO2gQzsGMsNXwTz59R/YXtfD/OfhGeHQfYSX5coIiIiIiJSbxRgG7jwkEBevK4vV/ZP4q9f5XDHwaspveINdz3s8+fBp7+G8lJflykiIiIiIuJ1CrCNQNWE4gcu6MyM1Tlc/mkz9l3zBfScAl89Bs+Ngl2rfF2miIiIiIiIVynANhJVE4qfntqbdTkHmfj8ajYO/CNM+S8U7IZ/jIQvH4WKcl+XKiIiIiIi4hUKsI1M1YTiknIPlzw9j7kB/eD2BdD5IvjsN66teO83vi5TRERERESkzinANkI9E6N46weDaB0ZyrUvLOS1NQXwvRdh8ouwfxs8MxTm/Q08Fb4uVUREREREpM4owDZSbaOb8b/KCcX3TV/FHz9aj6fLxW41NuVc+PhB+Nd4yNvi61JFRERERETqhAJsIxZRY0Lx019s5s7/LqM4OBamTINJz8DutfD0YFj4HHg8vi5XRERERETkjCjANnLHTii+4rmvyS0shYwr4AfzIWkAzLgX/jMJ8rf7ulwREREREZHvTAG2CTh2QvGkv89l4+5DENkGrnoTLnwcdiyBpwfB0v+Atb4uWURERERE5LQpwDYhVROKi8sqJxRvygVjoM/1cNtcaNkD3r0Dpl0Oh3b5ulwREREREZHTogDbxPRMjOLt22tMKF6U5Q5Et4Nr34Oxf4Cts+Hv/WHl/7QaKyIiIiIijYYCbBN03AnFHgt+fjDgNrh1DsSlwps3wevXQGGur0sWERERERE5JQXYJioiJJAXjp1QXFa5L2xcKtwwE0b/CjZ85FZj173ny3JFREREREROyasB1hgz1hjzjTFmkzHm/uMcv8cYs9YYs9IY86kxJtmb9ZxtAo83obigxB3084chP4JbZkNEa3jtKnjzFji837dFi4iIiIiInIDXAqwxxh/4OzAO6AJcYYzpcsxpy4A+1toewBvAI96q52x17ITii5+ay6Y9h6pPSOgCN38Gw++H1dPhqYGwcZbvChYRERERETkBb67A9gM2WWu3WGtLgVeBiTVPsNZ+bq0tqvz2a6CtF+s5q1VNKD5c6uHipyonFFfxD4SRP4ObPoGQKHhlMrx7JxQf9F3BIiIiIiIix/BmgG0DbK/xfXblYydyI/ChF+s56x07ofj1RduPPqF1L/j+bBh8Nyx7GZ4eDFu/9E2xIiIiIiIix/BmgDXHeey4e7YYY64C+gCPnuD4LcaYxcaYxXv37q3DEs8+NScU/3T6Sh6pmlBcJSAYznvIDXnyD4R/XwQzfgqlhb4rWkREREREBO8G2Gwgscb3bYGdx55kjBkNPABMsNaWHO+FrLX/sNb2sdb2iY+P90qxZ5OaE4qfOnZCcZXEfm67nf63wsJn4S/d4J3bYcPHUH7c/5tERERERES8ylh73EXRM39hYwKADcC5wA5gEXCltXZNjXN64YY3jbXWbqzN6/bp08cuXrzYCxWffay1/POrrfzuw3VkJEbx3DV9iGse/O0TM+fD4udhw0woOQjBEZA2BjpfBCmjISis/osXEREREZEmyRizxFrb57jHvBVgK3/wBcDjgD/wgrX2t8aYh4HF1tp3jTGfAN2BnMqnZFlrJ5zsNRVg695Hq3dx92vLiA8P5sXr+pLSIvz4J5aXwJbZsO5dWP8BHM6DgFBIORe6THShNiSyfosXEREREZEmxWcB1hsUYL1jxfZ8bvz3YkrKK3jmqt4MTok7+RMqyiFrHqx9F9a/D4dywC8QOoxwK7OdxkPYKV5DRERERETkGAqwUivZ+4u48V+L2by3gN9d3J3L+iae+kkAHg/sWOxWZte+C/mZYPwgeTB0ngCdL4SI1t4tXkREREREmgQFWKm1g8Vl3P7KUr7amMsPRnTk3vPT8fM73kDpE7AWdq2Cde+5QLt3vXu8bV+3Mtv5Iojp4J3iRURERESk0VOAldNSVuHhl++uYdqCLMZ3b8WfLutJSKD/d3uxvRtckF33LuSscI8ldHdBtssEiO8E5jQCsoiIiIiINGkKsHLarLU899UWfv/h+pNPKD4d+zPd9bJr34XtCwALsanVK7OteynMioiIyP+3d6dBcpz3fcd/T8+9972LxQ0SB08IFEVKlEhRF03JomTHh6hSbNmW5MiJY/uFEztJVeJS8sJO4iR27EpKBx3Lli3ZcqQQsk7aEklRoniDIgkCIAEQWGBPYI/ZY85+8qJ7ZnpmZ3ax4O7Ozu73UzU1PU8/3dPD5iz2t/+nnwawxRFgcdW++cKwfutLzykWDukX3rxbv3jHbvW1xl//jpMjXpg9flQ686hk81L7Tj/MfsC7D61zlVVfAAAAAA2LAIvX5fjwjP7ooVP61ksjijiOfurIoD5x5z7t769xu52Vmr8snfiGN8z41X+U8hmppd+byfi6+6Q9d0qhyOq8FwAAAIANjQCLVXF2Yk6f+/4Z/e3T55XKunrHwV594s59ess13TKrNfQ3NSOd+rZXmT31HSk7J8U7pIPv866Z3fcOKbIKFWAAAAAAGxIBFqtqci6jv3z8Nf35D89qYjajGwbb9Kt37dP7btqmSMhZvTfKLngV2Zce9Cq06Wkp2iLtv8erzO6/R4q1rN77AQAAAKg7AizWRCqb11efvaDPPHpar47PabA9rl9+617df9tOtcZXechvLiOdfcSrzL7899LcuBSKSde+y7tm9uC9UqJzdd8TAAAAwLojwGJNua7V906O6dOPnNbjpy+rNRbWh2/fpV+6Y48GOxJr8IZ56dzj/u15jkozFyQnLO29y6vMHnq/1NK3+u8LAAAAYM0RYLFunh+a0mcePaOv/3hYRtL7b96mj9+5Tzdub1+bN7RWuvBM6V6zl09LMtKut3jXzB56v9Sxc23eGwAAAMCqI8Bi3Q1NzuvPHjurLz5xTnOZvN56bbc+fuc+3X2gd/UmfKpkrTT2knfN7PGj0tiLXnvfDd5teXbeJu24Teq+hvvNAgAAABsUARZ1M72Q1RefOKc/e+ysRmZSOtDfoo/fuU8ffMOgYuE1vs/rxCvSy0elM49IQ09J6RmvPdEl7XiT99j5Jmn7G6XYKt0SCAAAAMDrQoBF3WVyrr72/EV9+pHTenkkqd7WmH7pjj36yO271NEUXfsDcF1p4oR0/glp6Anp/JPea0kyjtR3vR9oqdICAAAA9USAxYZhrdVjr1zSpx89rUdOjisRCelDb9qpX3nrXu3qblrfg1mYlIae9gPtE9KFpxdXaXe+yQu029/ILXsAAACAdUCAxYZ0fHhGn330jB48dkF51+reGwf0iTv36ciuOt0Ox3Wl8Ze9QDv0ZJUq7Q3Sjlup0gIAAABriACLDW10JqX/84Oz+svHX1MyldOb9nTqE3fu07uv65fj1DkgUqUFAAAA1hUBFg1hNp3T3zx5Xg88dkZDkwva29Osj71tr37mlh1KRNd4wqcr5eal8ROl62iHnpAmTnrrClXaQqDdeZvUtY8qLQAAALACBFg0lFze1TdfHNFnHjmtY0PT6mqO6p++ebd+8S271dMSq/fhLTZ/2avMnveHHgertE3dgRmPb5MGb6FKCwAAACyBAIuGZK3VE2cu6zOPntFDx0cVDTv6mVt26ON37tU1vRs4BFKlBQAAAK4aARYN75WxWX3u+2f0d88MKZNz9e7r+vSJO/fptr1dMo0Q/sqqtE9419Vmkt46qrQAAABAEQEWm8bEbFqf/+Fr+osfntXkfFaHd7Tr43fu03tvHFA45NT78K6cm/dmPC4MOx56srxK23+DV6EdPOLdo7b3IKEWAAAAWwIBFpvOQiavv3tmSJ/7/hmdmZjT9o6EPva2vfr5N+1USyxc78O7OktVaSWpY5fUe53UF3j0HJAiifodMwAAALDKCLDYtPKu1UPHR/XZR0/rybOTaouH9ZE379Yv3bFH/W3xeh/e6+Pmpcmz0thxafy49zz2slepdbNeH+NInXv8Ku2hUrDt3i+Fo/U8egAAAOCqEGCxJTxzblKfffS0vvnCiEKO0QcOb9cn7tqrQwNt9T601ZXPSpdPS2MveYF27CVvOPKlVyWb9/o4YanrmlKg7T3khdyufVKoQSvUAAAA2BIIsNhSzl2a1wOPndGXnjyvhWxet+/t0gffsF3vvXFAnc2buCqZS0sTpwIVWz/cTp6V5H/PQ1Fv2HGwWtt3ndSxR3Ia6BpiAAAAbFoEWGxJU/MZfeFH5/Tlp4d0ZmJOYcfozv09uu/woN5zfb9a45F6H+L6yMxLEyfKq7Vjx6Xp86U+4YQ3UVSwWtt3SGrfyS1+AAAAsK4IsNjSrLV68eKMjh67qK89P6wLUwuKhh2982Cf7js8qHce6lMiGqr3Ya6/1Ix3v9ri9bXHvXCbHC71ibaWgm1xOPJ1UusAwRYAAABrggAL+FzX6tnzkzp6bFhfe35YE7NpNUVDes/1/brv5kHdeaBHsfAWDLNBC5OLq7VjL0nzl0p94u2BiaP8am3f9VJzT/2OGwAAAJsCARaoIu9a/ej0JR19/qK+8cKIpuazaouHde+NA7rv8KDesq+7se4tu9Zmx8urtYVrbVPTpT5NPYGZkK+V2rZL7dulth1euKVqCwAAgGUQYIFlZHKuHntlQkePXdS3XxrVbDqnnpao3nfTNt13eFBv3NUpxyF8LWKtlBwJVGv9mZHHX5Yys+V9QzGpbVBq3xEIttu914W2eDshFwAAYIsjwLNXxycAACAASURBVAIrkMrm9b0TYzp6bFgPHR9VOudqW3tc77/ZC7M3bW+XIWQtzVppblyaHpJmLkjTF6SZIf/Zf50cLt32pyDasjjcBqu47dulaHN9PhMAAADWBQEWuEqz6Zz+4fiojh67qIdPjiubt9rd3aT7bh7UfYcHdXCgtd6H2LjyOWl21A+5wXAbCL1zY4u3i3cEgu2O8nDbtt2r8oZj6/95AAAAsCoIsMAqmJrP6FsvjujosWH94NUJuVY60N9SDLN7eqgMrrpcWpq5WLuKOzPkTTpVqblv6Spuy4AUCq//5wEAAMCyCLDAKhtPpvWNF4Z19NhFPXnWC1A3bW/XBw4P6idv3qbBjkSdj3ALycx5IXf6fPUq7syFxdfjmpB3K6Bq4bbNfzT3SM4Wn5EaAACgDgiwwBq6OLWgv39+WEefv6jnh7wZed+0p1P3HR7Ue2/cpt5WhrPWlbXeTMlVq7iBoJtPl29XCLmtA1LrNv/hL7cFXsc7mHgKAABgFRFggXVydmJOX3v+oh48dlEnR2flGOmOa3p03+FtuveGbWpvitT7EFGNtd59bqeHvEdy2H+MeNXd5Ij3OjW1eNtwIhByB7xrcKuF3mjT+n8uAACABkSABergxEiyGGZfuzSvSMjorv29uu/woN59fb9aYlyD2XCyC6VgmxyWZobLXxfacguLt423BwLtYPXQ29IvhfgjBwAA2NoIsEAdWWv14wvTOnrsor72/LCGp1OKhR2967o+3XfzoN5xqE/xCNdabhrWSumZinB7sSL0jkizI5Kbq9jYSM291YcqB0NvU7fkOHX5eAAAAGuNAAtsEK5r9fS5SR09dlFf//GwJmYzao6GdM8NA7rv8Da97dpeRcMEky3Bdb1hy4vCbUXonRtfvK0TCVyfW6Wi29LvPRKdBF0AANBwCLDABpTLu3r89GUdPXZR33hhWDOpnDqaInrvjQO67+ZB3b6vWyGHyYG2vFzGux/uchXd9PTibZ2wd0uhVj/QtvSVwm3Zcj/X6AIAgA2DAAtscJmcq0dPjevosYv69kujms/k1dMS00/c0K+7D/bpjmu61cw1s1hKZs4fmjzqP495y8Vn/zE3Lll38fbR1kCo7fMruVUCb3MvtxcCAABrigALNJCFTF7fPTGmB5+7qEdOjWs+k1ckZHTr7i7dfbBXbz/Yq4P9rTLcugVXw817Q5eLoXasRuAdq17VNY7U1FNRxa0ReGNt3GIIAACsGAEWaFDpXF5Pn53UwyfH9fDJcb08kpQk9bfF9PYDvXr7gT697doebs+DtZGZ94YvB6u4ydHFQXd2VHKzi7cPx/1QW6OaWxja3NwnhaPr//kAAMCGRIAFNomR6ZQe8cPso6fGNZPKyTHSkV2dfqDt1U3b2+Vw7SzWk7XSwmR5VbdW4F24XH0fiU6vstvUJSW6/NddUqLDe93ktwXXRZqo8AIAsAkRYIFNKJd3dWxoSt874QXa54e84Z5dzVHdtb9Hbz/Yqzv396qnJVbnIwUCchnvOtzZimHLyRFvaPPCZWl+0gvEC5el7HztfYVigaDbWXosCsEVwTfMdwIAgI2MAAtsAROzaX3/1IQePjmuR06O69JcRpJ00/Z2rzp7sFdHdnYoHOK2Kmgg2ZQfZv1AO3+5tLwwGXg9Wb4un6m9z0izH2Y7S+G2WtANrot3SCEmUgMAYD0QYIEtxnWtXrw4o4dPjunhk+N65tyU8q5Vazyst13bUwy029oT9T5UYPVZ61Vu5y9XCbqXpYWpGusmJZuvvd9YuzekuVaFt9DeFHjNRFYAAKwYARbY4qYXsvrBKxPF4cYjMylJ0sH+Vr39oHft7K17OhULc3sUbGGuK2WSFeF2skrFtyL4pqZq79MJBwJudynwFsNutedOKcTEbACArYsAC6DIWquTo7PF6uyTZyaVybtKREK645ruYqDd3d1c70MFGoOb96q6xSHONZ4r25Ya5hxrKwXdpu7qIbcy/EabqfYCADYFAiyAmubSOT1++pIePjmu750Y17nL3qQ5e7qbikON37yvW01Rrv8DVo21UmauStidrBF+L3nr0jO19xmKLg66S1Z6/VmeHUZeAAA2FgIsgCt2dmJO3zvhVWd/ePqSUllX0bCj2/d2FW/Vc21fiwyVHmD95bO1Q27VEOwHXzdXY4dGirdJ8XbvEWsvLRcfbVXa2r0qcbydAAwAWHUEWABXJZXN68mzl/Wwf+3sqbFZSdJge7w41PiOa3vUFud6PWDDstar3BZD7WR56F2YklLTXp/UdPljqYpvQbT1ysNutQfX+wIAKhBgAayKC1MLeuTkuB4+Ma7HXplQMp1TyDF6467OYqC9flubHIfqLLApuHkpnawebCvbqj3SM5J1l36PSNMyQbeyraPUN9HBfX0BYBMiwAJYddm8q2dem9TDJ73q7IsXvUpNT0tMdx3o0Vuv6dEtuzu1p7uJ4cbAVmWtlJmtEm6DAXiJCnBqeonhz75I8zKzO3cunvwq1i453BMbADYqAiyANTeWTOnRkxN6+OS4Hj01rsn5rCSpqzmqIzs7dMvuTh3Z1aHDOzrUHGNCKABXwFopu1C9slu4hVHlkOjg0GjV+B3HOIvv3xu8r2+tSa8i8XX9+ACwVRFgAayrvGt1aiypZ16b0jPnJvXsuUm9Oj4nSXKMdHCgTbfs6tAtuzqp0gJYG67rBdylJr0qW+cH4ex87X1GmgLBt3OZGZ79IBzvoNoLACtEgAVQd1PzGT17fkrPvjapZ85N6bnzU5pNe0MDqdIC2DCyqeozOy91q6OFySWu9TXetbqVITfWJkWbvCHQ0SbvPr6F5UiTFG0JLDd7z5EmwjCALYEAC2DDqazSPnNuUqcDVdpDA226Zbdfpd3Vqd1UaQFsVK7rD2uuMstz1SA8KaWnpcy85GZX9l6FIBv1Q25heSVBuLIt2iyF4xI/YwFsEARYAA1haj6jZ8+VAu1z56Y0l8lLkrqbozqyq0NH/EB7eGe7mqJUaQE0uHxWysx5Q5czc4HleSk7t7gtM1ux3t+usBzsZ/MrOBATCLjLhONIXApFvVsghWKB5ag3K3RhuewR8dcF+oYq+lJdBuBbKsDy2x+ADaOjKap3HOrTOw71SfKqtCdHk16gfW1Kz56b1EPHxyRJIcfo0ECrfx2tV6nd1UWVFkCDCUX8IcYdq7tfa6V8ZpnwO1cehLN+n+Kyv+3cRPl+cqmVV46vhBOuHoxrhuJq6yv6hiv6h+NVKtLN5WE9xK/HwEZGBRZAQ5mcy+i581RpAaCuXNcLsfmMlMt4z/mMV1HOp0vLuXRFe7bUt2y7TJX2Qt/AdrmKfQTft2xdevl7ENcSilYZit28uCJds32J9eHo6p4HYJOiAgtg0+hspkoLAHXnOJIT86qfsXofTA1uviL0pr1JurLVKs+VQ7ertM9PSFMV7bnUyo7JiVxhCK5SIQ5Wmp2IVyledtl/FJad0Nr8twbWERVYAJvO5FxGz56fLE4Qdex8qUrb0xLVG3aWAu3NO6jSAgCukpsvv4Z5uRC8kvC81C2drpZxqgfbspAbGMpdc3mZfVQuh+NSJOE/mqo/hxNcB40iKrAAtpTO5qjeeahf7zzUL8mr0p4YSZYNO37o+Kgkr0p73bbW4mzHt+zq1M6uBFVaAMDynJAUb/Meq811S0E2M1c+rNrNXcFyVsrnai8XhoCXLWf9fQSWs/PV26st63UWxsJxP+xWhtxqwbdW23JBmRm3Gx0VWABb0lJV2vZERAf6W3SgvzXwaFF3y0YdJwcAwAZQGLZdDLaBa5cLQ66z81J2IfC8UKUtuK6yT2A5t3B1x7lUGA5XhmA/VIdj3rpwLPA6Xr4cqXhdeA5FCc0rRAUWACosVaV9aXhGJ0eSOnrsomZSueI23c3RYpjd39+qgwOtOtDXqvamSL0+BgAAG4cTkhw//K0H1/VDcUWovdqAnJ6VZscD6+e88L3Sa52rKYbdKgG3ZgCu7FMtQAeeI4nybWNtm3JW7c33iQDgKoQco+sH23T9YGkYmLVWY8m0TowkdXI0qVOjszoxmtSXnx4qVmslqa81poMDrdrf54XbAwOt2t/XotY4wRYAgDXjOP49ipskda/d+xRuS5VLlQJtNlX+uvgcfATWZReq9A1sMz+xuD2b8gL51c6oLSM190gtA1Jrf+DZf7QOlJ7X648Oq4AACwA1GGPU3xZXf1tcdx3oLbZba3VxOqWTfrA94Yfbv3riNaWypX9kBtvjOjDgDUHe39eigwOturavhUmjAABoJMb4lc06XUqUz11ZSC4L1ilpYVJKjkizo97z6IvS7Jhk84vfI9ZeCreFYFsZclv6pXh73YdD81sUAKyQMUbbOxLa3pEo3s5HklzX6vzkvE6OzurkaNJ/zOoHr15SJuf620o7OhM62N+q/f5w5AP9rbqmt0XxCLc3AAAAFUJhKdQixVpe/77cvDR/yQ+1o9LsSHnInR2Vhp701lW7xjgcD4Tb/orqbiDsNnWv2W2bmMQJANZYLu/qtcvzOuUHWq9im9Tp8TnlXO9nsGOk3d3NxUC7v79VB/tbtbenWdEwtxUAAADryFopPRMIuaNeuC0uB55T04u3NyGpubdKyA2G3T4v8FapbDOJEwDUUTjk6JreFl3T26J7byy1Z3Kuzl6a8yq1I164PTmW1HdeGpWfaxV2jPb2NPuhtqVYud3T3aRwiGALAADWgDHecOF4u9R7YOm+2QVvaHKwipsc8cLt7JiUHJaGn5Pmxqtfz5voXBxyl0CABYA6iYad4m16dHOpPZXN6/T4nE6NJf0JpGb1wsVpff2FYRUGzURDjvb1NhdnRS7sZ2dXk0IOU/UDAIB1EklInbu9x1LyOW+yqkLIrTaM+bUfeK+XwBBiAGgQC5m8XhkrDUEuXGN7Yap0jUo84oXig/5tfg4NtOngQKt6W7mHLQAAaADWyjgOQ4gBoNEloiHdtKNdN+1oL2ufTeeKgfbEyKxOjM7ouyfG9LdPDxX7dDdHdXCgEGpbdXCgTQf6mREZAABsMMvMcsxvLgDQ4FpiYR3Z1akjuzrL2idmvXvYvjyS1ImRGZ0YSeqLT5zXQtabPt8YaVdXkw72l0LtwQGurwUAABsXARYANqmelph6ro3prdf2FNvyrtX5y/N+qE3qxOiMXh5J6qHjpYmjomGneN/aQrA9NNCqvtaYTJ3v/QYAALY2roEFACiV9a6vLVRrX/Yrt+PJdLFPZ1NEB/pbdd22tuJw5AP9rWqJ8bdQAACweriNDgBgSfFISDdub9eN28uvr708l9HL/vDjwnDkv3nqvOYz+WKfnV0JHexv86u1XtV2b08zw5ABAMCqI8ACAGrqao7qjmt6dMc1pWHIrms1NLlQDLYvj3rh9rsnxpT3xyFHQ46u6WsphtpCsB1oizMMGQAAXDUCLABgRRzHaFd3k3Z1N+meGwaK7emcf5ufQLX2h69e0leevVDs0xYPF2/tUwi1BwZa1RaP1OOjAACABkOABQCsilg4pBsG23XDYPkw5On5rFetHU0WJ4/66rMXlEznin22dyS0r7dZA21xbWuPa1tHQgPtcQ22e89t8TCVWwAAQIAFAKyt9qaIbt/Xrdv3dRfbrLW6MLUQuM1PUucuz+vk6LjGkmlVzi/YFA15wba9EGzjGmhP+GE3rm1tCbUlCLkAAGx2BFgAwLozxmhHZ5N2dDbpXdf1l63L5l2NJ9Manl7Q8HRKI9MpXZxKaWRmQRenUvr+qQmNJVPF2/4UNEVDGmiPF4PutvZ4WRV3W3tc7YkIIRcAgAZGgAUAbCiRkKPBjoQGOxI1++TyrsaSaQ1PpzQ8vaCR6VRxeXg6pcdemdDozOKQm4iEisF2W7CC2x7XQFtCgx2EXAAANjICLACg4YTLQm5n1T65vKvx2bRXvQ2E28LyD16tHnLjEaesglutotvRRMgFAKAeCLAAgE0pHCoE0aUrueOz6cBQZb+aO5PS8NSCHn/1kkaT6eLtgQoKIbcw6dRgR0LbOxPa0ZnQjs4mDXbEFQuH1vojAgCw5RBgAQBb1pWE3Lxry67J9cLugi76offx05c0UqWS29ca80Ntk7Z3eOF2e2dCOzsT2t7RpESUgAsAwEoRYAEAWELIMRrwhw8fqdEnl3c1MpPS0OSCLkwueM9T8xqaXNCx81P65gvDyubLE253c7RYtfUCrh90u7zXrdwbFwCARQiwAAC8TuGQU5xVuZq8azWWTAXC7YKGJr2A+/JwUg8dH1Mm55Zt056IlIfbQNjd2dnEbYMAAFsSARYAgDUWckxxqPKtexavd12ribn0ooB7YXJBZybm9P1XJjSfyZdt0xILBwJuYtFw5a7mKAEXALDpEGABAKgzxzHqa42rrzWuI7sWz6psrdXkfNYPuPN+wC085vXEmctKpnNl2yQiIW2vEXB3dibU0xKT4xBwAQCNhQALAMAGZ4xRV3NUXc1R3bSjvWqf6YXFAffC5IKGpub1/NCUJuezZf2jIUfbO72ZlHtaY+ppiaq3Naaelph6W2PqbfGWu1uiioSc9fiYAAAsiwALAMAm0J6IqD0R0fWDbVXXz6VzZUOTCxXc0ZmUfjw0pfFkWnMVw5QLOpsixXBbCLjecin09rXG1NUcVZiwCwBYQwRYAAC2gOZYWAf6W3Wgv7Vmn4VMXhOzaY0l05qYTWvcfy4tZ3TMD7uV1+RKkjFSZ1PUq962Rr2w2xJTT2vpuRB6u5oIuwCAlSPAAgAASVIiGtLOribt7Ko+m3LQfCaniWRG47MpjSczZYG38PzsOS/sLmSrh92upvJhyz0t0bIKb+G5qzmqENfrAgBEgAUAAFehKRrWru6wdnUvH3bn0rnygDubKVV3k2mNz6b12mtzGk+mlcq6i7Z3jNTVXKreFqq5nU1RtSciakuEved4RG3+UOrWeJhrdwFgEyLAAgCANdUcC6s5Ftbu7uYl+1lrNZfJF0PtRKCiOz5bqvKemfDCbjq3OOyWvW80VAy0XrgNq81f9oJvYV24+LrQ1hwNcRsiANiACLAAAGBDMMaoJRZWSyysPT3Lh91U1tVMKqvphaxmFrKB5VyxbTrQfnEqpePDSc2kskqmckvuP+QYtcXDiwJwsNIbDMDBfu2JiKJhqr8AsBYIsAAAoOEYY5SIhpSIhtTfFl/x9nnXajaVKwbcyrBbDMGBgDw8vaAZf5vMMtXfeMRZNKy5rNIbj6glHlZr3AvsrfGwWuOR4nJzNMx9egGgCgIsAADYckKOUXtTRO1NkavaPpXNL6r6BsNuZQgenUnp1FhW0/NZJdM5Wbv0/o2RWqLhipDrh95Yoc271rclHlZbxevWeFitsYjiEYeh0AA2FQIsAADACsUjIcUjIfVdRfXXda3mMjklUznNpnNK+kOaC69nU35b2m9L5ZRMZzU1n9H5y/N+e7bqhFeVwo5RSyAAF8NvvDwEl4XkYkW49Joh0QA2CgIsAADAOnIc44XJ+NVVfwuyeVdzfshN+qF3tvDaD7mzxZBc6jMyk1JyrBSes/llysGSYmGnGHKbY2ElIt7w7Xgk5C1Xvo46/nO4+Doe6FfYJu4vM2M0gCtFgAUAAGhAkZCjjqaoOpqiV70Pa63SObcYchdXf7OlSrHfNp/OaSGb12w659/6KK+FbF4LmbxSWVeZ/PKV4Uphx5QF2rKg6z/Hg8G4St9CQG4qLFesj4UZTg1sBgRYAACALcoYUxwO3dMSW5V95vKuUjnXD7SlcLvgL6cCy+V9XC1kc4G+rlKZvMaT6cV9s/llryNe/Fm9SnI05CgW8QJtNOwoFg4ul17Hwo5ikYr+Ia+tfJuK7SOhQL/yPoRo4PUjwAIAAGDVhEOOWkKOWmJr92tmoXJcGZBTxSC8OCzPZ/JK5/JKZ12lc64yOdd7HVieTed0aTajTN4t9s3kXX+bvNwVhuZqoiFnUeCtDNCl5VIAjoQcRcJGEcdRJOQoHDKK+s+RkKOI/xwOOYqGjMKOo0jYUcQxioQdhZ1Cv1LfSLXtHUPIxoZGgAUAAEBDCVaOO9bpPa21yrnWD7teoC0uZ11l8qVwXAjGhUcxLFcE4mpBeiGb19RCpmzf6Vxe2bxVJu8qm3dXXH1eqUghAFcE3WBgDgblxWHa29br44XiUMgL3yHHFNeFHeOv84J2yA/ZhT4hf99ev9Lyoj6O8ddVLPv9CeSby5oGWGPMvZL+SFJI0mettb9fsT4m6fOS3ijpkqQPWWvPruUxAQAAACtljCkGuubVGW191fKuVdYPs7m8v+xa5fy2bKHNfy728dtyrheWc25lP1eZ/OL9BN8jm3O97QP9FrKV/Vxlc+Xvk/OPbzWq2CsVKoRjJxCcAyHXWxcMxaV+oULIdowc423nGK/NcYxCgbbC+4SMF9hDFW1OYF+h4D6d8nWV+y/sq9gWfJ+K/VW2GSPvvf33d4yKx1q5rlGsWYA1xoQk/amk90gakvSkMeZBa+1LgW4fkzRprb3WGHO/pD+Q9KG1OiYAAACg0XnhxKtANxrXD7N51yrrusrn/WfXKpcvBd2c/7qwLpuv0cd1S21+u9e/oo+/bXBf2bxV3vWCeT5f3i8Yvl3rtbnW2z7vWuVrtOXz/nOgba0r5qulEG4LQbcy+Ib8anZhnTGFwC2/j/HbSyG52rpCSA8uO6b0Hs4yFfO1rMDeJukVa+1pSTLGfFHSByUFA+wHJf2ev/xlSX9ijDHWNsppBgAAAHClHMco6lf7Emq8AH41XHdxqC0EebciDC/ZFgzKbsWjSptr5Qfo0uvgsmtL7+Va7zhdu3idtfK3Ke+3eJ2V66q0XGVd3np/QCiss7bw36O0XX6ZMv1aBtjtks4HXg9Jur1WH2ttzhgzLalb0sQaHhcAAAAArAvHMXJk1IAF87oxv1173VreNbpa7bcyTl9JHxljftUY85Qx5qnx8fFVOTgAAAAAQGNZywA7JGln4PUOSRdr9THGhCW1S7pcuSNr7aettbdaa2/t7e1do8MFAAAAAGxkaxlgn5S03xiz1xgTlXS/pAcr+jwo6aP+8s9K+keufwUAAAAAVLNm18D617T+uqRvybuNzgPW2heNMZ+S9JS19kFJn5P0F8aYV+RVXu9fq+MBAAAAADS2Nb0PrLX265K+XtH27wPLKUk/t5bHAAAAAADYHNZyCDEAAAAAAKuGAAsAAAAAaAgEWAAAAABAQyDAAgAAAAAaAgEWAAAAANAQCLAAAAAAgIZAgAUAAAAANAQCLAAAAACgIRBgAQAAAAANgQALAAAAAGgIBFgAAAAAQEMgwAIAAAAAGgIBFgAAAADQEAiwAAAAAICGQIAFAAAAADQEAiwAAAAAoCEQYAEAAAAADYEACwAAAABoCMZaW+9jWBFjTFLSiXofB1Zdj6SJeh8EVhXndPPhnG5OnNfNh3O6OXFeNx/OaW27rbW91VaE1/tIVsEJa+2t9T4IrC5jzFOc182Fc7r5cE43J87r5sM53Zw4r5sP5/TqMIQYAAAAANAQCLAAAAAAgIbQiAH20/U+AKwJzuvmwzndfDinmxPndfPhnG5OnNfNh3N6FRpuEicAAAAAwNbUiBVYAAAAAMAWtGEDrDHmXmPMCWPMK8aY362yPmaM+ZK//kfGmD3rf5S4UsaYncaY7xpjjhtjXjTG/GaVPncbY6aNMc/5j39fj2PFyhhjzhpjfuyfs6eqrDfGmD/2v6vPG2Nuqcdx4soYYw4GvoPPGWNmjDG/VdGH72oDMMY8YIwZM8a8EGjrMsZ8xxhzyn/urLHtR/0+p4wxH12/o8ZSapzT/2KMedn/+foVY0xHjW2X/FmN+qlxXn/PGHMh8HP2fTW2XfL3ZdRHjXP6pcD5PGuMea7GtnxXl7EhhxAbY0KSTkp6j6QhSU9K+rC19qVAn38u6WZr7SeNMfdL+mlr7YfqcsBYljFmm6Rt1tpnjDGtkp6W9FMV5/RuSb9trX1/nQ4TV8EYc1bSrdbaqvcx8//R/ZeS3ifpdkl/ZK29ff2OEFfL/1l8QdLt1trXAu13i+/qhmeMuUvSrKTPW2tv9Nv+s6TL1trf93/Z7bTW/k7Fdl2SnpJ0qyQr7+f1G621k+v6AbBIjXN6j6R/tNbmjDF/IEmV59Tvd1ZL/KxG/dQ4r78nadZa+1+X2G7Z35dRH9XOacX6P5Q0ba39VJV1Z8V3dUkbtQJ7m6RXrLWnrbUZSV+U9MGKPh+U9Of+8pclvcsYY9bxGLEC1tpha+0z/nJS0nFJ2+t7VFgnH5T3A9xaax+X1OH/QQMb37skvRoMr2gc1tpHJF2uaA7+2/nnkn6qyqY/Iek71trLfmj9jqR71+xAccWqnVNr7bettTn/5eOSdqz7geF1qfFdvRJX8vsy6mCpc+rnlZ+X9NfrelCbyEYNsNslnQ+8HtLisFPs4//gnpbUvS5Hh9fFH+59RNKPqqx+izHmmDHmG8aYG9b1wHC1rKRvG2OeNsb8apX1V/J9xsZ0v2r/A8t3tTH1W2uHJe8Pi5L6qvThO9u4fkXSN2qsW+5nNTaeX/eHhj9QY7g/39XGdKekUWvtqRrr+a4uY6MG2GqV1MqxzlfSBxuMMaZF0t9J+i1r7UzF6mck7bbWHpb0PyV9db2PD1flrdbaWyS9V9K/8IfNBPFdbUDGmKikD0j62yqr+a5ubnxnG5Ax5t9Jykn6Qo0uy/2sxsbyvyRdI+kNkoYl/WGVPnxXG9OHtXT1le/qMjZqgB2StDPweoeki7X6GGPCktp1dcMvsE6MMRF54fUL1tr/W7neWjtjrZ31l78uKWKM6Vnnw8QKWWsv+s9jkr4ib0hT0JV8n7HxvFfSM9ba0coVfFcb2mhhCL//PFalD9/ZBuNPtPV+SR+xNSY3uYKf1dhArLWj1tq8tdaV9BlVP198VxuMn1n+iaQv1erDd3V5GzXAPilpvzFmr18FwY7uYgAABANJREFUuF/SgxV9HpRUmBnxZ+VNYMBfnTYof7z/5yQdt9b+txp9BgrXMRtjbpP3/+el9TtKrJQxptmflEvGmGZJ90h6oaLbg5J+0XjeLG/SguF1PlSsXM2/EPNdbWjBfzs/Kun/VenzLUn3GGM6/WGL9/ht2ICMMfdK+h1JH7DWztfocyU/q7GBVMwV8dOqfr6u5PdlbCzvlvSytXao2kq+q1cmXO8DqMafSe/X5f2DGZL0gLX2RWPMpyQ9Za19UF4Y+gtjzCvyKq/31++IcQXeKukXJP04MG34v5W0S5Kstf9b3h8ifs0Yk5O0IOl+/iix4fVL+oqfZcKS/spa+01jzCel4nn9urwZiF+RNC/pl+t0rLhCxpgmebNa/rNAW/Cc8l1tAMaYv5Z0t6QeY8yQpP8g6fcl/Y0x5mOSzkn6Ob/vrZI+aa39uLX2sjHmP8r75ViSPmWtZYTTBlDjnP4bSTFJ3/F/Fj/u36FhUNJnrbXvU42f1XX4CKiixnm92xjzBnlDgs/K/3kcPK+1fl+uw0dAhWrn1Fr7OVWZW4Lv6sptyNvoAAAAAABQaaMOIQYAAAAAoAwBFgAAAADQEAiwAAAAAICGQIAFAAAAADQEAiwAAAAAoCEQYAEAWEPGmLwx5rnA43dXcd97jDHcIxAAsGVsyPvAAgCwiSxYa99Q74MAAGAzoAILAEAdGGPOGmP+wBjzhP+41m/fbYz5B2PM8/7zLr+93xjzFWPMMf9xh7+rkDHmM8aYF40x3zbGJPz+v2GMecnfzxfr9DEBAFhVBFgAANZWomII8YcC62astbdJ+hNJ/8Nv+xNJn7fW3izpC5L+2G//Y0kPW2sPS7pF0ot++35Jf2qtvUHSlKSf8dt/V9IRfz+fXKsPBwDAejLW2nofAwAAm5YxZtZa21Kl/aykd1prTxtjIpJGrLXdxpgJSdustVm/fdha22OMGZe0w1qbDuxjj6TvWGv3+69/R1LEWvufjDHflDQr6auSvmqtnV3jjwoAwJqjAgsAQP3YGsu1+lSTDiznVZrf4icl/amkN0p62hjDvBcAgIZHgAUAoH4+FHj+ob/8A0n3+8sfkfR9f/kfJP2aJBljQsaYtlo7NcY4knZaa78r6V9L6pC0qAoMAECj4a+xAACsrYQx5rnA629aawu30okZY34k7w/KH/bbfkPSA8aYfyVpXNIv++2/KenTxpiPyau0/pqk4RrvGZL0l8aYdklG0n+31k6t2icCAKBOuAYWAIA68K+BvdVaO1HvYwEAoFEwhBgAAAAA0BCowAIAAAAAGgIVWAAAAABAQyDAAgAAAAAaAgEWAAAAANAQCLAAAAAAgIZAgAUAAAAANAQCLAAAAACgIfx/6DkgwbpRmu0AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1152x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_history([('baseline', baseline_history),\n",
    "              ('l2', l2_model_history)])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Kx1YHMsVxWjP"
   },
   "source": [
    "As you can see, the L2 regularized model has become much more resistant to overfitting than the baseline model, even though both models have the same number of parameters."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "HmnBNOOVxiG8"
   },
   "source": [
    "### Add dropout\n",
    "\n",
    "Dropout is one of the most effective and most commonly used regularization techniques for neural networks, developed by Hinton and his students at the University of Toronto. Dropout, applied to a layer, consists of randomly \"dropping out\" (i.e. set to zero) a number of output features of the layer during training. Let's say a given layer would normally have returned a vector [0.2, 0.5, 1.3, 0.8, 1.1] for a given input sample during training; after applying dropout, this vector will have a few zero entries distributed at random, e.g. [0, 0.5,\n",
    "1.3, 0, 1.1]. The \"dropout rate\" is the fraction of the features that are being zeroed-out; it is usually set between 0.2 and 0.5. At test time, no units are dropped out, and instead the layer's output values are scaled down by a factor equal to the dropout rate, so as to balance for the fact that more units are active than at training time.\n",
    "\n",
    "In tf.keras you can introduce dropout in a network via the Dropout layer, which gets applied to the output of layer right before.\n",
    "\n",
    "Let's add two Dropout layers in our IMDB network to see how well they do at reducing overfitting:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "OFEYvtrHxSWS"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 25000 samples, validate on 25000 samples\n",
      "Epoch 1/20\n",
      "25000/25000 - 3s - loss: 0.6039 - accuracy: 0.6658 - binary_crossentropy: 0.6039 - val_loss: 0.4575 - val_accuracy: 0.8600 - val_binary_crossentropy: 0.4575\n",
      "Epoch 2/20\n",
      "25000/25000 - 2s - loss: 0.4365 - accuracy: 0.8168 - binary_crossentropy: 0.4365 - val_loss: 0.3283 - val_accuracy: 0.8822 - val_binary_crossentropy: 0.3283\n",
      "Epoch 3/20\n",
      "25000/25000 - 2s - loss: 0.3343 - accuracy: 0.8749 - binary_crossentropy: 0.3343 - val_loss: 0.2868 - val_accuracy: 0.8859 - val_binary_crossentropy: 0.2868\n",
      "Epoch 4/20\n",
      "25000/25000 - 2s - loss: 0.2753 - accuracy: 0.9017 - binary_crossentropy: 0.2753 - val_loss: 0.2740 - val_accuracy: 0.8880 - val_binary_crossentropy: 0.2740\n",
      "Epoch 5/20\n",
      "25000/25000 - 2s - loss: 0.2308 - accuracy: 0.9195 - binary_crossentropy: 0.2308 - val_loss: 0.2788 - val_accuracy: 0.8876 - val_binary_crossentropy: 0.2788\n",
      "Epoch 6/20\n",
      "25000/25000 - 2s - loss: 0.2032 - accuracy: 0.9306 - binary_crossentropy: 0.2032 - val_loss: 0.2933 - val_accuracy: 0.8834 - val_binary_crossentropy: 0.2933\n",
      "Epoch 7/20\n",
      "25000/25000 - 2s - loss: 0.1804 - accuracy: 0.9410 - binary_crossentropy: 0.1804 - val_loss: 0.3033 - val_accuracy: 0.8828 - val_binary_crossentropy: 0.3033\n",
      "Epoch 8/20\n",
      "25000/25000 - 2s - loss: 0.1593 - accuracy: 0.9439 - binary_crossentropy: 0.1593 - val_loss: 0.3266 - val_accuracy: 0.8822 - val_binary_crossentropy: 0.3266\n",
      "Epoch 9/20\n",
      "25000/25000 - 2s - loss: 0.1422 - accuracy: 0.9498 - binary_crossentropy: 0.1422 - val_loss: 0.3528 - val_accuracy: 0.8803 - val_binary_crossentropy: 0.3528\n",
      "Epoch 10/20\n",
      "25000/25000 - 2s - loss: 0.1275 - accuracy: 0.9560 - binary_crossentropy: 0.1275 - val_loss: 0.3705 - val_accuracy: 0.8789 - val_binary_crossentropy: 0.3705\n",
      "Epoch 11/20\n",
      "25000/25000 - 2s - loss: 0.1160 - accuracy: 0.9580 - binary_crossentropy: 0.1160 - val_loss: 0.3763 - val_accuracy: 0.8812 - val_binary_crossentropy: 0.3763\n",
      "Epoch 12/20\n",
      "25000/25000 - 2s - loss: 0.1077 - accuracy: 0.9608 - binary_crossentropy: 0.1077 - val_loss: 0.4056 - val_accuracy: 0.8797 - val_binary_crossentropy: 0.4056\n",
      "Epoch 13/20\n",
      "25000/25000 - 2s - loss: 0.0990 - accuracy: 0.9626 - binary_crossentropy: 0.0990 - val_loss: 0.4224 - val_accuracy: 0.8781 - val_binary_crossentropy: 0.4224\n",
      "Epoch 14/20\n",
      "25000/25000 - 2s - loss: 0.0935 - accuracy: 0.9643 - binary_crossentropy: 0.0935 - val_loss: 0.4491 - val_accuracy: 0.8770 - val_binary_crossentropy: 0.4491\n",
      "Epoch 15/20\n",
      "25000/25000 - 2s - loss: 0.0854 - accuracy: 0.9665 - binary_crossentropy: 0.0854 - val_loss: 0.4765 - val_accuracy: 0.8776 - val_binary_crossentropy: 0.4765\n",
      "Epoch 16/20\n",
      "25000/25000 - 2s - loss: 0.0821 - accuracy: 0.9677 - binary_crossentropy: 0.0821 - val_loss: 0.4774 - val_accuracy: 0.8765 - val_binary_crossentropy: 0.4774\n",
      "Epoch 17/20\n",
      "25000/25000 - 2s - loss: 0.0789 - accuracy: 0.9679 - binary_crossentropy: 0.0789 - val_loss: 0.4895 - val_accuracy: 0.8771 - val_binary_crossentropy: 0.4895\n",
      "Epoch 18/20\n",
      "25000/25000 - 2s - loss: 0.0747 - accuracy: 0.9695 - binary_crossentropy: 0.0747 - val_loss: 0.5093 - val_accuracy: 0.8752 - val_binary_crossentropy: 0.5093\n",
      "Epoch 19/20\n",
      "25000/25000 - 2s - loss: 0.0725 - accuracy: 0.9699 - binary_crossentropy: 0.0725 - val_loss: 0.5196 - val_accuracy: 0.8738 - val_binary_crossentropy: 0.5196\n",
      "Epoch 20/20\n",
      "25000/25000 - 3s - loss: 0.0704 - accuracy: 0.9694 - binary_crossentropy: 0.0704 - val_loss: 0.5718 - val_accuracy: 0.8749 - val_binary_crossentropy: 0.5718\n"
     ]
    }
   ],
   "source": [
    "dpt_model = keras.models.Sequential([\n",
    "    keras.layers.Dense(16, activation='relu', input_shape=(NUM_WORDS,)),\n",
    "    keras.layers.Dropout(0.5),\n",
    "    keras.layers.Dense(16, activation='relu'),\n",
    "    keras.layers.Dropout(0.5),\n",
    "    keras.layers.Dense(1, activation='sigmoid')\n",
    "])\n",
    "\n",
    "dpt_model.compile(optimizer='adam',\n",
    "                  loss='binary_crossentropy',\n",
    "                  metrics=['accuracy','binary_crossentropy'])\n",
    "\n",
    "dpt_model_history = dpt_model.fit(train_data, train_labels,\n",
    "                                  epochs=20,\n",
    "                                  batch_size=512,\n",
    "                                  validation_data=(test_data, test_labels),\n",
    "                                  verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "SPZqwVchx5xp"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_history([('baseline', baseline_history),\n",
    "              ('dropout', dpt_model_history)])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "gjfnkEeQyAFG"
   },
   "source": [
    "Adding dropout is a clear improvement over the baseline model.\n",
    "\n",
    "To recap: here are the most common ways to prevent overfitting in neural networks:\n",
    "\n",
    "* Get more training data.\n",
    "* Reduce the capacity of the network.\n",
    "* Add weight regularization.\n",
    "* Add dropout.\n",
    "\n",
    "And two important approaches not covered in this guide are data-augmentation and batch normalization."
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [
    "fTFj8ft5dlbS"
   ],
   "name": "overfit-and-underfit.ipynb",
   "private_outputs": true,
   "provenance": [],
   "toc_visible": true,
   "version": "0.3.2"
  },
  "kernelspec": {
   "display_name": "tensorflow-gpu/2.0.0-beta-py36",
   "language": "python",
   "name": "tensorflow_gpu_2.0.0-beta-py36"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}