Transformer_concise.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Transformer_concise.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyN2U4jcSxH3KNYSS4a3S+4N",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/wenxuan0923/My-notes/blob/master/Transformer_concise.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RMupv3G0DwEp",
        "colab_type": "text"
      },
      "source": [
        "## Sentiment Analysis with Transformer (Concise)\n",
        "\n",
        "This note covers a Transformer model for sentiment prediction problem using Keras and Tensorflow. This is a concise version of the step-by-step guide in <a target='_blank' href='#'>this</a> note. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "I9bgzQdH1S4P",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "con# Set up\n",
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "from tensorflow.keras import layers\n",
        "from tensorflow.keras.datasets import imdb\n",
        "from tensorflow.keras.layers import Activation\n",
        "from tensorflow.keras.preprocessing.sequence import pad_sequences"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "A_LrETsowqdh",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "vocab_size = 20000\n",
        "maxlen = 200\n",
        "(x_train, y_train), (x_val, y_val) = imdb.load_data(num_words=vocab_size)\n",
        "x_train = pad_sequences(x_train, maxlen=maxlen)\n",
        "x_val = pad_sequences(x_val, maxlen=maxlen)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TjlG3RZ514Py",
        "colab_type": "code",
        "outputId": "b7cbe0f4-5423-431f-841c-096ad98bafeb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 90
        }
      },
      "source": [
        "print(x_train.shape)\n",
        "print(y_train.shape)\n",
        "print(x_val.shape)\n",
        "print(y_val.shape)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(25000, 200)\n",
            "(25000,)\n",
            "(25000, 200)\n",
            "(25000,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YlZgnQ42yKMo",
        "colab_type": "text"
      },
      "source": [
        "The source code of `layers.Layer` class is <a target='_blank' href='https://github.com/keras-team/keras/blob/7a39b6c62d43c25472b2c2476bd2a8983ae4f682/keras/engine/base_layer.py#L36'>here</a>\n",
        "\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bp-GGQ-4v2i_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class Embedding(layers.Layer):\n",
        "  def __init__(self, maxlen, vocab_size, embed_dim):\n",
        "    super().__init__()\n",
        "    self.maxlen = maxlen\n",
        "    self.word_embed = layers.Embedding(vocab_size, embed_dim)\n",
        "    self.position_embed = layers.Embedding(maxlen, embed_dim)\n",
        "  \n",
        "  def call(self, inputs):\n",
        "    word_emb = self.word_embed(inputs)\n",
        "    positions = np.array(range(self.maxlen))\n",
        "    position_emb = self.position_embed(positions)\n",
        "    return word_emb + position_emb"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ER0p6FFmv2nS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class MultiHeadAttention(layers.Layer):\n",
        "  def __init__(self, maxlen, num_heads, embed_dim):\n",
        "    super().__init__()\n",
        "    self.maxlen = maxlen\n",
        "    self.num_heads = num_heads\n",
        "    self.projection_dim = embed_dim // num_heads\n",
        "    self.query_dense = layers.Dense(embed_dim)\n",
        "    self.key_dense = layers.Dense(embed_dim)\n",
        "    self.value_dense = layers.Dense(embed_dim)\n",
        "    self.combine_heads = layers.Dense(embed_dim)\n",
        "\n",
        "  def split_heads(self, x):\n",
        "    x = tf.reshape(x, (-1, self.maxlen, \n",
        "                       self.num_heads, \n",
        "                       self.projection_dim)) \n",
        "    x = tf.transpose(x, perm=[0, 2, 1, 3])\n",
        "    return x\n",
        "  \n",
        "  def calc_attention(self, Q, K, V):\n",
        "    dimension_k = tf.cast(K.shape[-1], tf.float32)\n",
        "    scores = tf.matmul(Q, K, transpose_b=True)/tf.math.sqrt(dimension_k)\n",
        "    weights = Activation('softmax', name='self-attention')(scores)\n",
        "    attention = tf.matmul(weights, V)\n",
        "    attention = tf.transpose(attention, perm=[0, 2, 1, 3])\n",
        "    concat_attention = tf.reshape(attention, (-1, self.maxlen, \n",
        "                                              self.num_heads*self.projection_dim))\n",
        "    return concat_attention\n",
        "\n",
        "  def call(self, inputs):\n",
        "    Q = self.query_dense(inputs)\n",
        "    K = self.key_dense(inputs)\n",
        "    V = self.value_dense(inputs)\n",
        "\n",
        "    Q = self.split_heads(Q)\n",
        "    K = self.split_heads(K)\n",
        "    V = self.split_heads(V)\n",
        "\n",
        "    concat_attention = self.calc_attention(Q, K, V)\n",
        "    outputs = self.combine_heads(concat_attention)\n",
        "    return outputs\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_2yr_5IAv2qA",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class TransformerEncoder(layers.Layer):\n",
        "  def __init__(self, maxlen, num_heads, vocab_size, \n",
        "               embed_dim, ff_dim, dropout_rate):\n",
        "    super().__init__()\n",
        "    self.emb = Embedding(maxlen, vocab_size, embed_dim)\n",
        "    self.attention = MultiHeadAttention(maxlen, num_heads, embed_dim)\n",
        "    self.ff = keras.Sequential([layers.Dense(ff_dim, activation=\"relu\"),\n",
        "                                layers.Dense(embed_dim),])\n",
        "    self.norm1 = layers.LayerNormalization(epsilon=1e-6)\n",
        "    self.norm2 = layers.LayerNormalization(epsilon=1e-6)\n",
        "    self.dropout1 = layers.Dropout(dropout_rate)\n",
        "    self.dropout2 = layers.Dropout(dropout_rate)\n",
        "\n",
        "  def call(self, inputs):\n",
        "    embeded_inputs = self.emb(inputs)\n",
        "    multiHead_output = self.attention(embeded_inputs)\n",
        "    out1 = self.dropout1(multiHead_output)\n",
        "    out1 = self.norm1(out1 + embeded_inputs)\n",
        "    feed_forword_output = self.ff(out1)\n",
        "    out2 = self.dropout2(feed_forword_output)\n",
        "    out2 = self.norm2(out2 + out1)\n",
        "    return out2"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Qis5FualvSxO",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "embed_dim = 32       # Dimension of Embedding layers\n",
        "num_heads = 2        # Number of heads for multi-head attention model\n",
        "ff_dim = 32          # Dimension of feed forward network\n",
        "dropout_rate = 0.1"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eOPEMEm0-DtF",
        "colab_type": "code",
        "outputId": "2fff4299-7894-4ac3-bebf-13d81a63e828",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 754
        }
      },
      "source": [
        "# Sentiment Prediction Model\n",
        "inputs = keras.Input(shape=(maxlen,))\n",
        "Encoder = TransformerEncoder(\n",
        "    maxlen, num_heads, vocab_size, \n",
        "    embed_dim, ff_dim, dropout_rate)\n",
        "x = Encoder(inputs)\n",
        "x = layers.GlobalAveragePooling1D()(x)\n",
        "x = layers.Dropout(0.1)(x)\n",
        "x = layers.Dense(20, activation=\"relu\")(x)\n",
        "x = layers.Dropout(0.1)(x)\n",
        "outputs = layers.Dense(2, activation=\"softmax\")(x)\n",
        "model = keras.Model(inputs=inputs, outputs=outputs)\n",
        "keras.utils.plot_model(model, \"sentimen_model.png\", show_shapes=True)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAn8AAALhCAYAAAAuIsLIAAAABmJLR0QA/wD/AP+gvaeTAAAgAElEQVR4nOzde1TU5do//vcoh5lBjipIIMpBUdRWlj1LPNYm3SkmHjhZ7CJXPh4y0OjJpEhExZC2sVCpXVtd+6mUkywwEfUxYisrD7XNdFO6BSMFElBUDoMyMPfvD38zX0cOznCYAeb9Wmv+8J5rPvf1uWeGufwc7lsihBAgIiIiIlOQMcDYGRARERGR4bD4IyIiIjIhLP6IiIiITAiLPyIiIiITYmbsBKhvCwoKMnYKREQmxdfXF2+//bax06A+jEf+qEsyMzNRVlZm7DSIOu306dM4ffq0sdPoU8rKypCZmWnsNEzS6dOncerUKWOnQX0cj/xRl61duxbBwcHGToOoU9RHrzMyMoycSd+Rnp6OkJAQjpkR8GwLdQce+SMiIiIyISz+iIiIiEwIiz8iIiIiE8Lij4iIiMiEsPgjIiIiMiEs/oiIusHhw4dha2uLb775xtip9EorVqyARCLRPMLCwlrFHD9+HOvXr4dKpcLChQvh5uYGqVQKFxcXBAQE4MKFC3r3GxcXBx8fH9jY2MDS0hJeXl549913UV9f3yq2sLAQU6dOhVwuh7OzM9atW4f79+/rHXfw4EEkJCSgpaVF63XZ2dlaYzBkyBC994eoO7D4IyLqBkIIY6fQ6zk4OCAvLw+XL1/G7t27tZ7bsGEDkpOTER0dDZVKhZMnT2Lfvn2oqalBYWEhGhsbMWPGDFRUVOjVZ35+PlavXo3S0lLcvHkT8fHxSEpKajVlSlFREWbPng0/Pz9UV1cjKysLe/bswcqVK/WOmz9/PqRSKfz8/HDnzh1Ne0BAAMrKynDixAnMnTtXr/0g6laCqAsAiLS0NGOnQdRpgYGBIjAw0NhpdCuFQiF8fX17bPtpaWlC35+P5cuXCxcXlzaf27p1qxg9erRobGwUQgihVCrFvHnztGLOnj0rAIgtW7bo1a+/v79obm7WagsODhYAxLVr1zRtISEhwt3dXahUKk1bYmKikEgk4tdff9U7TgghIiIihK+vr1Aqla3yioyMFIMHD9ZrX4Ton59XMrh0HvkjIupndu/ejaqqKmOnoZPi4mLExMRg48aNkEqlAAAzM7NWp889PDwAACUlJXpt/9ChQxg4cKBWm/p0q0KhAAA0NzcjNzcXM2fOhEQi0cTNmTMHQgjk5OToFacWGxuL8+fPIykpSa+ciXoaiz8ioi4qLCyEm5sbJBIJdu7cCQBISUmBlZUV5HI5cnJyMGfOHNjY2MDV1RX79+/XvDY5ORlSqRSOjo5YsWIFnJ2dIZVKMWXKFJw5c0YTFxERAQsLCwwbNkzT9uabb8LKygoSiQQ3b94EAKxZswZRUVEoKSmBRCKBl5cXAODIkSOwsbHBli1bDDEkOktOToYQAvPnz+8wrrGxEQBgY2PT5T7Ly8shk8ng7u4OALh69Srq6+vh5uamFefp6QkAmmsNdY1Ts7e3x8yZM5GUlMTLAqhXYfFHRNRF06ZNw/fff6/VtmrVKqxduxaNjY2wtrZGWloaSkpK4OHhgWXLlkGpVAJ4UNSFh4dDoVAgMjISpaWlOHfuHJqbmzFr1ixcv34dwIMi6dFlFHft2oWNGzdqtSUlJeGll16Cp6cnhBAoLi4GAM3NByqVqkfGoLNyc3Ph7e0NuVzeYdzZs2cBPBjrrlAoFMjPz8eyZctgYWEBALhx4wYAwNraWitWKpVCJpOhsrJSr7iHTZw4EeXl5fj555+7lDdRd2LxR0TUw6ZMmQIbGxsMHToUoaGhaGhowLVr17RizMzMMHbsWFhaWsLHxwcpKSmoq6vD3r17uyUHf39/1NbWIiYmplu21x0aGhrw22+/aY6ctaWyshKpqamIjIyEr6/vY48QPk58fDycnZ2xefNmTZv6Tt1HTw8DgLm5ueaoo65xDxs1ahQA4OLFi13Km6g7mRk7ASIiU6I+2qQ+8teeSZMmQS6X49KlS4ZIyyiqqqoghOjwqJ+vry8aGhoQHByMzZs3w9zcvNP9ZWVlIT09HceOHdM6eqe+1rC5ubnVa5qamiCTyfSKe5h639o6KkhkLCz+iIh6KUtLS1RXVxs7jR5z7949AA/2sz2Ojo7YvXs3xo0b16W+UlNTsX37dhQUFOCJJ57Qek59HWVtba1Wu0KhwL179+Ds7KxX3MPUBaF6X4l6AxZ/RES9kFKpxJ07d+Dq6mrsVHqMujB6dDLkhw0dOhR2dnZd6mfHjh04evQo8vPzMWjQoFbPu7u7w9raGr///rtWu/p6ySeffFKvuIc1NTUBQJtHBYmMhcUfEVEvVFBQACEEJk+erGkzMzN77OnivsTR0RESiQR3795tN6YrK6YIIfDee+/h9u3byM7OhplZ2z95ZmZmmDt3Lk6cOAGVSoUBAx5cDp+XlweJRKK5zlDXuIep983JyanT+0HU3XjDBxFRL6BSqXD79m00NzfjwoULWLNmDdzc3BAeHq6J8fLyQk1NDbKzs6FUKlFdXd3qKBTwYCWNiooKlJaWoq6uDkqlEnl5eb1uqhe5XA4PDw+UlZW1+XxxcTGcnJwQEhLS6rnQ0FA4OTnh3Llz7W7/l19+wbZt2/DFF1/A3Nxca2k1iUSCjz/+WBMbExODyspKbNiwAQ0NDTh16hQSExMRHh4Ob29vvePU1Ps2YcIEnceFqKex+CMi6qKdO3fi2WefBQCsW7cOAQEBSElJwSeffALgwenAq1ev4osvvkBUVBQA4MUXX8SVK1c027h37x4mTJgAmUyG6dOnY/To0fjuu++0rodbtWoVnn/+eSxZsgTe3t7YtGmT5nSir6+vZlqYlStXwtHRET4+Ppg7dy5qamoMMg6d4e/vj6KiojbvlO1obrympiZUVVW1mlhZ19c/aty4cTh69CiOHTuGwYMHY/HixVi6dCk+/fTTTsWp/fDDD3BxcWnzlDCRsfC0LxFRF61evRqrV69u1b5q1Sqtf6vn+GuLtbV1u0fA1BwcHJCfn9+qfdu2bVr/njhxIkpLS7Xa5syZ0+pGhd7grbfeQkpKCg4cOICwsDCt50aNGtXuXbKZmZl47rnnMGLEiHa3PX78eL0KwBkzZmhNrN3VuFu3buHbb7/F5s2btVYEITI2HvkjIuoFOrrpob9obGzE0aNHceXKFc2NEF5eXoiLi0NcXBzq6+t12k5LSwuys7NRV1eH0NDQnky5S2JjY/HUU08hIiICwIMjkRUVFSgsLNTcJEJkDCz+iIjIIGpqavDiiy9i9OjRWLp0qaZ9/fr1CAoKQmhoaIc3f6gVFBTgwIEDyMvLe+zKIMayfft2nD9/HocPH9bMTZiTkwMXFxdMnz4dubm5Rs6QTBmLPzKow4cPw9bWtkt38PUGzz33XKuLx9WPtqaS6Mjp06cxduxYDBgwABKJBE5OTlqrD/QGBw4cgIeHh2Yfhw0b1uoUHXVOdHQ09u7di7t378Ld3R2ZmZnGTqlHfPbZZxBCaB5fffWV1vNbtmxBREQEtm7d+tht+fn54euvv9Za57g3ycnJwf3791FQUAB7e3tN+4IFC7TGQL0eM5Gh8Zo/MihTWNxc37VHJ0+ejF9//RUvvvgijh49isuXL3d5XrPutnjxYixevBheXl64efOmZo1T6rr4+HjEx8cbO41eYfbs2Zg9e7ax0+iygIAABAQEGDsNonbxyB8ZlL+/P+7evYuXXnrJ2KmgsbERU6ZM6dRrpVIpamtrtf4XL4TA8uXL8e6773ZzpobXlbEhIqLejcUfmazdu3ejqqqqU689cuSI1tqgAHD9+nX8+9//xp/+9KfuSM+oujI2RETUu7H4I4MpLCyEm5sbJBIJdu7cCQBISUmBlZUV5HI5cnJyMGfOHNjY2MDV1RX79+/XvDY5ORlSqRSOjo5YsWIFnJ2dIZVKMWXKFK0pFyIiImBhYaF1LdCbb74JKysrSCQSzTU2a9asQVRUFEpKSiCRSODl5dXl/fvoo48QGRmp1XbkyJFOT6zb18fm5MmT8PHxga2tLaRSKSZMmICjR48CAN544w3N9YOenp746aefAACvv/465HI5bG1tcfDgQQAP7uz88MMP4ebmBplMhieffBJpaWkAHkxxIpfLYW1tjaqqKkRFRcHFxQWXL1/uVM5ERCZBEHUBAJGWlqZz/PXr1wUAsWPHDk3b+++/LwCIb7/9Vty9e1dUVVWJ6dOnCysrK9HU1KSJW758ubCyshK//PKLuHfvnigqKhLPPvussLa2FteuXdPEvfLKK8LJyUmr38TERAFAVFdXa9oWL14sPD09O7PbrZSVlQkfHx/R0tKi1X7o0CFhbW0t4uLiHruNP//5zwKAuH37tqatt42Np6ensLW1ffyACCEyMjJEbGysqKmpEbdu3RKTJ08WgwcP1upj4MCBory8XOt1L7/8sjh48KDm3++8846wtLQUmZmZ4vbt2yI6OloMGDBA/PDDD1pjFBkZKXbs2CEWLVokfv31V51yFEKIwMBAERgYqHM8CZGWlib482Ec/LxSN0jnkT/qNaZMmQIbGxsMHToUoaGhaGhowLVr17RizMzMMHbsWFhaWsLHxwcpKSmoq6vD3r17jZT1Ax999BHeeustzVqfav7+/qitrUVMTEyXtt8XxyYwMBAbNmyAvb09HBwcMH/+fNy6dQvV1dUAHqxC0dLSopVfbW0tfvjhB8ydOxfAg1UvUlJSsHDhQixevBh2dnb44IMPYG5u3mq/PvroI6xevRoHDhzAmDFjDLejRER9DO/2pV7JwsICAB67iP2kSZMgl8tx6dIlQ6TVpoqKChw8eBCJiYkG6a8vjc3D1HOdqScz/tOf/oTRo0djz549iI6OhkQiQWpqKkJDQzFw4EAAwOXLl6FQKDB+/HjNdmQyGYYNG9at+5WZmckVGDqBY2YcgYGBxk6B+jgWf9TnWVpaao4mGUNCQgKWLVsGqVRqtBzaY8yxyc3NRWJiIoqKilBbW9uqWJVIJFixYgXefvttfPvtt3jhhRfwv//7v/j66681MQ0NDQCADz74AB988IHW652dnbst18mTJ2Pt2rXdtr3+7tSpU0hKStJce0mGo14vmqgrWPxRn6ZUKnHnzh24uroapf8bN25g3759vfIGA0OPzYkTJ/Cvf/0La9euxbVr17Bw4UIsWrQIe/bswRNPPIEdO3a0mgYnPDwc0dHR+Pvf/47hw4fDxsZGa63WoUOHAnjwg7dmzZoey93V1RXBwcE9tv3+KCkpiWNmBBkZGcZOgfoBFn/UpxUUFEAIgcmTJ2vazMzMHntKtLskJCQgLCwMDg4OBulPH4Yem3/961+wsrICAFy8eBFKpRKrVq2Ch4cHgLZPEdrb2yMkJASpqamwtrbGsmXLtJ4fPnw4pFIpzp8/3yM5ExGZIt7wQX2KSqXC7du30dzcjAsXLmDNmjVwc3NDeHi4JsbLyws1NTXIzs6GUqlEdXU1fv/991bbcnBwQEVFBUpLS1FXV6d3UVRZWYk9e/Z0eLowLy+v01O96MtYY6NUKlFZWYmCggJN8efm5gYAOH78OO7du4crV65oTTvzsJUrV+L+/fs4dOhQq8m/pVIpXn/9dezfvx8pKSmora1FS0sLysrK8Mcff+g7REREBPBefeoa6DHVy44dO8SwYcMEACGXy8X8+fPFrl27hFwuFwDEqFGjRElJifj888+FjY2NACBGjBgh/vOf/wghHkxnYm5uLlxcXISZmZmwsbERCxYsECUlJVr93Lp1Szz//PNCKpUKd3d38dZbb4n/+Z//EQCEl5eXZuqTc+fOiREjRgiZTCamTZsmbty4ode+v/322yIsLKzDmMOHDwtra2uxefPmdmNOnz4txo0bJwYMGCAAiGHDhoktW7b0qrH59NNPhaenpwDQ4SMrK0vT17p164SDg4Ows7MTQUFBYufOnQKA8PT01Jp+RgghJk6cKNavX9/m+Ny/f1+sW7dOuLm5CTMzMzF06FCxePFiUVRUJBISEoRMJhMAxPDhw8WXX37Z4fvRFk6doT9O9WI8/LxSN0iXCGECi61Sj5FIJEhLSzPItT8rVqxARkYGbt261eN99TV9fWz8/f2xc+dOuLu7G7zvoKAgALyWSh/p6ekICQkxibW6ext+XqkbZPC0L/Up6mlCqLW+NDYPn0a+cOECpFKpUQo/IiJTxOKPCMClS5c0y4119AgNDTV2qv3CunXrcOXKFfznP//B66+/jk2bNhk7JephK1as0PouhYWFtYo5fvw41q9fD5VKhYULF8LNzQ1SqRQuLi4ICAjAhQsX9O43Li4OPj4+sLGxgaWlJby8vPDuu++ivr6+VWxhYSGmTp0KuVwOZ2dnrFu3Dvfv39c77uDBg0hISGj1H7Ls7GytMRgyZIje+0PUHVj8UZ8QHR2NvXv34u7du3B3d0dmZma3bn/MmDEQQjz2kZqa2q39doeeHpueIJfLMWbMGLzwwguIjY2Fj4+PsVMiA3BwcEBeXh4uX76M3bt3az23YcMGJCcnIzo6GiqVCidPnsS+fftQU1ODwsJCNDY2YsaMGaioqNCrz/z8fKxevRqlpaW4efMm4uPjkZSUpDl9qlZUVITZs2fDz88P1dXVyMrKwp49e7By5Uq94+bPnw+pVAo/Pz/cuXNH0x4QEICysjKcOHFCs4oNkVEY51pD6i+g59q+RL1Nb7iAXqFQCF9f3z7TR2du+Fi+fLlwcXFp87mtW7eK0aNHi8bGRiGEEEqlUsybN08r5uzZswKA2LJli179+vv7i+bmZq224OBgAUDrxqOQkBDh7u4uVCqVpi0xMVFIJBKttaJ1jRNCiIiICOHr6yuUSmWrvCIjI7XWutZVb/i8Up/HtX2JiIxt9+7dqKqq6vN9dEZxcTFiYmKwceNGzSo5ZmZm+Oabb7Ti1PNFlpSU6LX9Q4cOaZYLVFOfblUoFACA5uZm5ObmYubMmVrzUc6ZMwdCCOTk5OgVpxYbG4vz588jKSlJr5yJehqLPyIiPQkhsH37dowdOxaWlpawt7fHggULtNYbjoiIgIWFBYYNG6Zpe/PNN2FlZQWJRIKbN28CANasWYOoqCiUlJRAIpHAy8sLycnJkEqlcHR0xIoVK+Ds7AypVIopU6ZozZfYlT4A4MiRIwabh7I9ycnJEEJg/vz5HcY1NjYCAGxsbLrcZ3l5OWQymeYmo6tXr6K+vl4zP6Wap6cnAGiuNdQ1Ts3e3h4zZ85EUlIS74ymXoXFHxGRnmJjY7F+/Xq8//77qKqqwokTJ3D9+nVMnz4dlZWVAB4UNY9OgbRr1y5s3LhRqy0pKQkvvfQSPD09IYRAcXExIiIiEB4eDoVCgcjISJSWluLcuXNobm7GrFmzcP369S73Afy/O8RVKlX3DY6ecnNz4e3tDblc3mHc2bNnAQDTpk3rUn8KhQL5+flYtmwZLCwsADxYphEArK2ttWKlUilkMpnmPdU17mETJ05EeXk5fv755y7lTdSdWPwREemhsbER27dvx6JFixAWFgZbW1tMmDABn332GW7evInPP/+82/oyMzPTHF308fFBSkoK6urqsHfv3m7Zvr+/P2praxETE9Mt29NXQ0MDfvvtN82Rs7ZUVlYiNTUVkZGR8PX1fewRwseJj4+Hs7MzNm/erGlT36n76OlhADA3N9ccddQ17mGjRo0C8GDJQ6Legmv7EhHpoaioCPX19Zg0aZJW+7PPPgsLC4t2l7HrDpMmTYJcLtc6vdyXVVVVQQjR4VE/X19fNDQ0IDg4GJs3b4a5uXmn+8vKykJ6ejqOHTumdfROfa1hc3Nzq9c0NTVBJpPpFfcw9b61dVSQyFhY/BER6UE9dcegQYNaPWdnZ4e6uroe7d/S0hLV1dU92oeh3Lt3D8CDfWqPo6Mjdu/ejXHjxnWpr9TUVGzfvh0FBQV44okntJ5TXzNZW1ur1a5QKHDv3j04OzvrFfcwdUGo3lei3oDFHxGRHuzs7ACgzSLvzp07cHV17bG+lUplj/dhSOrCqKPVaYYOHaoZ887asWMHjh49ivz8/DaLdnd3d1hbW+P333/XaldfG/nkk0/qFfewpqYmAGjzqCCRsbD4IyLSw/jx4zFo0CD8+OOPWu1nzpxBU1MTnnnmGU2bmZmZ1lJ2XVVQUAAhBCZPntxjfRiSo6MjJBIJ7t69227Mo1O+6EMIgffeew+3b99GdnY2zMza/skzMzPD3LlzceLECahUKgwY8OBy+Ly8PEgkEs11hrrGPUy9b05OTp3eD6Luxhs+iIj0IJVKERUVhaysLHz11Veora3FxYsXsXLlSjg7O2P58uWaWC8vL9TU1CA7OxtKpRLV1dWtjhoBD1a+qKioQGlpKerq6jTFnEqlwu3bt9Hc3IwLFy5gzZo1cHNzQ3h4eLf0kZeXZ9SpXuRyOTw8PFBWVtbm88XFxXByckJISEir50JDQ+Hk5IRz5861u/1ffvkF27ZtwxdffAFzc/NWyzV+/PHHmtiYmBhUVlZiw4YNaGhowKlTp5CYmIjw8HB4e3vrHaem3rcJEyboPC5EPY3FHxGRnjZs2ID4+HjExcVhyJAhmDlzJkaOHImCggJYWVlp4latWoXnn38eS5Ysgbe3NzZt2qQ5/efr66uZsmXlypVwdHSEj48P5s6di5qaGgAPrhObMGECZDIZpk+fjtGjR+O7777Tukauq30Ym7+/P4qKitq8U7ajufGamppQVVXVamJlXV//qHHjxuHo0aM4duwYBg8ejMWLF2Pp0qX49NNPOxWn9sMPP8DFxaXNU8JExsLTvkREepJIJHjnnXfwzjvvdBjn4OCA/Pz8Vu3btm3T+vfEiRNRWlraKs7a2rrdo2Ld0cecOXNa3bxgaG+99RZSUlJw4MABhIWFaT03atSodu+SzczMxHPPPYcRI0a0u+3x48frVQDOmDFDp7u1dY27desWvv32W2zevFlrRRAiY+ORPyKiXqqjGyH6osbGRhw9ehRXrlzR3Ajh5eWFuLg4xMXFob6+XqfttLS0IDs7G3V1dQgNDe3JlLskNjYWTz31FCIiIgA8OBJZUVGBwsJCzU0iRMbA4o+IiAyipqYGL774IkaPHo2lS5dq2tevX4+goCCEhoZ2ePOHWkFBAQ4cOIC8vLzHrgxiLNu3b8f58+dx+PBhzdyEOTk5cHFxwfTp05Gbm2vkDMmUsfgjIuploqOjsXfvXty9exfu7u7IzMw0dkpd9tlnn0EIoXl89dVXWs9v2bIFERER2Lp162O35efnh6+//lprTePeJCcnB/fv30dBQQHs7e017QsWLNAaA/Xay0SGxmv+iIh6mfj4eMTHxxs7DYObPXs2Zs+ebew0uiwgIAABAQHGToOoXTzyR0RERGRCWPwRERERmRAWf0REREQmhMUfERERkQnhDR/UZadOnTJ2CkSdpp5EOT093ciZ9B3q7zzHzPDKysrg6upq7DSoj5MIfaY/J3oEZ60nIjKswMBAZGRkGDsN6rsyeOSPuoT/d6C+KD09HSEhIfz8EpFJ4jV/RERERCaExR8RERGRCWHxR0RERGRCWPwRERERmRAWf0REREQmhMUfERERkQlh8UdERERkQlj8EREREZkQFn9EREREJoTFHxEREZEJYfFHREREZEJY/BERERGZEBZ/RERERCaExR8RERGRCWHxR0RERGRCWPwRERERmRAWf0REREQmhMUfERERkQlh8UdERERkQlj8EREREZkQFn9EREREJoTFHxEREZEJYfFHREREZEJY/BERERGZEBZ/RERERCaExR8RERGRCWHxR0RERGRCWPwRERERmRAWf0REREQmhMUfERERkQlh8UdERERkQlj8EREREZkQFn9EREREJsTM2AkQEfWksrIyvPbaa2hpadG03b59G9bW1njuuee0Yr29vfG3v/3NwBkSERkWiz8i6tdcXV3x+++/o6SkpNVz//znP7X+PWPGDEOlRURkNDztS0T93quvvgpzc/PHxoWGhhogGyIi42LxR0T93iuvvILm5uYOY8aNGwcfHx8DZUREZDws/oio3/P09MSTTz4JiUTS5vPm5uZ47bXXDJwVEZFxsPgjIpPw6quvYuDAgW0+19zcjKCgIANnRERkHCz+iMgkLFmyBCqVqlX7gAEDMHnyZIwcOdLwSRERGQGLPyIyCc7Ozpg6dSoGDND+szdgwAC8+uqrRsqKiMjwWPwRkcn4y1/+0qpNCIFFixYZIRsiIuNg8UdEJiMwMFDrur+BAwfihRdegKOjoxGzIiIyLBZ/RGQy7O3tMWvWLE0BKIRAWFiYkbMiIjIsFn9EZFLCwsI0N36Ym5tjwYIFRs6IiMiwWPwRkUmZP38+LC0tAQAvvfQSBg0aZOSMiIgMi8UfEZkUKysrzdE+nvIlIlMkEUIIYyfRkfT0dISEhBg7DSIiIqLH6uVlFQBkmBk7A12lpaUZOwUi6idaWlqQlpaGl19+2dip9FuffPIJAGDt2rVGzqTvOHXqFJKSkvh710ep37++oM8Uf8HBwcZOgYj6kYULF0IqlRo7jX4rIyMDAP926yspKYlj1of1leKP1/wRkUli4UdEporFHxEREZEJYfFHREREZEJY/BERERGZEBZ/RERERCaExR8REfVahw8fhq2tLb755htjp9KvHT9+HOvXr4dKpcLChQvh5uYGqVQKFxcXBAQE4MKFC3pvMy4uDj4+PrCxsYGlpSW8vLzw7rvvor6+vlVsYWEhpk6dCrlcDmdnZ6xbtw7379/XO+7gwYNISEhAS0uL3vmaEhZ/RETUa/WBCXP7vA0bNiA5ORnR0dFQqVQ4efIk9u3bh5qaGhQWFqKxsREzZsxARUWFXtvNz8/H6tWrUVpaips3byI+Ph5JSUkICgrSiisqKsLs2bPh5+eH6upqZGVlYc+ePVi5cqXecfPnz4dUKoWfnx/u3LnT+UHp70Qvl5aWJvpAmkRE9JDAwEARGBho7DS6lU/kinAAACAASURBVEKhEL6+vj22fWP83m3dulWMHj1aNDY2CiGEUCqVYt68eVoxZ8+eFQDEli1b9Nq2v7+/aG5u1moLDg4WAMS1a9c0bSEhIcLd3V2oVCpNW2JiopBIJOLXX3/VO04IISIiIoSvr69QKpV65dwVfaheSeeRPyIiIh3s3r0bVVVVxk6j2xQXFyMmJgYbN27UzHtpZmbW6hS7h4cHAKCkpESv7R86dAgDBw7UahsyZAgAQKFQAACam5uRm5uLmTNnQiKRaOLmzJkDIQRycnL0ilOLjY3F+fPn+8yky4bG4o+IiHqlwsJCuLm5QSKRYOfOnQCAlJQUWFlZQS6XIycnB3PmzIGNjQ1cXV2xf/9+zWuTk5MhlUrh6OiIFStWwNnZGVKpFFOmTMGZM2c0cREREbCwsMCwYcM0bW+++SasrKwgkUhw8+ZNAMCaNWsQFRWFkpISSCQSeHl5AQCOHDkCGxsbbNmyxRBD0q2Sk5MhhMD8+fM7jGtsbAQA2NjYdLnP8vJyyGQyuLu7AwCuXr2K+vp6uLm5acV5enoCgOZaQ13j1Ozt7TFz5kwkJSXx0oE2sPgjIqJeadq0afj++++12latWoW1a9eisbER1tbWSEtLQ0lJCTw8PLBs2TIolUoAD4q68PBwKBQKREZGorS0FOfOnUNzczNmzZqF69evA3hQAD26nNquXbuwceNGrbakpCS89NJL8PT0hBACxcXFAKC5sUClUvXIGPSk3NxceHt7Qy6Xdxh39uxZAA/ej65QKBTIz8/HsmXLYGFhAQC4ceMGAMDa2lorViqVQiaTobKyUq+4h02cOBHl5eX4+eefu5R3f8Tij4iI+qQpU6bAxsYGQ4cORWhoKBoaGnDt2jWtGDMzM4wdOxaWlpbw8fFBSkoK6urqsHfv3m7Jwd/fH7W1tYiJiemW7RlKQ0MDfvvtN82Rs7ZUVlYiNTUVkZGR8PX1fewRwseJj4+Hs7MzNm/erGlT36n76OlhADA3N9ccddQ17mGjRo0CAFy8eLFLefdHZsZOgIiIqKvUR5LUR/7aM2nSJMjlcly6dMkQafVaVVVVEEJ0eNTP19cXDQ0NCA4OxubNm2Fubt7p/rKyspCeno5jx45pHb1TX2vY3Nzc6jVNTU2QyWR6xT1MvW9tHRU0dSz+iIjIpFhaWqK6utrYaRjVvXv3ADwYi/Y4Ojpi9+7dGDduXJf6Sk1Nxfbt21FQUIAnnnhC6zn1tZa1tbVa7QqFAvfu3YOzs7NecQ9TF4TqfaX/h8UfERGZDKVSiTt37sDV1dXYqRiVujDqaDLkoUOHws7Orkv97NixA0ePHkV+fj4GDRrU6nl3d3dYW1vj999/12pXX1P55JNP6hX3sKamJgBo86igqWPxR0REJqOgoABCCEyePFnTZmZm9tjTxf2No6MjJBIJ7t69225MV1ZVEULgvffew+3bt5GdnQ0zs7bLDTMzM8ydOxcnTpyASqXCgAEPbkXIy8uDRCLRXGeoa9zD1Pvm5OTU6f3or3jDBxER9VsqlQq3b99Gc3MzLly4gDVr1sDNzQ3h4eGaGC8vL9TU1CA7OxtKpRLV1dWtjjABgIODAyoqKlBaWoq6ujoolUrk5eX1yale5HI5PDw8UFZW1ubzxcXFcHJyQkhISKvnQkND4eTkhHPnzrW7/V9++QXbtm3DF198AXNzc0gkEq3Hxx9/rImNiYlBZWUlNmzYgIaGBpw6dQqJiYkIDw+Ht7e33nFq6n2bMGGCzuNiKlj8ERFRr7Rz5048++yzAIB169YhICAAKSkp+OSTTwA8ONV39epVfPHFF4iKigIAvPjii7hy5YpmG/fu3cOECRMgk8kwffp0jB49Gt99953WtW6rVq3C888/jyVLlsDb2xubNm3SnCr09fXVTAuzcuVKODo6wsfHB3PnzkVNTY1BxqGn+Pv7o6ioqM07ZTuaG6+pqQlVVVWtJlbW9fWPGjduHI4ePYpjx45h8ODBWLx4MZYuXYpPP/20U3FqP/zwA1xcXNo8JWzqJKKXz36Ynp6OkJAQTtJIRNSHqNdvzcjIMFoOK1asQEZGBm7dumW0HPRh6N+74uJijB07Fnv37kVYWJjOr1OpVHjuuecQHh6OpUuX9mCGnXfr1i24urpi8+bNmv8Y9LQ+VK9k8MgfERH1Wx3d0GDqvLy8EBcXh7i4ONTX1+v0mpaWFmRnZ6Ourg6hoaE9nGHnxcbG4qmnnkJERISxU+mVWPwZyP379xEZGYlhw4ZBLpfjyJEjxk7JpL3xxhuwtraGRCLB+fPnDdbvc8891+raF/WjrTvhOhIaGtruth59HDp0qIf2SHf94Ttw4MABeHh4dDjWI0eONHaaWoz1Wae+Yf369QgKCkJoaGiHN3+oFRQU4MCBA8jLy3vsyiDGsn37dpw/fx6HDx/u0tyE/RmLPwP561//iiNHjuDSpUtISkrS+X9Z1DP+/ve/44svvjB2Glo6s3TSsWPHcOfOHSiVSvzxxx8AgPnz56OpqQkNDQ2oqqrCsmXLujvVTukP34HFixfj6tWr8PT0hK2tLYQQEEKgubkZCoUClZWVve4HsTd+1g0hOjoae/fuxd27d+Hu7o7MzExjp9RrbdmyBREREdi6detjY/38/PD1119rrYXcm+Tk5OD+/fsoKCiAvb29sdPptfrlVC+NjY3w8/NrtSakMWVnZ2PSpEmws7PDf//3fxs7HTISqVSK2traVutTrlixotX6oo8jkUgwderUVsWGRCKBubk5zM3NIZfL8cwzz3Q57+7Qn78DAwcOhEwmg0wmw+jRo42dDuHBUmLx8fHGTqPPmD17NmbPnm3sNLosICAAAQEBxk6j1+uXR/52796NqqoqY6ehpaysjIefexmJRGLwPo8cOdKq8Lt+/Tr+/e9/409/+pNe29q/f79OR5mWL1+OefPm6bXtnmAq34Hs7Gxjp9CKMT7rRNR79bvib82aNYiKikJJSQkkEgm8vLywbds2yOVyWFtbo6qqClFRUXBxccHly5dx8uRJ+Pj4wNbWFlKpFBMmTMDRo0cBACkpKbCysoJcLkdOTg7mzJkDGxsbuLq6Yv/+/Vr9/vOf/8R//dd/QS6Xw8bGBhMmTEBtbS3+7//+D15eXvjjjz/wj3/8Q+vaLiEEtm/frll03N7eHgsWLNBac7K93FeuXAkrKysMGDAAzzzzDJycnGBubg4rKys8/fTTmD59OoYPHw6pVAo7Ozu8++67Wvm2tLTgww8/hJubG2QyGZ588kmkpaV12Ofly5d1fh862r4+4woAX375JSZNmgSpVAorKyuMHDkSmzZt0nkM1XGJiYnw9vaGpaUlbG1t8T//8z965d0d49KWjz76CJGRkVptR44c6da5w/gdMPx3QN/xA/r/Z52IegnRy6WlpQl901y8eLHw9PTUanv//fcFABEZGSl27NghFi1aJH799VeRkZEhYmNjRU1Njbh165aYPHmyGDx4cKvXffvtt+Lu3buiqqpKTJ8+XVhZWYmmpiYhhBD19fXCxsZGJCQkiMbGRnHjxg2xaNEiUV1drdmOk5OTeO2117Ry+vDDD4WFhYX48ssvxZ07d8SFCxfE008/LYYMGSJu3Ljx2Nw3bNggAIgzZ86IhoYGcfPmTfHiiy8KACI3N1dUV1eLhoYGERERIQCI8+fPa7b5zjvvCEtLS5GZmSlu374toqOjxYABA8QPP/zQYZ+60nX7HY2rEEJ88sknAoDYunWruHXrlqipqRF/+9vfxCuvvKL3GEokEvHXv/5V3L59WygUCrFr1y4BQPz0008GG5dHlZWVCR8fH9HS0qLVfujQIWFtbS3i4uJ03tYff/whAIiAgIA2n+d3oPu+A56ensLW1lZrXyIjI8XFixfbHXdT+6wHBgaKwMBAneOpc7931Hv0ofcvvddn2d3FX2NjY4evjY+PFwBEVVVVu69T/yEtLi4WQgjx73//WwAQhw4dane7j/7wKRQKMWjQIBEaGqoVd/bsWQFA60e/vdzVP3x1dXWatn/84x8CgNaPkHqbqampQgghGhsbhVwu1+pboVAIS0tLsWrVKr3Gqy2d3f6j49rU1CTs7OzE888/r7X95uZmkZSUpPMYKhQKIZfLxaxZs7Ti9u/fr/WD2NPj0pbVq1eLTz/9tFu2pWvxx+9A199rT09PAaDVo6Piz9Q+6yz+9NeHigdqQx96/9L75Q0fXaG+JqmjuaEsLCwAQLMWpIeHBxwdHREWFobIyEiEh4c/drqHoqIi1NfXY9KkSVrtzz77LCwsLHDmzJlO5a/Orbm5WdOm3id1vpcvX4ZCocD48eM1MTKZDMOGDWt1CqkzOrv9R8f1woULuHPnDv785z9rxQ0cOBCRkZH48ccfdRrD4uJiKBQK+Pn59UjenVVRUYGDBw8iMTGx27fdFfwO6PZe29ra4s6dO5p/r1mzRu8c+/tnvaysDOnp6d22vf7u1KlTAMAx66PU719fYPLFX25uLhITE1FUVITa2tpOLe4tk8mQn5+P9957D1u2bEFcXByCg4Oxd+9ezRJBj1L/aLQ1t5udnR3q6ur0zkNXDQ0NAIAPPvgAH3zwgdZzzs7OvWb7tbW1AB6MR1t0HUP1+o5Dhw41SN66SkhIwLJlyyCVSrt92/rgd6B73uukpKRO59NfP+unT59uc21Y6hjHjHpav7vhQx/Xrl3DwoULMWzYMJw5cwZ3795FQkJCp7Y1btw4fPPNN6ioqMC6deuQlpamtXD1o9R/5Nv6gbtz5w5cXV07lYcu1D8Mn3zyiWaeMvWjO/7n0l3bf+KJJwAAN2/ebPN5XcdQXVzdv3/fIHnr4saNG9i3bx9WrVrVrdvVF78DPf9e66K/ftYDAwNbbZ+P9h/qG26MnQcfXXv/+gKTLv4uXrwIpVKJVatWwcPDA1KptFNTIlRUVOCXX34B8OCP6tatW/H0009r2toyfvx4DBo0CD/++KNW+5kzZ9DU1NSjc7Op74Dsqdn+u2v7I0eOhIODA44dO9bm87qO4fjx4zFgwAD885//NEjeukhISEBYWBgcHBx6vK+O8DvQ/e/1H3/8gddff12v1/TnzzoR9T79svhzcHBARUUFSktLUVdX1+5pLDc3NwDA8ePHce/ePVy5cqVT1xlVVFRgxYoVuHTpEpqamvDTTz/h999/x+TJk9t9jVQqRVRUFLKysvDVV1+htrYWFy9exMqVK+Hs7Izly5frnYeupFIpXn/9dezfvx8pKSmora1FS0sLysrKNKtE9IbtW1paIjo6GidOnEBERATKy8uhUqlQV1eHX375RecxHDp0KBYvXozMzEzs3r0btbW1uHDhAj7//HODjotaZWUl9uzZg7Vr17Ybk5eX161TvbSH34Hue6+FEGhsbMSBAwdgY2Oj12v762ediHop0ct15u6Zc+fOiREjRgiZTCamTZsm3n77bSGTyQQAMXz4cPHll19qYtetWyccHByEnZ2dCAoKEjt37hQAhKenp3jvvfeEXC4XAMSoUaNESUmJ+Pzzz4WNjY0AIEaMGCH+85//iNLSUjFlyhRhb28vBg4cKJ544gnx/vvvi+bmZlFaWiomTpwoAAgzMzPx9NNPi8zMTCGEECqVSiQmJopRo0YJc3NzYW9vLxYuXCguX76syS8hIaHN3JOSkjS5jRw5Upw8eVJ89NFHwtbWVgAQTk5O4uuvvxapqanCyclJABD29vZi//79Qggh7t+/L9atWyfc3NyEmZmZGDp0qFi8eLEoKipqt099dLT9Xbt26TSuajt37hQTJkwQUqlUSKVSMXHiRLFr1y6dx1AIIerq6sQbb7whBg8eLAYNGiSmTZsmPvzwQwFAuLq6ip9//tkg4yKEEG+//bYICwvrMObw4cPC2tpabN68+bHbq62tFTNmzBAODg4CgBgwYIDw8vISW7Zs0cR0lDu/A7q/11lZWe3e6fvw44MPPhBCCJP+rPNuX/31obtFqQ196P1LlwghRM+Vll2Xnp6OkJAQ9PI0iYjoIUFBQQCAjIwMI2fSd/D3rm/rQ+9fRr887UtEREREbWPxRzq5dOkSJBLJYx+hoaHGTtWgOC5ERNTXsPgjnYwZM0anW91TU1ONnapBcVyIqLc4fvw41q9fD5VKhYULF8LNzQ1SqRQuLi4ICAjAhQsX9N5mXFwcfHx8YGNjA0tLS3h5eeHdd99FfX19q9jCwkJMnToVcrkczs7OWLduXZvTDukap4uEhASMGTMGMpkMVlZWGDNmDGJiYjRzZ+qzHwcPHkRCQkKHE9z3Fyz+iIiI+rgNGzYgOTkZ0dHRUKlUOHnyJPbt24eamhoUFhaisbERM2bMQEVFhV7bzc/Px+rVq1FaWoqbN28iPj4eSUlJmms61YqKijB79mz4+fmhuroaWVlZ2LNnD1auXNmpOF2dPHkSy5Ytw7Vr11BZWYlNmzYhISEBgYGBeu/H/PnzIZVK4efnp7V6T79kwLtLOqUP3T1DRET/v95wt69CoRC+vr59po/O/t5t3bpVjB49WrMOs1KpFPPmzdOKUa8D/fAsALrw9/cXzc3NWm3BwcECgLh27ZqmLSQkRLi7uwuVSqVpS0xMFBKJRPz66696x+lq4cKFrdafDgoKEgBERUWF3vshhBARERHC19dXKJVKvXLpQ/VKOo/8ERFRv7R7925UVVX1+T46UlxcjJiYGGzcuFGzwouZmRm++eYbrTgPDw8AQElJiV7bP3ToEAYOHKjVNmTIEACAQqEA8GAd7dzcXMycOVNrkvg5c+ZACIGcnBy94vSRlZXVaolMFxcXANA6pavLfqjFxsbi/PnzXVqysbdj8UdERL2CEALbt2/H2LFjYWlpCXt7eyxYsACXLl3SxERERMDCwgLDhg3TtL355puwsrKCRCLRLJG3Zs0aREVFoaSkBBKJBF5eXkhOToZUKoWjoyNWrFgBZ2dnSKVSTJkyRWty8670AQBHjhwxyCTtAJCcnAwhBObPn99hXGNjIwDoPQF5W8rLyyGTyeDu7g4AuHr1Kurr6zWTxqt5enoCgOZaQ13juurKlSuws7PDiBEj9NoPNXt7e8ycORNJSUl9YdqWTmHxR0REvUJsbCzWr1+P999/H1VVVThx4gSuX7+O6dOno7KyEsCDYic4OFjrdbt27cLGjRu12pKSkvDSSy/B09MTQggUFxcjIiIC4eHhUCgUiIyMRGlpKc6dO4fm5mbMmjUL169f73IfADQ3DKhUqu4bnHbk5ubC29sbcrm8w7izZ88CAKZNm9al/hQKBfLz87Fs2TJYWFgAeLBWOQBYW1trxUqlUshkMs17p2tcZyiVSpSXl2Pnzp04fvw4duzYoclP1/142MSJE1FeXo6ff/650zn1Ziz+iIjI6BobG7F9+3YsWrQIYWFhsLW1xYQJE/DZZ5/h5s2brZao6wozMzPN0UUfHx+kpKSgrq4Oe/fu7Zbt+/v7o7a2FjExMd2yvfY0NDTgt99+0xw5a0tlZSVSU1MRGRkJX1/fxx4hfJz4+Hg4Oztj8+bNmjb1nbqPnlYFAHNzc81RR13jOmP48OFwdXVFbGwstm3bhpCQEL3342GjRo0C8GD98/6IxR8RERldUVER6uvrMWnSJK32Z599FhYWFp1ac1pXkyZNglwu1zq93BdUVVVBCNHhUT9fX19ERkZiwYIFyMvLg7m5eaf7y8rKQnp6Oo4ePap19E59zV1zc3Or1zQ1NUEmk+kV1xnXr19HVVUV9u3bh3/84x+YOHFiu9ditrcfD1OPaVeORvZmZsZOgIiISD21xqBBg1o9Z2dnh7q6uh7t39LSEtXV1T3aR3e7d+8egAe5t8fR0RG7d+/GuHHjutRXamoqtm/fjoKCAjzxxBNaz6mvjXx0bj2FQoF79+7B2dlZr7jOMDc3x9ChQzF79my4u7tj9OjRmulcdN2Ph6kLUfUY9zcs/oiIyOjs7OwAoM0i786dO3B1de2xvpVKZY/30RPUBUpHkxIPHTpUM7adtWPHDhw9ehT5+fltFufu7u6wtrbG77//rtWuvgbyySef1Cuuq7y8vDBw4EAUFRXptR8Pa2pqAoAuHY3szXjal4iIjG78+PEYNGgQfvzxR632M2fOoKmpCc8884ymzczMDEqlstv6LigogBACkydP7rE+eoKjoyMkEgnu3r3bbsw333yjmfpEX0IIrFu3DhcvXkR2dna7BZOZmRnmzp2LEydOaN3kkpeXB4lEornOUNc4Xd26dQsvv/xyq/YrV66gpaUFw4cP12s/HqYeUycnJ71y6itY/BERkdFJpVJERUUhKysLX331FWpra3Hx4kWsXLkSzs7OWL58uSbWy8sLNTU1yM7OhlKpRHV1daujSQDg4OCAiooKlJaWoq6uTlPMqVQq3L59G83Nzbhw4QLWrFkDNzc3hIeHd0sfeXl5BpnqRS6Xw8PDA2VlZW0+X1xcDCcnpzZvfggNDYWTkxPOnTvX7vZ/+eUXbNu2DV988QXMzc1brVn+8ccfa2JjYmJQWVmJDRs2oKGhAadOnUJiYiLCw8Ph7e2td5wu+VlZWeHYsWPIz89HbW0tlEolfvrpJ7z22muwsrLC22+/rfd+qKnHdMKECe3235ex+CMiol5hw4YNiI+PR1xcHIYMGYKZM2di5MiRKCgogJWVlSZu1apVeP7557FkyRJ4e3tj06ZNmtNzvr6+milbVq5cCUdHR/j4+GDu3LmoqakB8OA6rgkTJkAmk2H69OkYPXo0vvvuO61r57rah6H4+/ujqKiozTtlO5qjrqmpCVVVVR1OrKzPHHfjxo3D0aNHcezYMQwePBiLFy/G0qVL8emnn3YqTpf8pFIppk6dijfeeAMuLi6wtrZGUFAQRo4cidOnT2P8+PF674faDz/8ABcXl247Fd3bSEQvn8EwPT0dISEh/XaiRSKi/ki9ZmpGRoaRM9G2YsUKZGRk4NatW8ZOpZXO/N4VFxdj7Nix2Lt3L8LCwnR+nUqlwnPPPYfw8HAsXbq0M+n2KGPmd+vWLbi6umLz5s2IiorS+XV9qF7J4JE/IiIyKR3dINHXeHl5IS4uDnFxcVrLmXWkpaUF2dnZqKurQ2hoaA9nqD9j5xcbG4unnnoKERERBu/bUFj8ERER9WHr169HUFAQQkNDO7z5Q62goAAHDhxAXl7eY1cGMQZj5rd9+3acP38ehw8f7tKciL0diz8iIjIJ0dHR2Lt3L+7evQt3d3dkZmYaO6Vus2XLFkRERGDr1q2PjfXz88PXX3+ttXZxb2Ks/HJycnD//n0UFBTA3t7eoH0bGuf5IyIikxAfH4/4+Hhjp9FjZs+ejdmzZxs7jT4rICAAAQEBxk7DIHjkj4iIiMiEsPgjIiIiMiEs/oiIiIhMCIs/IiIiIhPSZ274UE8YSkREvd/p06cB8G+3PtRLinHM+qb2ltnrjXr9Ch+nTp3C9u3bjZ0GEfUjN27cwE8//YQ5c+YYOxUi6md626o2bcjo9cUfEVF360PLMBERdTcu70ZERERkSlj8EREREZkQFn9EREREJoTFHxEREZEJYfFHREREZEJY/BERERGZEBZ/RERERCaExR8RERGRCWHxR0RERGRCWPwRERERmRAWf0REREQmhMUfERERkQlh8UdERERkQlj8EREREZkQFn9EREREJoTFHxEREZEJYfFHREREZEJY/BERERGZEBZ/RERERCaExR8RERGRCWHxR0RERGRCWPwRERERmRAWf0REREQmhMUfERERkQlh8UdERERkQlj8EREREZkQFn9EREREJoTFHxEREZEJYfFHREREZEJY/BERERGZEBZ/RERERCaExR8RERGRCTEzdgJERD1JqVSivr5eq62hoQEAcPv2ba12iUQCOzs7g+VGRGQMLP6IqF+rqamBi4sLWlpaWj3n4OCg9e/nn38e+fn5hkqNiMgoeNqXiPo1JycnzJgxAwMGdPznTiKRYMmSJQbKiojIeFj8EVG/95e//OWxMQMHDsSiRYsMkA0RkXGx+COifm/x4sUwM2v/KpeBAwfixRdfxODBgw2YFRGRcbD4I6J+z8bGBnPmzGm3ABRCICwszMBZEREZB4s/IjIJYWFhbd70AQAWFhaYN2+egTMiIjIOFn9EZBLmzZsHuVzeqt3c3BwLFy6ElZWVEbIiIjI8Fn9EZBKkUikWLVoEc3NzrXalUolXXnnFSFkRERkeiz8iMhkvv/wylEqlVpuNjQ1mzZplpIyIiAyPxR8RmYwXXnhBa2Jnc3NzLFmyBBYWFkbMiojIsFj8EZHJMDMzw5IlSzSnfpVKJV5++WUjZ0VEZFgs/ojIpCxZskRz6tfJyQnTpk0zckZERIbF4o+ITMqUKVPg4uICAHj11Vcfu+wbEVF/02rG07KyMnz//ffGyIWIyCCeffZZlJeXY/DgwUhPTzd2OkREPSY4OLhVm0QIIR5uSE9PR0hIiMGSIiIiIqKe8UiZBwAZ7S522UYwEVG/kZmZicDAQGOnQQYSFBQEAMjIyDByJn2H+mAQ64G+qaODebzYhYhMEgs/IjJVLP6IiIiITAiLPyIiIiITwuKPiIiIyISw+CMiIiIyISz+iIiIiEwIiz8iIiIdHT58GLa2tvjmm2+MnUqvd/z4caxfvx4qlQoLFy6Em5sbpFIpXFxcEBAQgAsXLui9zbi4OPj4+MDGxgaWlpbw8vLCu+++i/r6+laxhYWFmDp1KuRyOZydnbFu3Trcv3+/03G6SEhIwJgxYyCTyWBlZYUxY8YgJiYGtbW1eu/HwYMHkZCQgJaWlk7l0hEWf0RERDrinHe62bBhA5KTkxEdHQ2VSoWTJ09i3759qKmpQWFhIRobGzFjxgxUVFTotd38/HysXr0apaWluHnzJuLj45GUlKSZx1GtqKgIs2fPhp+fH6qrq5GVlYU9e/Zg5cqVnYrT1cmTJ7Fs2TJcbwh8OgAAIABJREFUu3YNlZWV2LRpExISElpNLaXLfsyfPx9SqRR+fn64c+dOp/Jpl3hEWlqaaKOZiIiozwoMDBSBgYHGTqNbKRQK4evr22Pb72w9sHXrVjF69GjR2NgohBBCqVSKefPmacWcPXtWABBbtmzRa9v+/v6iublZqy04OFgAENeuXdO0hYSECHd3d6FSqTRtiYmJQiKRiF9//VXvOF0tXLhQs99qQUFBAoCoqKjQez+EECIiIkL4+voKpVKpVy4dvH/pPPJHRETUB+3evRtVVVXGTkNLcXExYmJisHHjRkilUgCAmZlZq9PkHh4eAICSkhK9tn/o0CEMHDhQq23IkCEAAIVCAQBobm5Gbm4uZs6cCYlEoombM2cOhBDIycnRK04fWVlZmv1Wc3FxAQCtU7q67IdabGwszp8/j6SkJL3zaQ+LPyIiIh0UFhbCzc0NEokEO3fuBACkpKTAysoKcrkcOTk5mDNnDmxsbODq6or9+/drXpucnAypVApHR0esWLECzs7OkEqlmDJlCs6cOaOJi4iIgIWFBYYNG6Zpe/PNN2FlZQWJRIKbN28CANasWYOoqCiUlJRAIpHAy8sLAHDkyBHY2Nhgy5YthhiSVpKTkyGEwPz58zuMa2xsBADY2Nh0uc/y8nLIZDK4u7sDAK5evYr6+nq4ublpxXl6egKA5lpDXeO66sqVK7Czs8OIESP02g81e3t7zJw5E0lJSd122QGLPyIiIh1MmzYN33//vVbbqlWrsHbtWjQ2NsLa2hppaWkoKSmBh4cHli1bBqVSCeBBURceHg6FQoHIyEiUlpbi3LlzaG5uxqxZs3D9+nUAD4qn4OBgrT527dqFjRs3arUlJSXhpZdegqenJ4QQKC4uBgDNzQEqlapHxuBxcnNz4e3tDblc3mHc2bNnATwY065QKBTIz8/HsmXLYGFhAQC4ceMGAMDa2lorViqVQiaTobKyUq+4zlAqlSgvL8fOnTtx/Phx7NixQ5OfrvvxsIkTJ6K8vBw///xzp3N6GIs/IiKibjBlyhTY2Nhg6NChCA0NRUNDA65du6YVY2ZmhrFjx8LS0hI+Pj5ISUlBXV0d9u7d2y05+Pv7o7a2FjExMd2yPX00NDTgt99+0xw5a0tlZSVSU1MRGRkJX1/fxx4hfJz4+Hg4Oztj8+bNmjb1nbqPnlYFAHNzc81RR13jOmP48OFwdXVFbGwstm3bhpCQEL3342GjRo0CgP+PvXuPi7LM/8f/GmGGgeFoAhKIclARxUNpCWriWj5KVkUTxXSTWl1PLZBsudrHRFRKbZEHqPXJdaks4qAuaImaH6LsW9rBAyx2ABKPBQgIKCADc/3+8MesIwwwMDCD83o+Hv7RPe/7ut73yd7eh+tCXl5ep3O6F4s/IiIiPWu+e9N850+bsWPHwsrKCj/99FNPpNWtSktLIYRo866fv78/IiIiEBwcjKysLEil0k73d/DgQaSlpeHYsWMad++a37lrbGxssU5DQwMsLS11iuuMK1euoLS0FMnJyXj//fcxZswYre9natuOezXv067cjbyXuV5aISIiok6xsLBAWVmZodPosvr6egB3t0cbJycn7N27F8OHD+9SXykpKYiLi0NOTg4efvhhjd+a35e8f2y92tpa1NfXw8XFRae4zpBKpXB0dMS0adPg4eGBIUOGqIdz6eh23Ku5EG3ex13F4o+IiMhAlEolbt68CTc3N0On0mXNBUpbgxI7OjrC3t6+S/0kJibi2LFjyM7OhrW1dYvfPTw8YGNjg0uXLmksb34vcuTIkTrFdZW3tzfMzMyQn5+v03bcq6GhAQC6dDfyXnzsS0REZCA5OTkQQmD8+PHqZebm5u0+LjZGTk5OkEgkqKqq0hpz+PBh9dAnuhJCYM2aNcjLy0NGRobWgsnc3BzTp0/Hl19+qfHhS1ZWFiQSifo9w47GdVR5eTmee+65FssLCgrQ1NSEAQMG6LQd92rep87OzjrlpA2LPyIioh6iUqlQWVmJxsZG5ObmIjIyEu7u7ggLC1PHeHt7o6KiAhkZGVAqlSgrK2txdwoA+vbti+vXr6O4uBg1NTVQKpXIysoy2FAvVlZW8PT0xNWrV1v9vbCwEM7Ozq1+/BAaGgpnZ2ecOXNGa/sXLlzAtm3bsGfPHkilUkgkEo0/b731ljp2/fr1KCkpwYYNG3D79m1888032L59O8LCwjB06FCd4zqSn0KhwPHjx5GdnY3q6moolUqcPXsWixcvhkKhwOrVq3XejmbN+9TPz09r/7pg8UdERNQBO3fuxLhx4wAAa9aswaxZs7B7927s2LEDwN3HhL/++iv27NmDqKgoAMDTTz+NgoICdRv19fXw8/ODpaUlJk2ahCFDhuDzzz/XeE9u5cqVmDJlChYsWIChQ4di06ZN6sd9/v7+6mFhVqxYAScnJ/j6+mL69OmoqKjokf3QlqCgIOTn57f6pWxbY9Q1NDSgtLS0zYGVdRnjbvjw4Th27BiOHz+Ohx56CM8++yxefPFFvP32252K60h+crkcEyZMwJIlS+Dq6gobGxuEhIRg0KBBOHXqFEaMGKHzdjT77rvv4OrqqrdH0RJxXxZpaWmYP38+5y8kIqIHRvOcqenp6QbLYfny5UhPT0d5ebnBctBFZ+qBwsJCDBs2DElJSVi0aFGH11OpVAgMDERYWBhefPHFzqTbrQyZX3l5Odzc3LB582b1Pyo6oo3jl847f0RERD2krY8hHgTe3t6IiYlBTEyMxnRmbWlqakJGRgZqamoQGhrazRnqztD5RUdHY/To0QgPD9dbmz1W/L311lvql0HfeecdndYdN24czMzMMHr0aIPlQMZB2zE8cuQI7OzsWswf2R1UKhV27NiBgICADsUvWbIENjY2kEgkOHfuXKf7/eWXX/DXv/4Vw4cPh62tLWQyGRwdHeHj44M5c+bg3//+tzq2N11vycnJkEgkHd6fvdWBAwfg6emp8W6PVCqFq6srFi5ciB9//LHb+jb266a1fSORSCCTyeDk5ITAwEBs374dlZWV3Z4ndd3atWsREhKC0NDQNj/+aJaTk4MDBw4gKyur3ZlBDMGQ+cXFxeHcuXM4cuRIl8ZEbEHcJzU1VbSyWC8KCgoEAPH222/rvO7UqVPFqFGjDJoDGYfWjuEnn3wibG1txaFDh7q1719++UVMmDBBANDpfPz4448FAHH27NlO9ZuUlCRkMpmYOHGiOHr0qKisrBT19fWiqKhIHD58WAQFBYlly5ZprNNbrregoCDh5eUlAIiCgoIu92nsvLy8hJ2dnRBCiFu3bolDhw4Jd3d3YW1tLX766adu67c3XDf37huVSiUqKyvF559/LsLCwoREIhEuLi7iu+++61QOc+fOFXPnzu3Uuvqwdu1aIZPJBAAxaNAgkZ6ebrBcOqqr9cCxY8fEmjVr9JiRacnIyBCxsbGisbGxU+u3cfzSetU4fxKJxNApkJEKCgrq0L8wu+L8+fOIiYnBihUrcPv27R57L/bUqVNYsmQJJk2ahM8++wzm5v+9bD09PeHp6QlfX19s27ZNr/32xPVWXl6OCxcuYNOmTVi0aBE++OADxMTEdHu/xkKhUGDGjBloamrC7NmzkZiYiJ07d/ZY/8Z83UgkEtjb2yMwMBCBgYEICgrC/PnzERQUhF9++QV2dnbdmre+xcbGIjY21tBp9Khp06Zh2rRphk6j15o1axZmzZrVLW1322NfIQTS09Px7rvv6q1Nvd7yNBHdcRxMQWv7bdSoUThw4AAWLlzY5gj2relKIbVlyxY0NTXhjTfe0Cj87uXp6an3Vxl64npLS0tDUFAQZs6cCblcjn379vWaj830eW099thjAID//Oc/XW7LkPR93dxr7ty5CAsLQ2lpKV/bIeoivRR/TU1NiI2NxdChQ2FpaYl+/frBw8MDsbGxmDdvXpvrCiEQFxennujawcEBwcHBrc5zWFhYCB8fHygUCvVn8l999ZVGzMmTJ+Hr6ws7OzvI5XL4+fnh2LFj+tjMNtseNmwYJBIJ+vTpg0cffRS1tbUAgFdffVUd/9577wG4u79ef/11uLu7w9LSEiNHjkRqaioAYNu2bbCysoKNjQ1KS0sRFRUFV1dX/Pzzz+1uW0ePQ1v9d1RCQgLkcjmcnJywfPlyuLi4QC6XIyAgAKdPn9aI7egx1uVcuNdXX30Fd3d3SCQS9V2T3bt3Q6FQwMrKCpmZmXjmmWdga2sLNzc3fPzxxxrrd+X8bY0QAtu3b8fQoUNhYWEBOzs7vPLKKy3ijh492u54XA0NDThx4gT69u2rMQhsZxnb9ZacnIw5c+bAxsYG06ZNQ3FxMU6ePKkRYwrXVvPcovcWR6Z23XRE81h4WVlZ3dI+kcnQ4RmxVlu2bBFmZmYiMzNT1NbWih9++EE4OzuLwMBAjbjW3jl5/fXXhUwmE/v27RM3b94Uubm54pFHHhH9+vUTv//+uzpu6tSpwtPTU1y8eFEolUrxn//8Rzz++ONCLpeLX375RR2Xnp4uoqOjRUVFhSgvLxfjx48XDz30UJs5dFRbbTc2NopBgwYJd3f3Fs/nX375ZbFjxw71f//tb38TFhYWYv/+/aKyslKsW7dO9OnTR/0uy2uvvSYAiIiICJGYmCjmzJkjfvzxx3a3raPHob3+O2rZsmVCoVCICxcuiPr6epGfny/GjRsnbGxsxOXLl9VxHT3GHY1r7RheuXJFABCJiYnqZc378f/+7/9EVVWVKC0tFZMmTRIKhUI0NDTovN/u9fjjj2t9d+m1114TEolE/OMf/xCVlZWitrZW7Nq1q8U7f5988omwsbERMTExWvv55ZdfBAAxfvx4rTHaGPv1dunSJeHo6Ki+Xvbt2ycAiD//+c8acQ/atXXve23Nmrf9lVde0flYPSjXjbZ9c6/q6moBQAwYMEBrjDaGfuevN+rObwCo+7X1zp9eir9x48aJxx57TGPZX/7yF9GnTx9x584d9bL7//Kpra0V1tbWIjQ0VGPdb7/9VgDQ+J9iay+g5+bmCgDib3/7m9bcYmNjBQBRWlraag5dcX/bO3bsEABEWlqaOub27dvC3d1dVFVVCSGEqKurE1ZWVhrbXFtbKywsLMTKlSuFEP/9y7eurk6n/jtyHDrSf0ctW7asxV/U3333nQAgNm7cqG67I8dYl3NB1/+J3bsfm4uwwsJC9bKOnr/30vY/sdraWmFlZSWeeuopjeWd/eDj+++/FwDEk08+qdN6Qhj/9fbGG2+IF154Qf3fVVVVwsLCQtja2ora2lqN2Afp2rr/g4/9+/cLZ2dn4eTkJK5evapez5Sum9b2jTYSiUTY29u3GdMaFn+6Y/HXu3X7Bx/19fWQy+Uay5qamiCVSmFmZqZ1vfz8fNy6dQtjx47VWD5u3DjIZLIWjw/v5+fnBzs7O+Tm5mqNaX5vqTvGVrq/7SVLliA6Ohrx8fHqAUU//PBDBAcHw9bWFgDw888/o7a2Vj3SN3B3oub+/fu3+5imvf47chz02X9rxo4dCysrK3VbHT3GXT0XOkomkwGAxryZnT1/W1NYWIja2lpMnTq168kC6jkfb9++3ervaWlpWLNmDYqLiwEAPj4++OKLL+Dk5NQi1tiut+TkZI0X4G1tbTFt2jQcPnwYmZmZGuNpPWjXVlVVFSQSCczMzNC/f39Mnz4dGzZsUM95amrXTUc1fzDSfMx1derUKfX5Q+1rnlKM+6x30jbNHqCnd/6mT5+OH374AZmZmairq8P333+PjIwM/PGPf2zzL4GbN28CQKuTGtvb26OmpqbdvqVSqcZfSJ9++ikCAwPh6OgICwsLvPrqq53Yota117a1tTX+8pe/4Ouvv8a3334LAHj77bc1BmZs/p/4//zP/2iMZ3Xp0iX1u0yd7b8jx6Er/XeUhYUFysrKAHT8GOvjXOiszp6/rWm+2BwdHfWS28CBA2FhYYHCwsJWf583bx4uXryIgQMHwtnZGT/++GOrhR9gXNfbf/7zH+Tl5WHGjBka52HzeHMffPCBRvyDdm3Z2dlBCIHGxkZcvXoV//rXvzBw4ED176Z23XTUL7/8AuDuP3KIqPP0cucvOjoaP/zwA8LCwnDr1i24uLhg3rx57U4sbW9vDwCt/gV18+ZNuLm5tbl+Y2MjKioq4O7uDgC4fPkyZs+ejTlz5uBf//oXHn74YSQmJuqlAOxo2+Hh4YiPj8eOHTuwYsUKDBgwAF5eXurfm4uCHTt2IDIyUq/9d+Q4dLb/jlIqlRrHrqPHuKvnQld09vxtTfOdkDt37uglN7lcjieffBKffvopTp061aWPPozpevvoo4+wYMECJCcnayyvrKyEq6srjh8/jt9//x39+/dX/2ZK15apXTcddfToUQDAM88806n1x48fb9Dp3Xqb5unBuM96p+bj1xq9FH/5+fkoKipCWVmZ1qEoWjNixAhYW1vj+++/11h++vRpNDQ04NFHH21z/c8//xwqlQqPPPIIACAvLw9KpRIrV66Ep6cnAP2NVdbRtt3c3DBv3jykpqbi+vXr2LBhg8bvAwYMgFwu13mmh47035Hj0Nn+OyonJwdCCHWR0tFj3NVzoSs6e/62ZsSIEejTpw+++OILrFixQi/5bdy4EcePH8crr7yC7OzsTg/BYizXmxACKSkp+PDDD1v85uDggJCQEHzwwQdITk7G6tWr1b+Z0rVlatdNR/z+++/YsWMH3NzcjHLuV6LeRC+PfV966SW4u7t3eB6/ZnK5HFFRUTh48CA+/PBDVFdXIy8vDytWrICLiwuWLVumEd/Q0ICqqio0NjbizJkzCA8Px8CBA9Wf/zffkThx4gTq6+tRUFCgt3dedGk7KioKjY2NqKysxB/+8IcW2/zCCy/g448/xu7du1FdXY2mpiZcvXoVv/32W5f678hx6Gz/2qhUKlRWVqKxsRG5ubmIjIyEu7u7+ph09Bjrei7oU2fP39Y4Ojri2Wefxf79+7F3715UV1cjNze31bHgsrKy2h3qBQAeffRR7Nu3Dz/88AMCAwNx9OhR/Pbbb2hsbMSlS5ewb98+VFRUtJubsVxvX3/9NWxtbTFhwoRWf28umu9/9AuYzrVlatfNvYQQuHXrFlQqFYQQKCsrQ2pqKiZMmAAzMzNkZGR0+p0/Ivr/6fB1iFbZ2dnioYceEgDUf6RSqRg2bJg4cOCAEEKIf/zjH8LZ2VkAEAqFQsyZM0cIcXcKn+3bt4vBgwcLqVQqHBwcxOzZs8XPP/+s0UdSUpKYMmWKcHJyEubm5uKhhx4SCxYsEJcuXdKIW7Nmjejbt6+wt7cXISEhYufOnQKA8PLyEpGRka3m0FFttX3v0CZCCDFlyhTxz3/+s9V27ty5I9asWSPc3d2Fubm5cHR0FM8++6zIz88XW7duFZaWlurhDPbt29fh/jtyHNrrXxfLli0TUqlUuLq6CnNzc2FrayuCg4NFUVGRRlxHj3FH4lo7jxITE0X//v0FAGFlZSVmzpwpdu3aJaysrAQAMXjwYFFUVCTeffddYWtrKwCIgQMHqocs6eh+++abb8SECROEi4uLOq5///4iICBAfPHFF+q4mpoasWTJEvHQQw8Ja2trMXHiRPH6668LAMLNzU2cP39eCCHEkSNHhI2Njdi8eXOH9vfFixdFZGSkGD58uFAoFEIulwsPDw8xadIk8fe//118+eWXbe4nXY5Fd11vf/7zn4VCoRDm5uZi1KhR4syZMxrtbdq0SWP/urq6il27dmnE9NZr6//9v/8nhgwZol7fxcVFhISEaD3epnTdHDp0SIwcOVJYWVkJmUwm+vTpIwCov+x97LHHRExMjCgvL9e6v9rDr311x699e7duH+pl165dIjIyUmPZnTt3xMsvvywsLCxaDNtA3aOnj8OyZctE37599dqmIfD8pfbwHGmpt+0TFn+6Y/HXu3XrUC+///47wsPDW7znIpPJ4O7uDqVSCaVSCUtLy652RW0w1HHojiF0ehLPX2oPz5GWuE+Iercuv/NnaWkJqVSKvXv3oqSkBEqlEtevX8c///lPvP766wgNDTXa9zN++uknjSEZtP25d7wxY6WP4/Ag7Y+O6s3nL/UMniMtcZ+QIZw4cQJr166FSqXC7Nmz4e7uDrlcDldXV8yaNavNMUi12bp1K3x8fGBpaQmFQgEfHx+sX78e1dXVGnExMTHw9fWFra0tLCws4O3tjVdffVXjnddDhw5h69atveOmiA63CbX68ssvxZNPPilsbW2FmZmZsLOzEwEBAWLXrl1CqVR28oYl6aonj8PatWuFTCYTAMSgQYNEenq6XtvvSTx/qT08R1rqbfuEj311Z0yPfV9//XUxY8YMUV1dLZRKpXjooYfEyZMnxe3bt8Wvv/4qnnrqKWFnZyeuXbumU7tBQUHirbfeEqWlpaKmpkakpaUJqVTaYpamyZMni127dony8nJRXV0tUlNThVQqFU8//bRGXHx8vJg8ebKorKzs8jZ3Vbe/80dERGTMjKH4q62tFf7+/r2mD2OpB9544w0xZMgQ9ZSDSqVS/PGPf9SIaZ7ScMuWLTq1PXv27BbTPYaEhAgA4vr16+plQUFBLeYWnzdvngDQ4oPP8PBw4e/vb/B/BLVV/OllqBciIiJq2969e1FaWtrr++hJhYWFWL9+PTZu3KgeRN/c3Fw9G1Cz5jE6i4qKdGr/4MGDLaYpbJ5m8d5Hup988kmLmWv69esHAC1m8ImOjsa5c+cQHx+vUy49icUfERFRK4QQiIuLw7Bhw2BhYQEHBwcEBwdrzNUcHh4OmUymMRvNqlWroFAoIJFIcOPGDQBAZGQkoqKiUFRUBIlEAm9vbyQkJEAul8PJyQnLly+Hi4sL5HI5AgICNMaa7EofwN2ZUToypqgxSkhIgBACM2fObDOurq4OAPTyrmlBQQHs7e01plxszbVr12BpaQkPDw+N5Q4ODpg8eTLi4+MhhOhyPt2BxR8REVEroqOjsXbtWrz22msoLS3Fl19+iStXrmDSpEkoKSkBcLc4mTdvnsZ6u3btwsaNGzWWxcfHY8aMGfDy8oIQAoWFhQgPD0dYWBhqa2sRERGB4uJinDlzBo2NjXjqqadw5cqVLvcB/HdUBpVKpb+d00M+/fRTDB06FFZWVm3GNc/5PXHixE71o1Qqce3aNezcuRMnTpxAYmIiZDKZ1vja2lpkZ2dj6dKlrcaNGTMG165dw/nz5zuVT3dj8UdERHSfuro6xMXFYc6cOVi0aBHs7Ozg5+eHd955Bzdu3Gh11p7OMjc3V99d9PX1xe7du1FTU4OkpCS9tB8UFITq6mqsX79eL+31lNu3b+PixYsac3jfr6SkBCkpKYiIiIC/v3+7dwi1GTBgANzc3BAdHY1t27ZpnRO3WWxsLFxcXLB58+ZWfx88eDCAu9NHGiMWf0RERPfJz8/HrVu3MHbsWI3l48aNg0wm09vUoa0ZO3YsrKysNB4vm6LS0lIIIdq86+fv74+IiAgEBwcjKyur03OfX7lyBaWlpUhOTsb777+PMWPGaH138uDBg0hLS8OxY8dgY2PTakxzzs13iI1N98/GTURE1MvcvHkTAGBtbd3iN3t7e9TU1HRr/xYWFigrK+vWPoxdfX09gLv7QhsnJyfs3bsXw4cP71JfUqkUjo6OmDZtGjw8PDBkyBDExsa2+GgjJSUFcXFxyMnJwcMPP6y1veYBzpu3wdiw+CMiIrqPvb09ALRa5N28eRNubm7d1rdSqez2PnqD5gKqrUGTHR0d1cdKX7y9vWFmZob8/HyN5YmJiTh27Biys7Nb/UfBvRoaGgDAaGe54WNfIiKi+4wYMQLW1tb4/vvvNZafPn0aDQ0NePTRR9XLzM3NoVQq9dZ3Tk4OhBAYP358t/XRGzg5OUEikaCqqkprzOHDh9VDs+iqvLwczz33XIvlBQUFaGpqwoABAwDc/ep7zZo1yMvLQ0ZGRruFHwB1zs7Ozp3Krbux+CMiIrqPXC5HVFQUDh48iA8//BDV1dXIy8vDihUr4OLigmXLlqljvb29UVFRgYyMDCiVSpSVleHSpUst2uzbty+uX7+O4uJi1NTUqIs5lUqFyspKNDY2Ijc3F5GRkXB3d0dYWJhe+sjKyuqVQ71YWVnB09MTV69ebfX3wsJCODs7t/pxRmhoKJydnXHmzBmt7SsUChw/fhzZ2dmorq6GUqnE2bNnsXjxYigUCqxevRoAcOHCBWzbtg179uyBVCptMeXpW2+91aLt5pz9/Pw6s+ndjsUfERFRKzZs2IDY2FjExMSgX79+mDx5MgYNGoScnBwoFAp13MqVKzFlyhQsWLAAQ4cOxaZNm9SP+/z9/dVDtqxYsQJOTk7w9fXF9OnTUVFRAeDue2F+fn6wtLTEpEmTMGTIEHz++eca77p1tY/eKigoCPn5+epx/O7V1hh6DQ0NKC0tRWZmptYYuVyOCRMmYMmSJXB1dYWNjQ1CQkIwaNAgnDp1CiNGjGi3H22+++47uLq6YuTIkTqv2xMk4r6tSktLw/z58412YEIiIiJdhYSEAADS09MNnImm5cuXIz09HeXl5YZOpQVjqAcKCwsxbNgwJCUlYdGiRR1eT6VSITAwEGFhYXjxxRe7McOWysvL4ebmhs2bNyMqKqpH+75XG8cvnXf+iIiIDKitDxpMnbe3N2JiYhATE6Mx3VpbmpqakJGRgZqaGoSGhnZzhi1FR0dj9OjRCA8P7/G+O4rFHxERERmttWvXIiQkBKGhoW1+/NEsJycHBw4cQFZWVrszg+hbXFwczp07hyNHjnR6zMGewOKPiIjIANatW4ekpCRUVVXBw8MD+/fvN3RKRmvLli0IDw/HG2+80W7s1KlT8dFHH2nMhdwTMjMzcefOHeTk5MDBwaFH+9YVx/kjIiIygNjYWMTGxho6jV5j2rRpmDZtmqHT0GrWrFmYNWuWodMBD7IMAAAgAElEQVToEN75IyIiIjIhLP6IiIiITAiLPyIiIiITwuKPiIiIyISw+CMiIiIyIVq/9pVIJD2ZBxERUbfj/9t0x3324GlR/AUEBCA1NdUQuRAR9YhvvvkG8fHx/LuOiExSi7l9iYgedMYwZykRkYFwbl8iIiIiU8Lij4iIiMiEsPgjIiIiMiEs/oiIiIhMCIs/IiIiIhPC4o+IiIjIhLD4IyIiIjIhLP6IiIiITAiLPyIiIiITwuKPiIiIyISw+CMiIiIyISz+iIiIiEwIiz8iIiIiE8Lij4iIiMiEsPgjIiIiMiEs/oiIiIhMCIs/IiIiIhPC4o+IiIjIhLD4IyIiIjIhLP6IiIiITAiLPyIiIiITwuKPiIiIyISw+CMiIiIyISz+iIiIiEwIiz8iIiIiE8Lij4iIiMiEsPgjIiIiMiEs/oiIiIhMCIs/IiIiIhPC4o+IiIjIhLD4IyIiIjIhLP6IiIiITIi5oRMgIupOZWVl+Pe//62x7PvvvwcAvPvuuxrLbWxssGDBgh7LjYjIECRCCGHoJIiIusudO3fg5OSEW7duwczMDADQ/NeeRCJRxymVSixevBjvvfeeIdIkIuop6XzsS0QPNAsLC8ydOxfm5uZQKpVQKpVobGxEY2Oj+r+VSiUA4LnnnjNwtkRE3Y/FHxE98J577jk0NDS0GWNvb48//OEPPZQREZHhsPgjogfelClT4OjoqPV3qVSKRYsWwdycr0ET0YOPxR8RPfD69OmDhQsXQiqVtvq7Uqnkhx5EZDJY/BGRSViwYIH63b77Pfzww/D39+/hjIiIDIPFHxGZhMceewwDBw5ssVwmk2Hx4sUaX/4SET3IWPwRkcn405/+1OLRb0NDAx/5EpFJYfFHRCZj4cKFLR79ent7w8/Pz0AZERH1PBZ/RGQyfHx84Ovrq37EK5VK8cILLxg4KyKinsXij4hMyvPPP6+e6aOxsZGPfInI5LD4IyKTsmDBAjQ1NQEAHnnkEXh4eBg4IyKinsXij4hMiru7Ox5//HEAwOLFiw2cDRFRz+Nw9tQtQkJCDJ0CkVZ37tyBRCLB8ePH8eWXXxo6HaJW+fv7Y/Xq1YZOgx5AvPNH3WL//v24evWqodMgapWbmxucnZ0hl8v10h7Pd92dOnUKp06dMnQaRuvUqVP45ptvDJ0GPaB454+6zcsvv4x58+YZOg2iVhUWFsLb21svbUkkEp7vOmp+OpCenm7gTIwTn55Qd+KdPyIySfoq/IiIehsWf0REREQmhMUfERERkQlh8UdERERkQlj8EREREZkQFn9EREbiyJEjsLOzw+HDhw2ditE7ceIE1q5dC5VKhdmzZ8Pd3R1yuRyurq6YNWsWcnNzdW5z69at8PHxgaWlJRQKBXx8fLB+/XpUV1drxMXExMDX1xe2trawsLCAt7c3Xn31Vdy6dUsdc+jQIWzdulU9mwyRMWHxR0RkJIQQhk6hV9iwYQMSEhKwbt06qFQqnDx5EsnJyaioqMBXX32Furo6PPHEE7h+/bpO7Z48eRJLly7F5cuXUVJSgk2bNmHr1q2YO3euRlx2djZeeuklFBcX48aNG4iNjUV8fLzG8CwzZ86EXC7H1KlTcfPmTb1sN5G+sPgjIjISQUFBqKqqwowZMwydCurq6hAQEGDoNFp48803kZKSgrS0NNjY2AC4OxPGxIkTYWVlBQ8PD2zZsgVVVVV47733dGpbJpNh1apVcHR0hLW1NUJCQhAcHIzPPvsMv/32mzrO2toay5YtQ9++fWFjY4N58+Zh9uzZOHr0KK5cuaKOi4iIwKhRozB9+nQ0NjbqZfuJ9IHFHxERtbB3716UlpYaOg0NhYWFWL9+PTZu3KiencXc3LzFY3JPT08AQFFRkU7tHzx4sMWsL66urgCg8Uj3k08+gZmZmUZcv379AAC1tbUay6Ojo3Hu3DnEx8frlAtRd2LxR0RkBL766iu4u7tDIpFg586dAIDdu3dDoVDAysoKmZmZeOaZZ2Braws3Nzd8/PHH6nUTEhIgl8vh5OSE5cuXw8XFBXK5HAEBATh9+rQ6Ljw8HDKZDP3791cvW7VqFRQKBSQSCW7cuAEAiIyMRFRUFIqKiiCRSNQDYh89ehS2trbYsmVLT+ySFhISEiCEwMyZM9uMq6urAwDY2tp2uc+CggLY29tj4MCBbcZdu3YNlpaW8PDw0Fju4OCAyZMnIz4+no/1yWiw+CMiMgITJ07E119/rbFs5cqVePnll1FXVwcbGxukpqaiqKgInp6eWLp0KZRKJYC7RV1YWBhqa2sRERGB4uJinDlzBo2NjXjqqafUjyITEhJaTEG3a9cubNy4UWNZfHw8ZsyYAS8vLwghUFhYCADqjxdUKlW37IP2fPrppxg6dCisrKzajPv2228B3N2nnaFUKnHt2jXs3LkTJ06cQGJiImQymdb42tpaZGdnY+nSpa3GjRkzBteuXcP58+c7lQ+RvrH4IyLqBQICAmBrawtHR0eEhobi9u3buHz5skaMubk5hg0bBgsLC/j6+mL37t2oqalBUlKSXnIICgpCdXU11q9fr5f2dHH79m1cvHgRXl5eWmNKSkqQkpKCiIgI+Pv7t3uHUJsBAwbAzc0N0dHR2LZtG+bPn99mfGxsLFxcXLB58+ZWfx88eDAAIC8vr1P5EOkbiz8iol6m+e5S850/bcaOHQsrKyv89NNPPZFWtyotLYUQos27fv7+/oiIiEBwcDCysrIglUo71deVK1dQWlqK5ORkvP/++xgzZozW9x8PHjyItLQ0HDt2TP0Byv2acy4pKelUPkT6Zm7oBIiIqPtYWFigrKzM0Gl0WX19PYC726ONk5MT9u7di+HDh3epL6lUCkdHR0ybNg0eHh4YMmSIejiXe6WkpCAuLg45OTl4+OGHtbZnaWmpsQ1Ehsbij4joAaVUKnHz5k24ubkZOpUuay6g2ho02dHREfb29nrt19vbG2ZmZsjPz9dYnpiYiGPHjiE7OxvW1tZtttHQ0ADgv9tAZGh87EtE9IDKycmBEALjx49XLzM3N2/3cbExcnJygkQiQVVVldaYw4cPq4dm0VV5eTmee+65FssLCgrQ1NSEAQMGALg7EPeaNWuQl5eHjIyMdgs/AOqcnZ2dO5Ubkb6x+CMiekCoVCpUVlaisbERubm5iIyMhLu7O8LCwtQx3t7eqKioQEZGBpRKJcrKynDp0qUWbfXt2xfXr19HcXExampqoFQqkZWVZbChXqysrODp6YmrV6+2+nthYSGcnZ1b/TgjNDQUzs7OOHPmjNb2FQoFjh8/juzsbFRXV0OpVOLs2bNYvHgxFAoFVq9eDQC4cOECtm3bhj179kAqlUIikWj8eeutt1q03Zyzn59fZzadSO9Y/BERGYGdO3di3LhxAIA1a9Zg1qxZ2L17N3bs2AEAGDlyJH799Vfs2bMHUVFRAICnn34aBQUF6jbq6+vh5+cHS0tLTJo0CUOGDMHnn3+u8Z7cypUrMWXKFCxYsABDhw7Fpk2b1I8j/f391cPCrFixAk5OTvD19cX06dNRUVHRI/uhLUFBQcjPz1eP43evtsbQa2hoQGlpKTIzM7XGyOVyTJgwAUuWLIGrqytsbGwQEhKCQYMG4dSpUxgxYkS7/Wjz3XffwdXVFSNHjtR5XaLuIBEcdZK6gUQiQWpqaosxxYgeRMZwvi9fvhzp6ekoLy83WA66aJ4HNz09vcPrFBYWYtiwYUhKSsKiRYs6vJ5KpUJgYCDCwsLw4osv6pxrV5SXl8PNzQ2bN29WF+0d0Zn9Q9RB6bzzR0T0gGjrY4gHgbe3N2JiYhATE6Mx3VpbmpqakJGRgZqaGoSGhnZzhi1FR0dj9OjRCA8P7/G+ibRh8UdERL3G2rVrERISgtDQ0DY//miWk5ODAwcOICsrq92ZQfQtLi4O586dw5EjRzo95iBRd2DxR0ZpyZIlsLGxgUQiwblz5wydjlFQqVTYsWMHAgICtMYolUrExsbC29sbMpkM9vb2GDFiBIqLi3Xq68CBA/D09GzxMrtMJoOTkxMCAwOxfft2VFZWdnGrSB/WrVuHpKQkVFVVwcPDA/v37zd0St1qy5YtCA8PxxtvvNFu7NSpU/HRRx9pzGfcEzIzM3Hnzh3k5OTAwcGhR/smag+LPzJK//znP7Fnzx5Dp2E0CgoK8MQTT2D16tWora3VGjd//nx88MEH+Oijj1BbW4sff/wRXl5eHX5E1uzZZ5/Fr7/+Ci8vL9jZ2UEIAZVKhdLSUqSlpcHDwwNr1qzB8OHD8f3333d186iLYmNjcefOHQghcPHiRcydO9fQKXW7adOm4c033zR0GlrNmjULa9euhZmZmaFTIWqBxR9RD6irq2vzjl1bzp8/j7///e9YsWIFRo8erTUuJSUFGRkZSE9Px+OPPw5zc3O4uLggMzNT/aViV0gkEtjb2yMwMBBJSUlIS0tDSUkJgoKCOvT4zdh15RgREfUmLP7IaEkkEkOnoDd79+7VOjdoe0aNGoUDBw5g4cKFbU5t9fbbb+ORRx7psbHE5s6di7CwMJSWluKdd97pkT67U1eOERFRb8Lij4yCEALbt2/H0KFDYWFhATs7O7zyyisaMdu2bYOVlRVsbGxQWlqKqKgouLq64ueff4YQAnFxcRg2bBgsLCzg4OCA4OBgjQntExISIJfL4eTkhOXLl8PFxQVyuRwBAQE4ffp0i3zaay88PBwymUzjXaJVq1ZBoVBAIpHgxo0bAIDIyEhERUWhqKgIEokE3t7eet9/DQ0NOHXqVJt3BpsdPXpUbwP1Ng8enJWVBYDHiIioVxBE3QCASE1N7XD8a6+9JiQSifjHP/4hKisrRW1trdi1a5cAIM6ePasRB0BERESIxMREMWfOHPHjjz+K119/XchkMrFv3z5x8+ZNkZubKx555BHRr18/8fvvv6vXX7ZsmVAoFOLChQuivr5e5Ofni3HjxgkbGxtx+fJldVxH21u4cKFwdnbW2Jbt27cLAKKsrEy97NlnnxVeXl467cPWPP7442LUqFEtll+8eFEAEKNHjxaBgYGif//+wsLCQvj4+IidO3cKlUqljv3kk0+EjY2NiImJabc/Ly8vYWdnp/X36upqAUAMGDBAvcwUj5Gu5zsJMXfuXDF37lxDp2G0uH+oG6Xxzh8ZXF1dHXbs2IEnn3wSq1evhr29PSwtLdG3b1+t67z55pt46aWXcODAAQwcOBBxcXGYM2cOFi1aBDs7O/j5+eGdd97BjRs38O6772qsa25urr5b5Ovri927d6OmpgZJSUnqfHRpzxg0f9Dh6OiILVu2ID8/HyUlJQgODsZLL72E5ORkdWxQUBCqq6uxfv36Lvfb/EV2TU1Ni994jIiIjJO5oRMgKiwsRG1tLaZOndqp9fPz83Hr1i2MHTtWY/m4ceMgk8laPC6839ixY2FlZaV+XNjV9gyh+V3A4cOHa3y0sHHjRrz99tt49913sXDhQr33e/v2bQghYGtr22acKRyj+fPntzqvLLXtQXq3V99M4attMgwWf2RwzZOeOzo6dmr9mzdvAgCsra1b/GZvb9/qXan7WVhYoKysTG/t9TQXFxcAUL/D1kwmk2HgwIEoKirqln5/+eUXAICPj0+bcaZwjCIjI+Hv79/j/fZWzXMWv/zyywbOxDg17x+i7sDijwxOLpcDAO7cudOp9e3t7QGg1f/h37x5E25ubm2ur1QqNeK62p4hWFtbY/Dgwbhw4UKL3xobG2FnZ9ct/R49ehQA8Mwzz7QZZwrHyN/fn3NZ66B5zlrus9ZxTl/qTnznjwxuxIgR6NOnD7744otOr29tbd1isOHTp0+joaEBjz76aJvr5+TkQAiB8ePH69yeubk5lEplp/LWt/nz5+Ps2bP49ddf1ctqa2tx6dKlbhn+5ffff8eOHTvg5uaGF198sc1YHiMiIuPB4o8MztHREc8++yz279+PvXv3orq6Grm5uR1+aV8ulyMqKgoHDx7Ehx9+iOrqauTl5WHFihVwcXHBsmXLNOJVKhUqKyvR2NiI3NxcREZGwt3dXT1siS7teXt7o6KiAhkZGVAqlSgrK8OlS5da5Ni3b19cv34dxcXFqKmp6ZZiZPXq1Rg4cCDCwsJw+fJllJeXY82aNairq8Pf//53dVxWVpZOQ70IIXDr1i2oVCoIIVBWVobU1FRMmDABZmZmyMjIaPedPx4jIiIjYtCPjemBBR2HvqipqRFLliwRDz30kLC2thYTJ04Ur7/+ugAg3NzcxPnz58XWrVuFpaWlemiRffv2qddXqVRi+/btYvDgwUIqlQoHBwcxe/Zs8fPPP2v0s2zZMiGVSoWrq6swNzcXtra2Ijg4WBQVFWnEdbS98vJyMWXKFCGXy4WHh4f461//Kl555RUBQHh7e6uHJjlz5owYOHCgsLS0FBMnTtQYiqQ933zzjZgwYYJwcXERAAQA0b9/fxEQECC++OILjdgrV66IBQsWCAcHB2FhYSEee+wxkZWVpRFz5MgRYWNjIzZv3qy1z0OHDomRI0cKKysrIZPJRJ8+fQQAIZFIhL29vXjsscdETEyMKC8v11jPVI+Rruc7cSiT9nD/UDdKkwghhKEKT3pwSSQSpKamGt37PMuXL0d6ejrKy8sNnQpp0RuPkbGe78YsJCQEAN9t04b7h7pROh/7kslpamoydArUDh4jIqLuw+KPqIf99NNPkEgk7f4JDQ01dKpERuvEiRNYu3YtVCoVZs+eDXd3d8jlcri6umLWrFnIzc3Vuc2tW7fCx8cHlpaWUCgU8PHxwfr161FdXa0RFxMTA19fX9ja2sLCwgLe3t549dVX1YOtA8ChQ4ewdetW/kOGjBKLPzIZ69atQ1JSEqqqquDh4YH9+/cbJA8fHx8IIdr9k5KSYpD8DMlYjhEZtw0bNiAhIQHr1q2DSqXCyZMnkZycjIqKCnz11Veoq6vDE088gevXr+vU7smTJ7F06VJcvnwZJSUl2LRpE7Zu3dpisOXs7Gy89NJLKC4uxo0bNxAbG4v4+Hj1o1oAmDlzJuRyOaZOnaoel5LIWLD4I5MRGxuLO3fuQAiBixcvcvR8I8Rj1Dl1dXUaM7v01j464s0330RKSgrS0tJgY2MD4O4YixMnToSVlRU8PDywZcsWVFVV4b333tOpbZlMhlWrVsHR0RHW1tYICQlBcHAwPvvsM/z222/qOGtrayxbtgx9+/aFjY0N5s2bh9mzZ+Po0aO4cuWKOi4iIgKjRo3C9OnT0djYqJftJ9IHFn9ERL3c3r17UVpa2uv7aE9hYSHWr1+PjRs3qgeHNzc3x+HDhzXiPD09AUDnmW0OHjyobreZq6srAGg80v3kk09gZmamEdevXz8Ad8fWvFd0dDTOnTuH+Ph4nXIh6k4s/oiIepgQAnFxcRg2bBgsLCzg4OCA4OBg9dzFABAeHg6ZTIb+/furl61atQoKhQISiUQ9lV9kZCSioqJQVFQEiUQCb29vJCQkQC6Xw8nJCcuXL4eLiwvkcjkCAgI05j3uSh/A3RledBkzsqsSEhIghMDMmTPbjKurqwOAdsef7IiCggLY29tj4MCBbcZdu3YNlpaW8PDw0Fju4OCAyZMnIz4+Hhxcg4wFiz8ioh4WHR2NtWvX4rXXXkNpaSm+/PJLXLlyBZMmTUJJSQmAu4XO/UPH7Nq1Cxs3btRYFh8fjxkzZsDLywtCCBQWFiI8PBxhYWGora1FREQEiouLcebMGTQ2NuKpp55SP5rsSh/Af7/KVqlU+ts5bfj0008xdOhQWFlZtRn37bffAgAmTpzYqX6USiWuXbuGnTt34sSJE0hMTIRMJtMaX1tbi+zsbCxdurTVuDFjxuDatWs4f/58p/Ih0jcWf0REPaiurg5xcXGYM2cOFi1aBDs7O/j5+eGdd97BjRs3OjyzTUeYm5ur7y76+vpi9+7dqKmpQVJSkl7aDwoKQnV1NdavX6+X9tpy+/ZtXLx4EV5eXlpjSkpKkJKSgoiICPj7+7d7h1CbAQMGwM3NDdHR0di2bRvmz5/fZnxsbCxcXFywefPmVn8fPHgwACAvL69T+RDpG4s/IqIelJ+fj1u3bmHs2LEay8eNGweZTKbxWFbfxo4dCysrK43Hy71FaWkphBBt3vXz9/dHREQEgoODkZWVBalU2qm+rly5gtLSUiQnJ+P999/HmDFjtL7vePDgQaSlpeHYsWPqD1Du15xz811dIkMzN3QCRESmpHnYD2tr6xa/2dvbo6amplv7t7CwQFlZWbf20R3q6+sB3M1fGycnJ+zduxfDhw/vUl9SqRSOjo6YNm0aPDw8MGTIEPVwLvdKSUlBXFwccnJy8PDDD2ttz9LSUmMbiAyNxR8RUQ+yt7cHgFaLvJs3b8LNza3b+lYqld3eR3dpLqDaGjTZ0dFRvX/1xdvbG2ZmZsjPz9dYnpiYiGPHjiE7O7vVQv5eDQ0NAP67DUSGxse+REQ9aMSIEbC2tsb333+vsfz06dNoaGjAo48+ql5mbm4OpVKpt75zcnIghMD48eO7rY/u4uTkBIlEgqqqKq0xhw8fVg/Noqvy8nI899xzLZYXFBSgqakJAwYMAHD3S+01a9YgLy8PGRkZ7RZ+ANQ5Ozs7dyo3In1j8UdE1IPkcjmioqJw8OBBfPjhh6iurkZeXh5WrFgBFxcXLFu2TB3r7e2NiooKZGRkQKlUoqysDJcuXWrRZt++fXH9+nUUFxejpqZGXcypVCpUVlaisbERubm5iIyMhLu7O8LCwvTSR1ZWVo8N9WJlZQVPT09cvXq11d8LCwvh7Ozc6scZoaGhcHZ2xpkzZ7S2r1AocPz4cWRnZ6O6uhpKpRJnz57F4sWLoVAosHr1agDAhQsXsG3bNuzZswdSqbTFtIxvvfVWi7abc/bz8+vMphPpHYs/IqIetmHDBsTGxiImJgb9+vXD5MmTMWjQIOTk5EChUKjjVq5ciSlTpmDBggUYOnQoNm3apH506O/vrx6yZcWKFXBycoKvry+mT5+OiooKAHffMfPz84OlpSUmTZqEIUOG4PPPP9d4b66rffSkoKAg5Ofnq8fxu1dbY+g1NDSgtLQUmZmZWmPkcjkmTJiAJUuWwNXVFTY2NggJCcGgQYNw6tQpjBgxot1+tPnuu+/g6uqKkSNH6rwuUXeQCI46Sd1AIpEgNTW1xRhiRA8iYzzfly9fjvT0dJSXlxs6lVY1z4Obnp7e4XUKCwsxbNgwJCUlYdGiRR1eT6VSITAwEGFhYXjxxRd1zrUrysvL4ebmhs2bNyMqKqrD63Vm/xB1UDrv/BERPaDa+jiiN/L29kZMTAxiYmI0pltrS1NTEzIyMlBTU4PQ0NBuzrCl6OhojB49GuHh4T3eN5E2LP6IiKjXWLt2LUJCQhAaGtrmxx/NcnJycODAAWRlZbU7M4i+xcXF4dy5czhy5Einxxwk6g4s/oiIHjDr1q1DUlISqqqq4OHhgf379xs6Jb3asmULwsPD8cYbb7QbO3XqVHz00Uca8xf3hMzMTNy5cwc5OTlwcHDo0b6J2sNx/oiIHjCxsbGIjY01dBrdatq0aZg2bZqh09Bq1qxZmDVrlqHTIGoV7/wRERERmRAWf0REREQmhMUfERERkQlh8UdERERkQvjBB3Wbb775xtApEPUYnu+6aZ7yLC0tzcCZGKerV6/Czc3N0GnQA4ozfFC3kEgkhk6BiKhXmzt3Lmf4oO6Qzjt/1C34bwoyZmlpaZg/fz7PUyIySXznj4iIiMiEsPgjIiIiMiEs/oiIiIhMCIs/IiIiIhPC4o+IiIjIhLD4IyIiIjIhLP6IiIiITAiLPyIiIiITwuKPiIiIyISw+CMiIiIyISz+iIiIiEwIiz8iIiIiE8Lij4iIiMiEsPgjIiIiMiEs/oiIiIhMCIs/IiIiIhPC4o+IiIjIhLD4IyIiIjIhLP6IiIiITAiLPyIiIiITwuKPiIiIyISw+CMiIiIyISz+iIiIiEwIiz8iIiIiE8Lij4iIiMiEsPgjIiIiMiEs/oiIiIhMCIs/IiIiIhPC4o+IiIjIhLD4IyIiIjIhLP6IiIiITAiLPyIiIiITwuKPiIiIyISYGzoBIqLudPXqVSxevBhNTU3qZZWVlbCxsUFgYKBG7NChQ/G///u/PZwhEVHPYvFHRA80Nzc3XLp0CUVFRS1+++KLLzT++4knnuiptIiIDIaPfYnogff8889DKpW2GxcaGtoD2RARGRaLPyJ64C1cuBCNjY1txgwfPhy+vr49lBERkeGw+COiB56XlxdGjhwJiUTS6u9SqRSLFy/u4ayIiAyDxR8RmYTnn38eZmZmrf7W2NiIkJCQHs6IiMgwWPwRkUlYsGABVCpVi+V9+vTB+PHjMWjQoJ5PiojIAFj8EZFJcHFxwYQJE9Cnj+Zfe3369MHzzz9voKyIiHoeiz8iMhl/+tOfWiwTQmDOnDkGyIaIyDBY/BGRyZg7d67Ge39mZmZ48skn4eTkZMCsiIh6Fos/IjIZDg4OeOqpp9QFoBACixYtMnBWREQ9i8UfEZmURYsWqT/8kEqlCA4ONnBGREQ9i8UfEZmUmTNnwsLCAgAwY8YMWFtbGzgjIqKexeKPiEyKQqFQ3+3jI18iMkUSIYQwdBJE2mibkYGIyFjNnTsX6enphk6DSJt0c0NnQNSeyMhI+Pv7GzoNeoA0NTUhNTUVmZmZPL90tGPHDgDAyy+/bOBMjFPz/iEyZiz+yOj5+/tj3rx5hk6DHjCzZ8+GpaUlzy8dNd/R4j5rHe/4UW/Ad/6IyCTJ5XJDp0BEZBAs/oiIiIhMCIs/IiIiIhPC4o+IiIjIhLD4IyIiIjIhLP6IiLroyJEjsLOzw+HDhw2ditE7ceIE1q5dC5VKhdmzZ8Pd3R1yuRyurq6YNWsWcnNzdW5z69at8PHxgaWlJRQKBXx8fLB+/XpUV1drxOUk8RAAACAASURBVMXExMDX1xe2trawsLCAt7c3Xn31Vdy6dUsdc+jQIWzduhVNTU1d3lYiY8Xij4ioizhWfsds2LABCQkJWLduHVQqFU6ePInk5GRUVFTgq6++Ql1dHZ544glcv35dp3ZPnjyJpUuX4vLlyygpKcGmTZuwdetWzJ07VyMuOzsbL730EoqLi3Hjxg3ExsYiPj4eISEh6piZM2dCLpdj6tSpuHnzpl62m8jYsPgjIuqioKAgVFVVYcaMGYZOBXV1dQgICDB0Gi28+eabSElJQVpaGmxsbADcHcNz4sSJsLKygoeHB7Zs2YKqqiq89957OrUtk8mwatUqODo6wtraGiEhIQgODsZnn32G3377TR1nbW2NZcuWoW/fvrCxscG8efMwe/ZsHD16FFeuXFHHRUREYNSoUZg+fToaGxv1sv1ExoTFHxHRA2Tv3r0oLS01dBoaCgsLsX79emzcuFE9vqK5uXmLx+Senp4AgKKiIp3aP3jwYItxG11dXQFA45HuJ598AjMzM424fv36AQBqa2s1lkdHR+PcuXOIj4/XKRei3oDFHxFRF3z11Vdwd3eHRCLBzp07AQC7d++GQqGAlZUVMjMz8cwzz8DW1hZubm74+OOP1esmJCRALpfDyckJy5cvh4uLC+RyOQICAnD69Gl1XHh4OGQyGfr3769etmrVKigUCkgkEty4cQPA3akQo6KiUFRUBIlEAm9vbwDA0aNHYWtriy1btvTELmkhISEBQgjMnDmzzbi6ujoAgK2tbZf7LCgogL29PQYOHNhm3LVr12BpaQkPDw+N5Q4ODpg8eTLi4+P5WJ8eOCz+iIi6YOLEifj66681lq1cuRIvv/wy6urqYGNjg9TUVBQVFcHT0xNLly6FUqkEcLeoCwsLQ21tLSIiIlBcXIwzZ86gsbERTz31lPpRZEJCQovp1Hbt2oWNGzdqLIuPj8eMGTPg5eUFIQQKCwsBQP3xgkql6pZ90J5PP/0UQ4cOhZWVVZtx3377LYC7+7QzlEolrl27hp07d+LEiRNITEyETCbTGl9bW4vs7GwsXbq01bgxY8bg2rVrOH/+fKfyITJWLP6IiLpRQEAAbG1t4ejoiNDQUNy+fRuXL1/WiDE3N8ewYcNgYWEBX19f7N69GzU1NUhKStJLDkFBQaiursb69ev10p4ubt++jYsXL8LLy0trTElJCVJSUhAREQF/f/927xBqM2DAALi5uSE6Ohrbtm3D/Pnz24yPjY2Fi4sLNm/e3OrvgwcPBgDk5eV1Kh8iY8Xij4iohzTfXWq+86fN2LFjYWVlhZ9++qkn0upWpaWlEEK0edfP398fERERCA4ORlZWFqRSaaf6unLlCkpLS5GcnIz3338fY8aM0fr+48GDB5GWloZjx46pP0C5X3POJSUlncqHyFiZGzoBIiJqycLCAmVlZYZOo8vq6+sB3N0ebZycnLB3714MHz68S31JpVI4Ojpi2rRp8PDwwJAhQ9TDudwrJSUFcXFxyMnJwcMPP6y1PUtLS41tIHpQsPgjIjIySqUSN2/ehJubm6FT6bLmAqqtQZMdHR1hb2+v1369vb1hZmaG/Px8jeWJiYk4duwYsrOzYW1t3WYbDQ0NAP67DUQPCj72JSIyMjk5ORBCYPz48epl5ubm7T4uNkZOTk6QSCSoqqrSGnP48GH10Cy6Ki8vx3PPPddieUFBAZqamjBgwAAAdwfiXrNmDfLy8pCRkdFu4QdAnbOzs3OnciMyViz+iIgMTKVSobKyEo2NjcjNzUVkZCTc3d0RFhamjvH29kZFRQUyMjKgVCpRVlaGS5cutWirb9++uH79OoqLi1FTUwOlUomsrCyDDfViZWUFT09PXL16tdXfCwsL4ezs3OrHGaGhoXB2dsaZM2e0tq9QKHD8+HFkZ2ejuroaSqUSZ8+exeLFi6FQKLB69WoAwIULF7Bt2zbs2bMHUqkUEolE489bb73Vou3mnP38/Dqz6URGi8UfEVEX7Ny5E+PGjQMArFmzBrNmzcLu3buxY8cOAMDIkSPx66+/Ys+ePYiKigIAPP300ygoKFC3UV9fDz8/P1haWmLSpEkYMmQIPv/8c4335FauXIkpU6ZgwYIFGDp0KDZt2qR+HOnv768eFmbFihVwcnKCr68vpk+fjoqKih7ZD20JCgpCfn6+ehy/e7U1hl5DQwNKS0uRmZmpNUYul2PChAlYsmQJXF1dYWNjg5CQEAwaNAinTp3CiBEj2u1Hm++++w6urq4YOXKkzusSGTOJ4OiVZMQkEglSU1NbjHFGpA/GcH4tX74c6enpKC8vN1gOumieBzc9Pb3D6xQWFmLYsGFISkrCokWLOryeSqVCYGAgwsLC8OKLL+qca1eUl5fDzc0NmzdvVhftHdGZ/UPUw9J554+IyMDa+hjiQeDt7Y2YmBjExMRoTLfWlqamJmRkZKCmpgahoaHdnGFL0dHRGD16NMLDw3u8b6LuxuKPHmhLliyBjY0NJBIJzp07Z+h09KK+vh4+Pj74n//5nxa/JScnY9y4cbCxscHAgQPxwgsv4Pfff9e5jwMHDsDT07PFe1EymQxOTk4IDAzE9u3bUVlZqY9NIhOwdu1ahISEIDQ0tM2PP5rl5OTgwIEDyMrKandmEH37/9i796ioq71/4O+vDDAMdxUQIYqLkRBmHS1BSz0cfVJSU1GxOB2OvzzeOqD5FCKRSoCaPciiNJ+McK1zUi7qQVNRVxGZJzNd5iXMChSPSnIR5TYjDM7+/eGaeZqG28hlgHm/1po/zv7u796f+TIrP2fv7947NTUVZ8+exaFDhx54z0Gi3ozJH/VrH3/8MbZv327qMLpUfHw8fvrpJ4Py7OxsvPzyy5gzZw6uX7+Offv24dixY5gyZQqam5uN6mP27Nm4fPkyfH194ejoCCEENBoNKioqkJOTA29vb8TGxiIwMBCnT5/uqq9mdlavXo3MzEzU1NTA29sbu3fvNnVI3So5ORnR0dFYv359u3VDQ0Px6aef6p1n3BP27duHxsZGFBYWwtnZuUf7JuopTP6I+pBvvvkGP/zwQ4vX/vd//xdDhw7FG2+8AUdHR4wcORKvv/46zp49i5MnT3a6b0mS4OTkhAkTJiAzMxM5OTkoLy9HWFhYh0ZyyFBKSgoaGxshhMCVK1cQHh5u6pC63eTJk7FhwwZTh9GqGTNmIC4uDhYWFqYOhajbMPmjfk+SJFOH0CVUKhXeeOMNg9MKtK5duwZ3d3e976vd46ylLUE6Kzw8HFFRUaioqMC2bdu6vH0iIuoeTP6oXxFCYNOmTfD394e1tTUcHR3xxhtvGNS7d+8e3n77bXh5ecHGxgYjRoxAdnY2AGDr1q2wtbWFQqHAvn37MGXKFDg4OMDT0xO7du3Sa+err77C008/DYVCAQcHBwQFBaG2trbdPh5EfHw8li1bBhcXlxav+/j4GJxjqn3fz8fHR1d2+PDhLtvzTbsPXX5+vq6sLz5bIiJzwuSP+pWEhATExsZi0aJFKC8vx82bN7Fq1SqDeqtWrcK7776LzZs349dff8W0adPw0ksv4fTp01i6dClWrFgBlUoFe3t7ZGdno6SkBD4+Pli4cKHulIWGhgZMnz4d4eHhqK6uxi+//IJHH31UdyRUW30Y69///jdKSkpaPMlAa/Xq1bh58ybef/991NXVoaioCGlpafiv//ovvZMitCtLNRqN0XH83siRIwEAly9f1pX1tWdLRGR2BFEvBkBkZ2d3qK5SqRQKhUJMmjRJr3zXrl0CgPj++++FEEKoVCqhUChERESE3r3W1tZi6dKlQggh4uPjBQChUql0dbZs2SIAiOLiYiGEED/88IMAIA4cOGAQS0f66CilUilGjRolrl+/LoQQorKyUgAQ8fHxBnXfeustAUD38fT0FNeuXTOqv9/y9fUVjo6ObdaRJEk4OTkJIfreszXm90X3hYeHi/DwcFOH0Wvx+VAfkCMzScZJ1A2Ki4uhVCoRGhraZr2ffvoJSqVSt/M/cP/g9iFDhuDSpUut3mdlZQUAutEpHx8fuLq6IjIyEjExMYiKisIjjzzSqT5asnr1avztb39r9+zT+Ph4fPzxx/jiiy/wzDPPoKKiAqtWrUJwcDC++eYb3ft/XamhoQFCCDg4OADoe88WAE6cOGH0PeZMe+RZTk6OiSPpna5fvw5PT09Th0HUNlOnn0RtgREjM4cOHRIAREZGhl7570f+/v3vf+uNjv32M2bMGCFEy6NT27dvFwDEjz/+qCv74YcfxAsvvCBkMpmQJEnMmzdPKJXKDvXREV9//bUIDQ0VGo1GV9bSyF9ZWZmwsLAQb731lt79NTU1YsCAAeLvf/97h/v8rfZG/s6cOSMAiMmTJwsh+tazFUK02g4//HTmw5E/6uVy+M4f9RtyuRwA0NjY2GY97YKJzZs3Qwih9zF2FCgwMBCfffYZysrKEBsbi+zsbLz33ntd1kdGRga++OILDBgwQLfRsrbt5ORkSJKE06dP45dffsG9e/cwdOhQvfsdHBwwcOBAFBUVGfW9Ourw4cMAgClTpgDoW89WKzs726Adflr/hIeHIzw83ORx9NaPOWzXQ30fkz/qNx5//HEMGDAAX331VZv1HnroIcjl8k6f+FFWVoaLFy8CuJ/0rF+/Hk899RQuXrzYZX1kZmYa/ONSWVkJ4P40rxACo0aN0k0z/frrr3r319XVobq6ulumfG/evInNmzfD09NTd+5qX3q2RETmiskf9RsuLi6YPXs2du/ejYyMDNTW1uL8+fP46KOP9OrJ5XL89a9/xa5du7B161bU1tbi3r17uH79ukHy1JaysjIsXrwYly5dQlNTE77//ntcvXoVY8aM6bI+Osrb2xsTJ07E9u3bcezYMahUKly7dg2LFi0CAPy///f/dHXz8/ON2upFCIH6+npoNBpd8pmdnY2xY8fCwsICeXl5unf++uOzJSLqdwRRLwYYtxqzrq5OvPrqq2LQoEHCzs5OjBs3Trz99tsCuL/y9dy5c0IIIRobG0VsbKzw8vISMplMuLi4iNmzZ4uioiKxZcsWoVAoBAAxbNgwUVJSIj766CPh4OAgAIiHH35Y/Pzzz6K0tFSEhIQIZ2dnYWFhIYYOHSri4+NFc3Nzu310RmurfauqqsTy5cuFn5+fsLa2FnZ2dmLs2LHiX//6l169Q4cOCXt7e5GUlNRqH/v37xcjRowQCoVCWFlZiQEDBggAupW9Tz/9tEhMTBS3bt0yuLcvPVtjf1/E1azt4fOhPiBHEkIIk2WeRO2QJAnZ2dmYO3euqUOhfoi/L+PNmTMHAJCbm2viSHonPh/qA3I57UtERERkRpj8EfWwS5cu6VbutvWJiIgwdahERNQPMfkj6mGPPfZYh7aMyMrKMnWoRF3u888/R1xcHDQaDWbOnAkvLy/I5XJ4eHhgxowZOH/+vNFtJiYmIiAgAA4ODrC2toafnx/efPNN1NfXG9Q9fvw4xo4dC4VCAXd3d8TGxuptD7V//35s3LhRdwwiUX/E5I+IiHrEmjVrkJ6ejtWrV0Oj0eDrr7/Gzp07UV1djePHj0OlUuG5555DWVmZUe0WFBTgtddeQ2lpKaqqqpCSkoK0tDTd+3daRUVFmDx5MkJDQ1FZWYm9e/fik08+wZIlS3R1pk+fDrlcjtDQUNy5c6dLvjdRb8Pkj4jIRFQqFUJCQvp8Hx2xYcMGZGVlIScnB/b29gCA4OBgjBs3DgqFAt7e3khOTkZNTQ127NhhVNt2dnZYtGgRBg4cCHt7e8ydOxczZ87E4cOHce3aNV29d955B0OGDMG6detga2uL4OBgxMbGYseOHXpHA8bExOCJJ57A1KlT0dzc3CXfn6g3YfJHRGQiGRkZqKio6PN9tKe4uBgJCQlYt26d7iQemUyGzz77TK+ej48PAKCkpMSo9g8cOAALCwu9ssGDBwMAlEolAKC5uRkHDx7E+PHjIUmSrt6UKVMghMC+ffv07l+7di3Onj2LtLQ0o2Ih6guY/BERdZAQAqmpqRg+fDisra3h7OyMF198UW/UKDo6GlZWVhgyZIiubNmyZbC1tYUkSaiqqgIALF++HCtXrkRJSQkkSYKfnx/S09Mhl8vh6uqKxYsXw93dHXK5HCEhITh58mSX9AHcP5bPmI2+Oys9PR1CCEyfPr3NeiqVCgB0m4Z3xo0bN2BjYwNvb28AwOXLl1FfXw8vLy+9er6+vgBg8K6hs7Mzxo8fj7S0NHBHNOpvmPwREXXQ2rVrERcXh/j4eFRUVODYsWO4du0ann32WZSXlwO4n+j8ft/ALVu2YN26dXplaWlpmDZtGnx9fSGEQHFxMaKjoxEVFQWlUomYmBiUlpbizJkzaG5uxqRJk3RTmJ3pA4BuMYNGo+m6h9OGgwcPwt/fHwqFos163333HQBg3LhxnepPqVSioKAACxcuhJWVFYD7xxEC0E05a8nlctjY2Oj+fr/15JNP4saNGzh37lyn4iHqbZj8ERF1gEqlQmpqKmbNmoXIyEg4OjoiKCgI27ZtQ1VVlcExgp0hk8l0o4sBAQHYunUr6urqkJmZ2SXth4WFoba2FgkJCV3SXlsaGhpw5coV3QhbS8rLy5GVlYWYmBgEBwe3O0LYnpSUFLi7uyMpKUlXpl3R+/vpYQCwtLTUjTr+1rBhwwAAFy5c6FQ8RL2NzNQBEBH1BUVFRaivr8eoUaP0ykePHg0rKyu9admuNmrUKCgUCr3p5b6ioqICQog2R/2Cg4PR0NCAuXPnIikpCZaWlg/c3969e5GTk4OjR4/qjfJp3zVsaQFHU1MTbGxsDMq1Mbc0KkjUlzH5IyLqAO22H3Z2dgbXnJycUFdX1639W1tbo7Kyslv76A53794FcD/+1ri6uiIjIwOBgYGd6isrKwupqakoLCzE0KFD9a5p34+sra3VK1cqlbh79y7c3d0N2tMmhNrvQNRfMPkjIuoAJycnAGgxybtz5w48PT27rW+1Wt3tfXQXbQLV1qbJLi4uuuf7oN5//30cOXIEBQUFLSbo3t7esLe3x9WrV/XKte9BjhgxwuCepqYmAGhxVJCoL2PyR0TUAY8//jjs7Oxw+vRpvfKTJ0+iqakJf/jDH3RlMpkMarW6y/ouLCyEEAJjxozptj66i6urKyRJQk1NTat1fr/lizGEEFi1ahVu376NvLw8yGQt/7Mmk8kwdepUHDt2DBqNBgMG3H/lPT8/H5IktfieoTZmNze3B46PqDfigg8iog6Qy+VYuXIl9u7di3/+85+ora3FhQsXsGTJEri7u2PRokW6un5+fqiurkZeXh7UajUqKysNRpwAYODAgSgrK0NpaSnq6up0yZxGo8Ht27fR3NyM8+fPY/ny5fDy8kJUVFSX9JGfn99jW70oFAr4+Pjg+vXrLV4vLi6Gm5sb5s2bZ3AtIiICbm5uOHPmTKvtX7x4Ee+++y62b98OS0tLgzOy33vvPV3dhIQElJeXY82aNWhoaMCJEyewadMmREVFwd/f36BtbcxBQUHGfm2iXo3JHxFRB61ZswYpKSlITEzE4MGDMX78eDzyyCMoLCyEra2trt7SpUsxceJEzJ8/H/7+/njnnXd0U4fBwcG6LVuWLFkCV1dXBAQEYOrUqaiurgZw/x2zoKAg2NjY4Nlnn8Wjjz6KL7/8Uu+9uc720ZPCwsJQVFTU4oratvbQa2pqQkVFhcEGzB29//cCAwNx5MgRHD16FIMGDcLs2bOxYMECfPjhhy3WP3XqFDw8PFqcEibqyyTB3SupF5MkCdnZ2QZ7mhF1hd74+1q8eDFyc3Nx69YtU4fSIu15ubm5uR2+p7i4GMOHD0dmZiYiIyM7fJ9Go8GECRMQFRWFBQsWGB1rZ9y6dQuenp5ISkrCypUrO3zfgzwfoh6Wy5E/IqJepq3FEX2Rn58fEhMTkZiYiPr6+g7dc+/ePeTl5aGurg4RERHdHKGhtWvXYuTIkYiOju7xvom6G5M/IiLqdnFxcZgzZw4iIiLaXPyhVVhYiD179iA/P7/dk0G6WmpqKs6ePYtDhw51as9Bot6KyR8RUS+xevVqZGZmoqamBt7e3ti9e7epQ+pSycnJiI6Oxvr169utGxoaik8//VTv/OKesG/fPjQ2NqKwsBDOzs492jdRT+FWL0REvURKSgpSUlJMHUa3mjx5MiZPnmzqMFo1Y8YMzJgxw9RhEHUrjvwRERERmREmf0RERERmhMkfERERkRlh8kdERERkRrjJM/VqkiRhzJgxffJAe+r9du/ezd+Xkb799lsA0DtnmP7Pt99+izFjxnCTZ+rNcrnal3q18PBwU4dA/dDNmzfx/fff8/f1AJj0tW3MmDEIDg42dRhEbeLIHxGZnZycHMybN8+oc2GJiPoJHu9GREREZE6Y/BERERGZESZ/RERERGaEyR8RERGRGWHyR0RERGRGmPwRERERmREmf0RERERmhMkfERERkRlh8kdERERkRpj8EREREZkRJn9EREREZoTJHxEREZEZYfJHREREZEaY/BERERGZESZ/RERERGaEyR8RERGRGWHyR0RERGRGmPwRERERmREmf0RERERmhMkfERERkRlh8kdERERkRpj8EREREZkRJn9EREREZoTJHxEREZEZYfJHREREZEaY/BERERGZESZ/RERERGaEyR8RERGRGWHyR0RERGRGmPwRERERmREmf0RERERmhMkfERERkRmRmToAIqLupFarUV9fr1fW0NAAALh9+7ZeuSRJcHJy6rHYiIhMgckfEfVr1dXV8PDwwL179wyuDRw4UO9/T5w4EQUFBT0VGhGRSXDal4j6NTc3Nzz33HMYMKDt/9xJkoT58+f3UFRERKbD5I+I+r0///nP7daxsLDArFmzeiAaIiLTYvJHRP3e7NmzIZO1/paLhYUFnn/+eQwaNKgHoyIiMg0mf0TU7zk4OGDKlCmtJoBCCERGRvZwVEREpsHkj4jMQmRkZIuLPgDAysoKL7zwQg9HRERkGkz+iMgsvPDCC1AoFAbllpaWmDlzJmxtbU0QFRFRz2PyR0RmQS6XY9asWbC0tNQrV6vVePnll00UFRFRz2PyR0Rm46WXXoJardYrc3BwwKRJk0wUERFRz2PyR0Rm409/+pPexs6WlpaYP38+rKysTBgVEVHPYvJHRGZDJpNh/vz5uqlftVqNl156ycRRERH1LCZ/RGRW5s+fr5v6dXNzw7hx40wcERFRz2LyR0RmJSQkBB4eHgCAV155pd1j34iI+pvWt7wn6oScnBxTh0DUqtGjR+PGjRsYNGgQf6vUaz300EMIDg42dRjUD0lCCGHqIKj/kSTJ1CEQEfVp4eHhyM3NNXUY1P/kcuSPuk12djbmzp1r6jCIWrR7926Eh4d3SVuSJPH3bqQ5c+YAAJObVmifD1F34MsuRGSWuirxIyLqa5j8EREREZkRJn9EREREZoTJHxEREZEZYfJHREREZEaY/BERERGZESZ/RES9xKFDh+Do6IjPPvvM1KH0ep9//jni4uKg0Wgwc+ZMeHl5QS6Xw8PDAzNmzMD58+eNbjMxMREBAQFwcHCAtbU1/Pz88Oabb6K+vt6g7vHjxzF27FgoFAq4u7sjNjYWjY2Nuuv79+/Hxo0bce/evU59T6LuwOSPiKiX4J77HbNmzRqkp6dj9erV0Gg0+Prrr7Fz505UV1fj+PHjUKlUeO6551BWVmZUuwUFBXjttddQWlqKqqoqpKSkIC0tzWDPvaKiIkyePBmhoaGorKzE3r178cknn2DJkiW6OtOnT4dcLkdoaCju3LnTJd+bqKsw+SMi6iXCwsJQU1ODadOmmToUqFQqhISEmDoMAxs2bEBWVhZycnJgb28PAAgODsa4ceOgUCjg7e2N5ORk1NTUYMeOHUa1bWdnh0WLFmHgwIGwt7fH3LlzMXPmTBw+fBjXrl3T1XvnnXcwZMgQrFu3Dra2tggODkZsbCx27NiBS5cu6erFxMTgiSeewNSpU9Hc3Nwl35+oKzD5IyIiAxkZGaioqDB1GHqKi4uRkJCAdevWQS6XAwBkMpnBNLmPjw8AoKSkxKj2Dxw4AAsLC72ywYMHAwCUSiUAoLm5GQcPHsT48eP1jrGcMmUKhBDYt2+f3v1r167F2bNnkZaWZlQsRN2JyR8RUS9w/PhxeHl5QZIkfPDBBwCArVu3wtbWFgqFAvv27cOUKVPg4OAAT09P7Nq1S3dveno65HI5XF1dsXjxYri7u0MulyMkJAQnT57U1YuOjoaVlRWGDBmiK1u2bBlsbW0hSRKqqqoAAMuXL8fKlStRUlICSZLg5+cHADh8+DAcHByQnJzcE4/EQHp6OoQQmD59epv1VCoVAMDBwaHTfd64cQM2Njbw9vYGAFy+fBn19fXw8vLSq+fr6wsABu8aOjs7Y/z48UhLS+O0PvUaTP6IiHqBcePG4ZtvvtErW7p0KVasWAGVSgV7e3tkZ2ejpKQEPj4+WLhwIdRqNYD7SV1UVBSUSiViYmJQWlqKM2fOoLm5GZMmTdJNWaanpxucP7xlyxasW7dOrywtLQ3Tpk2Dr68vhBAoLi4GAN3iBY1G0y3PoD0HDx6Ev78/FApFm/W+++47APefaWcolUoUFBRg4cKFsLKyAgDcvHkTAHRTzlpyuRw2NjYoLy83aOfJJ5/EjRs3cO7cuU7FQ9RVmPwREfUBISEhcHBwgIuLCyIiItDQ0ID//Oc/enVkMhmGDx8Oa2trBAQEYOvWrairq0NmZmaXxBAWFoba2lokJCR0SXvGaGhowJUrV3QjbC0pLy9HVlYWYmJiEBwc3O4IYXtSUlLg7u6OpKQkXZl2Re/vp4cBwNLSUjfq+FvDhg0DAFy4cKFT8RB1FZmpAyAiIuNoR6G0I3+tGTVqFBQKhd4ihL6qoqICQog2R/2Cg4PR0NCAuXPnIikpCZaWXG5xngAAIABJREFUlg/c3969e5GTk4OjR4/qjfJp3zVsaQFHU1MTbGxsDMq1Mbc0KkhkCkz+iIj6MWtra1RWVpo6jE67e/cugPvfpzWurq7IyMhAYGBgp/rKyspCamoqCgsLMXToUL1r2vcla2tr9cqVSiXu3r0Ld3d3g/a0CaH2OxCZGpM/IqJ+Sq1W486dO/D09DR1KJ2mTaDa2jTZxcUFTk5Onern/fffx5EjR1BQUAA7OzuD697e3rC3t8fVq1f1yrXvRY4YMcLgnqamJgBocVSQyBSY/BER9VOFhYUQQmDMmDG6MplM1u50cW/k6uoKSZJQU1PTap3OnIwihMCqVatw+/Zt5OXlQSZr+Z9HmUyGqVOn4tixY9BoNBgw4P6r8/n5+ZAkqcX3DLUxu7m5PXB8RF2JCz6IiPoJjUaD27dvo7m5GefPn8fy5cvh5eWFqKgoXR0/Pz9UV1cjLy8ParUalZWVBqNYADBw4ECUlZWhtLQUdXV1UKvVyM/PN9lWLwqFAj4+Prh+/XqL14uLi+Hm5oZ58+YZXIuIiICbmxvOnDnTavsXL17Eu+++i+3bt8PS0hKSJOl93nvvPV3dhIQElJeXY82aNWhoaMCJEyewadMmREVFwd/f36BtbcxBQUHGfm2ibsHkj4ioF/jggw8wevRoAEBsbCxmzJiBrVu3YvPmzQDuTydevnwZ27dvx8qVKwEAzz//PH755RddG3fv3kVQUBBsbGzw7LPP4tFHH8WXX36p957c0qVLMXHiRMyfPx/+/v545513dNORwcHBum1hlixZAldXVwQEBGDq1Kmorq7ukefQlrCwMBQVFbW4oratPfSamppQUVFhsAFzR+//vcDAQBw5cgRHjx7FoEGDMHv2bCxYsAAffvhhi/VPnToFDw+PFqeEiUxBEtx1krqBJEnIzs422FOMqD/qDb/3xYsXIzc3F7du3TJZDMbQnpebm5vb4XuKi4sxfPhwZGZmIjIyssP3aTQaTJgwAVFRUViwYIHRsXbGrVu34OnpiaSkJF3S3hEP8nyIOiiXI39ERP1EW4sh+gM/Pz8kJiYiMTER9fX1Hbrn3r17yMvLQ11dHSIiIro5QkNr167FyJEjER0d3eN9E7WGyR/1Sq+++irs7e0hSRLOnj1r6nB6BY1Gg82bNyMkJKTF6xMmTDB4T0n7aWnVYlv27NkDHx8fg3asrKzg6uqKCRMmYNOmTbh9+3ZXfDWiDouLi8OcOXMQERHR5uIPrcLCQuzZswf5+fntngzS1VJTU3H27FkcOnSoU3sOEnU1Jn/UK3388cfYvn27qcPoNX755Rc899xzeP3113UHzBvD2GOuZs+ejcuXL8PX1xeOjo4QQkCj0aCiogI5OTnw9vZGbGwsAgMDcfr0aaPjoa61evVqZGZmoqamBt7e3ti9e7epQ+pWycnJiI6Oxvr169utGxoaik8//VTvPOOesG/fPjQ2NqKwsBDOzs492jdRe5j8EfUAlUrV6ohde86dO4dVq1ZhyZIlGDlyZKv15HI5amtrIYTQ+yxatAhvvvnmg4auI0kSnJycMGHCBGRmZiInJwfl5eUICwvr0AhMb9eZv5GppaSkoLGxEUIIXLlyBeHh4aYOqdtNnjwZGzZsMHUYrZoxYwbi4uJaPAaOyNSY/FGvJUmSqUPoMhkZGaioqHige5944gns2bMHL7/8cpunGxw+fNjgsPlr167hhx9+wB//+McH6rst4eHhiIqKQkVFBbZt29bl7fe0zvyNiIj6EiZ/1CsIIbBp0yb4+/vD2toajo6OeOONN/TqvPvuu1AoFLC3t0dFRQVWrlwJDw8P/PTTTxBCIDU1VXeovbOzM1588UW9M03T09Mhl8vh6uqKxYsXw93dHXK5HCEhITh58qRBPO21Fx0dDSsrK73ppGXLlsHW1haSJKGqqgoAsHz5cqxcuRIlJSWQJAl+fn7d8QhbtGHDBsTExOiVHT58uMv2atPuH5efnw+AfyMioj5BEHUDACI7O7vD9ePj44UkSeJ//ud/xO3bt4VSqRRbtmwRAMT333+vVw+AiImJEe+//76YNWuW+PHHH8Xbb78trKysxD/+8Q9x584dcf78efHUU0+JwYMHi5s3b+ruX7RokbC1tRUXL14Ud+/eFUVFRWL06NHC3t5e/Oc//9HV62h7L7/8snBzc9P7Lps2bRIARGVlpa5s9uzZwtfX16hn2JJnnnlGPPHEEx2qe/36dREQECDu3bunV37gwAFhb28vEhMT223D19dXODo6tnq9trZWABAPPfSQrswc/0bG/t5JiPDwcBEeHm7qMHotPh/qRjkc+SOTU6lU2Lx5M/70pz/h9ddfh5OTE2xsbDBw4MBW79mwYQNee+017NmzBw8//DBSU1Mxa9YsREZGwtHREUFBQdi2bRuqqqrw0Ucf6d0rk8l0o0UBAQHYunUr6urqkJmZqYvHmPZ6qw0bNuDvf/+77vgprbCwMNTW1iIhIaHTfWhXZNfV1bXYP/9GRES9D8/2JZMrLi6GUqlEaGjoA91fVFSE+vp6jBo1Sq989OjRsLKyMpgu/L1Ro0ZBoVDopgs7215vUFZWhv3792PTpk3d2k9DQwOEEHBwcGiznjn8jTZv3swNeY3w7bffAvi/zYxJ37fffqt3JjNRV+LIH5mc9txLFxeXB7r/zp07ANDiXnZOTk4tjkr9nrW1NSorK7usPVPbuHEjFi5cCLlc3q39/PzzzwCAxx57rM16/BsREfUeHPkjk9MmKI2NjQ90v5OTEwC0+A/+nTt34Onp2eb9arVar15n2zO1mzdvYufOnfjpp5+6va/Dhw8DAKZMmdJmPXP4G61YsYLHGRqBx5e1jSOi1J048kcm9/jjj2PAgAH46quvHvh+Ozs7g82GT548iaamJvzhD39o8/7CwkIIIXRTLMa0J5PJoFarHyju7rJx40ZERka2+c5kV7h58yY2b94MT0/Pds9L5d+IiKj3YPJHJufi4oLZs2dj9+7dyMjIQG1tLc6fP9/hl/blcjlWrlyJvXv34p///Cdqa2tx4cIFLFmyBO7u7li0aJFefY1Gg9u3b6O5uRnnz5/H8uXL4eXlpdu2xJj2/Pz8UF1djby8PKjValRWVuLq1asGMQ4cOBBlZWUoLS1FXV1dtyUj5eXl+OSTT7BixYpW6+Tn5xu11YsQAvX19dBoNBBCoLKyEtnZ2Rg7diwsLCyQl5fX7jt//BsREfUiJl1sTP0WjNz6oq6uTrz66qti0KBBws7OTowbN068/fbbAoDw9PQU586dExs3bhQ2Nja6rUX+8Y9/6O7XaDRi06ZNYtiwYcLS0lI4OzuLmTNnip9++kmvn0WLFglLS0vh4eEhZDKZcHBwEC+++KIoKSnRq9fR9m7duiUmTpwo5HK58Pb2Fn//+9/FG2+8IQAIPz8/3dYkZ86cEQ8//LCwsbER48aN09uKpD0nTpwQY8eOFe7u7gKAACCGDBkiQkJCxFdffaVX9/XXXxeRkZFttnfo0CFhb28vkpKSWq2zf/9+MWLECKFQKISVlZUYMGCAACAkSRJOTk7i6aefFomJieLWrVt695nr38jY3ztxK5P28PlQN8qRhBDCVIkn9V+SJCE7O7vXvQO1ePFi5Obm4tatW6YOhVrRF/9GvfX33pvxnb+28flQN8rltC+ZnXv37pk6BGoH/0ZERN2HyR9RD7t06RIkSWr3ExERYepQiXqtzz//HHFxcdBoNJg5cya8vLwgl8vh4eGBGTNm4Pz580a3mZiYiICAADg4OMDa2hp+fn548803UV9fb1D3+PHjGDt2LBQKBdzd3REbG6u3Y8H+/fuxceNG/h8Z6pWY/JHZWL16NTIzM1FTUwNvb2/s3r3bJHE89thjEEK0+8nKyjJJfKbUW/5G1LutWbMG6enpWL16NTQaDb7++mvs3LkT1dXVOH78OFQqFZ577jmUlZUZ1W5BQQFee+01lJaWoqqqCikpKUhLSzPYdqWoqAiTJ09GaGgoKisrsXfvXnzyySdYsmSJrs706dMhl8sRGhqq25eSqNcw0cuG1M+BL8CTGTH1712pVIrg4OA+1ceDLmhYv369ePTRR4VKpRJCCKFWq8ULL7ygV+e7774TAERycrJRbYeFhYnm5ma9srlz5woAeudKz5s3T3h7ewuNRqMr27Rpk5AkSfz4449690dHR4vg4GChVquNioULPqgb8WxfIqK+LiMjAxUVFX2+j/YUFxcjISEB69at020OL5PJ8Nlnn+nV8/HxAQCUlJQY1f6BAwdgYWGhVzZ48GAAgFKpBAA0Nzfj4MGDGD9+PCRJ0tWbMmUKhBDYt2+f3v1r167F2bNnkZaWZlQsRN2JyR8RUQ8TQiA1NRXDhw+HtbU1nJ2d8eKLL+rOLgaA6OhoWFlZYciQIbqyZcuWwdbWFpIkoaqqCgCwfPlyrFy5EiUlJZAkCX5+fkhPT4dcLoerqysWL14Md3d3yOVyhISE6J173Jk+gPsnvBizZ2RnpaenQwiB6dOnt1lPpVIBQLv7T3bEjRs3YGNjA29vbwDA5cuXUV9fDy8vL716vr6+AGDwrqGzszPGjx+PtLQ0CG6uQb0Ekz8ioh62du1axMXFIT4+HhUVFTh27BiuXbuGZ599FuXl5QDuJzq/3zpmy5YtWLdunV5ZWloapk2bBl9fXwghUFxcjOjoaERFRUGpVCImJgalpaU4c+YMmpubMWnSJFy7dq3TfQD/typbo9F03cNpw8GDB+Hv7w+FQtFmve+++w4AMG7cuE71p1QqUVBQgIULF8LKygrA/ZNtAMDe3l6vrlwuh42Nje7v91tPPvkkbty4gXPnznUqHqKuwuSPiKgHqVQqpKamYtasWYiMjISjoyOCgoKwbds2VFVVdfhkm46QyWS60cWAgABs3boVdXV1yMzM7JL2w8LCUFtbi4SEhC5pry0NDQ24cuWKboStJeXl5cjKykJMTAyCg4PbHSFsT0pKCtzd3ZGUlKQr067o/f30MABYWlrqRh1/a9iwYQCACxcudCoeoq4iM3UARETmpKioCPX19Rg1apRe+ejRo2FlZaU3LdvVRo0aBYVCoTe93FdUVFRACNHmqF9wcDAaGhowd+5cJCUlwdLS8oH727t3L3JycnD06FG9UT7tu4bNzc0G9zQ1NcHGxsagXBtzS6OCRKbA5I+IqAdpt/2ws7MzuObk5IS6urpu7d/a2hqVlZXd2kd3uHv3LoD78bfG1dUVGRkZCAwM7FRfWVlZSE1NRWFhIYYOHap3Tft+ZG1trV65UqnE3bt34e7ubtCeNiHUfgciU2PyR0TUg5ycnACgxSTvzp078PT07La+1Wp1t/fRXbQJVFubJru4uOie74N6//33ceTIERQUFLSYoHt7e8Pe3h5Xr17VK9e+BzlixAiDe5qamgCgxVFBIlNg8kdE1IMef/xx2NnZ4fTp03rlJ0+eRFNTE/7whz/oymQyGdRqdZf1XVhYCCEExowZ0219dBdXV1dIkoSamppW6/x+yxdjCCGwatUq3L59G3l5eZDJWv7nUSaTYerUqTh27Bg0Gg0GDLj/6nx+fj4kSWrxPUNtzG5ubg8cH1FX4oIPIqIeJJfLsXLlSuzduxf//Oc/UVtbiwsXLmDJkiVwd3fHokWLdHX9/PxQXV2NvLw8qNVqVFZWGow4AcDAgQNRVlaG0tJS1NXV6ZI5jUaD27dvo7m5GefPn8fy5cvh5eWFqKioLukjPz+/x7Z6USgU8PHxwfXr11u8XlxcDDc3N8ybN8/gWkREBNzc3HDmzJlW27948SLeffddbN++HZaWlgbHLb733nu6ugkJCSgvL8eaNWvQ0NCAEydOYNOmTYiKioK/v79B29qYg4KCjP3aRN2CyR8RUQ9bs2YNUlJSkJiYiMGDB2P8+PF45JFHUFhYCFtbW129pUuXYuLEiZg/fz78/f3xzjvv6KYOg4ODdVu2LFmyBK6urggICMDUqVNRXV0N4P47ZkFBQbCxscGzzz6LRx99FF9++aXee3Od7aMnhYWFoaioqMUVtW3todfU1ISKigqDDZg7ev/vBQYG4siRIzh69CgGDRqE2bNnY8GCBfjwww9brH/q1Cl4eHi0OCVMZAqS4K6T1A0kSUJ2drbBHmJE/VFv/L0vXrwYubm5uHXrlqlDaZH2vNzc3NwO31NcXIzhw4cjMzMTkZGRHb5Po9FgwoQJiIqKwoIFC4yOtTNu3boFT09PJCUlYeXKlR2+70GeD1EH5XLkj4ion2prcURf5Ofnh8TERCQmJqK+vr5D99y7dw95eXmoq6tDREREN0doaO3atRg5ciSio6N7vG+i1jD5IyKiPiMuLg5z5sxBREREm4s/tAoLC7Fnzx7k5+e3ezJIV0tNTcXZs2dx6NChTu05SNTVmPwREfUzq1evRmZmJmpqauDt7Y3du3ebOqQulZycjOjoaKxfv77duqGhofj000/1zi/uCfv27UNjYyMKCwvh7Ozco30TtYdbvRAR9TMpKSlISUkxdRjdavLkyZg8ebKpw2jVjBkzMGPGDFOHQdQijvwRERERmREmf0RERERmhMkfERERkRlh8kdERERkRpj8EREREZkRnvBB3UKSJFOHQETUp4WHh/OED+oOudzqhbpFdna2qUMgatWJEyeQlpbG3yn1ag899JCpQ6B+iiN/RGR2cnJyMG/ePPA/f0Rkhni2LxEREZE5YfJHREREZEaY/BERERGZESZ/RERERGaEyR8RERGRGWHyR0RERGRGmPwRERERmREmf0RERERmhMkfERERkRlh8kdERERkRpj8EREREZkRJn9EREREZoTJHxEREZEZYfJHREREZEaY/BERERGZESZ/RERERGaEyR8RERGRGWHyR0RERGRGmPwRERERmREmf0RERERmhMkfERERkRlh8kdERERkRpj8EREREZkRJn9EREREZoTJHxEREZEZYfJHREREZEaY/BERERGZESZ/RERERGaEyR8RERGRGWHyR0RERGRGmPwRERERmREmf0RERERmRGbqAIiIulNlZSX+9a9/6ZWdPn0aAPDRRx/pldvb22P+/Pk9FhsRkSlIQghh6iCIiLpLY2MjXF1dUV9fDwsLCwCA9j97kiTp6qnVavzlL3/Bjh07TBEmEVFPyeW0LxH1a9bW1ggPD4dMJoNarYZarUZzczOam5t1/1utVgMAXnrpJRNHS0TU/Zj8EVG/99JLL6GpqanNOk5OTvjjH//YQxEREZkOkz8i6vcmTpwIFxeXVq9bWloiMjISMhlfgyai/o/JHxH1ewMGDMDLL78MS0vLFq+r1Wou9CAis8Hkj4jMwvz583Xv9v3e0KFDERwc3MMRERGZBpM/IjILTz/9NB5++GGDcisrK/zlL3/RW/lLRNSfMfkjIrPx5z//2WDqt6mpiVO+RGRWmPwRkdl4+eWXDaZ+/fz8EBQUZKKIiIh6HpM/IjIbjz32GAICAnRTvJaWlvjrX/9q4qiIiHoWkz8iMiuvvPKK7qSP5uZmTvkSkdlh8kdEZmX+/Pm4d+8eAOCpp56Ct7e3iSMiIupZTP6IyKx4eXnhmWeeAQD85S9/MXE0REQ9j9vZU682Z84cU4dA/VBjYyMkScLRo0dx7NgxU4dD/UxwcDBef/11U4dB1CqO/FGvtnv3bly/ft3UYVA/4+npCTc3Nxw4cIC/LyN9++23+Pbbb00dRq/17bff4sSJE6YOg6hNHPmjXm/FihWYO3euqcOgfqa4uBjDhg3j78tI2tH43NxcE0fSO3G2gvoCjvwRkVny8/MzdQhERCbB5I+IiIjIjDD5IyIiIjIjTP6IiIiIzAiTPyIiIiIzwuSPiKiTDh06BEdHR3z22WemDqXX+/zzzxEXFweNRoOZM2fCy8sLcrkcHh4emDFjBs6fP290m4mJiQgICICDgwOsra3h5+eHN998E/X19QZ1jx8/jrFjx0KhUMDd3R2xsbFobGzUXd+/fz82btyoOwWGqD9i8kdE1ElCCFOH0CesWbMG6enpWL16NTQaDb7++mvs3LkT1dXVOH78OFQqFZ577jmUlZUZ1W5BQQFee+01lJaWoqqqCikpKUhLSzPYdqWoqAiTJ09GaGgoKisrsXfvXnzyySdYsmSJrs706dMhl8sRGhqKO3fudMn3JuptmPwREXVSWFgYampqMG3aNFOHApVKhZCQEFOHYWDDhg3IyspCTk4O7O3tAdw/CWPcuHFQKBTw9vZGcnIyampqsGPHDqPatrOzw6JFizBw4EDY29tj7ty5mDlzJg4fPoxr167p6r3zzjsYMmQI1q1bB1tbWwQHByM2NhY7duzApUuXdPViYmLwxBNPYOrUqWhubu6S70/UmzD5IyLqRzIyMlBRUWHqMPQUFxcjISEB69atg1wuBwDIZDKDaXIfHx8AQElJiVHtHzhwABYWFnplgwcPBgAolUoAQHNzMw4ePIjx48dDkiRdvSlTpkAIgX379undv3btWpw9exZpaWlGxULUFzD5IyLqhOPHj8PLywuSJOGDDz4AAGzduhW2trZQKBTYt28fpkyZAgcHB3h6emLXrl26e9PT0yGXy+Hq6orFixfD3d0dcrkcISEhOHnypK5edHQ0rKysMGTIEF3ZsmXLYGtrC0mSUFVVBQBYvnw5Vq5ciZKSEkiSpNvI+vDhw3BwcEBycnJPPBID6enpEEJg+vTpbdZTqVQAAAcHh073eePGDdjY2MDb2xsAcPnyZdTX18PLy0uvnq+vLwAYvGvo7OyM8ePHIy0tjdP61O8w+SMi6oRx48bhm2++0StbunQpVqxYAZVKBXt7e2RnZ6OkpAQ+Pj5YuHAh1Go1gPtJXVRUFJRKJWJiYlBaWoozZ86gubkZkyZN0k1ZpqenGxxBt2XLFqxbt06vLC0tDdOmTYOvry+EECguLgYA3eIFjUbTLc+gPQcPHoS/vz8UCkWb9b777jsA959pZyiVShQUFGDhwoWwsrICANy8eRMAdFPOWnK5HDY2NigvLzdo58knn8SNGzdw7ty5TsVD1Nsw+SMi6kYhISFwcHCAi4sLIiIi0NDQgP/85z96dWQyGYYPHw5ra2sEBARg69atqKurQ2ZmZpfEEBYWhtraWiQkJHRJe8ZoaGjAlStXdCNsLSkvL0dWVhZiYmIQHBzc7ghhe1JSUuDu7o6kpCRdmXZF7++nhwHA0tJSN+r4W8OGDQMAXLhwoVPxEPU2MlMHQERkLrSjUNqRv9aMGjUKCoVCbxFCX1VRUQEhRJujfsHBwWhoaMDcuXORlJQES0vLB+5v7969yMnJwdGjR/VG+bTvGra0gKOpqQk2NjYG5dqYWxoVJOrLmPwREfVC1tbWqKysNHUYnXb37l0A979Pa1xdXZGRkYHAwMBO9ZWVlYXU1FQUFhZi6NChete070vW1tbqlSuVSty9exfu7u4G7WkTQu13IOovmPwREfUyarUad+7cgaenp6lD6TRtAtXWpskuLi5wcnLqVD/vv/8+jhw5goKCAtjZ2Rlc9/b2hr29Pa5evapXrn0vcsSIEQb3NDU1AUCLo4JEfRmTPyKiXqawsBBCCIwZM0ZXJpPJ2p0u7o1cXV0hSRJqampardOZk1GEEFi1ahVu376NvLw8yGQt/7Mmk8kwdepUHDt2DBqNBgMG3H/lPT8/H5IktfieoTZmNze3B46PqDfigg8iIhPTaDS4ffs2mpubcf78eSxfvhxeXl6IiorS1fHz80N1dTXy8vKgVqtRWVlpMIoFAAMHDkRZWRlKS0tRV1cHtVqN/Px8k231olAo4OPjg+vXr7d4vbi4GG5ubpg3b57BtYiICLi5ueHMmTOttn/x4kW8++672L59OywtLSFJkt7nvffe09VNSEhAeXk51qxZg4aGBpw4cQKbNm1CVFQU/P39DdrWxhwUFGTs1ybq1Zj8ERF1wgcffIDRo0cDAGJjYzFjxgxs3boVmzdvBnB/OvHy5cvYvn07Vq5cCQB4/vnn8csvv+jauHv3LoKCgmBjY4Nnn30Wjz76KL788ku99+SWLl2KiRMnYv78+fD398c777yjm44MDg7WbQuzZMkSuLq6IiAgAFOnTkV1dXWPPIe2hIWFoaioqMUVtW3todfU1ISKigqDDZg7ev/vBQYG4siRIzh69CgGDRqE2bNnY8GCBfjwww9brH/q1Cl4eHi0OCVM1JdJgrtXUi8mSRKys7MN9jgj6gq94fe1ePFi5Obm4tatWyaLwRja83Jzc3M7fE9xcTGGDx+OzMxMREZGdvg+jUaDCRMmICoqCgsWLDA61s64desWPD09kZSUpEvaO+JBng9RD8vlyB8RkYm1tRiiP/Dz80NiYiISExNRX1/foXvu3buHvLw81NXVISIiopsjNLR27VqMHDkS0dHRPd43UXdj8kdERN0uLi4Oc+bMQURERJuLP7QKCwuxZ88e5Ofnt3sySFdLTU3F2bNncejQoU7tOUjUWzH5o37t1Vdfhb29PSRJwtmzZ00dTpe4e/cuHnvsMbz11lt65Wq1Gm+//TZ8fHxgZWUFDw8P/Pd//3eL71m1Z8+ePfDx8TF4ed7Kygqurq6YMGECNm3ahNu3b3fV1zJLq1evRmZmJmpqauDt7Y3du3ebOqRulZycjOjoaKxfv77duqGhofj000/1zjPuCfv27UNjYyMKCwvh7Ozco30T9RQmf9Svffzxx9i+fbupw+hS8fHx+OmnnwzKly9fjk2bNiElJQW3bt3Cp59+iu3bt+PVV181uo/Zs2fj8uXL8PX1haOjI4QQ0Gg0qKioQE5ODry9vREbG4vAwECcPn26K76WWUpJSUFjYyOEELhy5QrCw8NNHVK3mzx5MjZs2GDqMFo1Y8YMxMXFtXgMHFF/weSPqA/55ptv8MMPPxiUX758Gdu2bcMrr7yCiIgI2NvbY8KCX+yJAAAOFUlEQVSECYiOjsbOnTvx448/drpvSZLg5OSECRMmIDMzEzk5OSgvL0dYWFiHpvGIiKh3YPJH/Z4kSaYOoUuoVCq88cYbSEtLM7h26tQpaDQaPPPMM3rlzz//PADgyJEjXR5PeHg4oqKiUFFRgW3btnV5+0RE1D2Y/FG/IoTApk2b4O/vD2trazg6OuKNN94wqHfv3j28/fbb8PLygo2NDUaMGIHs7GwAwNatW2FrawuFQoF9+/ZhypQpcHBwgKenJ3bt2qXXzldffYWnn34aCoUCDg4OCAoK0p0d2lYfDyI+Ph7Lli2Di4uLwTXtaQW/P4Zq2LBhAKA38nf48OEu2/BXuwlxfn6+rqwvPlsiInPC5I/6lYSEBMTGxmLRokUoLy/HzZs3sWrVKoN6q1atwrvvvovNmzfj119/xbRp0/DSSy/h9OnTWLp0KVasWAGVSgV7e3tkZ2ejpKQEPj4+WLhwoe6IrYaGBkyfPh3h4eGorq7GL7/8gkcffVR3HmhbfRjr3//+N0pKSvDSSy+1eP2xxx4DAIPp3UGDBgEAKisrdWXabUU0Go3RcfzeyJEjAdyfdtbqa8+WiMjsCKJeDIDIzs7uUF2lUikUCoWYNGmSXvmuXbsEAPH9998LIYRQqVRCoVCIiIgIvXutra3F0qVLhRBCxMfHCwBCpVLp6mzZskUAEMXFxUIIIX744QcBQBw4cMAglo700VFKpVKMGjVKXL9+XQghRGVlpQAg4uPj9eo9//zzYuDAgeKLL74QKpVK/PrrryInJ0dIkiReeOEFo/rU8vX1FY6Ojm3WkSRJODk5CSH63rM15vdF94WHh4vw8HBTh9Fr8flQH5DT8gnYRH1QcXExlEolQkND26z3008/QalU4vHHH9eV2djYYMiQIbh06VKr91lZWQGAbnTKx8cHrq6uiIyMRExMDKKiovDII490qo+WrF69Gn/729/g4eHRZr2srCzExsbilVdeQXV1Ndzd3fHMM89ACKEbAexqDQ0NEELAwcEBQN97tgAwb968Fs+Vpbb1l3dpu4M5rNqmvo3JH/Ub2kPYW3on7rcaGhoAAG+99ZbBXnnu7u4d7s/GxgYFBQVYtWoVkpOTkZiYiLlz5yIzM7PL+jh+/DguXLiA1NTUdus6OjoaLLz49ddfsWvXLgwdOrTDfRrj559/BvB/08596dlqLV++HMHBwUbfZ660ZxavWLHCxJH0TtrnQ9SbMfmjfkMulwMAGhsb26ynTQ43b96M5cuXd6rPwMBAfPbZZ6isrERqaio2bNiAwMBA3XFUne0jIyMDX3zxhW5Bx28lJycjOTkZp06dwqhRo1q8/9SpUwCAiRMnPnAMbTl8+DAAYMqUKQD61rPVCg4O5tnRRtCeWctn1jKe6Ut9ARd8UL/x+OOPY8CAAfjqq6/arPfQQw9BLpd3+sSPsrIyXLx4EcD9pGf9+vV46qmncPHixS7rIzMzE0IIvY928UZ8fDyEEK0mfgCwfft2eHt7Y/z48Z2KoyU3b97E5s2b4enpiQULFgDoW8+WiMhcMfmjfsPFxQWzZ8/G7t27kZGRgdraWpw/fx4fffSRXj25XI6//vWv2LVrF7Zu3Yra2lrcu3cP169fx6+//trh/srKyrB48WJcunQJTU1N+P7773H16lWMGTOmy/owxtNPP42rV6+iubkZpaWl+O///m98/vnnyMjI0L1TB9zflsWYrV6EEKivr4dGo9Eln9nZ2Rg7diwsLCyQl5ene+evvz5bIqJ+xWRrTYg6AEauxqyrqxOvvvqqGDRokLCzsxPjxo0Tb7/9tgAgPD09xblz54QQQjQ2NorY2Fjh5eUlZDKZcHFxEbNnzxZFRUViy5YtQqFQCABi2LBhoqSkRHz00UfCwcFBABAPP/yw+Pnnn0VpaakICQkRzs7OwsLCQgwdOlTEx8eL5ubmdvvojNZW+06aNEk4OTkJmUwmnJ2dRVhYmDh16pTB/YcOHRL29vYiKSmp1T72798vRowYIRQKhbCyshIDBgwQAHQre59++mmRmJgobt26ZXBvX3q2xv6+iKtZ28PnQ31AjiSEECbLPInaIUkSsrOz+X4RdQv+vow3Z84cAHy3rTV8PtQH5HLal4iIiMiMMPkj6mGXLl2CJEntfrSrWon6s88//xxxcXHQaDSYOXMmvLy8IJfL4eHhgRkzZuD8+fNGt5mYmIiAgAA4ODjA2toafn5+ePPNN1FfX6+rs3//fmzcuFF34g2ROWHyR9TDHnvsMYMVvC19srKyTB0qUbdas2YN0tPTsXr1amg0Gnz99dfYuXMnqqurcfz4cahUKjz33HMoKyszqt2CggK89tprKC0tRVVVFVJSUpCWlqabkgWA6dOnQy6XIzQ0FHfu3Onqr0bUqzH5IyIyEZVKhZCQkD7fx4PYsGEDsrKykJOTA3t7ewD391wcN24cFAoFvL29kZycjJqaGuzYscOotu3s7LBo0SIMHDgQ9vb2mDt3LmbOnInDhw/j2rVrunoxMTF44oknMHXqVDQ3N3fl1yPq1Zj8ERGZSEZGBioqKvp8H8YqLi5GQkIC1q1bp9ucXSaT4bPPPtOr5+PjAwAoKSkxqv0DBw7AwsJCr2zw4MEAAKVSqVe+du1anD17FmlpaUb1QdSXMfkjIuogIQRSU1MxfPhwWFtbw9nZGS+++KLemcLR0dGwsrLCkCFDdGXLli2Dra0tJElCVVUVgPvHyq1cuRIlJSWQJAl+fn5IT0+HXC6Hq6srFi9eDHd3d8jlcoSEhODkyZNd0gdw/2QWY/Z67Grp6ekQQmD69Olt1lOpVACg20eyM27cuAEbGxt4e3vrlTs7O2P8+PFIS0sDN78gc8Hkj4iog9auXYu4uDjEx8ejoqICx44dw7Vr1/Dss8+ivLwcwP3E5vdbx2zZsgXr1q3TK0tLS8O0adPg6+sLIQSKi4sRHR2NqKgoKJVKxMTEoLS0FGfOnEFzczMmTZqkm7LsTB8AdIscNBpN1z0cIxw8eBD+/v5QKBRt1vvuu+8AAOPGjetUf0qlEgUFBVi4cKHehudaTz75JG7cuIFz5851qh+ivoLJHxFRB6hUKqSmpmLWrFmIjIyEo6MjgoKCsG3bNlRVVRmcJNMZMplMN7oYEBCArVu3oq6uDpmZmV3SflhYGGpra5GQkNAl7RmjoaEBV65cga+vb6t1ysvLkZWVhZiYGAQHB7c7QtielJQUuLu7IykpqcXrw4YNAwBcuHChU/0Q9RUyUwdARNQXFBUVob6+3uAs5dGjR8PKykpvWrarjRo1CgqFQm96ua+qqKiAEKLNUb/g4GA0NDRg7ty5SEpKgqWl5QP3t3fvXuTk5ODo0aO6hSW/p41FO3pL1N8x+SMi6gDtdiB2dnYG15ycnFBXV9et/VtbW6OysrJb++gJd+/eBXD/+7TG1dUVGRkZCAwM7FRfWVlZSE1NRWFhIYYOHdpqPRsbG73Y6P+3d/+w7PxhHMDfTUj8GxDaiEaQhhAViYVBEEkHhi5EJYbGQg2NpAsWUlQiYjEaujBQpF00Fm4TISKEqQNJYyBtVS1a2t/wSxu+paqqPfp+jXef+zxPL5f06fXu+dBfx+KPiCgOhYWFAPBukXd/fw+5XP5jsQOBwI/HSJVwoRWruXJpaWnkfCdqeXkZu7u72Nvbe7dgf83v97/JjeivY/FHRBSHhoYGFBQU4Pj4+M32w8ND+P1+NDc3R7ZlZWUhEAgkLbYgCAiFQmhpafmxGKkilUohkUjg9Xo/HPNvy5evCIVCGB8fh8fjgdVqRVbW519z4VxkMlnCcYl+E77wQUQUh5ycHBgMBmxvb2N1dRUPDw84Pz+HTqdDWVkZhoeHI2MVCgXcbjesVisCgQDu7u5wfX0dNWdxcTFubm5wdXUFn88XKeaCwSA8Hg+en59xdnaGsbExVFRUQKvVJiWG3W5PW6uXvLw8VFdXw+l0vrvf4XBAJpOhv78/ap9Go4FMJsPJycmH819eXmJhYQErKyvIzs6OWjZxcXEx6phwLkqlMsFPRfS7sPgjIorT1NQUTCYTjEYjSkpK0N7ejsrKSgiCgPz8/Mi40dFRdHZ2YmBgALW1tZiZmYn8pdja2hpp2aLT6SCVSlFfX4/u7m643W4A/z97plQqkZubi7a2NtTU1GB/f//Nc3LfjZFOPT09uLi4iPTxey1Wrz2/34/b21vYbLYPxyTSq+/o6Ajl5eVobGz88rFEv5EkxK6WJGISiQTr6+tRPc2IkkGM19fIyAgsFgtcLle6U3lXeH1ci8WS8BwOhwN1dXUwm80YHByM+7hgMIiOjg5otVoMDQ0lHP81l8sFuVyO2dlZGAyGb8+XjPND9MMsvPNHRCQysV6G+AsUCgWMRiOMRiMeHx/jOubl5QVWqxU+nw8ajSZpuUxPT6OpqQl6vT5pcxKJHYs/IiJKuYmJCfT19UGj0cR8+SNMEARsbW3Bbrd/ujJIvJaWlnB6eoqdnZ1v9RIk+m1Y/BERicTk5CTMZjO8Xi+qqqqwubmZ7pR+1NzcHPR6Pebn5z8d29XVhbW1tTfrGX+HzWbD09MTBEFAUVFRUuYk+i3Y6oWISCRMJhNMJlO600gplUoFlUqV8rhqtRpqtTrlcYnEgHf+iIiIiDIIiz8iIiKiDMLij4iIiCiDsPgjIiIiyiB84YNE7+DgIN0p0B/G6+trwkuhbWxspDkTcXI6nZDL5elOgygmrvBBoiaRSNKdAhHRl/T29nKFDxIzC+/8kajxtwkREVFy8Zk/IiIiogzC4o+IiIgog7D4IyIiIsogLP6IiIiIMsh/8uVRs6OhhLMAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 46
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3NcN1gYNv2u7",
        "colab_type": "code",
        "outputId": "6dc47a38-d57d-4c88-d4a0-16cb47913ea1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 110
        }
      },
      "source": [
        "model.compile(optimizer=\"adam\", \n",
        "              loss=\"sparse_categorical_crossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "\n",
        "history = model.fit(\n",
        "    x_train, y_train, \n",
        "    batch_size=32, \n",
        "    epochs=2, \n",
        "    validation_data=(x_val, y_val)\n",
        ")"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/2\n",
            "782/782 [==============================] - 94s 120ms/step - loss: 0.3837 - accuracy: 0.8226 - val_loss: 0.3213 - val_accuracy: 0.8632\n",
            "Epoch 2/2\n",
            "782/782 [==============================] - 94s 120ms/step - loss: 0.2011 - accuracy: 0.9233 - val_loss: 0.3170 - val_accuracy: 0.8671\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}