DenseNet.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "from mxnet import autograd, gluon, nd, init, context\n",
    "from mxnet.gluon import nn\n",
    "from mxnet.gluon.data.vision import transforms\n",
    "from IPython import display\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Animator(object):\n",
    "    def __init__(self, xlabel=None, ylabel=None, legend=[], xlim=None,\n",
    "                 ylim=None, xscale='linear', yscale='linear', fmts=None,\n",
    "                 nrows=1, ncols=1, figsize=(3.5, 2.5)):\n",
    "        \"\"\"Incrementally plot multiple lines.\"\"\"\n",
    "        self.fig, self.axes = plt.subplots(nrows, ncols, figsize=figsize)\n",
    "        if nrows * ncols == 1: self.axes = [self.axes,]\n",
    "        # use a lambda to capture arguments\n",
    "        self.config_axes = lambda : Animator.set_axes(\n",
    "            self.axes[0], xlabel, ylabel, xlim, ylim, xscale, yscale, legend)\n",
    "        self.X, self.Y, self.fmts = None, None, fmts\n",
    "    \n",
    "    @staticmethod\n",
    "    def set_axes(axes, xlabel, ylabel, xlim, ylim, xscale, yscale, legend):\n",
    "      \"\"\"A utility function to set matplotlib axes\"\"\"\n",
    "      axes.set_xlabel(xlabel)\n",
    "      axes.set_ylabel(ylabel)\n",
    "      axes.set_xscale(xscale)\n",
    "      axes.set_yscale(yscale)\n",
    "      axes.set_xlim(xlim)\n",
    "      axes.set_ylim(ylim)\n",
    "      if legend: axes.legend(legend)\n",
    "      axes.grid()\n",
    "    \n",
    "    def add(self, x, y):\n",
    "        \"\"\"Add multiple data points into the figure.\"\"\"\n",
    "        if not hasattr(y, \"__len__\"): y = [y]\n",
    "        n = len(y)\n",
    "        if not hasattr(x, \"__len__\"): x = [x] * n\n",
    "        if not self.X: self.X = [[] for _ in range(n)]\n",
    "        if not self.Y: self.Y = [[] for _ in range(n)]\n",
    "        if not self.fmts: self.fmts = ['-'] * n\n",
    "        for i, (a, b) in enumerate(zip(x, y)):\n",
    "            if a is not None and b is not None:\n",
    "                self.X[i].append(a)\n",
    "                self.Y[i].append(b)\n",
    "        self.axes[0].cla()\n",
    "        for x, y, fmt in zip(self.X, self.Y, self.fmts):\n",
    "            self.axes[0].plot(x, y, fmt)\n",
    "        self.config_axes()\n",
    "        display.display(self.fig)\n",
    "        display.clear_output(wait=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def show_images(X, nrows, ncols):\n",
    "    _, axes = plt.subplots(nrows, ncols)\n",
    "    axes = axes.flatten()\n",
    "    for ax, img in zip(axes, X):\n",
    "        ax.imshow(img)\n",
    "        ax.get_xaxis().set_visible(False)\n",
    "        ax.get_yaxis().set_visible(False)\n",
    "    return axes "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_dataloader(batch_size, resize=None, shuffle=True):\n",
    "    transform_list = []\n",
    "    if resize is not None:\n",
    "        transform_list.append(transforms.Resize(resize))\n",
    "    transform_list.append(transforms.ToTensor())\n",
    "    transformer = transforms.Compose(transform_list)\n",
    "    \n",
    "    train = gluon.data.vision.datasets.FashionMNIST(train=True)\n",
    "    train = train.transform_first(transformer)\n",
    "    train_iter = gluon.data.DataLoader(train, batch_size, shuffle, num_workers=4)\n",
    "    \n",
    "    test = gluon.data.vision.datasets.FashionMNIST(train=False)\n",
    "    test = test.transform_first(transformer)\n",
    "    test_iter = gluon.data.DataLoader(test, batch_size, shuffle=False, num_workers=4)\n",
    "    \n",
    "    return train_iter, test_iter, train, test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "class DenseBlock(nn.Block):\n",
    "    def __init__(self,num_convs, growth_rate, **kwargs):\n",
    "        super(DenseBlock, self).__init__(**kwargs)\n",
    "        self.network = nn.Sequential()\n",
    "        for _ in range(num_convs):\n",
    "            self.network.add(self.conv_block(growth_rate))\n",
    "    def conv_block(self, num_channels):\n",
    "        blk = nn.Sequential()\n",
    "        blk.add(nn.BatchNorm())\n",
    "        blk.add(nn.Activation('relu'))\n",
    "        blk.add(nn.Conv2D(num_channels, kernel_size=3, padding=1))\n",
    "        return blk\n",
    "    def forward(self, X):\n",
    "        for blk in self.network:\n",
    "            Y = blk(X)\n",
    "            \n",
    "            X = nd.concat(X, Y, dim=1)\n",
    "        return X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "class DenseNet(nn.Block):\n",
    "    def __init__(self, **kwargs):\n",
    "        super(DenseNet, self).__init__(**kwargs)\n",
    "        self.network = self.get_network()\n",
    "    def __getitem__(self, key):\n",
    "        return self.network[i]\n",
    "    def __len__(self):\n",
    "        return len(self.network)\n",
    "    def transition_block(self, num_channels):\n",
    "        blk = nn.Sequential()\n",
    "        blk.add(nn.BatchNorm())\n",
    "        blk.add(nn.Activation('relu'))\n",
    "        blk.add(nn.Conv2D(num_channels, kernel_size=1))\n",
    "        blk.add(nn.AvgPool2D(pool_size=2, strides=2))\n",
    "        return blk\n",
    "    def get_network(self):\n",
    "        net = nn.Sequential()\n",
    "        net.add(nn.Conv2D(64, kernel_size=7, strides=2, padding=3))\n",
    "        net.add(nn.BatchNorm())\n",
    "        net.add(nn.Activation('relu'))\n",
    "        net.add(nn.MaxPool2D(pool_size=3, strides=2, padding=1))\n",
    "        num_channels = 64\n",
    "        growth_rate = 32\n",
    "        num_convs = [4, 4, 4, 4]\n",
    "        for i, c in enumerate(num_convs):\n",
    "            net.add(DenseBlock(c, growth_rate))\n",
    "            num_channels += c * growth_rate\n",
    "            if i != len(num_convs)-1:\n",
    "                num_channels //= 2\n",
    "                net.add(self.transition_block(num_channels))\n",
    "        net.add(nn.BatchNorm())\n",
    "        net.add(nn.Activation('relu'))\n",
    "        net.add(nn.GlobalAvgPool2D())\n",
    "        net.add(nn.Dense(10))\n",
    "        return net\n",
    "    def forward(self, X):\n",
    "        X = self.network(X)\n",
    "        return X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "batch_size = 4\n",
    "train_iter, test_iter, train, test = get_dataloader(batch_size, resize=96, shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(4, 1, 96, 96)\n",
      "(4,)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAASsAAADrCAYAAAAxO7C0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOy9SY8syZqe95iZm48x5XjyjFV1qu6t7maTYA9kEwIoQaIWlLTQRoL+AH+A/oj2XGolSIAALbRUQ4AgcSFRZDe71QP7Dn1rOnWGnCJj8NHMtDB3D484earOVHWrbscLJDIzBncLD/fXv/H9hHOOPfbYY48fOuSvewF77LHHHq+DPVntsccePwrsyWqPPfb4UWBPVnvsscePAnuy2mOPPX4U2JPVHnvs8aNA8CYvDkXkYrLvai17vAEWXJ07505+3ev4TcD+vP7h4JvO6zciq5iMPxL/7P2sao93wh+7//mzX/caflOwP69/OPim83rvBu6xxx4/CuzJao899vhRYE9We+yxx48Ce7LaY489fhTYk9Uee+zxo8CerPbYY48fBfZktccee/wo8EZ1Vm8NIbb/f0cNLRFFiDBERCEi2HwE5xw0DdQNrmnAWpBys4bhfq0FpRBKgVIQakQS45IIF2lE1SCKCsoKV5a4ssIVJa6u3mnte+yxx9vhuycrIUDsGnD2nQhLhCFyOsGNU2zYfoR2eyKvPMnkuX+s278UYB04279eaA1a4yKNSyPqg4RypqnGEr126JuGcF4hb3Lkco2FPVntscevCd+PZQUIKXD2/aiSiijEjVOaw4wmUf4xBzgIlhq1DBCwsayU3FhWZkNWhBqXRNhEU08i1nc061NJceII55L4XJImikhLAkCUJSzey0fYY4893hDfG1l9K4Twrl0Q9C5eZ/Wgg9ZCEjghvAV0pCmmiiYBBLjW1YyvNPFVTHQRIcp6ex9S9q9DCZpRSD0OqMaSYiYpj6A8snBcUs9D6rGkmgbEB4rwTkx0PSG4PkNWDVQ1oqpxRQl5gS1LXFl+v8dsjz3+DuF7I6tvtaqERI5HiMkYO0mpDmLKmaacSupMYANwClwATQwmdZjYYSMLyvkf6QiehaTPNNnXivDGICuLqi2icTRZQJN4l7RJJPlhS1CHFjsribKKw7TkKF3xYppxc5Bys9Asc4WoJbKOUGVMeAPhtSOeW+KLCv1ihXx+gdmT1R57fGf47snKOcDe8tg2hFKIUUZzPKY4TVidKdZnguKOQRxUqMAQhoZI10x0Q6gMSVCTBhVjXTIOChJV879/9VPOJzNMGBBfCPTKES4tsnYUB4py5i2rciYoTi3BvTV//+wZn06eEcsaLQyRrDmvx5yXIy7KjMoqQmkIpAHgr57fYfF0RPxUkX0ZM1GSeF3A+cV3fTT32OPvLL4fy2qXnDpXTEiE9AFwmSXYaUZxJ2FxX7F6CM0HOb/78Gv+YPY5x3rBSXDDkVoCYJzEIkllyZlacSIFIxnxL8Mr/qfgD3hiT7BaEc4FToEqHcWhpGjFJ8oTw/TBnP/kwd/w3xz+3/xBqLixBdfWsnABa6tZuZDCaWJRc6aW3FOOiYz5H09P+F9Of48/mT7EBgmq0oSXI0QUgXU4Y7YC+Xvs8YPF4Frsz1n4QZ6332vMSkSR/x2GiFDD4YzmdEJ+GnmX7EBQTR311KKOSz48veQno+fcDa/JZEkoTE9SBkHtArBwicG4kszlRLLm0fiK63sJqyDDxAonJeHCUY+hHrVfyKgmiyoMkqfNlF/J56xcQOE0hdPUTlG7AOMkBknhFAtXgS0Yy5yfjp6zvBfxS3XMi0nC6v6M7Hd/j+TcED1bo55e4PIcV9W4qvKlFHvs8QOC0CEySxBZhksiRFHhisKX6VQVGOPDN9b8upcK7ItC99hjjx8Jvj/Lqs32AcgshTSheHTA5acRi8cW+WBFFDVMw5pxVHInWfBhesGH8TlnwTVKWBRuy6qqnGotH0HhNNo2hKLhcXZOcNfwF/qMSzlFNhqEoB47TGtZJVlFpiusEzxtZt5qQ/SWG4AcxNpqp1hYhRGGmVrz99KvONArHo9P+PzuAU8+nvD8KiP6RczsZ2Nm1iIvA1ivvVu4t6z2eB/YLbCGt3bZhA4Q0wnmeEIziQiWNXK+RswXCCG8N1BVW97hS8XVt63tO3Ihv183MNT+jzTBHGQs74fMf8vw+7/3C/7F3f+TqnW9aqeIZc1Y5mSiQgtDseWieaIqrMYiWdmQa+FQWLQw/HbyhJ/GXzPTOf+H+YRFPkUYST1yqLEvZzgcrRmHBY1TPK8n1E7171fC/9aiIRY1FknhNFioRcNMFsyir/j96AvqkaQ4VtRO8bPqjP9u8p8ytzPS5ylRbRDWIop9WcMe7wG3EVX3+FsQhAg1dpxRnqTkx4r4ShEpQdB4908Arh7cZPv41i37G67tLdfzbfgeLSuJmIwBqO4fsLoXcfORILm/5I8O/pY/jC5ZWNcHtw0ChUO3Fs9tUMJhHVgk1kHhNJksSWVJLGoeJy/48nDGX1YBeRKjJyWnMx+gfzC+5jBcMw4KJI7aqd6iAm9VKSRWSAzWkxVgEGTUpLIhFY5YOLRwaBwPg1/yr+59xh9fpZwvIybTgOyLCNk0sFp9xwd4j994dB0Z74JhciuOqY5TlncDVvcETezrGNO8RhqDcxZRVbj6mzf5feF7rWA3xxMAbj6Kuf4JNB/m/PbRJamseGYkKxewthE3NgboXT+J7d0/8EQSdrWdWAwS4wQdp9UuQAnLTK35nclT0qDm+jRhFubMdA7AUbgklVVrQXkrqrOqJBYl/J2hczW18HeYbg2VlRTCEPc/3lb+B6MvefbxmL+MzygPUmyQMb0ewbPn3/1B3uM3H7uE9aYWjJC+TEhJ3DijONas7gnyxxUuCJGNIpzH6KJGNAaXF6+3r+G6fvBu4GscwPLQZwMXDwXytxb8o3tf8Tg9J5MlL0zGyoWsbMTa+td1pLHtnjX9c7GwWCGpnPIfxXniklgKGzJTa34v/YzfS2/XoO9iX8Zt1t7tr4N1crdKjNoF7dossagZy4pMNGgBv5/8LSdnN3xxdMh/H/0TbpZTxr9IXu/Y/QDTxXv8mnHbdfUufbVSIJT03SFpRH4kyR80/PTDp/xNc5dgpUlfBAQ3GlEGvtH/dfEdn7/frWUlBCLQCCURSUJx5HdX3DP88w9/xj8e/7K3ap6bMYXVFC6ktBrZWiodUWnREMu6D3or0dUx4d01LAZ/YGsXYLFksuSennOmDFMZMrcV1+3bFm0d1dpGrGxE7fx7u/3atmTBOOGtrPZ7GLqKxkpi6csfrCgZy5rHwZrfDUtG4oKvPjzgf/3sDymPE+Is86ngptkuY/gefP099uggggCRpYg0pThKKQ4Fycmaf3ryc57cTCgvpxQHCr2M0bWBpf7B3EzfH1l1ZuDgA8kkQR4eYGdj7xs/8Be6OlhzJ7xhLIu+nqkjC4UllT4Y3RFSR1CF1dQiQGGRrYsIPlNXtTGnrqunRnFjY7RpqF3OhahZO03h/Ede2Ygav+8OBkll/XZ27SmDBOf31e27W19hNVZIChcwlhU1BmTJNMhxRxXXP4k5Ln6KPl8jL65pOpfQue/FfN7jB4jXceXexd17BeTRIebuIcWdhPkHmuKO5SgpWZuQUVyyPDAsHgVYFTMOJFHdINe5z2i3N9tvxevE1d7i87w5WX3T3X/ncRFF2OMp6wcjFg8U63v+4r57eMPd8JqxzFnYxFtUVvcWVCg2RWhSWKyTfaZwbdVWxg7AOOFjWq011MEYiXWShUzQoumLPIGeHLt9dNvpiFMKRyianrSME73lFgJKGCS23+faRUgsFWsMBZKGqco5Olpy9WGIMDGTzwNSa+H5eXu8zK3HbY/fcOxezO8SNH9Da9xNR6weZcw/UqweWORpwTgqWZqIUVgRH+WsSXAyQBqNvsmQ55EvYQAwZrO/b1v3jjRUr7zi7FvdoN/OsnrNAyTiiOI4YfFQsfgI9D2fEfvJ7AVnga9KX7uIqg2sp7IklvWWuwf4uJRRba1TjBaGWNQYIVHC9gRkkFgn2t/+sS7+1aEjJqDdTkXY7QdJ7RRrG3kilBCLqn9uQ4ZVT5jGSdY2onCaygW9m6jIOQyW/OTgBf/uI81cjhE2ILxKkdJ/UW43GLbH3x0ML+TXORG+qVzgDQirmSYs7yoWP2k4uDdnmhQcRGtyEzINc+yBYJXlPOMQvQrIvg6J0zbm6hyvTS1dK91Lj1mclW918n+3MatQU80C8jsC8WjJh8eXANyN5m0R5ssF9HVbPzUkFdu32AysJiQSH0saklT3ev8asfW/FBbcZhsWS03QZ/52LbN+X20AvtuPd/k0xgqsk31SwCC4MbG3DvGE+jg7p7kj+TMjyS/G1JOQuFU37e8ye8vq7x7e5k71KlL6hlCCCALkKEOMRrg0Zv4gpjgBPS05SHO0NDRWMa99Bn4a5ox1yeU0o04DrxcXqI0m3HAtcIuw5vAjtoKYHWk5+06adm9OVq8yAW85iC7UlBNJcafhjx5+ycPkCoD70ZWvOreaalCMCVANXDU1IKyqjS1pYfpYFXgCqZ1qA+J+TQqHFN7GsU6idtzKXXQu4dAiA3piAvo1DddY4HsIC6cprW5LHAyh8WuMZc1H0QsOgxXWCf7ky4xqLEnitkfSOR8HeJVpvSex30zcVi/1OnGr3f9vO2d2SEuEIRwfUt+Zkp9GLB4qykPDLCuIVIN1grzRrJqQLKhIg4osqMjSknWSYiLh5b+DACdlS07tNXSb9cTLclBb/79Ds/T7LV3Y3XmoqaY+2/CfH/1ZH4vqiKlwuiUTX5JgkNRW9gWhwEsWVkdE/v+Ny1c71b+vIyeFZbeeVN0SON+NYw33OSSpDrVTlE63pQ+K2gaULYFp26BFghSWiSz4SfSUTFREsubfHj2iGkegveMp6sZ/Sut8/Gq3A36fHfzNxdt+r7vqCK96vj13RBBgDkcsH8QsHknyOxZ5VDFNCtKgYlHF5I2maAJ0ajhSNffia2bpMcvwAKsFTgc+i7hFihvFlJeXYDdew3v87O9GVt1Ch4uSyvccBQFmHFGP4M50wR/EX7Cw/iJdubAPrHdul24VFXx2cEMayrmtzB940pLC9kRlEBjElhXVkVL32yBfIiq/rY2Ldxte/fiGqPzvtmzCBpRCsxYRE1lwJHPuKccq/oLpbE1xFOPuHvnPN49xyxX2Zon7gXS27/HjRzcERSiFOJhSnCYs70uWjxvUtGI6XpMEviy9cZKiCViXIWVUYhFoYQikpS8/fKXV93o6CMP1OGP6WQhuZybCt+HtA+yvgMxS5GSMG6XkZzH1yJEENUXrLoEvG+hcwF10JACgoSef3ZYb07p+HRQOLevBdrYPpHXyJStrF932OuvNlyvcUsLQvgZhUNJ5q1h6i0v1mUWfwVzYkGuRY53k7uSGv3445fwPDwAYPRkTP1kgjcFcDwZR7LWw9oBXNC3bb31eRClilCGylOrejOV9xeqh5fjhNVHQoJWhtoqLImNVheSVpmkkizLiazlB4ni2GBGsBXplEKsct8592cJg/86YV7uBbSy2l6EZjXBJhOykwMvSyycZ4/sP3bffrN/eshr4q8PovkgT7MGE+iRlfaJoxoY0qPpGZPDu38pGvbWjhP/B0bt0EodsewOH6CypIVFpYZCDUgbrtmNPHdGZHXJUt+Q2dmNaQ9IbBuq1aNCA6e4udvv9pm1+XrmQuW0wCB5mV3z+4IBL43skm1RzYEckLxK4nm/HIPZEtUeHV7l+t3k2tHGq6Zj6eMTqfszqPuj7K/7Duz9naSJeFCMui4xFGVLWmrpWGCNZlyFSOJwTrJYx6RL0osHeLKCuccNhK+0J7+wt1tVgPUIHiPEYezCimSaodYVclohV7idQFaUXrHwNz+LdY1ZCApsdiSSmPkhY3wnJjwUia0iDinpIVlb3fyO8qybx9VTGSWqr0HKzTYXrCar7Xbc9e0r4AtGoLWeonQKxiT/ZW2JOHbpDuqmWd+1v+1JmETakJ6GvwzJOgmyfs5vXWecTCAuRoHDULuCT9DnVvYC/HXk38Et7RnQdkCTx9sL2RLXHq7ATnPe1S3LLGhdRSDNLKU4jVncl5d2aP7r/Ff/1wb/mz8qH/Gv3EV+vJqyLCNNIrJVYIygLjbWCqlHYhUYvHMGywi6XL5+TfevPt5CMUrg0pj5MKY414bVEK4lyzieYrHttYcp3IitnHUJus7odpazPQuYfSdYfNBwfLTkJly+5fB0hWCeo2mX0ZNCWEtg2FsUghiUBhCGibrfj+ibkTd9g+34n0bLuY1Vyp+9vF8Mm5q5cwQ4C8LuuZeUCnzFsA//Ixgf4u/3jXcGVC6ldQCwa7kZz8P3cfDY5oZxo7DhBpunGJO7uMu95OOweP1LsWlU7hCXjaDPsN9SYh6csHmfcfCBZPzRkR2sCYfnr6i6fl0dclClVozBG+MLpWoIREFgi3TCJS/LjnJtPMmw4YfzoH6MKS5AbZGEIbgrvGs4XuLL00k/t4GG/ID+JCikx04z1/YzVmaI4FqRfS0aAul7jqtoXmb5mOcPbkdWuGTg4mGYcsT5VrD9oOPvggo+nF5xF862eOiUcUjiM67JxG/fPP297wqltsEUCWjRb1eVdjdQWCbnABwql6d1MX/KwcRWHsSVgIEmzYflNIWhn1W2yhlVb6d5Zex2R9mUTwrUJg4CV9Z8vlSWn4Q0j5duJ/tX4I5pMY0YRQZYhytIPUi13MoP+QO/dw980vOn3+ZJ1YwHlJ5SPMv/QKGF9P+HmQ8nqk4qz+1dMowIpLH++esBXxYzLIqOsNdZKnBFQS7ACIRxxWHOW3TCLcr5KplzdS5nfhOjrAD3XRNeO5DwmfZqiA4WYLxHjDJvGNGNfkuOUwAUSJwXlQcDyvmJ911Ef1YAmXAbEAF2f7GvWnL29ZdUfuO2ixiYLKE7g7IML/uOzn3GslxwHN70KZweFxbJpb+ncOovoicpvXVDbACXrPi4Vy5pY1MSy6l3EYStNv4+BiF7YuokdGXWk1e23e31HNn2WcaD53pEPsEVWQzfSZy43B7+wmgLtFRpkTSZLtPb7no1zFtmIehSgswQHiKbZfPpXxAX3+A3A6wjZfZskS/fydugvQH2YsjxTrB41/P2ffMk/PfoZT8spL6oRf3Vzxk0ZsypD6lrhjMC1VpVo/D5TXXM/vuY0XCCP/PWzNhH/z9WH/M2LE66ep5RfBDgVMSlGKGMxBxn1NKaatR6SFlgNNhCUB4LVPYd8sObxyRW/Wt0j/dpfb66qvDfxnVpWuwetHdEupMAkinpk+WhyyT9Iv+jLDmoXDALdGxO2LwEYBNXlIKjexaSgtb5QGOsLMpWNB+UOXU/fhnC6q74jj47kgF7Ur3tfF/+CbTnjXctNCru1/tviXUN3s3N1AS8no3LG0msEnY0WnB8esjrT6OUhwfkS4Sys1+/2nezxmw0h+nIAmcTY4wPyhz5pszoLWD2E+CTn8eich/qSeZNSGM1lnvZEZRqFbdqbuwAXOHTYMA1zTsMFn0TPOAluOJFrlHAc6wUH0cf8RXLGCzXDSY0JR8RXCeXEz/asx+2NVYFVfrv1xGGOK05mS+6mN/wyPsMp5aelv4ELCO8lwL6RgUFrTChwAQTS9BdyR0Zd3KprndmtGIeNe9aJ4u1uY2Fi1jZkaSJyo8mNprEK2wa1x0HJWBccBGssgrUJWZmI0gQE0hBJb1lFsvFrsgGlCajbDKJ1AosgUTWhbEhUTSorUlWRSl/c2RFRKqutwlSDQCKReGvOtkoNnSvXqzW077+bzPnr05z5JxlOJYwTRVI3m/mDzm6yLXur6jcLt6iUbD33DRBKIZLETy2fTVh/NOH6E38pLz4yqOOSk/GKVRPxZ+uHfJ4fclFkVI2iaX9M5V0/JIikQSjHLMs5itYcBCtiWWGdZN721oai4VFySX2gCJXh+XTE84cxMpe4wOGCdtgw+HPetQmnwKGTGmMlz/Ixai1RlUM0Bmvs9ti6b8HbqS4MD6iQXq8q9pXZVgtc4IikIRSmr6fyltVOK8vA1TJO9hZVR1SRrPtG4rnJWNqQhYl5Xo55lo+5KhLWZYgDbHtRH2ZrTtMFNhE0TnFeZlwUGYsyIlQGrfw+AmlprG+QLuqA2iifFbECawVR2DCKS1Jde/ILc8a64Fh7hVEvnVyB8M3StiNe0cXJfK3JbUWlnRX2QXzJ49MLfm4lc5Ugm5DwItscXyHeqT1hjx843rLFSgQBIo4RYz8UePFAcfMTfxP+9Le+4jj27V2LxpcpzKuYVRVS1oHP/tXSx6kAERuCsCGKGo6SNSfhgiO19EIBCG5sTI0ilSX3wysOghW/M/qa5Z2oNRhCcqOprKIyLWHWEasqZF35uJhWBucEl+sEWQpUaaExG6J6zfP6/bTb6DYbkMQ0kcAFlkhu5FWGZQtD9I3GeGsG2oyfsERtfCdrta26ivfrOuHJasrXizHLeYLLtz9CNVNYJ4hVQ+MkT1cTzhcZxTpEKocQXXOl9eRUS++3N94KEo0AKygjyyJNiJKaUVKySiKOYq8acayXnkjbItTu8/X1X6Jr9WlFAQfW49DFfBBe8IeHIbMo59/wiPV1yuSzcENve3La4zZojcgSzDSjOPZ1VGcfeWv8v330x6xsxL9ZfcifXj/g2XJEbVprykhMI/s4FdIhA0sc1xxla+7EC471kplaEYualY38ddfONvhQvyCTJYey4Eg5DmSMFoovmyVfNglfNb7Y+bPqmM/LQ54XYy6KjGXtC09XeYQqBKp2uLp+4yb+dyYroQPkbIo9mlAeJeSnAjWuyYLNNJeu2XhYhDksvtwt/uzcMf+cbXWsQm6amBfFiPN1ymoZw0Kjii5Q6QUVmiQgrwMqq2hs22tofNbDtFkPwJuq3d8WhBM44XDKeXM2sAjZqjFYQeO8MF9uNIXSW9XuXWC+q583yN7lNb1r2epxuaivx4plzd3wmljWPDma8mySUKcBsQ437Qj7NpwfJt51JNarukDkTldHd0FL1WunizvHrD8+YvFQs3woqB8WTCMfB/1ZecbCxHxZzFjXIVUTUNeerGwtcVb481taVNwwGeWcjpZ8MLrkXjQnlWU/QaqbJlXYtmFfaBY2YSVz1m7FSuYoAU+ahOdmzIvGx83O6xGXVcqLYsSyCjFWoqQjDBvK2FFlEmYTVF545dzXHAL89qoL3TEPAux0RH43Y3U3oDh2ZFnBWPmD18uqOAnD6vDBxY5s+ou/0zzvh0C0KFur6rLIWK5j7FKjFxJVCJwEJx0ugLpUlLWmMoEnmEZhTWs51QLRmr+yarctXft+EAGerKRDBA4hXSvmKaiNj2/lRm81TcO2teQ/g1dr6IL/nStYWO0trtZ0ikXNQ33Bh+ELPpsd8eX4Dk0qEaHGNQ0Cbz5vHXMxINs9vn+8pIIwyNK+KrP3Sh2q7WJOpPLZ30733Ji+5lIohQg1Igwxx2NuPtRc/Y7j4PEF95OCw8gnZX5RnLCoY57nY1ZV6GNUdYApFVTS34RDiwoMo6zgbLzgk/ELPk2fthn2mqLt4e06TWqnwMHCeV2rsYpZu4iFXCOF5UUz4UUz5rwlqxfVmPNixHWeUNQBWhkCZUnCmjxyNInATBOC9cQnk6z9jshqB0IHNJOI/CRg+VDQnFScjn1cp4PpJtMMdaVa9BpTYlPBDi9Xnq9NyE2VcLVOqNYatZTohUCVXfbBB/absaSuFUXrP1sr/QVvBLKUqNyfKEHuSc6nWMFp12YwHGiLCGxfnGqcwDhB2QQUSlPaYGNZdRnENvvYBdm7gtHOquxh6a2u02DBmbrhWNX8avw5fzz5LZpEI3SrdwU+Y9If7HdQlNzj3bGjOuB/iVcnQW4rT9itn9ttpWkbfoEtRQ6hJCIMEWlCeRSxfAQf/M7X/ItH/xeflcc8KWcAfLmecVWmfc+faaQPplcSWUhs4pDKEScVh9maD0eX/E76hL8ff9GT03AuwbC0aJjkWqiEa5WiheFFM2bepFw1vnziqkqYlzGLdeQNhRgi3RApw4vEUmcBzThELVOEc75X8DXw7jErHVJPQtanguJ+zcHxgrupNyc75U3rJLUNsF2mD3or67Z2mK5y3deTb1/sUniLx2qwIeA81znl069OOaR0PdFIaZHKYLTCNQLb1pMY40B6orLa4UKHC20fcFTKIaVFKUsUGLS0/g4xqKPy1pVHP4Gnr56/nVi2iKv9PJmQnAQ3JOOS9WnE7KMHyMUaMV9ir+e4urp1W3t8z7hFh+qtxOS2iG6HxKzb8kD6EoWTY8zplPI44fpjTXVW8cnknPvBFS+aMbnxMeHzfMSiDCkqTVV6ssJ518/GFjVqGI9yDrM1j0ZX3IuuOQyWxK3sN9DGXzeGhWy7SPpWOGW2EmOl1Vw1KZeVJ6t55RNfTR30nokSjlRXEFvqEeTHGlmmaGMRq/VrzdV8d8sqUFRjSXnkOLl/zePZBQ/ia7Qw/cw9P0B0Y1lBS1i8rObZweuuB302UQpHHNRkUUWRagojKQlQpfBSFsLhFIjEEIc1saqxTpKFNVUT4KzEKIcJW8st8YFw18WntEWHBh02xGGNkg4lvXUVKkOmK0a6JAsqnzzotOHb9Urh0GzL03TYbZjuCKtwmrXVXAs/y/BwtObp3TFXvzsh+zomfqIR6/WGrN7HkMs93g27Wdpv0pd65awCuy2vMiAvnPWunxReP0p74bvm/iHzT1IWjyTrRw2Hpzdoafib6oxfFce8KEcA3BQR6yLcuH4WhLbI2N94D8ZrzrIFD9LrnqgUjoWNWbnQE1XvATkkZit0ARBJX84zU2sMgvOWLK9bslpWnixtI0CAEA6tDJOwQKcV9USTH0mCMkRWCcE85HXw7pZVEFCNJPVRwz89+wWn4YKRKohlvSm4bAs/+yD6oOobBjLEO03DnQpn1Q4tzVTFJCqojL/YS2Vp6vbO0SJKK7KoIlQGMKS6oo4lQjiaUNG0hXDWKGjbfqTyX2Qc1mRhTaqr3jIDCKXZKCgGJVHb/tNNvAFfhzKEvaVw1Ah+1Y0AACAASURBVH+uzWcsrOZGxMTte++N5jy9N+W6iLFBiCpS9Nfh9l1nnyH89WNLHePdZKl3JVa8pdXa7KqNX0UR+VnM/GMJv7vgt48vOIzWhLLhV8UxT4sJF7kninURUpcBtgig9SJEZEiSiqNszb1szofpBR/HzzkK/HRyieXGxn1gfSOD1KqhCHoZJOMksWiYyJyxzPs2utxorgsf01oWEU2tfNaxrb0KpWEclIzSkqtpTHEcEBQKfRMSxNtzEl6Fd9OzEhIXaeqRIJoV/AfjnxPLisKGPdHULniJjKQwfZvKbW7gsBG5i/1oYciCklmYt1ZWwyKMqBqFc742yjnBOC0ZhSWJasXFtN9+pAyN3TTy9NlG4RD4uqtUe1nXsS43jcxOoqUhUTWJrEhUO9CirWRXgr6mqlNhGBLSsJKdft/+76o9RjfOf1kPkysu7mR8Jg5ZFSnpeYgOXy752OMHgvd54+isqzZKIEKNSFNElmBHKcu7iuKDiv/y8V/xk+QZ5/WYizrjV+sjnq4mzNeeKKpC40rVt9C4wKG04SDN+XhyzqPkkg+icx6Hz8lExcLG3Ni4H3qy27LWT5AS/rxOVdmX7YTCUDhJYTXLOuI69+ohZaFpSuWTWsLinO871NKQhjXzrKaeSKqJpMkULni9QapvRVYi0P5gxhF2mmJiCALbT3upnfICe23tkZdwabaC0X17yqBKHTYu07BhGQFTlSNDx0iV3DQJqzhkkUZUNqCxEmMljZNMw5xpWDDVOQpvjeW6oL7FzRxCS0MoG6L2x79mQ2iR9ENWOxM4Fv7Lkp1SAxLTjrpXyF5JYjjIwn8+u6UtXzjNwvov+Sya8/dmirEu+fNnH1GNJaPg/ZTC7fEaeJN5d2/qjg+nKYuu33PQ3yetb0hOYkSWYqcZxWlGfqLJTySLx5bxobewz+sxz6oJz4sR5/mIi1VKkXtXyq0DnyQKLSIzhFHD6XTJh5MLPk5fMA3WxKJmbSNqobasqY3n0yXF/DAU8OdtV/PYhXguzIgXzZhf5Ud8uZixuPLWHU1bx9WWBzWNZFlFPC9GvlC0VqhXxHS/CW9HVjrwrD9KaSYRJoI4aC/wtsaoG6LgP6hDq5cj/sPGXznowRtKtXSPSeWHnx5rH9ArXOB/tzUgjfUk2ZGNH+fl+mJT8BbbVvYOtjKQwz7EDp0F2FXVa2EIRdM2RW/GcXVzBBWSynm3sBty0W2nI+BhTKt2AQvj939PX3MnmPNp+pQ/P7lHNY79ZJE9vnvckukDXp3h233d7ut3t7FbwrCzXWe81AuzCWaWkd9NmX8YsHro4MGao+mKe6MbwJcGPC9GPF1NuF4nFHmIWflzTZQSF1lUXDMZ5ZxkKx6NrvgoOefT+Ov+fPUkFL0kZAn0r7HtHE2AVJZt14YP7yxszMpGfF3N+HI142qeIecbL8BJBwqcc5hGsSpDzuWIvAy91ef4VuXeXbxamW6PPfbY4weEt3QDA0QaY8cJ9SjAxI5IN321eeFC1jaktgFaNsSi6afXdBhaLENtKi2aW+WGU1kStnIvfS0Wm0nN3pxVW4WYCi/L0g1F1a1FBLRzC/2doxs9X/X6VNsFqeDjUsOx9Z31B0PlBZ8dDIHKdTI4u1rxdkszq2itz0yWPNQXnKkVqXD8D8f/iJtxgtN7N/B7xW5ZwausrGEGr396+/WvlPd5RaxLaI0ZpxQnCYv7iptPLPc+fc5/9fDfsjYRpQtYNhHntXf/rlYJ+SrC5QpReAs8yAV1CGHY8GA657cnT3kQXvE4esaHwRXXNuLaplybjJWNtvTeYBCjcpLSRRRtzDlqFUuO1JJrm3JhRlw1GZ/nB5wvM5qbkPiqvS61w4YCE1sIwBpBXvrzvK6CjYvYf/DXM7He7kpIYszhhOIsZfFAUc9sX+4/LEWwnWid8IWSHXZn8xl8GQHOu1EMAu9DwbtaGJTQPUkMYz8dAWwVlQqvOyWx1GwUHwBqNq/v5Gs6ovJyNTta7O34eP+MbfM1t9SI7QTYgVcmErpYgXWiJ0qFIxSCaVRwPnXU92eExX1cnuNWa2xRvPp72eO94aX6qZfUOn35wTfVWfXPDTXJAz/5Ca0RUYjIUlwaY9OQ1WlKfhKQHwvyO47gNOcsuyEWNUtiruqUyyrjWT7mpogoS41rScpFbR+uBjWtOJ0seZRe8Si64Egt0RhWLmgz9MHWrMxXdUP4sIU/d7t4lQ9bJDyvJ/xifcIXywNWeQgWbNAmxgIwkcNFvm5RtrWPACow2FTQCEe11tSphNe8Ib+dZZXEVCcJywcBqwegDkuO41U/9KGDcX5q8m4hpBpk+zxJbUZlWSQag2n95aEeVIehUihsj3bvqmzLdqqzwlEK3cfGhtvYxXAs11AKeQuOl+5GQxgnttaz2d8gK+iC/u8uLlA70wbbNZKakS6pp5bFw4hJc+K1rgD2ZPXd4DaVi299z5vL9oggQGQpIk1xk4zyNCM/9nVH5YGgPLKYg5p0lnMyXhFKw9+WJ5yXI+Z1zLxKuM4TlusIk7flCaFFRv7mG4SGk+mSjyfnfJo+5cPwBQA1ihdm3HsSsJ2p7uO2g2uk80qGCruF0zxtpnyWH/GrxSHPb0Y0VQACTDtKwMYWF1lk0hCEhiAwhIFvucmSkjpssJkgL8ZUY4GNXi/j/VZk5dKY4jBgdU9Q3y95cDTnTnTTP2+dGLhq7YAHBsWRg2xgNxoex5ar2Clxlnbj3tm2alYLg5bNLdOZt2f5dRhaYLelZofPv6T66ba33z2nRUMlgm/UdPfbcy+d1ENrsZPJ6TKoK6nBQhZUMKtZPoiRTcwI0Kv8G/e1xztitwfzdUoTvkne5TbFT+1LEux0RHWasXgUsrwvyB806IOCs9mSx9NzZjqntAGVDfhifcBlmbKsItatekGdayglopa4xJBkvnD4zmTB4/EFv519zafRE+4HN7wwWev6pZtlDNw92NyoLao3BrqA+jBksrIBV03Gs2LMs/mYfB57t06ATVup89Ci0oYwqn14SNq+yDoLvUyTloa/XkXUoxgXfYelCzbRFDNBcafh/tkVH04uONKrNoPwshvUE0RHCjtNzVtjtWio2BBVl+3rfhorCaQldnWfnYPt7F4nObM7aXlYRmCd6DNzWpq+tmujFurQssG2pNURS7cvg0BjvAzzLVbaEP0ADCd21jDQdxebkg8kTHXOdLpmfl8ja0WQhwTPX694bo93wO4o9leMu+ox7CoYKOailK9AV37oL63756Zj6qOM6iBkfRKweCQoHtScPbzk/mjOB+klj5MXKCyfl0d8nh/yLB+zrjWrMqSqAppaQS0RbVZNhpZZ6m9kj8cX/DR7yuPoOfeDG8bScG19nHdhksFwFfMSYQ0z8p0hELZZ7pVNWNiYhUl4Usy4yFPKQvvmaIEXBNDtjT72nSBJVKGV3XgsVqJ0zUG05iBc88V4RpPE2Eih2inlfENn2duRVRhQjwT6oOTT2XPuRDekquwlJW5LhQ5dpt3ygCFMq7S5tb9WBXToDnYuZkcqqpWZMUKgndmQwDA+xrbFdxtBdaUOuyUOHWqn0IDFYQelEJ1r+KpJOEMXcxOP26SN+3gZvuzjVC/4ydEL/r0TLJsp0ZUiy3ZGdu3xftGRjVJ+QgtsjUx37eiojrh8QH0QZO+mIEcRIk0gCnFJRDNLqCaaeiS9BPBEUI2hnniXLzvImYQ+JvSiGpHbEOMEV1XKVZmyrrVXD2kbpoUAlOt7YZOo5jDxqgsP4isO1QotGgqnkNb15Qdd6cwQQ9IaJpA6dAWjL5ox5/WY83LEl6sZ6zL069BeIVQGFtFWqweBQSnbqpVIytoPUbWtJNNY+y6QJKy5SRx1FqBnU7/D56/+et7ODQwlTQanBwv+4fiLnq2LgSXUi+n1016+3b/3igXbKqJDgvGvkUjXksrgJtgNhpBIGAw77bbbba/D0Eraemwrc7f9xXaNm0Pi7ci1J62u7oqXVUKH8sfDNXSonaKwIUo67ug5/2AiOI2W/G/lpxRfjzCpftPSlD3eBC1RCR30Mi1CCE9c1nkFDOe8pIlrhRyd3YzB0trHpEYZbpLRTGOqacj6NKA4FhTHjmZiYVSTjEpO0oJU177vNCiprOK6Snm2nlCagNIoGqNojGyVbH2XBsIhAuu1qYQjiSqOIl8weje85iS46Wdozgdy4p0nMbzB+sdtH9roMu61C7ixMdcmY24SPsuPeZJPeL4es2obpREOGfkAugoMQWDbY+YI2hF9VRNQVYqmDHCVYh0aqlT5TpCg4SqEOpMw9b2N752sjJY0qePx9Jx/lPySa5tyY2Mum9HWYIhu0vJwlh5wK2kM/1etm9VZRhud8+1L1Q5iSH3hJualGFXt1GZy8s6+uvUM9baM255YA1CzUTTtBlPIllw79BlEwa2ENYxPDdHto1tr4TRn+pr7+gqdNTwvR/zpL3+KSYL3JO26x20QUngplqAlq27+nVJgDMKoVmPKeGkTDFjplXJpW2S0xo0S6oOE8igkP1IsH3pFkpP71zyaXHEvmfNhfMFIFcxNwtL4TN+TfMplnnK1Sqhr5Q29W0IMAhDK657LwDKOKk6jBUCv5glQ0YZP8P25u/2rsLnJdyQVi5pYNCxszLVLWZqYp+WUJ/mELxe++NNa0a9NaoPWPoAe6U3nh2mJ1RhJUwW4dYAsJGWmqa0ilr4H1ySOOpXYafrS2nbxRue+CBTq4IjVNMBklpNwyVSWVChubNxn44buV29Z7dgE3QitYaypw1CeQok2puR8gG7zXEsiTqGc9alTIduguN0yd7vu8e5A9msQtrWMvFsnETDoWxyO5ercTNgEJ318zLWlDLKNbXX6P0FvcXaV8YVzPmEwOAbd5+w+d+V8p7wRklhWyPa4OAkmlIRp+6V+u6LGHq8JESjU7BCODqjuTclPQ6pMIBtQlUM2DmFBOIcwIKzzEtgWcA7b9p/aUGC0oE79pJc689Nd6sOa+KBgGhcE0pKbkK/KGZFsuG5Slk3Iso64KlPyOvDunhM4P20E0Y5079c7KP92TlAaxWXltfu/qI84Cpb+3Gtv3ApHLLySCGwnnMBn0wsnvd56q3QyNynn9Yhn5YSrKuH5esy6DL0r1+7eAcIJrJXUg8hNZwUaI31Dc668llwpMEbQ2O1ruF3Et+LNbtRB4L/QscQlNUd6xVha5ta047Y2gfDO6umygV2gGoZjq1qLYkBasE1cneSxxGGF8KQ1fG1rqRgnqYXqJ+LoTtqCnUnMOxZWT2JYNC1RDkeB4bvOu+TqcJ11b177oayKjWXUqU10BalSeM33Wqit8o6t0WH4OpZ++o8VxKLGulYoMBQ+FgJ7snqfUAoOZ6w/PuTqU83iE4M6LGhWGpFLgpXPVgvHgKRE+xtsuKkvctrrRrnQp+7DqGEcl4yjirEuaKzkqkq4KFOsk6yakLIJqIyiapSXM/Is0C/PuVucf7f5XdYBV5U/Lz6vjiicZqbWzNSKEANtmU/YJaMG8dUaH3oonOayGXHZZFzUGdd1wnWVclPGvT5WXQX9fjfDeTwBWTtUNGlH0jcCVypEKZGFQFb4eYVD4m2PpzAvlwHt4o3IyilJc5RRTQRB0nAYLBlLRSpr1EDpsxPp8haNG1hXG9xmUcF2UelQ8rjXSxCbQHn3u5/55zYBQyvk1kj5Dq8qNVDt8dOYlyfSuA1hSban8gBIZ3uXrx70LKZtP6SvV/Gfo2C7pmTYH7nRavduqxFeHtm2cQqrBSJJXuer2uMN4AKFORqxeBhw8zs1/9nv/Tn/xcGf8u/WH/CXy7v86uaw72hwbaLHDVydOPS2sp+EVBIHddsU71P0tVU07TWxbjTzKumnv5j2wjZmm5C+qai7v9idwBlB1QRcld7i/jI/8Odl6LsiEFXbcWGQwo+R68652ikWNsEgWNmIZ/WEz/MDnq0nzMuYdaUpS+2zj62IXmdV+XhdO7fAbss02Ubgqlb5oRKo3FtVsgKajWXVxbW8xfrtZSJvRlaBoDyKqEcQRjVSOGpnKVzYtr2oPq5zGzZSwLez6KuIqwvQ98938aX2d0dY1olbux01DQj5SqLaxTDW9JIuVb/PjVXVTYnWwrTp3ZjLOiMQhnmQcmkyUln1DduHatV+vk0cbljO4LfreloLlcEmjnIsMadt1uTz1/ooe7wGnBJUs4jyQDA6WfHPpn/JP0/WnKj/j8NgyUQ/pLabOGM3W7IjsE6OaKRKJkGxPfwEwVWdctMk5E6zqqOtMgRn2yx3e0PqlHC7WFVHTMPYlRCuja/587lpJIvSl7V8JafUTpIbLyQwVWvidtblsKZQYimc7gPo8yblq2LGs/WEyzxlXerNMNRaItrJUKKV/hZ4pQi/SEFrdHYrfPkYS4dT20mxyiiEAdkA9XvWYLdasD5R1CNHoixLE/PEKC7NiIVJtkaxa9mmQ4V9yUe+dduDYPfuY7AhrA6KTanBsEq+Lx1oa5eMFH1NVNdys9nGN6/rtiGsm/c6rLDEomEsc2ZqTSgMz2o/qvtXiyMaJ70MsrSMdMkH6SUfx8/5OPJVxSsbcW1SFjZhbbfVEiN8Md5Y5RyHK8SkIj9NWD1qsyb/7zcufY83gFOCaiyxGoQVfFEf8ovma76o7/KsnnJdJS/dgIc3saYlstxobhpv+TbO95c21s8DWDchpdm4e4BP70uBbK2WfkzcVqmXu+Xv9rwNZC8bvC5biRgnWNchz/WYr6IZmaqIVEMgzEbjzW7qD1dNxMqEFI1m0RJp06qABIH1rl5gkNL1QXX/s/l7qCcH0EiJDRymkjghfSug9MNQRbSZ3WmcRDQCWVtE+e3S3W9GVgEUx4ImswTKMDcJT5opz5sxCxtvYjiD+iU54NtXlS/spvN3q8zVFkm53jLb7bnryMi3urTxsnbqbBc0H6JmezzYcJ3Dgs1Xrb0bxDpRhe+/ait9L8qMJzcTytLLKTsnvCV61/Fx/JzHwRyAFzaidgFzk/VyOt3xs7IVHBQVZ9Gc0bhgfRSzvL+XjHnfcAqqscREDmklX1cz/rK6w8/LOzwpp1y2LtYwINy5MLsYaqvVRlG337+xcpMh63TSpEMMzivnxED26mXrZEgS/oHWFbOCqvSXclUGLFSEUpZnwZhAWcKgIVSGSDWeWBD9OiujqI2kaoKtNQjh5xcEgUH1BOVQg/kGQ+uvc40BykbRNJZGSUygsDLASu8eBtoQSoNxksZKhPFJDLf+9jayN4xZQXnocFlDrBvWJuRpM2VhEtYm2sSpugxYp05wy4V+Wx1SX6qwo0HVW2s7Qn0ahxHCZ88GbQOde6WFoU31+b/bkoMhttqAaKvTB20w3bpkO3qre3/3OxQNqfCDHzv9rWUdsV5G2KVGlhJZCoo05MvRjHoW8CBo405NzqUsXypnGJJ9LGoOghWn4yU/PxiR53vl0PcNK/E9atqBFXxdTPiZvsMXxSEXZcay8i6W7Nw0MRhIIhx1K7PdWH8Bdhe/Hyy609cqHUpZpPS65KLfXjufsiU3bzFt3jckKW95WaT0N2PTKD8YAnDV5ppat5Xlqh29FQSmt4Ks8deIs+2PEUjt5b0DbZDSk5xWlqCdRdDNJOiPRdcyx2bdAGHgP3/ZKGrlqCUYpaAW6LBBtURfd25g7aD89gk3b0hWjnpiCMcVR8m6nd7qK2eHBZy7vXbDKlmgj928FLjuAn/29mX5oLbqSWMY0+r88doppHMvuc271lr3vldhmBAYZjJh2z2t2sGlcxuhhWGqcj4Zn1PfVXz2/BCexMQvBAjFk8mMvzq6y7PxXwCwdgotGg5bq6zbtnd5HdcmZWUj/mJ1n8/PD9AvAtKvX6NfbY83g4QmA5NaIu0tidL6kWvrJqQyrXS227Z4duNKndXk2h8pLUJsu3cdOQXSbr3fOPr33Zr9Y+MqDuNYUoJTm23Zrlh04Fp2P8P3SGl6S00rT04dMWlltlrEYEPUks22Nu05296PMJvrd+Mi+r7DMPCWlW3jXMIKRPNdzA1UDjWtmY5yTqIlx8GCscwxTvaNx9Ztj87q6pM6HaktiLa1RvjF1wTefBsMPYUd2Rfog+tbpRCDrKEVYsslNW1QfGilddnLjvR23cFhDZQdlFj0a2r/9lXnmmuRksqSmVrxafqUo3DJdR5zYxLSZw6dW4qTiF88OObJcbS1n5lak8lqy5K7sQmXzYiFifnZzQnVZczkmWD81bd/qXu8GZzy9VAutkTan6e+N1VRW0XdKGqj2syd6CvHO+zGmDoXyteTbrtRO/Mh6FRkhpYJ3N5D7R/bzRo6gsAMXtNaTlYM2nO21ypboowCQ6JrRmHJJCwI2prGQPjMdBfLyhu9RVCdCyx3buJDdLMNujV1rmWgNp6WsdKXLVhw7zvArpRlMl5zNlpwN55zFCyZtWTVyRjXbATwaF1ALQyZLHu3r5NRGQa4t1zC9r0wqGgf9PT1cwd7K6S1goTpCXPY59dlC3cnQg8Jyx/gHcWF1gLsij65pWWnbz7GE9hMrjmNF8Si5k8mD1mYI7KvK6IXa24+OOCr+ZQnzQEAs3ai7ZFa9sfAD8lQ3NiEuUn4Mj/gq6sp4bkie2pJP1+8yVe2x2vAKWhGFpn68IbE+RFwNvA1UI2irhXWKJ+Wb3as9LbYUgYWIUEqgxACpQxRYFAtOSjhtgaWOOfdsV2L6tYB3O1jvnp821ILpEW2BKKko2nd0LoGZ0WbXey25/oJyamuOU0XnMU33I+uB203Ph79vBpzzmirdQ7oLX//98bi6g+HcKybTcKoJ+TGEQYNQVvcbZ3wbqCxuPddZ2UaxfVVhrGSWPnBCUXkm5cLq/siSG9GbshmaWLmbVwLoLQBpQ1IVM1hsOI4uGEWrFnZiIWJe3XCV65jUAIxrDYHvJU2qDzvSOo2N3DosnYTo3v30nVaW9t3DoXrKbZTmeiUSWsXkMmKmcyZypo0qBAWgtwgz+dkT6a8+Nsx/3L2HwG+Q/5BdMXj6DmxrLloRsxNymWT8TfLU75YzLi4HmG/TJk8gfR5hbzck9V7h3C42BBFDSPtizenKvejo8KSIg7IpaaqoNN/dUb6pFxX0Q1Yq7z6QODVQTvr5na3T2xl0Yx52Qry/7+83I2L2Vpr0hK21pWWFqMEgbI0QZvVU9YH2IMGLQ2xaghVQ6xqZjrnQK9figd7q8kRqYbKvpzUGZLUq7ThjN1W3BWDyoXeknT44/ga2mBv1m5TCuKfxywPQ/5kHTG/k/BkMiNRfvBnKiumwZoj5WNZhdN8UR3xrJ7wdTFhWXuyWtQx61pzEOf89uQpd/ScMzXnOWPfCG2DviThm8oedt23rg5KCeOr0cVG34qd7XTb3yKxro4Kf/wkYitj2H2ZGsCpfsL0beuybKe33Tpn9EVBEyf87foRAP/+3l3u3b/kn5xkTIKCX6yP+WJ5wPObEevLlOAiILoSpM8c2dOG8PkKl+81rd47pEPGhjQuOYjX3A3n3NHz3qIIpGVRReS1Zl1pL9PS1h+54dxKBxiBbSRWOYyS1LJTZnC7YVR/Vg0ygDAgp66WaZj565Yru7iVa2NMFt1aVlE7uIXWnVXCMotyDsM1h3rV1wNG0g8B7oQHzuvR1to6LS3jRO/2bcWvWpIaZuS75ysbUJqAogkoaz99qlNccAyTYFsf61vxRmQVFHDw7y2rO5JVmfBzc8KyDjlK1tyJF9xrJzEfqiUPgzlrF/BVfcB5OeKX82Pm7VyxPA8xecDVtGAW5vzDDO6o3Jf8i1FfXOozeDsB+wE5dc3HHd0MO8p9wWqrYyUk1r18dxgSVucSWga9et37BxX4WhhqPJEx6GbvYBD9j3WDu+76/2fvTX/kSLf0vt+7xZZbVbGKZLOb3X0XXWkWjS1ZsGD4g2Hoj/TfYMD6bBmwYRi2MYAA2RKkGc147tZ9u5tNsrbcYn0Xf3gjIrOKRTZ5m83uusgHyMqsrMrMyIg3TpzlOc8pSX7/ktP1nMXvYzXw/C8zvmlO+U+m5UG25W9ePmb1ckryQjN/KSieeybPGsxVhVyVhPWGsDn02bxvCAEmsczSlpOk5JFZ8tRcAHG9pdJybXJWXcZKZ2x1wrZOaNH9ePb+jZzY9bh1giDkPrG77y0ZHt+1ISFWr3ti6MgU7/upBwMWhj4f6OkEfuQu5bpDC0eiYu5pbmqepNd8klzyuTkfW78SPKuQ8tv2IV80p7xsZ+Nm7IT4+upenw8eCN+SgJYxFycJdESu2fD/rVO0TlG2BtvP9RzD2LCTexo9q9ulz9fgnYyVbByzf1hjtgXCGbY+45tGcbmYcD3P2U4TmEaa/wO15drnfNk84DerU755fkSo4sfJWmJqQRXg+jRH4TlRipe+IQ6dUCMbPYqAvcM2jhWKWGkcq2u31R1urZbbiff9573wpG9BbB3+P8oTB2pnEF5ACISmwX71NXz19fh9zvw/w2U5vzk65VkxY/3NjPxrzfSrwOTblvyrNeH3X+G327fk3h/wxyCEmOJonIrEzaBxQVLIhpN+avFEN+RqQqY6rlSBEIEN0HiB74b1Iun7wRChN1w3zsE+DhIQBsO150lFixR6La1bJ+8ea3ygPyTaxupaz6ECyFRHpjpS6Uil5ciUnOo1T/QVT/R6FyUAW2dYu5yv6mN+vTq9kYcy0jE1DZnqyFWH9YquT3tEgYGA3svxehFGsqkLkcPlnBzDW2DUYYd4Xt6oeo7Judcfp3cLA51HLbekRjKXAl0rmosEWyR8O5nyVf6Iv57+gmTacjLfxgbL8xn6paG4EgyMBJ8EvAFUIOnliccvMSo3xKblURjsDcZilHfph0MMU2V2Y7B31cIBb2sAXzdLcGDrS+FvhHuXbkptYw7v6+UCXYLs/J3HQF9UzH+f4HVBmxUsLiA/9+TnluSiQmzKt0o8HvD9ICtJ8h8Lzo9zmOtsXAAAIABJREFU/vdHC/7m7DGfLy6Z6Ya5jjkdhSdXkWWthWeiW7ZZQtWZHSO8T2wPuSrnRCQFQ0/e3H2m6H/cYIXL0BNFd/dD8nyQBVZDglz6Pv/UkagdOz2RFusVlTMsu4xvqjl/y2PgzwEobazw1VazrlKqTQobg1pLfBbwiYfEk8xaHh2t+cXinMfpio1NqXxC5SLP70Yl8A4St5ae1NgbBQIhGLmI+zsiaIlI+oT8G7ih70ZdcI6wWqOdoyg70osUW2h8KnGpxCWCdmLopgmrYooIcHwdyK4CycpSncWPq84E9YOoLqil78dvCYZx8U1v1WQImHCTo3XnZo0J8Z7QuVdZHBUd+lzWna/9DtzUi99pbY3yNOxady7clBfdnG+bOctlwbwE0bk7jZW8uGb2a4XZFAQt0FuHWbfIVYUoa8Jm+1b8kwO+H3QVOPsPHdWppnyZcf4w5cXDOccnG352dMl8XkdPgyZWtnUzCkR2QY7tNsPj1quxvWZgsTsvsX3lbz/prHtDNIRzqk9cKxnPDSVicjyRlkS6vnVmqLLvqu1DXskIx2U3YdllXNQTllXGZpvRbQ2iVOitRG8FuoK0DBQNqCagG08zk7RzjZ1A9VhxaRy/WMBDsyIVBRcdND2HSvbVTdgX2IznXaJikS01AiXDyE8Lff4rVtj7IoOM7U7iLSaPvxsp1Drc5TViuUYoiVIKJeVOTVHIqJI4K/CTnvW7qRGrLaGsMH/5OQBeZzQnIFQgkQPbfZeka/0gw/qqcXkdr2OfMOrv+N87J9EwKJO+vvL46mvkWGXsgsKEmxrwS1vwu/IBX65PYGkw24Do7ja27uU5Yrmi+I0GKQjOg3N458CH6FX9ERNUDng3yE1N8X/9PcVHD5k8XbD5OGH7JOPyM01mLH8xfxaT0soyVc2NWZiZ7G5IWfsgKPuq9spmbFxK5Qx1fxsIprZ/jRaeVMfevYHDNOSEBtWGpC9eFaolld2NTorb8EGyshmt1yyrjOurCerblOnLWKjJLy3JZYt+ueo9dw8+rjeePKR5NKU6MxAU1XH0dp6aC1Sfntn0Ff0hXwV96mXvaiwJY1jaibCrfAYxsuBHFQsJXoudZ/UGvLvwpHcE7wh3cDwBZFUhygLVS5mEqsJvS3zboTa9G12lqErStVEyY78iJ0VA95QE2OlVwU0W+l1h4ZDsG/++115zdyD2dp5V/Gw1CvQNml3D9u3kigNN0NTOsG4SZBt5JPhb29r7xsG5767uvc2ElQO+F4LzuNUKlRhSo2Oe2xmE1zxrHvI/riYo5ceQzShHnnS9JExN0ie3B1mYIacTtdg8RjdMdfvK4Fxg9I7uKv+n0o4DTTqvuOoKuqCoXMLWJpQ2uaGHBdBaxXqTY9cGtVakK0l2DsW5J3/RxmLN9QZ/eY3bljcuhjpNSJQAMaE+SSg3hmWbsfb5qOOeSovdK1btZhbsclhaOhJ6Fr8IOH+zWrhqM1qnKcuUvAbVxov0d+HdjdVd44X2EJyHetfnE+omPsdOYEs3AV1BqKK7XAeDEuLGjLLB6HQ+ajvtT4WBff3o3XbcqBqGcOM5+abM3VvC9wZwlMK5tfgcgs4rWterPXpxkzExMvPewpMbFtGbRj0d8F4Rmha53JCGgGxykk1C/lLRfDnbpWUENBq2k4DLAq6IOR4AlVvStGOStSyyepzikquOqWrIZEc6zuILN7TW9mdeDtSdfVzYKZdtwUU94bIqWG0z2m2CKBWqlKg+16MqQVHF0FbVYEpHsvaYZYu+LnfphbbdrbE+oRS6DrmpMUaRrA1qo3hRznjeLUaKw02V3L2mbMKotGKDxAeHlRIRds9DDCPXbcqFneBWCXoDsnZxe74Df4Sxkn1Pn7jz5AnOQbPX6+PcOJl2CIdU7dFbiawkZZeMVns4gKm0Ywk4DmDsu9mJOYHhSnTbYN0wXLf+dtsTe5PxGomit8LDG95UGMbM78aPDaGhHTrKfRQWY39q79sYqtv/dwgFPwhCVeG9R2y2JBcJyTc50yzB52bIhBME+FzTzk2cVDNVtPO4frtpQjP1lAvL5ijFHwlmuiE1FQtdcqbXHKmSQjRMZMx/+SBpiV0X25D0skGTG61r0ZsyPCvnkYN3laMvY9Equwyky0Cy6r27qxZZW0TnonPQWURnCXUDTYPvLMHa2N5y6/wNdQNyiwqB9HqC2cZQ8tJOxgvzcG52cIMOtM9iT+TOCO/3ugLU1lA2CZtthr5WmDKgyhbfviZU28P7nz/gXTy3bieGpUI0cYPM1pJdK8q1ZFllbH2KRCJ7Ebt078tKEXaiep4obdwbqttG6rY+1XfN89vHXWHlayuQfTLfB3GjWjjE9ZU1NLVBdr2w2K1FIW43iN31EW8YS37AD4Ng44nM9g4u294MQT0pSB4cUywmcTDEcayQNQsZbycJda34so3EyEU651G+5qN0yUJXTFXNkSrHKTKD17L2Wez2sDmViyPtbJA0TvPb5QPOr2b4q4T0QsXQ7oUnf9mRXJTI8yg75J6/iN9jf7vf6ssHQtuBrME70qtj0kvN8sWEv579bNRl09JHZQmvRpmZAeNchP75zg3ySoxKC1frgnadIDea4kKQLh2iamPe7Dvw7sbqu67yt0Od0dX0UMb8jDlPmQDlWc5mG1txAAxREmWmajKpbrbR8Gp+aV/D/W2x/5q7tKz28bok5njVE4pCtkxkvFK6EBUar8qcbpkyWQt07RD27gPxnQbp4FF9eNw+uXcCU0QClYSuI2y2SCFIrEc2kexs1prsWtFcKtrnim5ecD7NeWHg77JASD2oACqO0hoVFAKxfceJKAU8kEsDDFrveiPINgKzhmQdSFeO9LJDLyvkuiL059a4pvYHtL72u3IzFAw+fjdAX26Zf5EQhOEPlx/33LDhNT3z3O8Y6DcoU31zMkEMkluxpUmALgVZDboM5BeO/Nsa8ZaTxv84z+p1uZM9QzV4D1GfuU9yb6OcjASSqqF4nLJcGzYuQyLIhGciGxaqHOPj/QGpr2t8Hhqj4+Pb0shvX+mD2zLG4RVvbaj4DFNshhHbhWjohKL1mm2VopcKswZd+puSrcHHfbL3+1vhkK/6cNi/0N5Od4Q+xVHV4OIF2GxihUwnBhJDkRtcpnGZwmWKoGPFyyvZl+ohDI1yYzdNPzknME7NEWF3rxqHqj2ycajaIssWUTVQ1YSm3RVqhm3eOwffhP3zM/gQIyLnEefXTKUkvS5ojnoNtf23G7bzzjft0x/Dbts3ZDagGo/sPGpdI5db/Ort+l3/CM/qLQ1Vf8CF9AQXX+c3kQ3sNxsQkuLTE9Q6ZWlzlJAYARPRMlMVPshe27zF4OIYa+FJcH07y25g6DCrcKfMsMek3ZNd2ceOMOpv/D4kDNUYar5KCI168446mDj/rPes6mConKGrNflSkC49qrSvhsTjvty7qh3w40Psr1tx40QGxuMUrI252bK89fq9ApAUkXi87928hfEAbuY497FnVNyt5243GI7n4Js+04fd+Qm7FA7gzs/h4hIlBcXb5lnfBnv70/uAf4fo4T1uxQEHHHDAD4cfZMBv6C328Hj3h/3HDtk4zFrwN9cf8b8dKdb+MSsXeR37AyT2Jy4Pcsj7U2FuD5t4Hcnzdm7rFW/rDknl29OdB95X52OVJpWWEz3lUm9Yu5zfXJ0iX6RMvgkUzzv0df1WZdkDfgLYy0sFH94cot+pjrcngvcjpxvHc9C/wR8J/vV50xAguB/9e+zj/Rmr8eANRupWgv0OyNqSrOGLFyf8m+O/6jV1Yj9gVGrUVC4ZO733S6D7uaW7jNPwGtiJhb3uNTuRPvHKew0SGPud6MMcOOsliXS8NFOOzJy1zbi4nDJ5IZj9oSH9doNYbvAD72ycDPlqaHEDr6GFHPCB8F3H56eO3ujGc/A7qmw/JWv0HXj/ntUto/Wmgy1bS7IMrM8z/vr4Z8zTKBlzmm5ovGbZ5Wy6dBTk3xfqv/E+r6EoDNwPLd0dxuqmFzaQPIfbq19LjP/j+sEALgiM9BSm5cI0rNsMcZ4weebJfndBuLzCt927eVY7Scf7eaLcd/yp7PPBS3yr/7sf+EHCQOCtdoKoGvJLR/G15mt1yte5w+Qdk7zFeknTaFyn8HZQUdx/8X5Csb+7o6P79nyzcfPGB2K3uWOV+g65ikGTZ5T+6O9VQOiANB7XSqbPJMWLlnB+iVutvnMfHHDAD4Z7ZIjeBj+csXoLhG1J/m3Nkc4xa41PNC5JaRLAQ+JAdjdLpHcO/nhNwWPkhfQckVccsPDqY3HHczee742a6O/HUnScc8H8C0f6ony9NzWEgsPj1/39T2yhHXDA98WPa6w2W/RXF8zXU4qvM4ISBC3jmOkAwgdEnFH05jd6E0v3beudd3nMd3zuzmjtyHdBAEqAC5irCnFxjXuTtMt3fZ+DoTrggFfwoxorv93it1v46qZo4n3D/nYfpPIOOOCHwYFndcABB9wLHIzVAQcccC9wMFYHHHDAvcDBWB1wwAH3AgdjdcABB9wLHIzVAQcccC9wMFYHHHDAvcDBWB1wwAH3AiK8A1taCPES+OKH25wD3gGfhRDOfuyN+FPAYV3/pPDadf1OxuqAAw444MfCIQw84IAD7gUOxuqAAw64FzgYqwMOOOBe4GCsDjjggHuBg7E64IAD7gUOxuqAAw64FzgYqwMOOOBe4GCsDjjggHuBg7E64IAD7gUOxuqAAw64FzgYqwMOOOBe4GCsDjjggHuBg7E64IAD7gUOxuqAAw64FzgYqwMOOOBe4GCsDjjggHuBg7E64IAD7gX0u/xzItKQMfmhtiVCjD/gLVRMhdaEzOBSiUtABOD2y4a3E695j9D/TQ73AdkIdB2QZUfouve6ze8Da67OD7LG7wcfZF2/BUSW4nKNS8AngPZo5Um1pbYaZxVYgbQgHP19iDffrzvvQUqClnglCAqCjDcEeA2ogNQeKeJrAhCCwDcKVYOuPKJz4B04z4dUE37Tun4nY5Ux4V+Kf/V+tuouSIWQApQCHwjOQfBvNADq5Izun3zC9S8ztp8IhI0HUdpoeIKEoHePB18yiN6weRA+/s0VAZcFXOGZ/kZz8neW6f/7Ffbrb25+qOgNk5D93e73O7dZ7FnJ93Tg/9fwPx00w98T3npdi1tXOzEsJr977nscX/XLf8zqnxyx/LmifOLhtOHByYbPF5f8+vKUq4sp+mVCciVIrwPZZSC7suhNh9q2AMh1hZ/ldMc5zYmhnUq6AlwusBm0xwF7bJk8KCnSlhAEzgs6p6j/YcHi7+Hkb0vMsyvCZkvYbPF1/e775o/cD29a1+9krH5wBE/wEiHCzQVwC3I2Q85nhPmE9vGM5ecpm6eC+olFOBGvPq246WUF4lVG91cTCcjQe1PxntShUkeRdpT1DNVohP+Y/HSBvN4Qlivc9TIeCCHiNgpJ8CEarDdsc/zQg979vcZw3Mff/c5gvQlSIZQCKZCzKWI+wy0m+EwjQvSM8J7ycUEzlwQJshXYWrOtEy7SCa3VCAEu83Sz6CZ5I2inBhEMwhf9Nh3hNbhU4E300Lwe1j74JCCMR4hACALrJJ1TdJ0iyEA3ldQPU4J+gGwWyMaiWotoOmj7KKPtCE1DaFuCtT/Irr4LPy1j1SM41z+426uSRwu6T0/Zfpyx+UhSPgn4JxVPH17TOkVrFa3VWCuxncZ3kuAEQntkb6y0dhjjSE1Hbiy56UiVJZGWTFn+k3rMMp3Rzg2TR0fMvyzIfivhenlrY31/d8dV9jaEOBis+467DNb+325DCIRSiMQgkoTw+IzysxmbJ5rmqA/pLEgb8CYaGATIFkQlqbYpL5Wn6xTIQMg8VkVPqe0vyC4N+KzfjsyBF9BJRCcQVkCI0QNB4FM/hoA+QOcUbauwrUZI6Kaw+UhRHymkC324GTBbj1lbVNkiVxVitYHlamesPsC6/mkZq9DHZTd+fxX+eMrmacb1LyXV0475ow1/cfYt//XR79i4jLXL2NqU6y7nsim4qnPq1pAaS6bjzi1My0la8ihd8TS75EiVNz7jNN3wHydP+ObBguvFhCATzHIGv7tj28SeV/Wmg3YwVH8auG2wvgPCaESWIpKE5uGE658bVr9ymIcVtlP4VkErkaVElwK9Faha4BOJNZqtzOL7yIAo4voV0QYhtePp8YpfzM8B+Kezr3jRzvn19ow/rI+4Whd0rcZVGtFKSD1KO2TvWTkncZ0iVAoUdLNA0ALhGL+jcJCsFOmVIrvWpEahQkBUNbwpRBTvN4/70zJWezmr4WqE1git4xdXCpRk8+mc9VNJ9XnLk08u+dXRS/757Ev+i/wLSp+y9Slbn/DcLviiOuUrecS1ypkmDVPTADBRLZ/ml/wsfcGvkuecqYo6qP6mYQpT1fD74gH/Xn3MppqTLScsLj4D68Bagu2TkJ0lWEvobB/Khvj8bRw8qz8d3GWwhpNbm+hJ9QYqTHLCJMMWCavPEjafeU5/fsk/O/uKZZezajM2bcq3l3O6iwzZKWQbvSu1UXgnCMYjUo9M3BgVJNpylNf8iwdf8q9mfwPAf5eXfGUb/s/85/zb9Of8Q3LGZZmzTVLa2sRoIrVkSYcbUi4AQRBSj03BHtFn3UH46J25TOBSic0ENpNkmcZkKfp6RqgqQlXjy/KVfTE+fg/r/qdjrIRAZikiTRFZSjia4RY57SLBThQuEdhM4BIoPxLUTzqOH675aLJiolpKn/D79hSPpA0aHwQbl7GyKcs2Y90kdF5SWQNAaVq0dBSqYa7i1eHaF6x8xtrlXLoJXVDkquPhfMOXP0t4nqasP/4YUwbMJpBsomus183ONW4aQt3gm70QdsxxHQzVvcftYsmNk1IilEIeLQiPHtCeFdQnhm4iaWcCO4HysSf9dMOvjl/yWXbJlS641gVLk7FtDVetomsEqoquk9kK2CpsJnF5wGUSn0mcVfhUUBvL2mZcuikAS3/NNmhckGjhmOgWCsiMpUwTsqQj05aJabFeoqRnJQKNCijlMcZhlKPpNHWZEDYa2Ul8As1RwBaC+kwhP1OoJiNZHzP9qiX98hJ++2XcD97d3Dd/cp6VkNFQTSeE+YTm8ZTykaF8KGmOwRYBN3OIwjKZ1Xw+3fJ0esVJUpJKS+kTvmxPkQSU8Ejh2diUTZeybRPKOo3xuY5fubGaRFrmes5CVbgguXBTli7nvJthvcQhmeiGX8zPOUorLh8VrP8s5eJygrowpJea7EKRnxvy5wYjJWJbxe/TdhD2vKuDobr/uCv0CwGkin9WCqEkLKZUT2esPtdsn4CdBfy8I5s1PF2s+dn8gj+ffsMjvWSmaha6YmlySpvQdIZtoyCoGBKWoCpQE0HXgXUK7wU2i2ur1I5Vl3HtYoL92kPp4wU5Vx1T05DpjplR2FyRSEsiHamydF4hRUCJwFp5jouKo7TiLNvwvJ7x65enVFuNdOB1wBeBNgmxMGU8QgVYG9p5ymm7QP0+7p8xjfee1/xPxlgJKRBpQpgV2KOc8pFh9Zmk/NQyebTl4/man88u+LPJM1LZoYhJwi4oruyEq65g41JSaclVRyFbKp+w7jKq1tBUBmclnY57sjEKLafMTMOxLumC4ryb8bydc9kW5KpjolrmuuLz7ILToxWP9ZIHasv/vP4r/o+X/4jfPjul/TrDZgpIka1DATQtQorvLA4ecI9wi64SK4HRWx6oK0JJMAa/KNh+pFn9IpD/YsXPFkt+OTvnLydfcaI3TGTDTPbevCtZ+5xrXbCyGdsuiQnvLkdsJHoL+bmnnYtY6QZckDjACWgSzaZLWfbG6tJlbEOCQ2KEY25qtHAY4dAyLsjh3KmcwUhHIh256fhkcs0vipf8KnvGf64+5qIqqM4LRAe+CHDUMZ9XPJ6teVSs+Flxwe/KB/zf3Z8z+TZl+jaV0e+BH9VYySxDLOYwm+BnBeWjguqBon4gqc4C7Zll8nDLp8dXPMrXnCVrpn3I5oPAB0EX4lVNioARnlx1zFTNVNV4BNXEoITnMukwymH6A2aU4yzbcJpsWOiSiWzYyIxEWmTPd+iCZONSjHSksmPiW5TwpLLjUb5mfZLyMsAmTemmivJsSrqckF4fk159hFpWiOs1frm6Gc8fcP+wT1e59ZyczQAQR3P8YsLm8ynlY4E/a/h4seQ025LKjrXP6DqNEXa80O5fIFdthvOSadGwcoq2TZGNRLWCdi6wObgMfNp7NyJgreSL62PW3Z8B8O+yT6mdYdVkrPpoQssduTTVllRZMtVhg+K6zlnXKXVr6JzCBYERjuftnNZqkAGfRMqPAJyXtF7hQzRMueoIRx2rT1Pyf/nnAJirCnG5xF9dvx1H6y3xoxorMSkIZyc0T6aUZ5rykaR6GLAPW4pFxemk4uPpkofphrmuODZbfE9D74LGsTNWRji8jIZqoSpmqiITHUY4ZrpmW6QAaBndZyUCR7rkYbLiSJVkomMt89Ez8wisV9i9A+ODZOsTfJCcJFs+X0hy03ExK9g+yKjWBnOtSK8N6ZWmeJFRfJUgrYO7ko+H0PB+4TV5GDGP+aLuoyOqxxnrTxXVY8fxyYbPppcoEWi84Q/1CUbsUgMbl/KsWvCynHC9KTDGMklbjosKrTznnaTtEoQX2EmIxmriwATQHgQ4q1heTbg+j9vw207uuIYdMXFuQrxpj8gdJrVkaezKqKqErjJQS76xEttTcDZdSmsVGI/Leh5igLZVbJqUTZKysSlKBCaLiu0nhpc2enfTr1OmXxhk2765WviO+HGNVZbRnhWsnhq2TwX1Rx0nT5b884df915UQyY7Uhl3rOppDV3QdEGNNxckRjqMdExVzUxVnKgNKDjRGz5JLseku9trh8xEx0xVzHuXvJANM1WzVhmN03RB0npN4zUeQe0NaxXLyA+TNQ+TNZ8VGdujlK1LuGoKvrg+Znk1obowdBONtAWT82zvS7//KskBHxC3j5eQ+Fls1akeZaw+VWw/9iSPS/7RyTn/uHjOeTfloptw1UyxXuKDxCPYdgnfLmeUqwx5bWhOWsyJ4+nsmuO0pO4061YivMZlkUslcouQu23wnURdGpLLuK7T68iLEr0D6BW4NCbIXaKwM003MbTTmNcSW43eSnQlaLzgBaClx/eEUaECIfVgPASBs4q602xtwsrGi/vj+ZqvniqWRQ6ASzWqKZg+21v37wE/qrEKeUp9YqgeCeqnLaePVvzV6Tf8t4t/YCZrHAIfYqK7C4q2N1KDR1V7g0cgCWNi3fTxuRKeTHRkoeNI7rwaf6t3Ww4GkD6cJJBIiwsC78UuJPSKysUDXMiWTMa82EJV1EbTec0yzcl0xx/SIy6KCVU7JT9XFNMCYZIdreGQzLrfECIWhKRAaI0v4rpoZ5L2CNzMkWnPpkv5bXXGi2bKVVNwWeYx1HIS5yRdq/Erg14rzErQiIR1krMscpT0eC9jP9/QeWE80ngE4L0guJ78aQV9wIBqQ+9RxXYyKaLxcp2IRFMvEE5i+6q4qgW6FD1NQtKZhOd6hhCBpjaEWiGciGeYD4QA1io6p2i8IpWWB9kWeRK4KuJF/2p1SvONZJqY93pB/nET7ImhnUmaB57TRys+ni2ZmZramz6JOIRf0SMawrHBm/IIun6HKeGRBHyIhm3rU5yQJMJhhCUTHYowuuGKQB10pCr4nLXPKH2KR6CIua8hBwbRiEXWr6Dxcbe5/rM2LqXxmq1N8UGwSGvUIvCHo5xmrnGzFDXJCW3Xc7LegkB6wE8XQkaip45kT5/36yGJbHK8YHud83fVY/4/dYbrFL6LxE/RCUQX28F0B6qJ4ZpqIVxLGpnxa3eGlAG7NqhNbC4OShCMiGGdCAQnCK0CJ3CFp0qix14/6N+viQZIOl5p7Fd1JH0GAbKL/bSI+L+yUtSrFLxA1BJdSYSNeTKfBYKMBivmjCVSeGa6QWaBSc9hfDk9xmaKkJi+IPF+1vuP61kZRTcV+NOGvzx9xkzHMm7pU6TwvVe1C5tUv9cHg9R5FY0YAkPAyMjurfvSLRKU8BTCkQlLJhxZb6wKAZfec+0L1j7j0k5HI6mlR+NJ5a7vaciV+SDoiFcWKQKN16y7jK1LqK0hUZajpOJRvubbozndXNNNDTrP44HznvDh2qkO+AEweFQiSxF53leDwWUCn3iEE8grjWwNqhGood2FnuhZgymjBxREGBURQCCsoqsyggqYNjLZZUfs88tETJshCE6CjRXCUDhUHhdVlre0rabeJoitQvXGZjBeu20Yqpu90oiI/YhKAUEj3J7XZaGd9WIARkSFhv5mhCM1JUdmF738+8lTXKYJRiOUigye9xBNfFhjdYtA51NNO4Pjkw3/zeI3NN7cCO8gei/A6Dkp4emCil5Vn1wfoAjRuAWN8xKJ7z0qTyYcM+ko+m2YyYQ6xCvB2uWcd9PxfTLZ3QgrBw+q9MlYgRyuLJUzLLuMZRtd/IfFmpOk5Gl2yd/MH7OcFtipIkwLRAixl6ppfsi9fMAPjcGzyjLCJMcWvbFK+0pdALOUZBeQXXm8EmMzsa5DJBWvLNIFbKGweWSGE6LB0KUgKNH388WPlB0jdQHi88L1xNGi4+nZFQB/cfSM6y7ny/UJL1ZTqnWKKDVhK9FCIJuoSKL6JehNvDkdDaqqe+WSVqBr0GWUofFKxKZoL244SFp6CtlSqB0dI5u0uCwjpAphoomJBuuOro53wIczVkIgp9EgyElBmBasPy7o5oGztCURFi8EDoHqNVsGYwHRWyq9Hg3FECKqvV5C2V++HAIXFHVIqEPH1qfxfTx0w/uFhpc+5doVrF3GxqW7bXWxuhjpEK5/So7elSJ220vhSBFk3mJ1SysViXRI4XFBkmnL+dyzeaww62OS8xT5QuA3mx92Xx/wg0IoichzwnyCPS5oFnEtNkcBcdQSvIgs9EaCHVjZAAAgAElEQVQibVRHcEk0CsJHT0laBSGqIri0NxhJrNp5FQ3HSArtDUvQgSzrMNrhi9jXB/BwvuFRvgbg2JTYoJgmDes0ic38KmCNwuV9vnYoaPaqIwyqIyrE5L0MhEqh1opkKZGtoJsH7Myh5i3TomGe1sxMPeZv96ucqbHUM6gfFxSrx8hNid9s8ev199rvB6XQAw444F7gg4aBchJ5GP70mO60YPtI0i0cR2mFEY5WaFTvYw5e1ZC7qjGUPSNdioAUHiM8SvqYZFcxZh+S8PsVw1oY6mAiT6r3vhI8L91sbLGpXNK/Pl52tHQoEdC9h7UPxc7jQ0IqLVZLtI/0iYHwV5iWMHVUDxW6SpgBadnssaC/n1t8wI8EpSBLcbOM9jihWfTcv7lnMauwXrIpNV0lkJ3ApcS+vhSCCjFPNKh36sh/wgSE8lFZQQVspeHCIFuB2fTrTwXytOW0KFHSk0hHoiwLU3OWRK9lqmpqbZibmm2WEIKgTVzsLbSKJO0o0o5p2pAqG/NPxPyTFH1VXXpebqdcXk1oVIouBXbqkbOO2aTmwaTkJC15YLbMVN2fi7u1PElb1jPP9qFGb49IXhqEdfA9PasPGAZKwjQaq+60YPskoT4TyHnHcRpJmZ1QdGJgpMcdkMqOLqjeoBiu2pxEORJpQXXIoEiJ1UBFwNMbKq9pCGSiow6G0qfc5pBf2ClLW7C2GVvbG6veR5YuoKXDiNiaYKS7YbikCGSyo3TJSCLtROy9Ung8gkK35PM6Nn52ClMZkov3yz054MNDKEXIU+w0oVkounn/h0XHx4sljdP8rkpw2wzbCOw0YPNAKBwyt5jEkiaWIm1RImBUbHnZx7frGat2hllqepohIvEc5TVPJktOkw2nZsNCvdoZ0SjDzNSUSTJWsJ2PaYwn0yW/nLzkV/m3HKmyVyhJKX0yvl4R+M/lR/wH/YRn4YhOGZh3FNOG0+mWs2zDo3TFQlcU8tX868S0uJmjemgwVYKwnmTz/df9h02wp3GHdDNNfSJpF568aJmotk+g+75vaeBLWTLRotA3LPeQ3LZeYYTHIXuqw17lsG9mHhDzWHLkWcWkefz6yR5rPf5t9z5dkBDABgkYJIFcdTjcmC+TwqOF76uPOy8sU5ZJ1lLPE7qNpJ0KfKLGXrID3eqeIjG4eUZ1Ztg8kdSn8UAW85pH2ZrGa67nORetojaakAREbjGpZZK3HBUVZ/mGmd6d6PtrzweJDZLNNKObKdpa4NPoeXVO0Tg90meGi+cQETgkksBUNbRpiZYudmEgsF5ynJQcmy1HqmQmK9oQ+1qboYJO33sYBJm2JFlHMxNkRccir1kk1UjniVX3YnzNgMbpqMwrI0cMJd5J/+t1+KDGKpjoNXUTSTsHN/VMs4ZctdFQichlUoSRE5XJjiREDytXHYlyI1EzVgRji0AnI5t9eK3aI4cCo6Ha36lKeArVcBzUaAyH5L31isZrmr7lpg7xymS9YmoaJroZjZUPd6f+UhUVSJO8o53EoRbBKL7/YTvgx4QwhnaRsH0s2T716IdRaeOjxYrTdIMPgnJuECKwnmQo5ZEyru+jouJRseZpfhVpOi4ZeXo+SLQAHzwT01JMGzYnmkrqmNyWsK5TnqsZbb8+H5g89q6KmAYx0uIRpNIy0zVaeCpnaL2ixtB6zXk3G2lA592Ul+2MqzYfnQCPoLKGsov6V3LWMM0bpklDpmJf48alVD6J3MZbK/qqzBG1QjWRpCpsAHePqAtCRq8CoMsF3Swgpx2ztGGqmj5HZVGY3rOyZLIjEx1eSArZjn17g+fjeuPRyXjgMtlBHw4aET0fyd07ySGQxLKr0oGZ2jHmAS7tBCw0XsUD7Uy8olk9HpyBhzXE/PsYmPDztGabG5oixaUKb+ShqnHfkRiaI039MFA8XfPRYgXAp5MrTk3My8hZ4DipWM4zWq9pe4/oQbblaX7FL7IXHKmSl3bOuZ1y0U7xe175VDc8mJR4L6iyFKUCUnqqxvAyTGisprQJ2zQlVx1zXY2vk8S0xbEsmaqGa1uwsZF20zjNeTNla1MqZ3jZTLmsCpZVRgiRluC9RPbNz4m25LllmjQcpRUT3YyUndbrvn3o5tqvqgRVC1TdM+pbF6fufE/8KJ6VzQV26phMGhZJRaGa0dIPuafoYfn+93grZEsibTRQoe+xEuFGS44BEB6JQBJDPbVHDLntBSXCoqQfqQkDU76WhkpE19j6uNDKztBa3ecYLI3XMQztk5T74aNHYISn0C3zrGGZu168/+BX3XcEo+kKQXvs+Kdnz3mQbgF4ki6ZyTo2z8uaT5Ir6hA9mct2wlWb8yDd8jBZcabXHKktbc/f28iOpucXuiDQ0nGSbTHKsSkS6tbQ2jjYoW01a+LFunGaRVKNBkOKcKOjw0hHLlsaoZEi9BdezUWYsLUJ55sJ2zKlKw3YniHqgcyTThum84bTfEuhW2am5shUY4531WZU1rxSgHIuXpCDiuvda4m6d2Fgf6IGDSHxZCYOZxjJnDDmnrqgWPmcNkT/KLbDJKOxMcIje05TrBgKSpeMRscHMRLWhj6+/apFIixt0JQ+ofHmBhG183EBVS65YdwE7GRgYTSW1qsx19B5hRaK0iV4BFp4ct1hEtsTAw/G6k8CA+u7v1gCNF5T+pS05x3NZMUMRk9/MBylS/lDd8JLO2PtM2pvRn2ptU2jVEyQ5DoWnz4qltTORO/ealqv+sEompVTaOHJ1K4tYmXjDIKtS2jdTkIJiI35VtM4zbpJKOuouiAqNQ6YkA5sJ+kSR2s1Wjpy1XFkKk7NGoUf+2RDEEjpx2ErAA+ON1zJwHKe0B5pXJJzVM7hq6+/1y7/sMaqTyx7AyQ+TpSRdkyE7xuGQVGhJBqMjcv6uF6M1bnYr7cT4fNBUPmExmkqFxdArjpyFUPIWL2IPQeFbKiDYWkLrmwx7vxhG2xveGzf/6T6z5E3jFXsTfTskqI2SBofr2LDFbLQLcZYOr3bBwfcYwixa1HZWw+N16xdrHplqqOQDZnokHtFoy7ElMWz9mhcyxBJyI3XXNQTXm4mCIhaWMWWX+YvqPvuji4onjdzvi4XPO+mVE1Cqi2zZKfztrUpz6o55+WEptNMs4bCdMxMTWM12y6hbM0NQyUriWpB9r2KeOgKjZvHi36uOhaq4kyv8UHysp2Nn6eEJ1Mds743MDs6p55dU39k+IfFQzZ1wfSb7HunPz6osfJqZ6xU6pgmTd8w7MbE98BMj71/emxzabym7aVatHCj+uFQmo2qDAmrLmPTpay7NBor3ZGpjolu+4RkdNmNcJQu5coWXLYFpb3Js9qnKAz3UgSE2DU0e2K7w2DUBuMlCbRiCGsDmbJkxtLqgTF8yFr9qUD3Xj0QE84uelEn2nMky1ErLRGOTHRc2inftnMu2wmtV0x0y0zXzHVN7TRXZc5mmSN1oJttOEvW/Ivit2OfrAuSvzUfs7UJz/ycpjaUiaHtm6m7oNi6hPNywuX1FNcomqnGz0omuqX1ajRUbWUIpUJVAl2BqgSqAV0FQGAXEuclWngmumGhS870itqbXb62j2AyZTnpz60HxZZHZsmZXvFvpn/F//Liv6Q5MuTfd19/z9e/E4LuVRQ0ZFnHPKnJVdvzknbtM5EvFfAi8pUy2UXjIAS5isnrfZnWLihs39gcr1Y+8qKk7+cAxkriEMvfxkA3GCkRIR4guedNaenJdIfoDeBQlRwIqjIIbpf5lIhJdi08iXJx0OR7iN0P+BEgY5+bUIpQpHG8uxwulP3Fx2lciJW4rU8pQ8okNKP8EDB6VqU1bLuU0iYsZUaiHM+2c9abHDYGpz3XVc7LdsZLO6cOu77ZP9QnPCvj//q1YWtStkW82NoQL5aTpKUqGhpp8F5wuZpwtS7wLl5UB9UG0UlkJ5CNQPVN1rqKChJqI9luM56Vc4x0LFRFZyKN6EGyYZWlZKpjkdScmC3HfTPzQpfMVMWRKpmoZiTCfl98WFJof8x8EpgkHQtTMVXNDVd6kA2Wwe9uPiBVGF3l/XBslDf2UV8HYnPl1DRo6ZmomBhc6Gokme5/lhGuN2iqzwX0BhWB7N9f9kZHEr2kpI/PB9VRIwReRD7WwGAfQtQhn55qG1nLN3uvD7gnEEYj8wyRZXSTBJcIUGFUlAXwQtIFycpmzFTGSmYUomHr097YRH5U5QylTdh0Cb6NhE3nBetNjl0lqFLiE8G6TPmmXPD36UeUPhlzUV9uj/nmcoG/TtBbSZcb1k3sba2cIVcdZ/mGiWm5qnNeXk9przL0UuFNIKSBYDz4qKgQexV7NYYmoNqArsFsBPV1wjfJAi08J2bLNk0xwnKst1BErlUmO6aqHlMsmejwQXLpply2k5gLc/dsFNd+zmqaNhyZWAlMhB3DwH3y5+BlKRkw2N7Y2BuJcE+UHR44UMDYTBxLujXHurxRcQTGxPzwnlZGr8wKjx8WH7FZGei9NDfmyxJpMUOuTUSl0iEElHte2RCq5rrbkwI54L4hSsJkhGmBnZqYd+09K3urwmyEZ60zFirl2heUPr2Rc2p9THCXTULT6TiEtJOErUZtFWYrcA6abcLzcspv0jOWXca6zdh0CefLaHzMUvatMIqqjTnXxmmOkkgxSKXlWbLgfDlFLxWz30M3lbRHIbLuJTGh3vXN1V2I3lUT0Eqgt6CXklpnvEimPM/nbIuUiWw4UZtRZWEkdO85HQ7JymWsbYq0AuHvmbEayvZBEV1HvY1Vuj0mOGH35Tv0+LwPEqNcn0iPNAVkHH8NPdepr77F1pioRzVVzShfsa+cALFKk/YUhFRZuiCR/qbrs98ztesX3IWSHjGGgYNXNZBbB9XStA9FgwqHMPCeQiQJYVrgj6e0c4VP4oTk2pkbWmfD42HC0i6nqsf/aZ2itjrSEWqN3xpkLaN+VJ83Ek5g15qX6Yy/BTZ1StNobKNhYzBriVlHz0hVkqqKYeDLeopHjFFLKi3BRy0rs4nhHR6CCaACLuulaHzUxlJNn6N1AdXG6dAuV3FbvULhmcimN057HSJ73SG1j724tY8VTEI890WagnM7tdx3FOP7wHpW8S6oQKbjwcxEhxKh51MRZ1UDMvjIj/L0LFmPgbHXTwYPPnpinXRo78fSKezyT43f6bXfxlDl0zKGmqrPTe33acW2HvBCjLmsrs9Ttb0Ri+8VRl6LHgipPeclk13MrwkOntU9hchS3NGE6lFO9UBii3jcV002DlkIMLbDDLlMYJRQKWRMTdigKJtkZ6iq6CENlTjZxtMgvVDYKuf5RRrVRTuB6bWmhI3KLkHF37vraKx+F075JpuTGkuRdLROYTuFLAKbT6JXZU875g+2FGnLtkloGkPdaNoXCf6FRDq5a5XpnUatHcdJyefJS2ay3uXQ+vB2v92t7ZV8AbRw+CTQFRJ1+oBQ13EIcFW/cyP/Bw4D473Xw5WnwQiL6VsFJJ5ERO1PGemdOBHF9BK5Y6PLECLnqg+/TOjDw1ukzEGgr/OaTuhonEYvLqBC32Xel5b1cJPuRuuBHfJkBFoYQzwt/ejJyd5L20/KD4t0yIvFBPuH298HvEdkKd0ipTpV1CcCl8d1tG4SbO/dOy/xXtDZqCKr+zTDXA9RRIMRDuslbafxpR4Nld70+aM+h0SIie5wJQhCIXzUuBIuRLJlL5rn9aD62UcYpaRKEkoTuEp8HNnVySh9/HFAHrc8Plnxj45ecpZsRqrP1iX8P8kn1G4St8WFyIfsUxdGOc6SDU/1ikwErr3j2mdRPjwMMxJ26RPX37T0hN5Y+eM5chNNjmiad+6N/bDGajhRVeQ/xZPZjgbqhl567zn5vsyf9HIxitA/H5PaQ05IyyiEN1QE3cCBkr1UTNBk7KqBg4FSg7Hx0SsaKAhxNhp3thMMxkj6PvEuPImMRm7wqHaelu0Nlh+5OQfcP4Q8pV1oqlNBexRwuUcIqJpkFMFzThKcwNq+XaW/8GnpOWHLRMaGfeclXadi/1wVDZXZRm9K+DAaLNUEdB3Qpdvx8wTYQlIvJO1CENKbMsXCiZ45DsGAzQJu5mEWVRM+ObrmV/MX/PPpFzw1F323SFTf/R/8f8+/236OPY/tMl4zrtfUWE70ls909OBcaFj3VXyH6DtIei36kZgdI5FgPHaisMc5GpCdhc0W7Lvpe/8ossZhoBSI7hUqQezpu5moG9URALWXgB/czkFfyguBlCGWkj3jQSh9Eo2SchT962eqQnk/loMjT77PRykbjVZv/KyPV4i7iKG77b6p6bMP2S9adMAZgZzG0U2hkgRrCc4dhkf8xCGqhvSqo3gRXY3GKTqg1R6leo20xKK1Y5Y1HGdxDPvDdM1M1XRBceGmlC5BiECSWKp5R5dLupPIHFebmIdK1qAJBBlVRrtC4FKBzQQ+jcqig0aW1wFhxU7j3Q7qo8PoLkdStEzyhtOi5ONiycfpNTNZk4ldZbxDsTA1+aKmemiQFmwRZW2SecNHkxVGOL6xDXWQ/Nae8of2Ac+7Rd/tYW5owg1tQ7++PkVuVMyFhYCIzYd/1DH4cTwrEZuAh/DP94GfJO48IxwGh+/7/W7nmwY3c4CRUbIYGXksw2PVt0JEOY0sjp3vj+qk1+GpvWEzhnUuVho9u8SoU2NOQik/crsG7BLq/hV2+z6UCAQVcKlE9CKEADQNoW4OQnw/cYSywlxWTKVAWoPwkiAVNleool9TWcvZZMNJWvIoXfHAbFnoclzDL7o5K5tFRnjaIkQYc0sT0/K7lyc03xZIKxFejAJ9QQm6aRT3czMXZ/hBP8GB2NPn9jyr1KOnHfNJzVFRMUsaprphZmqeZlec9n2JikDbV9O3PuXIlDycb/jDY4P1gmzScjIteVis+dnkAiMsX9g5L9yM3zUP+W11ykUzobQJlTXUtveseipGCILr6wnpSqKrEOkLvk+s/xHVwR8lwY6IScdEuBtKB7CjK0CsMMRZgXEzlfA4Qqw8hN1oriGxPQos7E3/8YjIfA+xoXLwfoZZgmuR743n8lFgTwa29MMh+jBQD8z2W3ONdpyqmKMaJu7sY/DaUAGXQJhFYyWcg+ARnSX4g7H6KSOUFfLFFWnZINsFXmexIX+hEJPIL5pnNT+fXvBZfs5CVTxQGzLZ8m13xPNuwYt2xtbGzop51nCU13w6veKfTL7ll+lz/nX2X/Fv3ee4qxzVgs2il2TzgH1gmZ1t+PnxJVo6lm3OdZVTtQZrJX7gB3pJXjQ8ma/4dHrFJ9lV35doyYRloaKW1US0ffU95prqYJjrKOxnz2Ll/aNixefFBR+nV2N64/fdKV+1J/ymPOOr7RHnm8kN+gVA8H1DtBPItcKsQFceaT34QPgjo4gfLWc1VTUzWcUEHbuy56i4EHalUYfAEI3XXaHWMC15KB2rWELsf4/M9g51w0sbBPuAV1p9Ih3h/2/vzZocSbMzvefb3B0OILZcqrKKvXBp9tjMcKhZeDFm0p/T/9C1TDf6C5RkRtOYyUw2MyRFdpPVzarsrFxiw+pw92/Txfe5AxGZ1WtFdmcljllkREYgAATgfvyc97znfeNYNQkSwFioJCML+zH1QGkYgPQhUQbuVoNShEQiNIJYZjfcsgBrOcYffsS2IwCi7zFaMZlr3EQTSkUf0sXneUjHxFU/Za47zkyDEZ6vdhe83s253tYIkQjRtel5XG35V9NX/NvJc/5DccVP5s/4yckTliclIHHTiK8DzBznFxv+7OKKf3fyAknk6+6MN+WMZTdJe6wZ5LdB8mjS8P3pLX88ueQTvRx13DxynLx7BG0oeeVOeeVOedOf8Lo7oXEFRnmmpudxueH75TWfmVv+ufuUn+2e8FVzzqvNnNW2otuUiEYh+uSDaAZF05iMMQjJHWdyFSmXHtFZRNcnh6ffQnny/SUrKcZpIDowly3zzHgNQdJmNxspAiZjUINy6JjA7kzz0m19MDQ+KSTsvKFUDn1AHnVRjaS9pZtwqtKG0qU7YRsK1qGiy07Pg/SMyyB9IZOOD6TF1EolPa3hNi7KUfp4TFYAMk+GBvJqTMJqZI+4QSonGp32BI/LzX/4EQNYSwTkakv1uoBYYzaK7jQnivmMr+dTnleRUEZiEUDH5N/XSMxW0J0H1s92/PHTa55VS2aqzXQcqFXClW4fzWlrhZxZ6mnPxTS1Yk/LzXicPSnWTHXHtixHuSRIF9Gp7nhkEofRxqQE0YSCJhSc622SMtKw8FP+e/M9/mH1KV/enqfBVHaLntUtM9PxeVlTSct/Wfwx/+3F5/jnNWYt0R2Uu70DdJpk5ooppn9ETNZj5a3F3LbI2w2x7Yi9TVyr3zDeb2WVF5mlDpyphrNMN9iyt8EqCBQicayKrPQ58FUG6oIayKO5JeuCZuXSnlUde050GhF7kYD2QaRv5wuWPiWrMusHNT6xi4dE1QeNHZjwyo/0hEECYyCR9iKCT3pWWnpq1e9tuw7a2pSk0tdCZI0fk9tXrZKtkxD3TXOP8QcW0ecLkQ8EQIVAvWyYzCaEKmM1lcaXEj+RuEriSk3QYHYR03j0xrP4UcGiqOApPCuWnKkGRaSJilr2fFKvuLxIrdWT+ZbPpku+V9+OMkelTM7ig1SLrOJb+Cgw6sENA6YbN+Wmn2ILBUXCUF/bU/5++YyfvnyKfzlJq2AqYasrL7mupyzrCaW0/OTqKfzTlE//a6BYdEgXkTYgfBixqDsrNbnVE9Yjdh3sWmKzSwMl+4deWbHnWQkVmasdtRBsB+nheOjVl/WqDpQ+71RUca/VDilZtd6w7CdoGQj0Y2W1pRyXk7ugWbp9srJB7yVl846XzcqgWoRxpWaonspM7hyAdD/I1Qg/LlcfTiiH8JnzhcgcmayYGpVCqOOy4AcRg0EtEG3/Tg88lT9EWSJnU0RdE0uD2O6I2wa/2XLh/pLmWUlA8IlZjoOebdQoAk+KDZ+drAhR8Ceza35cv+LH5dfj2s42lCjSxX6udpzJ5q01siakNZ+Fr1n6mhs35aqbcd1Nk4uyTAoQr+0JL5an+FcTZl/K/aSxivSi4GY+YXFSU0rH6nLG05/CyV//E/765u5Lc+/zQ8X7BdiHKjEI2lBgY0z0z8yvOgTXQxQ0sWTpp6x9hVXdOGodAfecIGaqo9cJRJ8oe4fJrqUfyZpT3XGu82a42tEJMxJH+4PVHinkmLxC/npQgNDR4/LKz/A8h5FtF/dVmUdSsn8ekmQgmZJVBiKNAinTxzG+OxEisbcgO4RzqfXJOI2IEeEFndPcuBmv4+noxvR368/4p9sn3NxOQcDycZV12e5WITZqLt0Jl+5k/N7hba7snK+7Uy7bGWtbjRNtn/cYtfTUsidEQV32LGae7lFatxGetGJTStbbijdtdir/VcvIv8ka2YcAsIuDZGVj4ql4EiBu8GOFIvP322BY+4prO03AuG4w0d+jLbhRHH9wnhmNIqJEkRaadZSc6h1zlZYvz9SWRpS0USNFNU70Di8Pg7Z68AIrYhLiyy9ZQCRqBCKv9Ei2vhx1tg4pDgPxTuQVhmDyG6slKPmtOH8c4w8novfQJzwWIfatD0BI5M/WaW7clMt+zsJOWPQTfn59QXM5pbhSRAUvvcDIwEWxpVZ9smmX/WhNt3ElK/e2StTrds6L1QnLdY1vNUIHpA5onTqXWvcsTTKIqLTDzDv6XqLWErNJLtDBSNqm4Go3S8OmXiId76YcDMfvoU7b/TZv9MoMGQ/5A8esxs3rIPJO0Z6PNGivKxFRpGS1DUkcb3hDBnDxsCU0pMpq+LmWIRlH5BhEwjTJdOJUJYGwoXzeypJK2hE3U5mKMCSqgWOVdrrkHQ2rYct8d7DMOsnaWfodfmpSBqLeJ6ugJdK83+L2GO8hYiD2PdE6hBR3F3djRASwTnFtp3y9O+Hl9oSb1RT7ZsLklaJ+FQkG1qLk6/KEL+sLzouGT8vVmKyu+xlf70647dIkMh7gVm9WM3bXE8yNxuzEOFXspp6FCtTGcl1MMSK56JzNdyyiwPkKs1aYbSQqQbdR3DYTSuWQvUD68M1Yk5AHFnPxnQKT6bWQv7UH3cOfKUKkP0TrUXVBqGHal5nf7KslHwX+QNca7hIsD7GgQbTPSEedsaJhT+8wEhcq7+hlELySaSH68P4G9YRDr7UhXJAjVqVlQLNn0ydViL3W9bCDqMTe/7CUDqUDwURcmdvAQh7bwO9iDFVD9G8XGCFNz3adSS4ztqR3g9Bb0j9XNn0US8HucsLfqs+4mG+5nk75QX2DjYqFndAHTcwX1H7cTxT0nUH0EpUF9UIBwQpinyZ9wwV4onoeldtRrO8rd4HdVditSOoMmf6zcyYx5CMpCX1DZfSrJny/zQTwMB42WQmBUCprAZVkX0aUTjITUykIwWOjxOakERB0EXr2J3+p3KheUEk7ckbeNQW5H8Pt3mUBP9z/oEFlAGRalM5PBiTjm3tHCmRwbs6VllEha74ncN8ccK+MyNNC42iL5O4D4CqFNhqhjsnqYwnhAqqDbVNw3SZp44lxFCdbXm0L7Npgp1mxs4HqlcJuZ3x9XnH7qGZ5PmFu2nHv7qRs2bm9QWnjk19h2paIEEU2KgFERIh0sTUicKp3lNLxtFyzdSWt07xykp0v0tJ9nToVG9JUXUT2NJv7CSuGX94GQvr57+Ds+8DJSoJSYAzCZMEyQJvk/FELRS8c26yAAKn96we9KlIF1Ek9UgP2/CtJj05fx708y1BV+SjvVEbAW0DluNQ8rsoEyqydNUQIYly9gexoE5NM6zANHKaF+l6COvQ/rGVPqT3bMuLLfF+FTLjVUZP9uxMDfvMNmIzwAdlH2GqumpppXrWZaMvtrMZONa6WyD5iNpFiDfEVNJ8a2l7ylZdcnGw5KTpOyx2VskA9kkLTU0jbEqFIx+mgXTX8TInARPWcqh2P9ZpKJlOLrWtt6rkAACAASURBVE9igLf2JE3nJxatAs6rEW8eOqV3Yk/3E9Hws/E1+c01rA7jwdtAoTWiMFCVI1ajVKASFiMUCndn9DporUOWOM4tVCktZTY9tSSH5HT7/f7gsJB8vxU8rKreSliH1AgxVGEH4noHgmrvvr+Yn5+jVn2SnBFhpDIMbaARHq080aT9QEjYVdT7Xv8YH3j8GoMS4ZMap+gFnTU5WXV8Nlnxej7j1azCzlILV6wixSZgNh4wRK1oi5KlChgZeFRtOTO75GuZ9/KULFA64CY+2cM5QdQRTECYQFVYZqbjVO84VVtOVMuF2lAJy8vpObenNW1e4akqS6EdrdV3Bk9vYU8xvv2330li3w6p4UGTlVAKUZWIuiac1GNlBXDtZ9z4NzSRveA++7asEikxVdKSDEN9MiTNSgrASHi7sjMWtmbrCqa650TvODcNCpAyaV8NSWmgPWxDmUTD2IPmNojRRWfnTfJmy71rIV0iiQo7JjOVJY4Pbb6GXcdhf3EwTT3GRxDvOmnv30QJfAmhCpzVOy6qhifVhk/LJc+mJ6wfVWz9lGBUvrhLpE2AOwKIYgTThwvlRKWEBwlGmRiLnUk6m47durBMTHJ4+mSy4k8mV/yguOJCbbAo1qHiJs5oQoEUgdI4tBZMy54679CsB20reFvp8/408HeY+P2yeOBkJRHGEGcT/Lwi05EQUbAOExZB0kVFf7AXqIgUGZiucjXlEWPyMsLRC5XWWaKkCQULW3PZzrhpa57W62SdTTMSPw+ng0OiG2RXB98zKSIuqlHMf7CLt16hZECbgBaWE9OOhqbDbmCt+rzrmBQUu6y3/aviKMT3HYxfcYJGJQlGICaex5NtTlQrHuukbNB6w3MR2cgpoJF9kjqOWhByK3f4EEYkA9JQpINJy4Cu0sbFRFmmaqiiduMS85lqOBEpuV2HOhUOLsnXSBGpy8zBygkuxKwA8eser78jNvVNcbzkH+MYx/gg4mExKymhLAh1gZ0bcgeGgLQKEMqD9ZSsZU5at6myKeSwjjBIBB/iWzYmm/alrbjaTVk0E5QMzE03YkbIBJIPMeztNaFMWtEHfoU+JjmZjS3HqmoQ3pubZMg61y1d0AknCMlauxQurT/IlnWoxrbP8zbeBceK6mOOqCS+AF1ZnlQbPqsWPDMLPjXLBEmIyEnR8lP1hIU9w64VfnGg2hn3x/Pgj5mcbFL3MFGWJ8WaZ2bBZ+aWz/WCC9lzJiXnqsbHgCPhWVehZxU9TSh5bU/Z+iRfM7R+U91T6x4XZTI7kQdA+f24z7N6gHhw6kI0mlBofCVG6oKWaeLWo+7wnGAgZe5jwH6+KZKnn6fUjtLYZEB6z8x0ANvv71Ap5LhjaIaJXjYkHSLm53sfdB/836QIBCOyfI0brcLHqSRJ2KwJqQeOMo5JOypxtJP/2GKwnZdZTjgolr5GisgmW89PlGVSWG5NIBiFL0TiShmgCBSFpzY9Zya1dzYoblxSn73tJ3Rej/SeNhquZcNUdsx9SzsMo6Lhhf2ML7pP+GL7hBfb0ySalwUnK53oN9XBRF1Efq+Ktg9fWSlFLCSulASzJ1Xen8rB3SkbMGo7D7bZb1VhIrHVT0xLXymM9FyUDRN1VyNqSFwDlQCSpnuIMm+nx2z4YJlm1+hW6vw8zahttU9SRca2UrJqfIHPSgoDq314zC4YgpCsfZWAUcmI3XlDcqk+rtt8dBGzEsjC1eOC/caX7HyBDSod6Zl+4Mt0zIQyIktPXfaclTueFmse6xVXdsbSpkT3YnPKGznnspvxVXHOSa66hoX7wRG6C5rrbsqLzSnXyyl2VSJKjzIBUzhcIVEyUCmbeIZBpIpqqJruJ60Y0oTwAeNhAXYhQCu8kfhSjG7EWg06Ve+wcieOsnU+g9g2t2uH7dWwnlOK/V7gVPfMdXJfvp8M96zy9CIPQP2hcUQp3ShQZqQZk+fgZAPk9k/RuATESxGTasM9lv3gdRhI7eq4kiMjYUhWhSCqI8/qY42AYOMKVrHa8/gydCBFRKhI0OBLkTTVi0BZOOZlx0XR8NSseKQ3ACy7tJJ2vcwVVjmhMifJCZxkERajIMSkH+ejoN0V2EWJWSrqpcCeatws4E8kQkRao+lNqtJE3O/2/srq6gHAdXjwykoQlSAYmVi0OicK5UcMCnirzVNC7KWNkePkbfz5QTunRGCuWippmfiCWvWUYs/dUkQQewG/obIaJo+Hj2+kp8SNSdQfcKyGfcFB86r1hk1f5mSVvNMGtdOhWhtU/If7VzKrhebXYdDXPorvfTwhXEBasJ3m9W4+YqM2SIxMpialcok1LhJk4LO8MWXAGJ8kvIEmFFy7GZf9nJtd2hG02wK8wBpDo/PyvBNEL5NWe1bwFFEgW0G1kpg1FOuIiIIoJK5QWKNpVEGpPDbI5FMYeLBE9OvE+9miFenEHDCrSjsu1IZHsqONihaV1RcSuC5hJFcCY/8NewG+wxjA9JnqRr1pedD6DanwUIJGjryrRCZd3tteT/eZDE9HmWNiNj2V7JxhZzVSlEn4LxSjRXiIIptepApOioiNCQcQKt7Roj/GxxWycxTLiHxT8FP9lOAk0UnoJaJ2FJVlOulGRrovI3ZOsv4yASEijS34YvWYV7s5UkRebE5ZrvPxayXCC7CSIFICwgmUS1bxwifVh9Ey3qX/+0KMHoFE8E6yaw3OJ9swvRWoLp9772LpH3YHD0RdeE+LzKSPXFHMTcdTteZUCmQ2SmijyoJ7YIQcE9VgNHoYh9rskGRixp8dVl33APWhqhrCk2y9B6xgcKgZTU+lR2eAf7iaAamycprOGqSAxiXJ2DYafMbBJGEktVbCYqNianqE+v1dmY7x+w/RWspVoLpUtHGCzh6BshfYuaI71fgThczHSSwCbiaJVUDrBGNsuoLbZoJzyW3ZW0lsU3ITTiBsTkpOInNiEpb0dZ8fzybZ4aDEaJo6mvBGkZKol7hOE3eaSQNyTFbvXrf5sKeBBxEFxExqq3XPE7XjXNbADhsCNspMCBUYFIq3jSEGbOm+JOGgF2UyaH7/dyD7EYp4Zx8Q9iz4Ub8dj5HZqVlESukJcb/0PMjG9E7R9yr19t6wcSWtMePtBlLrXO44U8mOqdb925XVsbr6qEK0HdWNpX4tUb3IBqUR1Ufax4KdV1gK5NSmk6YMBB2RE4c2HiEindXs1hWsNLqRqINzK8pcNWXjUzUmp2Saqvrhc/IDsFNJPxOp1TQHwJRNhq3CCXQjMZuIsuk8+F2lXn7bePhFZpHG86nE3APW9yNN/gQ+RkKusOayHfWnjPC0wXDNjLWfsA4VjS8TU1ySFBPe5XyTk1cXJYVwo8V10tPSY3V2+JyGVu7t+0rfK6RnWlhiFEzyrlVSIx1wsnBwX0mwv40mTVWiOPKsPuZQyWvQzgT9fHBhTialvkzJRvSC6AtwSTJGeIEP4FVATSKF9nSFxxcKHwYz06zbVjq8VbitQe7SjqHsQXYC1YtsQ59wq6jTY7o64qtsF6+HFRqg8qBicoC+KejODMV8jgyR2PeE3vI+/S4fvLKKIpWZUYm3qohAJJA1rLKDqxcRn7Grqeh5pDeojPm0sWDtJsmpw++5SwBShXfysQ59CdtgDjTdY05Ke64VDMkt4KIaAfaAGHlTAUGlLWfVjlI75qblomiYqW6/FH0wifQkocFtKLP+0OFr87u+usf40CKapKrQz6G/COkYyBc5EVL7plqZcaXUHgoPnVT4WqKVp9CO3inaqSAYRTHruThJF/Vn0xVrW3G1mbLZVtidRrQKkZPV0OFEldQYoo5QeqQJRCeJVibDVBUxE8usTqTs9dUF3aliejpDOgdSIHxIfpcDITTjVkJmGsO3nMgevg1UIiUsCcg9P2oIn3XYQ5QEIQhAiIm+UAvHE7VGEbn2qaK6cVOaUIws8oGSUEl7JxkO1vGDKkPCksKoVVVETx8H3fU4UhpCngYM+39DHOJmgxqoLwQz3fOkWFOrLqstHP5tAtC5uipyZcWx9fuIIxqNqwR2HonnPdIk+3ljPM26JC4NZicyoE366CN+IuisQonIxFj6UicKgvE8O1/x47PXAPz72VdcuTlfTJ/w1eacq82UtjPYTmMbjag8urJUlaXUKfEpEVEycLWZ0mxLwsYgZGQ66fnh2Q21tvzNxQndmcGflMimSm42bUf8ZbaX3/Iy84NTFxB7G+zhJFXvaAOH8BF8/nkhAkUM1LJj4dNoNiWpgX8lUdnwYVijGZIXpMfrBgWEXGEdLjIPcjC1Ssuag1DfoRmEfUe1Vso0Ph701m1QbHyVcCnZM5V7cwuPzEQ8s7foOsCsohSJj3aMjyKiUbgqSQ3PTndUxjEteqam54U6ZSlqXChIzJeUsBJtILkjGeWZaAsTRg7VH80W/MnkCoAfla+YqzZN0KNAi8CqLNm0JZ0xTCbJh/DJZMNU93c6Bi0Dr4CtF+jCM686TouWqeoRpScYQyhUcuzZSeJ9hduBghOSi8K3He8FYE98IvaV1R2elBiVOocIpD+1z6x1GxVGeEppOTdbNr5i40qcUAe/k1rJNE1MGlI2aoKQWKEIUd1pCYepXS17FOlzG8xozcWYsPYcrwGTOgTcbVRsbMnL7hSAZ+WS75fXXJhE1mtCiT10ZxYHUhu/ySb7Mb4boSW+EsTa8WS2ZaIT5jlVPYV03Exabk8mbNYV7rokXEuCFrhJRJpEpZnpjpnuoIJSOZ6VySwVYB0mtFmLqZSeqelG6k0jkyzNJ/Waz6sFM92NxGuPxAaFJHKlHUpGatPTec3WFcRWJXwtxrumEe8yi3iol+6hHyBhVam6EnIPsCcuVUouKiuF7lU+YzKTyIlq0KCaHpgwtMHwDspVEt/L6pwqpmQihcYHk/Sl8oubvp90s6a6o4963M0KUeCEoidZFw3665J4N9ES2IWS277mup3ig0SeRB6bDVOZDgQr9EhpgPQaRDlMbuSREPohxy8TnPuGCFriS9BTy7N6xVR3nOiWmep4VinsPG07fLW94B/Lp7RiivASPwkUhcMoz9y0nORNjcHReZBBGrh+w0ZGrfOqjfRU2vFksuF7k1u+X94wV7t8++RIrkh2daflbjzm+6BY9hOEze42gx1XPPg8Kh1n6sIDOQg++CIz5ErioA18W60zETBhX1WlNZW07Dys11RYlAx0wVDLPhmY5jsdFp6lTPhTJZJzbSvMnccaKiuLpjzgQQ1ho2IjynyfYmw593/TXfzKBcVtV3Ozrems5qRsaWvDVPQgUruZ/q69btagQDtWVsc28OMJmcwYJnXH9yc3oxZaLXtO5I5adpzIlp9OP8XFv+QfOo3dVcScrArpUmIrlnxmbvlUL1mFiiakY3YQlQSY6lRVlVJTKUcfFJ9XCz4vb/lecc2ZbFiFijYamlCOnUtTlqxcxWU/Y9VPWLYVohcI944k9A3HrpDiWx8UPnxlJUXyytP7yupdmNUgZ9znadw2arbRsA4V25BclfuoCHmxWWWHZD+apPoRVN+GcrzCNKGkCya1dpAs5QEvxViZHVZbe1JnSnqH+utD+CiwQaFFspx/l/TxwMjfO9z4sY0ccvW4b/V73GQ/xu8QgzLob/D+CR9QPeyakq92F5yZHY8KBTodf9d+hiLw0p4RoqCoHO2ZR00tSkQaV/C6Oxk3L6797M79hwyd1DJ5FtSyZ+NT8gm2GiXAt6HERzlOqg8n7DtvWLkJLhv2GuXH1bCoJWiVvBXG1+Bbe0V/abwfnlVmyYp7mFW4p5M+cK1slLRRsQ6T/EKW48rNMOFLKgmJW1KrHiPdSEfoghldbgfSpwvp94dy2XpFkOmNHTTfu2D2rjVZ5mVQXDiMhGOFUQLZh/vmX/tIVaM72IMUe1LrMUd92DHiNe9YP/mmX/ER1YHbaS53s+RFmb0GhkTSBc1VP0tqnVVHOBcUhUOrQOsMN11NHxRbV3JmmtEDAKDOK2eltMxEm5Vw59k5PFX5NuhUgQk1JqouGwpvXZmMI7ymz8lKZQOKgYIUlUDIX1Mt5FucCL6XymoA2KXc7+u9K8LAt0KwzuVpG9LHYaIaiJy1Sq635cFSdBLkK+8kqKHq8eEuWB6y9EzIU42hsnrbpuueoUV2YU4W899sCSYzlaF4q7IS+fM9sPIY3/1wAdVFRKO43CaFhEI6TnTLylWsXcXaljSuIEbB2aRlWth0bkRBYw0276Yu5YRrPeWk2HFmdkCCG871lqnsOFMNkCbSbTAj1NJlKk2HSRVVKGh8wcqlZNV6PVZVkBbwB26WLyRRqYS1/qqE9U0uOL9lPDx1IdMWoox3/q4ABMKIkY/OzFkNYbTcypXNgOuNmlaDPfw7Ep/K9IUgkgZPOJjGfVNiMcLjszFFkAITPF3GmO7XTYmLBTa/oaV2TMueWAhOTFKA8CROVUDe0eKC/QQwiqPiwsccYpwuy1z5FNx2NYtugguSqek5Lxtq3dNnyk4fNL1X2KDovKbz6RQu8+7quWnoguEmzkb45MZNWbkKGyUu5t8NSYttKAaGLqH1SSl3ELWsdDJsQcVxii1+3cRz6H7zLcT7oS6I9CEOKpRhtcbHeGflZQTbA1iZJoEJU0ps9cPK6jBRhShHKZih0hpkjwnp6pKY6vuJ5LC/V8suJaUY6IijV6DK7srktnN8LMRYVUkROS12nBY7tAw8q9IYeVBjTGs9Cps1r6SIhOF1OALrH3b8FpgVSuCNIBaBeZn8Ao1MjuQ7b1LFtKsQInJStpwXO/54cnVnkHTZz3nVzrnazdj2BbXp70gZrX01Vkk7b0aNrNQSRiZS50rL3xkW2aBovaF1BqM8te45NTsmyiKKkPcOA4QAPn8+wKziu16HD4YUSq4e4A4Il8iZEovHs1+3MQQq4SkFIB097X5/L0CI5Zg8uGdG6jOBKTkrR6RMYLskJrXRgX5wsL83SMsM2lptVvUcAHGdl5+HRLXnWe3bv0J5Tk072n8NJbiPEouiz0lrOGCAUYQQceRZffDxG56MUQqCAVEGZkVHpS06X3RdlGxtQdMWaJ1gg0/KFf968iLhnniUCHzRf4KNP+BqN2PXG7pKj+RlJQIrV/H17pRX2xOa3jApLJV2WfLbs1MFk0w99wfnUMiSxo011CRF36G9lCaASMMh4XKSesvgVB3899uHN96P6kKmLRwWEomgGXI7uL8qJLMIQSBSCU8tu1SdCJ2rJf0WvWpMWjlRldLmpWU9staHquZwyVgNjskHPJVBRrmUqZ8/xKxGrC0/nouKCseJ3vG96oZP9TKrOwQsKvNXNH3msYwxtoHD/48Z66OJDI0IHSgyKO6iZGEnLLoJy11F3xh8oZIWurR8pm+phM8X88A2lBjxPTqv6FpDMzGjv6WNio0vWfcVi2ZC1xq6UlMVdjTXlSKiRfK77ILGxdRSrmzFqi/Z7Eqcl5yVCZCvpE1dUQTpc1Xlfwkv4UNUChXDJFBCUGCyRk9yVvYYoaiI2AyOF1ltQR0sDg8n+UD2LKVM7VowiTaQb6uGSkuk2yaPQT+C9kMMiXHAm4Z2ME3sFIVQWKGyk867X3QtEh6mRVJ1LKWjlj31AWnV5zWf/XPPAHt+0+FAJvYYH00IH5EWwk5z006xQWK9wgfBal3jVgVqrQhl4Hl5xt9Vn/FYr8cLZiEcP919yk9WT7m6neNWBbe6Hh1pLooGJSKPqi2ldqlSs4ZtV/BqMacsJiyqCTdlMqnovKZzmp01rLYV/aZAbDS7ytPZNBGcmQ63NugdiD4geksc2sDDuF9NfUgmp5CqhyTwlZYlIZWXhkgpDDYGjIh4AkZAIQRSHIz3c0gRMmrVE2I5ft/nlnAA2weRvUralDCkGFVGByrD4X2OjHdh8UIShMQLyeaem83YPooIBLQQSBVHkt5Upo+hbT2Unxns5I/xEcSvojGEkITveslNM6FtTRK4axVyqygagdkIfKnYVRO+qB5zOkz6MhD/i+0Zz6/P8LcleiPpipLbKimF3kxqLoqGzycLatXT+IK/XXzGqqnobiv6yrNrDauqRAC91dhe43uJWBmKtaBYCdxE0oQpXwLTqkdtFLqJyNaB8xD8HqPKZhFC3kteH9Qic66skEmOQmeLq0RYAyMURkgqIhAxfLPraiJYDsRPnQwZMtDNYcLKVIGUfFJVNQgbD+D5nfsV+9tbobFjleVHesNhSDKgKJPQXylTYpzKbnRkXofqjoVYmmruwf1jfEfj12jnh8pKdIKmKfGrArWRmLVE70DtQDcRNxW4ieamnvLP1ZNUfWVMabmZ0N9UFAuJ3gr8RLGp0wV8MZtwUTQ8K5b8uPqaNhS8bE/5uXuEXmrCTmErjatSixdbhWwlphWYpaBcRMpFwM6SYkgja7q5yc8vIHsHzqU28Btwqbfs5b+leC+7gcFAKMLYBgJ0Ebpo8aMmeoq0ZhOwMa3bHLZTwMgKH+yzhhZQEu/IGw/2XeHe7w8VTvIJ9KMl/R2eVhiYvInN23uFloFCsqefD883yjuP1eR1h7QPuD94D7llb7Etjgz270YM08FfdhOV1m1iGZlUlp0X+GgS9y4KhAMEyD5SrATd65Kf9U/3nUYQiE6itxLVpg5EdhLbpL7j69UJMRNA174iIHixPcW2Gu1IZ7yIyMIjZSCYQJgIbKeIQiNCEusLOmu1NylJll2CLUKlUWUB1r532s3DJiul0ptTAuVd89Bt1KxDn8X38gIzYPPXbQbGDxULPG9XRSFzowZZ48SEl6kNyys6AzN4+B2AEjdO/aaip8j4VBsNS19za2tWrmRjS1yQVCopgZosEjiU5IP0TBvNuGe1DSXbvKs1tJljHI7/4hG3+s7Fr7jwRJVs6ZhZPjlZc6snbIsSW2uCKRAhqXuqLmLWIK3E3xSpOxGMhg4iZGFckbTVwzYd32tR07aGN5sZP588IkTBq9s5casTXKpAVJ667jipEmVHyUCMgl9Mz9mZEpBDo5KURq1CdenxfV0g2wnCOkSzeye48RB7gfAe2sBgkhOzLD3mIFk1wdBIlwH19GEBG1NS6qJKS8z3CJVjazXgUzAmqgEU91GkPcIox4nckKwOXW/SFLCnEg4jQlL1zE43KzdhY0s2fTliVkV0B88jJSob1Lin1cpEsBtWGAbOl7rf3B4T1EcbUUtCCWVt+Xy6YG461nXJuiu5iifYrsBsBLqBchkRNzElDgVBpz1bXyaHZl8mowdhQW2zpJE1uI1mYSoWRdobFDuFamRqW1REF55H04bPp0tOTMuF2TJXLX+t/5x/Fk/o3ATdZLx1l6o91QMC7Eyj2gLVGIS+lz4GmZgPcRrIsGqjQZs0OYM0kdvGgja2mIwBKbLw3uDNh7w77mefqA7bOrjrYuORmIOv9/tW5o6912A+YUi28cMqgo06O9YqXKY8HH7s/7T9gvN+b1HnJJmqLZWnkUOuHe9jpLD/zq/wMT6wiDJdvKvC8kfVgpXu2HnDdlKw6w3brcauUyUjt6C7iG4CvpK4EigEoYgEnToWX+cVrnz/shfgBMFGYieThlwQRB1xU2Bumc92PKtX/On0kqdmxad6ySO14epsxrov+XprCEYlVxyfjCd8Aa4SqKnE7QxyW95pA8X9llAkQYIPB2DPcZ+9PpAxf1UMnCV1wHLvD9yZwwGmpUTACsVctigZ7yh1NqFg6SZ3eFaldLSyJdU9mTlPWgDVMjBRll73uCK1e7XuKaTfA+wENAn7GtZ7BrA+ZM7W4eMNbWMMyQQA0m6g8MeM9Z2KX3OpWWSNqTPTcKKTbtv1yZQvW0NnJ0Qj6VvQjUJ1EjsVuBrcLCZ35tJDGRA6aadjs16aT5N3So8qAkr7JJukAlIGnsy2fD5d8Kf1Fc+KBVPZYYRLSg2q57RsuTlr6Sc6b55EnJPYNwWhSFpPeqfQpUEOldVQUQ0J64E8JN6bFVdag9u/gYdV07uKxkNHZSUCfkhSDMqGYpQr9gdyxVXGhyphxyngsKRpxN4YoguaU92MVZoRifxWiMSZ6lRe5izABUkhU2W4F+FLdN5B2njQXx+SsFE+YVvsq0EbFDGIEadKTiPx3WsKx/jw4hBc/6axvUimDUrG5Bwu9xDGzXTK7tzwKgraskC2CURXnaA/DYQzS33SclraUTtdisiqLWnaBLDbTqN1oKwsJ5OWedExMx1z03Fhtpybhsd6zRO9Zi5349NqY9KIe1xtaE6LcTdxajpClPx/5lNaJqhWUqwlZXkgEwO8ZXL6IcoaH7p3DDFoPEFuo/mmhBWG9MRA5R9arlFELy9mDuoKc9WOBhKtN3lCV7B1ZdZNz6sN0mdJmJysgAJ/sC/Yg7nb7kkRDzCv9BwHcH+gJ4AbW80+qhEvG9dtsoMJ5KngMU99+HEnSf0auI0YKivHXLbMVctc7mimBTbrqV2XNX1nsJ3Cdor6yZYfXNzyF2dfc6p240W6CQXPm3NeNicAXG9rKuN4XG/5ZJLki58VCz43t/xQX2cj4bT36qNgnYX3mlBSy55n1QpIi9E/mFzzR8UNRnj+N/6Kv+s+xy5L3EQSSoW+p7pwqBT6ECD7wwsnH+MYxzjGtxAPW1n5gG4juhFsVyW388SyXbr0OXGl7obKc9lhMdOiRkPSNuxVDEalhkHIL4PXg0/fYCzahGKUvghRQAb5CbzF4RrY7+Zgb0qKQMiVkQ1pdjksN0sR6YLhTX9C40teu1Mq4fI6Ub+/37w07YMkulTWQ7bwduGoafUxRUzTPeclG5+Meq/cHEnki+YJL5vT0RpeG09ZJSnvZycrPq+XXOgttezv0HEelxt0BkJPi5Za91wUDY+KDed6SyUsfVS88TOK3AkoErn62s9Yh4obN+PvN5/xs/UjXi1OMMbx8uyEH89rHpktq65KnoLx3vL9NyiFfniLzM6huoBZg1poVhfJkGHrEwfJ3PsjVZ4KyjhoBCc0kgAADpNJREFUWMkxSQ1Ey0GAL90+86ykTwxfJDakxeFD4b6d3y96jkvJMo6OOOPjEymyymIbzCgT02eKgssguRYBLdNi6c4btr7gdZyjRGSuWx6ZLRd6Oy5UD5hWiAKcROUVQt1lgP0923Af44Eim32OJ+o7cKskY51w0I0r2fgk49J6w5tmnrz7NiVCROpZx9P5hmf1iotiy9NizSzDHMCov/ZpseJpsU7fq+UIY9QZPA9I1n7Cwk/3xyOBPipu/Iylq7ns5/zj4ikvX5+hXpX0ReQnn5R0n2ie1UtumgnY7Oh8+Ce9xyX8B01W0QdU4zBbg1lJmm1KUhuXwMBD1rqPKVlJEh5k4h73aWMx0g8O+VIDz+oQrB9uN8ghb3yZJFp9mvq5vL9kcs+fHv+AkpBBcnNgB+9jcrnpvaIPmkI6KkCptAi6cSXrvsR6xdN6TVdplAicklQWh6VqHwXCCbJrEtJGpPVHBvt3LFLCkt94ERIBnFMsbM1tP+G2q1m1JZsmKS6IRhNMwJw2fH92y38+/WKcNktCwllD1vmXkZlqOVPJkflMNhRZSBIScH7tZqxDMgg+9ARog+HW1Vz1M667Ka+vT9EvSuZfgqskW1/xlTrHniuapkR28oBJ/z5eybvxsJVVDEiXW8GdoO1TUhmqHAW0ESyCNipskLQx6UMHEvtcEZnKbtSMtoNlfNaqGoiZhzF4DQ5mEhOVvNmGiR4kKdkB5PcILAcqpCQFBxdk1qJOKo0uJ0UtA6VyTJRl6wpebebcLGb4nWL5qKI9M8yyFvYd9jqQ1yCBgyvUsQ38eCK3gSGklZitLdl0JU1bpkS11uitJBSCzbziupuy9JO9em5WFxnaQB/lHbWPQnj60b+gYOmnXLkZl/2cq25KpSwz3TPXLS5IbuyUZV+x7kt8pzCdQLURESJmI2mXJa/lHL8qMK1IlvYu/fx9X2QfOFkNW9mkTJz/O0zXLNBExTZqbvyMF/acl/aM5+05j8yWH1ZX/NBc8j11zdfunBf2HB8lXdB3tNXvSxUP/CsjPI/NBiM9F0Wu5g5q2Fp1ya4rKrYh3FH2HAxPl7ai9WakLGjpmWYFxadmzU1fc7OcIr+sqK8Emx9ongOf10vOTZPVF2QmoUaijmQPSoIWRHmUNv4o4l67FIdVLcT+WtVLdCMpbwVRCNo44e/sZ7zazLmYNFyUDc+qJaV0o5IIJE31G5fY6r+Qj1j6yZicVn1FYwu2vaGzhtN6x3m140m1oZCOzuvxIqxKjz0J7D6RabWmSOoQ7aJCrxRmJTCbiN5FZO+TTMyAWcW0svOQ8fA8q5CysAhAuJtcfIQ2982v3Cl/2/wRP1095fnijM9Pl9SPe/6H6kt+qHtsXLHwNS/COV2Wehle5MHR+ZBo6pEUwnGhN5xm4fxDw9Q26kRPyN9fw7iEbKNm40u2rmTZT5IImrZUylEoz1y3PDVrPitu+e98jl8WnP0cTn7eASXrac3ySUVTFNSyHxeatQwgIxmyI+i0K3YU3/uORjzAru58H2J422xEWIneCMqbiHRpL7BvSq5uC67O55ydbWlOC06K3Z0hz+F9BASX7YzXmzmrbYVtdSKNukSbaU8K2hONFoGZ6bI5hCRGgSkc7VzT+kSviRJkK2Cnc6KCYhPRjUd0FsI9bsKQdb9l7fUhHjxZDXpz4h2VVYCx7bt0c362ecwXbx7jXtb87FPND2Y3FPPAUzXl0u+yqYMc7YqGdrKQjvKeBEuICXhP0i07KmFZhwkLXwNktcXEidqGcvw8+A3aqOiDYmeTeoKSgVr3qaVUHae64Uw1uCAxt4rTn/WU/+/PmJ//Oc1nmtbfnVpCwq/QkZwvkxa3ErwP6+1jvL/4RoD94H2O7/CaxAtUB8U6YpqAbiX9UmBnkt2nJbdeoFWgrxWF9BTKoUWg9Zom48Cd11yuZ2yXFXJhMI0YoQcRoQ+CjfZs6x2FcuNKmRAxqYmeCFwRiLskHaM6gWpyolpHTDMkK/fL1ULf9ff/jvHwi8xapgXMAjAp41bKUZC01i9kizHX2KjYOUN/W3HyL5J2M+Wv+REuKP6ns5+OTjf/rn7O2lesQ8WtnWKjyoB4Ag3vu9ueqYYT0WFEGN2RIVVRUqQp41aW43TEI5MzrW5wZTooXJRMlE0fsufaTvlqd0HjDP/1X77H/I1Ar/vktlsK/Nzz4/lr/k39YpRMbkLJRFtk4fFlpluYtNh6bAO/I5FF6Iav7ycqIQUhyxpLFaiUw5lU3Rfac/lEsRUlrlboXQLRD7cd2CkWq3Sx/XS25mm54dNyydpXLGyiA21dSSE9b5RnXUzoGw0+eRgIL1DnHZ9dLPnxyRueFGtWrmJha9b59zZlwbot6WuNcwrXK+xOgdCoFqrbiBhkjQ8qqbfE9z44PSuRTsSgwRuQOu/lKZelWSSFipzFDmuu6ZzG3ChOf+4obxXLOOX/9D/i9edzfjx/zV9Ov+I/Vs9ZhJJX7ozLLHYXslEEkBnBu+SbJhvmsmcuHErAOvRc5qfWRk0Ikk4YmuDH3x8kiC/0lkpanhRmL0OTS8Pnu3P++eYxt9czqi9Lpq8CatOBELiJQM0t/3H6L/yoeE2baRTXYkalLMr4sQ30hSBoMTJ/j/Edi3dVFiIt9ysVmOoeKSKVskyNwcjAoq7YPS5ptwa5UZi1QG9TspGtxK4KNsbDbM2n5ZJ/O/kFC1+zKFISW7qaqe6Ymo6byZRVW9I7jXMS5xSfXKz489M3/Kf5z7lQG165M97YE970c6a6Z21LTstkcupDGjKt2pJte0qxkEnjygaED4RvSEYPwbGC92Ry6k2qrFROVkV2T66lQaNQQtLGDTZIyhvB7B+uqU9rpJuy6if8ffc5/An8xfQX/Jtiwq1vOJOvqWTP2k/2tvJRMlc7TmRaXxgS1ZmUSCG4PMC0DmkQkHCvWnXUsqcSHZXqudCbA06LxEdBGwtWfcXtmzn1zwrmXwWmLzvEaksUSavoZLbjr6rnPFMFV6Hh0hdJOlk5isLTVHm52SiCOWJW39kYMatUVSFk8iQQKVlNVE8pXV6dEXxSreEs/cqL5pQv3jzGvqxRrUD6RHkRXtNOCkIUPDUr/qJ4xSIUXPtkmHqpT7jQW27KKdf1lIWdsOonyTjVGv7k9Ip/P/uKv6q+5FR6fqHWfCGfUkpL40u6MuG1h1JMV92M//u2JpRlmgK6d+ivv4d42GQVPGpnKVcF9lqxfp1K1f8y+SEz9Z/5YXU9sml/3j3h1VcXPH0T4foWteuY15qoKogFfy8+y+YO/w8+Ph73me7rVA17gx6RlqBlxESPirAIsxGzurKz5KmWiZ5SRKa641Tvku9f0HSZkLrzZnTKve1qfvblUyZfFsy/DMy+7jFXDbFtIUbMOnL5Zs7/uvxPPNZrbtyMazvly+aCf7z8hO2rKdVlNkddRszGEd0Drakf4/3Hu7hVBzZVemspb0sWL+b8H+LP7kzQtAwY7am0Y7Gr6G8rypWgWA3DGEFU0MaSfyqf8H8Vf46NiqWruXXpuL7pp2x9wcaW7JyhdZrW6lxdKf5FPkq3czMqafmX9hFf70652s2wQdI7jR8cw0VECtj1Bn1ZUN5GzMoimw56+xZm9VAV1fj6POSdx7ZDLrZMtUR1FdKmN2y7PuN/v/4PyNITfZJNEY3m9B8V8+ctsU28EXO55QRQXYnelfz94of8z589pa566rIfN8oH/tRgw/3YpAngVpecyYataigIvHJnvLTp0nXVzZIKaExcKiUSgF7rGVPdcd1NedPMWe4qttsKv9XIbSrLTy8F09eB+mWHuWkQyw2hS9jD7KWl/YeC/0X8j+nvc5LoBaJVFFeKs0uYXKYDun7TY6420HVvv3jH+HDim6Z+hzcJEfDo6zWnPy8R0dC+OL9zm07BrogsNKheMF9BeRspV/6OS5RuJFs35W+2f8Z/O/+Mvte4fG6FXiWmeZ4A3pfQfl5OeF4+4m/KPyUGQdxo1FZhcqspwl52e1AmlQ7OX0ZmL3rM9Rax2hB3O6I9VMANH7b4Xuwt4uoGvd0xva4pFmkzvH5TsntuiMoge9BtxOwik1dbzMsFPmdt8fqKYt2gr2dU11NmLwy7xzPaC1g+Crw875nOWyrjmBjLrOg4KVq2vuRMT1L1pUuamCZ/r+0pr7r0HK7aKZu+TFbcLlc6+X6mpuf54ozVzRR1oylvJOVNZHIdmFz26K1FbnPr17aErk8JVgqq50sec0p1bfCFGY1MVQ/Vrae6Sm84gFg3xKYh7NqHfBuO8b7iHfjU/mfpBI43t9Q/jZRvZoTa3Ll5KCSuUoQyTfD0LqB2Htm5THFJsEq1MJi1pr0qsLMCbaFIEmqZtJmNKXIu8UX2QTAgfQLupSdJJ28jZusxa58oRjEmTtHB8xchojYdcrklrjeEtiM6R7w/DXzgtbGHTVa2xy96WCxBCMztBQDm9SmzsynEuD/pt1uidfjeEp0l2ghtC9c3iFcF1eUZ1emck9OazQ9q1t9TbPuSjZXsKk9T9exKk0huIWFHg2FpIog6ruyMZZ6arPuSpivoncJaBVHQGU1nNG2hWS8n6DeG+qVg9jIwfdFivrrCPf8F6Rr5jpAKXrxmstpSfVkTS000imgUwgXksoGbBf7mNr9AR+b6RxODz8Bimc4H3t7/NVVFMZkg6glodXezQYosCifRqwqzmjC5MdhaomxE9ln/rQ+IzDAXPhKVwNUKV0tcJdC7iNl4zMbdTUDL1S91pQn8kg2bdxllPMCx/Z0i+AjxB3Dyv4uGcFyn+bjj2zxx/1AvcO/heYnfRKVSCHEJfPlwT+cYv0H8IMb45Pf9JL4LcTyu/6DiG4/r3yhZHeMYxzjG7yu+U23gMY5xjO9uHJPVMY5xjA8ijsnqGMc4xgcRx2R1jGMc44OIY7I6xjGO8UHEMVkd4xjH+CDimKyOcYxjfBBxTFbHOMYxPog4JqtjHOMYH0T8/8jQEHUI7KgTAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for X, y in train_iter:\n",
    "    print(X.shape)\n",
    "    print(y.shape)\n",
    "    show_images(X.squeeze(axis=1).asnumpy(), 2, 2)\n",
    "    break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "trainX, trainY = train[:]\n",
    "unique, counts = np.unique(trainY, return_counts=True)\n",
    "plt.bar(unique, counts)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "testX, testY = test[:]\n",
    "unique, counts = np.unique(testY, return_counts=True)\n",
    "plt.bar(unique, counts)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "conv61 (1, 64, 48, 48)\n",
      "batchnorm64 (1, 64, 48, 48)\n",
      "relu64 (1, 64, 48, 48)\n",
      "pool16 (1, 64, 24, 24)\n",
      "denseblock13 (1, 192, 24, 24)\n",
      "sequential80 (1, 96, 12, 12)\n",
      "denseblock14 (1, 224, 12, 12)\n",
      "sequential86 (1, 112, 6, 6)\n",
      "denseblock15 (1, 240, 6, 6)\n",
      "sequential92 (1, 120, 3, 3)\n",
      "denseblock16 (1, 248, 3, 3)\n",
      "batchnorm84 (1, 248, 3, 3)\n",
      "relu84 (1, 248, 3, 3)\n",
      "pool20 (1, 248, 1, 1)\n",
      "dense3 (1, 10)\n"
     ]
    }
   ],
   "source": [
    "ctx = context.gpu()\n",
    "net = DenseNet()\n",
    "net.initialize(init=init.Xavier(), ctx=ctx)\n",
    "x = nd.random.normal(shape=(1, 1, 96, 96), ctx=ctx)\n",
    "for i in range(len(net)):\n",
    "    layer = net[i]\n",
    "    x = layer(x)\n",
    "    print(layer.name, x.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "[[-0.0132034   0.09813713  0.04363178 -0.13270727  0.01422592  0.09507618\n",
      "  -0.14386451 -0.04419979 -0.05845089 -0.13445967]]\n",
      "<NDArray 1x10 @gpu(0)>\n"
     ]
    }
   ],
   "source": [
    "print(x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "[2.180623]\n",
      "<NDArray 1 @gpu(0)>\n"
     ]
    }
   ],
   "source": [
    "loss = gluon.loss.SoftmaxCrossEntropyLoss()\n",
    "print(loss(x, nd.array([1]).copyto(ctx)).sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 252x180 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "trainer = gluon.Trainer(net.collect_params(), 'adam', {'learning_rate':3e-3})\n",
    "epochs = 20\n",
    "animator = Animator(xlabel='epochs', xlim=[0, epochs], legend=['loss'])\n",
    "\n",
    "for epoch in range(epochs):\n",
    "    epoch_loss = 0\n",
    "    for X, y in train_iter:\n",
    "        X, y = X.as_in_context(ctx), y.as_in_context(ctx)\n",
    "        with autograd.record():\n",
    "            l = loss(net(X), y)\n",
    "        l.backward()\n",
    "        trainer.step(batch_size)\n",
    "        epoch_loss += l.sum().asscalar()\n",
    "    animator.add(epoch, (epoch_loss))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "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
}