trainCNN.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### import libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torchvision\n",
    "import torchsummary\n",
    "import matplotlib.pyplot as plt\n",
    "import preprocessing\n",
    "import cnn\n",
    "import network"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### set device"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cpu\n"
     ]
    }
   ],
   "source": [
    "# device config\n",
    "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
    "print(device)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### set hyperparameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "#in_channels = 1    # 28x28 = 784, size of MNIST images (grayscale)\n",
    "#hidden_size = 100\n",
    "num_classes = 24 # 2 letters are missing \"J\" and \"Z\"\n",
    "learning_rate = 0.001\n",
    "batch_size = 64\n",
    "num_epochs = 2"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### pre-preprocessing and splitting Dataset (training, validation, test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "train = r'data/sign_mnist_train.csv'\n",
    "test = r'data/sign_mnist_test.csv'\n",
    "train_data, val_data, test_data = preprocessing.make_dataset(train_path=train, test_path=test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### DataLoader"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_loader = torch.utils.data.DataLoader(dataset=train_data, batch_size=batch_size, shuffle=True)\n",
    "valid_loader = torch.utils.data.DataLoader(dataset=val_data, batch_size=batch_size, shuffle=True)\n",
    "test_loader = torch.utils.data.DataLoader(dataset=test_data, batch_size=batch_size, shuffle=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Initialize network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------\n",
      "        Layer (type)               Output Shape         Param #\n",
      "================================================================\n",
      "            Conv2d-1           [-1, 64, 24, 24]           1,664\n",
      "         LeakyReLU-2           [-1, 64, 24, 24]               0\n",
      "         MaxPool2d-3           [-1, 64, 12, 12]               0\n",
      "            Conv2d-4          [-1, 512, 10, 10]         295,424\n",
      "         LeakyReLU-5          [-1, 512, 10, 10]               0\n",
      "         MaxPool2d-6            [-1, 512, 5, 5]               0\n",
      "           Flatten-7                [-1, 12800]               0\n",
      "            Linear-8                  [-1, 256]       3,277,056\n",
      "         LeakyReLU-9                  [-1, 256]               0\n",
      "           Linear-10                   [-1, 24]           6,168\n",
      "================================================================\n",
      "Total params: 3,580,312\n",
      "Trainable params: 3,580,312\n",
      "Non-trainable params: 0\n",
      "----------------------------------------------------------------\n",
      "Input size (MB): 0.00\n",
      "Forward/backward pass size (MB): 1.61\n",
      "Params size (MB): 13.66\n",
      "Estimated Total Size (MB): 15.27\n",
      "----------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "# Create the network and look at it's text representation\n",
    "#\n",
    "# net = cnn.CNN().to(device)\n",
    "#torchsummary.summary(net, (1, 28, 28))\n",
    "\n",
    "net = network.network(in_channels=1, output_size=24, device=device)\n",
    "torchsummary.summary(net, (1, 28, 28))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### train the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 1 \tTraining Loss: 0.998712 \tValidation Loss: 0.218202\n",
      "Epoch: 2 \tTraining Loss: 0.121702 \tValidation Loss: 0.052718\n"
     ]
    }
   ],
   "source": [
    "trained_model, train_losses, valid_losses = cnn.train(\n",
    "    train_loader=train_loader,\n",
    "    valid_loader=valid_loader,\n",
    "    model=net, \n",
    "    num_epochs=num_epochs, \n",
    "    learning_rate=learning_rate, \n",
    "    device=device\n",
    "    )"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### training vs. validation loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f46d629bfd0>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABDAAAAJMCAYAAAAMpFaWAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAB7CAAAewgFu0HU+AACz0ElEQVR4nOzdd1hU5/Y24GeG3nvvICgKFlQU6WJJfkaNiYm9JkpyTFXTPCmeVJOY5KRoFHvXFGvU2GmKoggqYmXoIEWqdJj5/vBjjoQZBIShzHNfl9cZZ6+999omJzJr3nctgUQikYCIiIiIiIiIqAsTdnYCRERERERERESPwwIGEREREREREXV5LGAQERERERERUZfHAgYRERERERERdXksYBARERERERFRl8cCBhERERERERF1eSxgEBEREREREVGXxwIGEREREREREXV5LGAQERERERERUZfHAgYRERERERERdXksYBARERERERFRl8cCBhERERERERF1eSxgEBEREREREVGXxwIGEREREREREXV5LGAQERERERERUZfHAgYRERERERERdXmqnZ0Adbyqqipcu3YNAGBmZgZVVf5jJyIiIiIioo5TV1eH/Px8AICnpyc0NTWf+Jr8JKsErl27Bm9v785Og4iIiIiIiJRQbGwshg4d+sTX4RYSIiIiIiIiIuryuAJDCZiZmUlfx8bGwsrKqhOzISIiIiIiop4uJydHuhPg0c+kT4IFDCXwaM8LKysr2NradmI2REREREREpEzaqw8jt5AQERERERERUZfHAgYRERERERERdXksYBARERERERFRl8cCBhERERERERF1eSxgEBEREREREVGXxwIGEREREREREXV5LGAQERERERERUZfHAgYRERERERERdXlKW8DIy8vDX3/9hY8//hhPP/00TE1NIRAIIBAIMHfu3A655+7duzF27FhYWVlBU1MTjo6OmDVrFs6fP98h9yMiIiIiIiLqKVQ7O4HOYmFhobB7VVVV4YUXXsBff/3V6P20tDSkpaVh586dWL58OT766COF5URERERERETUnSjtCoxH2dnZYcyYMR12/ZdeeklavAgODsb+/fsRGxuLDRs2wMXFBWKxGB9//DHWr1/fYTkQERERERERdWdKuwLj448/xtChQzF06FBYWFggNTUVTk5O7X6fiIgI7Ny5EwAwfvx47Nu3DyoqKgCAoUOHYsKECRg8eDDS09Px7rvvYvLkyTA0NGz3PIiIiIiIiIi6M6VdgfGf//wHzzzzTIdvJfnmm28AACoqKli9erW0eNHA1NQUX3/9NQCgqKgIGzZs6NB8iIiIiIiIiLojpS1gKMKDBw9w6tQpAMDo0aNha2srM+65556Dvr4+AGDv3r0Ky4+IiIiIiIiou2ABowPFxsaiuroaABAYGCg3Tl1dHcOHD5eeU1tbq5D8iIiIiIiIiLoLFjA60I0bN6Sv+/Tp02xsw/G6ujrcuXOnQ/Pqqsqr65BXVtXZaRAREREREVEXpLRNPBUhIyND+lre9pEGdnZ2jc7r27dvi++TmZnZ7PGcnJwWX6sz7YpNxzd/38Lzg22wwN8Zzma6nZ0SERERERERdREsYHSgsrIy6Wtd3eY/jOvo6EhfP3jwoFX3ebT40V3V1ImxIToFNfVi7IrNwO6LGRjT1wKhgS7wsjfq7PSIiIiIiIiok7GA0YGqqv63HUJdXb3ZWA0NDenrysrKDsupqzp0JRs5Jf/785JIgGPXc3Hsei68HY0RGuiM4N7mEAoFnZglERERERERdRYWMDqQpqam9HVNTU2zsQ3NPgFAS0urVfd5dKuKLDk5OfD29m7VNRUtp6QS6ipC1NSLmxyLTS1EbGohXM11sTDAGRMH2kBdle1biIiIiIiIlAkLGB1IT09P+vpx20LKy8ulrx+33eSfHtdfozt4baQrXhxih03nUrH9fBrKquqaxNzJe4B3/riKlcdvYb6vE6YNs4e+plonZEtERERERESKxq+xO9CjhYXHNdp8dBVFT+hp0Rbm+pp476k+OPf+SPz7/9xhqa8pMy63tBpfHb0J369O46ujN5BbysklREREREREPR0LGB3o0UkiN2/ebDa24biqqip69erVoXl1dXqaalgQ4IzId4Px3QsD4GYhe0VKWXUd1kaI4Pf1abz7xxXczSuTGUdERERERETdHwsYHWjo0KHS5p0RERFy42pqanD+/Pkm5yg7dVUhnh9si2NvBWDT3KEY5mQsM662XoLfLmVi1PeReHnLJVxKLVRwpkRERERETy41NRUCgQACgQCbN2/usPsEBQVBIBAgKCiow+6hSI6OjhAIBJg7d25np0IdjAWMDqSnp4eQkBAAwMmTJ+VuI9m7dy9KS0sBAJMmTVJYft2FQCBAcB9z7An1wb5/jcDTHpYQyBlGcvJGLiavicHzv57D8ev3IBZLFJssEREREXUpjxYFnuQXEXU+FjCewObNm6X/QVu+fLnMmKVLlwIA6urqsGjRItTX1zc6XlBQgPfeew8AYGhoiJdffrlDc+7uBtkb4deZg3F6SRCmD7OXO40kLq0IC7fFYdQPEdgdm47qunqZcURERERERNQ9KO0UkujoaNy9e1f6+4KCAunru3fvNlmy1dblSCNHjsTUqVOxe/duHDx4EKNHj8Zbb70Fa2trXLt2DV988QXS09MBACtWrICRkVGb7qNsnEx18OUkT7w9yg1bzqVia0wqSmVMLhHll+P9vdfw3YnbmOfriBnDHGCgxcklRERERMrCxsYG165dk3t87NixyM7OhrW1NY4dO6bAzJpydHSERNLxK4jDw8M7/B5EHUFpCxjr16/Hli1bZB47e/Yszp492+i9J9lPtXHjRpSWluLIkSM4c+YMzpw50+i4UCjERx99hNDQ0DbfQ1mZ6Wlg6djeeDXIBXsuZmBDdAqyiiubxOWXVeObv29h1em7mD7MHvP9nGBloNUJGRMRERGRIqmpqcHDw6PZ4y2JI6LOxy0kCqClpYXDhw9jx44dGD16NMzNzaGurg47OztMnz4d0dHRcregUMvoaKhivp8Twt8Jwn+nDEQfSz2ZceU19VgXlQL/r89gyW9XcOseJ5cQERERERF1B0pbwNi8eTMkEkmLf8kyd+5c6fGWFCCmT5+O48ePIzc3F9XV1UhPT8eOHTvg4+PTzk+nvNRUhHh2kA2OvumPLfO9McLFRGZcnViCPy9nYux/IzFvUyzOi+4rZLkeEREREXUf/5zWcefOHbz22mtwdXWFtrY2BAIBUlNTpfE5OTlYvXo1Jk+eDFdXV+jo6EBDQwM2NjaYOHEi9uzZA7FYLPd+j5tCsnz58kZNRauqqvDtt9/Cy8sLenp60NPTg7e3N3755RfU1TXdXi3vuR6Xw4kTJzB+/HhYWlpCQ0MDTk5OePXVV+UOKXhUQUEB3nnnHbi5uUFLSwsWFhYYPXo09u3bB6BxX8FH/yw7wqFDhzB58mTY2tpCQ0MDJiYm8PHxwYoVK/DgwYNmzy0uLsYXX3wBHx8fGBkZQU1NDWZmZujbty8mTZqEX3/9FXl5eTLPPX36NKZNmwYnJydoaWlBW1sbjo6OGD58OJYuXYrTp093xOP2SEq7hYR6NoFAgEA3MwS6meFqZjHWRopw9FoOZA0lOXMrH2du5WOAnSFeCXDGmH6WUBGy0zQRERER/c+BAwcwY8YMlJeXyzxeX18PW1tbmQWK7OxsHDx4EAcPHsSGDRuwd+9e6OrqPlE+ubm5GDt2LK5cudLo/YsXL+LixYs4fvw49u/fD6Hwyb6zfv/99/H11183ei81NRVr1qzBn3/+iYiICLi7u8s898qVKxg9ejTy8/Ol71VVVeHkyZM4efIkFi5cqJAvc6uqqjB9+nRp0aRBYWEhzp8/j/Pnz+Pnn3/G4cOHMXDgwCbn37hxA6NGjUJ2dnaj9wsKClBQUIAbN25g//79qK+vx2uvvdYoZvHixfjhhx+aXDMtLQ1paWm4cOECNm/e3KgnI8nHAgb1eP1tDbFquhfS7pdjfVQKfruUgeq6pn+xXMkoxqs7LsPRRBsLApzxvJctNNVUOiFjIiIiIupK0tPTMXPmTGhra+Ojjz6Cv78/VFRUcPHiRWkhomE178iRI/H000/D09MTZmZmKCsrg0gkwrp16xATE4MTJ05g0aJFcvvxtdRzzz2HGzdu4I033sD48eNhbGyMW7du4bPPPsONGzdw6NAhrFu37on67K1btw7nzp1DYGAgQkND4ebmhuLiYmzduhVbt25Ffn4+5s+fj5iYmCbnFhUV4amnnpIWL2bMmIGZM2fCzMwMd+/exY8//oiwsLAmBZiOMGfOHGnxYsCAAViyZAnc3d1RWFiI3bt3Y/PmzcjOzkZISAiuXr0KGxubRufPmjUL2dnZUFNTw4IFC/D000/D0tISYrEY2dnZiI2NxZ9//tnkvn/99Ze0eNG/f3+8+uqrcHd3h4GBAUpKSnDz5k2cOHFC5p8fycYCBikNBxMdfPasB94a5YqtMWnYEpOK4oraJnGp9yvw732J+OHEbcwd4YiZwx1gqK3eCRkTERGRshOLJSiqqOnsNBTKSFsdwi62GjYlJQXW1taIiYmBvb299P1hw4ZJX6uoqODWrVvo1atXk/MDAwMxb948fPLJJ/j000+xbds2fPjhh3B1dW1zTg2rLB7dBuLl5YWxY8eib9++yM3NxerVq5+ogHHu3DksWLAAa9eulW5bAYCQkBCoq6tj/fr1OH/+POLj4zFo0KBG5y5fvhz37t0DAKxcuRJLliyRHhs8eDAmT56M559/HgcOHGhzfi1x+PBh/Pbbb9K8jxw5AnX1//1sP2bMGPj4+GDhwoUoLCzE4sWLsWfPHulxkUiEuLg4AMD333/fZIUFADz77LP44osvUFxc3Oj9hvs6ODjg7NmzTVbdNBSGCgsL2+VZlQELGKR0THQ18PZoN4QGOuP3S5lYFyVCZlHTySUFD2qw8vhtrA5PxtSh9njJ3wk2hpxcQkRERIpTVFGDwZ+f7Ow0FCruw1Ew0dXo7DSaWLFiRaPixT8JBAKZxYtHffzxx1i9ejUKCgpw8ODBRh/qW+v111+X2cPC2NgY8+bNw4oVK3D16lWUlJTAwMCgTfewsrLCzz//3Kh40WDp0qVYv349ACAqKqpRAaOqqkq6wsTLywuLFy9ucr6KigrWrl2LY8eOoaqqqk35tcSqVasAPJwys2nTpkbFiwYLFizAb7/9hpMnT2Lv3r3IycmBlZUVAEiLMAAQEBAg9z4CgQBGRkaN3ms418vLq9ktQ8bGxi1/ICWntE08ibTVVTFnhCPClwbhp2mD0M9aX2ZcRU09Np5NQcA3Z/DW7ngkZZcqOFMiIiIi6kzq6up44YUXWnVOw/aCW7duITExEYmJibhx4wZsbW0B4Im3TsyYMUPuscGDB0tfp6SktPkekydPhoaG7GJS7969pR/KRSJRo2NxcXEoKSkBAMyePVtmAQQALCwsMHbs2Dbn9zh1dXWIiIgAAIwePRp2dnZyYxcsWCA9Jzw8XPp+QyEDgMzGqs1pODcyMhLJycmtOpdkYwGDlJ6qihATBljjr9f9sP2lYfB3NZUZVy+WYH9CNv7vpyjM3hiLc3cLOLmEiIiISAm4urpCU1PzsXESiQTbt29HcHAwdHV1YWNjgz59+sDT01P6KyEhAQCeuGljnz595B579Bv9srKyDrkHAOmKg3/eIzExUfr60WKKLEOGDGljdo8nEolQUVEBoPF2H1kePf5o/k5OTvD39wcA/PDDD+jXrx8+/vhjnD59WnpteWbPng0AuH//Pjw8PDB16lRs2rQJd+/ebdPzEAsYRFICgQB+rqbY9tIw/PW6HyYOtJY7jSTydj6mr7+ACb+cxV9Xs1FXL38cFhERERF1b//cGiBLVVUVxo0bh1mzZiE8PByVlU23KD/qcccfR1tbW+6xRyeP1NfXd8g9Hr3PP+9RVFQkfW1ubt7sNczMzNqY3eM92lvCwsKi2VhLS0uZ5wHArl27pNNSkpKS8NlnnyEkJASGhoYIDAzEmjVrZG6DCQkJwS+//AItLS1UVVVhz549mD9/PlxdXWFra4tXXnlFIU1MexL2wCCSwcPGAD9OHYSlY3pjQ3QK9lzMQGVt0//4X8sqwWs742FvrI0F/k6YPNgOWuqcXEJERETtw0hbHXEfjursNBTKqAs2T1dRefzPd1988QWOHj0K4GFzxkWLFsHLywuWlpbQ0tKSftgPCAhAVFQUV/IqmLxtLC1hY2ODc+fO4dSpU9i7dy8iIiKQlJSE2tpaREZGIjIyEitXrsSRI0fg5ubW6NxFixbhhRdewM6dO3HixAmcPXsWJSUlyMrKwtq1axEWFoZly5bh888/f9JHVAosYBA1w85YG8sn9MObIa7Ydj4Nm8+lorC8aSfw9MIKfHTgOn44eQdzfBwx28cBRjpd7y9fIiIi6l6EQkGXbGhJjUkkEmlDSz8/P5w+fbrRKohHPbo6oad6dMVKXl5ekw/1j2oYs9oRHt1K82gzTlkePS6vqWZISAhCQkIAPNwWcvLkSYSFheH06dNITk7GlClTEB8f3+Q8c3NzvPXWW3jrrbcgFouRkJCAvXv3YtWqVSguLsYXX3yBoUOHYuLEiW15TKXCLSRELWCko443Qlxx9r2R+OxZD9gby15OV1hegx9O3saIFafxyYFEZBQ2vy+OiIiIiLq/wsJC6QfgF198UW7x4sGDB7h165YiU+sU/fr1k76+dOlSs7GPO/4knJ2dpdtgLly40GxsbGys9LWHh8djr21iYoIpU6bg1KlTmDBhAgAgISEBd+7cafY8oVAILy8vfP755zh16pT0/YaRq9Q8FjCIWkFLXQWzhjvgzNIgrJruhf62skdSVdbWY0tMGgK/PYPXd8UjMatEwZkSERERkaLU1dVJXzfX2HHDhg2ora1VREqdasiQIdLRrdu2bZO7XSY3NxfHjh3rsDxUVVURGBgIADhx4gQyMjLkxjasoFFRUZE5nrY5DasygNY1Z/Xy8pKuVnnSpq7KggUMojZQEQowrr8VDizyxa4FwxHUW3bzIbEEOHQlG8/8HI2Z6y8g6k4+9zsSERER9TBmZmYwNDQEAOzevRs1NU23HF+8eBEffvihgjPrHJqamtIJHJcvX8b333/fJEYsFiM0NFRm88v2tGjRIgBAbW0t5s+fL/OfzcaNG3H8+HEAwPPPP99odGpCQoJ0cowsEokEJ0+eBPCwz4ajo6P02J49e5pt1nrp0iXpliInJ6cWP5MyYw8MoicgEAjg42ICHxcT3MgpxbpIEQ5eyUaduGmRIvpuAaLvFqCvlT5CA50xztMKqiqsIRIRERF1d0KhEDNmzMCqVauQkJAAf39/vP322+jVqxdKSkpw5MgRrF69Grq6urC2tsbt27c7O+UOt3z5cvz++++4d+8eli5divj4eMyaNQtmZma4e/cufvzxR5w7dw7e3t7S7RtP0mhTnnHjxuGFF17A77//jpMnT2LYsGFYsmQJ3N3dUVRUhN27d2Pjxo0AHva++GexJSEhAfPmzcPQoUMxfvx4aWPW2tpapKSkYNOmTThx4gQAYOLEiY2KH++99x5eeeUVTJw4EQEBAXBzc4OOjg7u37+P6Oho/PzzzwAervpYsGBBuz97T8QCBlE7cbfSx/dTBmLJ2N7YGJ2CXbHpqKhpOrkkKacUb+5OwDd/38LL/k6YMtQO2ur8vyIRERFRd/bFF1/g7NmzSEhIQGxsLKZNm9bouLGxMf788098/PHHSlHAMDY2xt9//43Ro0cjPz8fO3bswI4dOxrFzJ07F/7+/tIChqamZofksnXrVtTV1WHfvn1ISEjArFmzmsRYW1vj8OHDsLGxkXmNixcv4uLFi3Lv4efnhw0bNjR5v7i4GFu2bMGWLVtknqepqYm1a9di8ODBLXwa5cavf4namY2hFj56pi9i3g/BO2N7w1RX9jSSrOJK/OdQEkasOI3vj99CwYNqBWdKRERERO3FwMAAZ8+exWeffQZPT09oampCV1cX7u7uWLp0Ka5cuYKAgIDOTlOhBgwYgKSkJCxZsgSurq7Q0NCAqakpgoODsXPnTmzatAmlpaXS+Ia+Ge1NU1MTe/fuxcGDB/Hcc8/B2toa6urqMDIywrBhw/DVV1/h1q1bGDhwYJNzp0+fjjNnzmDZsmXw9/eHk5MTtLW1oa6uDltbW0yYMAE7d+5EREREk+klkZGRWL9+PaZMmQJPT0+YmZlBVVUV+vr68PLywjvvvIOkpCTpdht6PIGEG/J7vMzMTNjZ2QEAMjIyYGtr28kZKZeq2nrsvZyFdVEipBSUy43TUBXihSG2WODvDAcTHQVmSERERETUOV5++WVs2LABtra2zTbZpO6nIz6HcgUGUQfTVFPB9GH2OLk4EGtmemGgnaHMuOo6MbafT0fwynAs2nEZVzKKFZonEREREZEiVVZW4sCBAwCA4cOHd3I21B2wgEGkICpCAZ7ysMK+f43Ab6E+COljLjNOLAEOX8vBxFVnMS3sPMJv5XFyCRERERF1O8nJyXJ/jq2vr8err74qHR86Z84cRaZG3RQ7BxIpmEAggLeTMbydjHE7twxhkSIcSMhCbX3T/7jHiO4jRnQffSz1sDDAGeMHWEONk0uIiIiIqBv47LPPEBsbi6lTp2LYsGEwNzdHZWUlrl69inXr1uHy5csAgJCQEIwbN66Ts6XugD0wlAB7YHR9OSWV2HQ2FTsvpONBdZ3cOGsDTcz3c8JUb3voarD+SERERERd19y5c+VO32jg6+uLAwcOwMTEREFZkaJ0xOdQFjCUAAsY3UdpVS12XkjHxugU5JXJn0qir6mKWT4OmDPCEeZ6HTNuioiIiIjoSdy6dQt//vknTpw4gbS0NOTn56O2thYmJiYYMmQIpkyZgqlTp0Io5ArjnogFDGoTFjC6n+q6ehyIz8bayGQk58ufXKKuKsTzXrZY4O8EZzNdBWZIREREREQkH6eQECkJDVUVvDjUDifeDsS62UMwxMFIZlxNnRi7YtMR8n0EXtkWh/j0IgVnSkREREREpBjcRE/UhQmFAozua4HRfS1wKbUQayNFOJGU2yROIgH+vn4Pf1+/B28nY4QGOCO4tzmEQkEnZE1ERERERNT+WMAg6iaGOBpjiKMx7uY9wLpIEfbFZ6GmXtwkLjalELEphXA118XCAGdMHGgDdVUutiIiIiIiou6NPTCUAHtg9Ex5pVXYdC4V28+noaxK/uQSS31NzPdzxDRve+hpqikwQyIiIiIiUlZs4kltwgJGz1ZWVYvdsRnYEJ2Ce6VVcuP0NFQxY7gD5vk6wkKfk0uIiIiIiKjjsIBBbcIChnKoqRPj4JVshEUm43buA7lxaioCTBpkg4UBzuhlrqfADImIiIiISFlwCgkRyaWuKsTkwbb4+80AbJw7BN5OxjLjausl+O1SJkZ9H4mXt1zCpdRCBWdKRERERETUemziSdTDCIUCjOxjgZF9LHA5vQhhESIcS7oHWWutTt7IxckbuRjsYITQAGeMcrfg5BIiIiIiIuqSuAKDqAfzsjfCmlmDcWpxIKYPs5c7jSQurQgLt8Vh1A8R2HMxHdV19QrOlIiIiIiIqHnsgaEE2AODGuSXVWPLuVRsjUlFaTOTS8z0NDDf1wnTh9nDQIuTS4iIiIiIqHXYxJPahAUM+qcH1XXYczEDG6JEyC6RP7lEV0MV07ztMN/PCVYGWgrMkIiIiIiIujMWMKhNWMAgeWrrxfjrajbWRohw816Z3DhVoQATBz6cXNLbkpNLiIiIiIioeR3xOZRNPImUmJqKEJMG2eLZgTaIvFOAtRHJOJd8v0lcnViCPy9n4s/LmRjZxxyhAc7wdjKGQMCGn0REREREpBhs4klEEAgECHQzw84Fw3HwNV+M628FecNITt/Mw5Sw85i0+hz+TsxBvZiLuIiIiIhaavny5RAIBHK/CAoKCoJAIEBQUNAT3afhHsuXL3+i67SHxz1zd7J582bps6SmpnZ2OkqHBQwiaqS/rSFWTffCmaVBmDXcARpyJpckZBTjle2XMer7COy8kI6qWk4uISIioq4nNDRU+oHzzJkzrTr31KlT0nNfe+21DsqQiFqKBQwiksnBRAefPeuBc++PxBshrjDUlj2NJKWgHMv2XYPf16fxy+k7KK6oUXCmRERERPLNnj1b+nrbtm2tOnf79u3S17NmzWq3nHoCrkSgzsACBhE1y0RXA4tHu+Hc+yOxfHxf2BjKnkZS8KAGK4/fxogVp/HpoSRkFVcqOFMiIiKipnx9feHi4gIA+OOPP1BZ2bKfUSorK/Hnn38CAHr37o1hw4Z1WI6PCg8Ph0QiQXh4uELupwjLly+HRCIB50fQk2IBg4haRFtdFXN9nRDxThB+mjYI/az1ZcZV1NRj49kUBHxzBm/vScCNnFIFZ0pERETUWMMqjLKyMhw4cKBF5+zfvx9lZQ+ntHH1BVHXwAIGEbWKqooQEwZY46/X/bD9pWHwdzWVGVcvlmBffBae/jEKczbG4lxyAavuRERE1ClmzZolbSDZ0m0kDXECgQAzZ87ssNyIqOVYwCCiNhEIBPBzNcW2l4bhr9f9MGGANVTkjC6JuJ2P6esuYMIvZ/HX1WzU1YsVnC0REREpMycnJ/j5+QEAjh8/jry8vGbjc3NzceLECQBAYGAgHBwcpMfOnz+PDz/8EEFBQbC0tIS6ujr09fXRt29fvPrqq0hKSnqiXFs6hWTHjh0ICgqCkZERdHV14eHhgU8++QTFxcUtuk9iYiI+//xzjB07Fra2ttDQ0ICuri5cXV0xZ84cnD9/XuZ54eHhEAgEmDdvnvQ9JycnaT+Mhl+PboFp6RSS1NRUvP322+jXrx/09PSgra0NV1dXhIaG4tq1a82e+8+pKxcvXsS0adOkz2ZjY4NZs2bhxo0bLfrzeVL5+fn48MMPMWjQIBgaGkJTUxOOjo6YNWsWoqOjH3v+6dOnMW3aNDg5OUFLSwva2tpwdHTE8OHDsXTpUpw+fVrmecXFxfjiiy/g4+MDIyMjqKmpwczMDH379sWkSZPw66+/Pvbf/y5NQj1eRkaGBIAEgCQjI6Oz06EeLP1+ueSTA4mS3h8ekTi895fcX/5fn5ZsPZciqaiu6+yUiYiISEmsW7dO+jPxf//732Zjf/jhB2nsxo0bpe9v2rRJ+r68XyoqKpJVq1bJvfYnn3wijZUlMDBQAkASGBgo83htba3k+eefl3t/FxcXiUgkkv7+k08+aXKNM2fOPPY5AEjef//9Np975syZFj+zRCKRbNmyRaKhodHsn+uXX34p9/xHn/fnn3+WqKqqyryOtra2JCIiQu51HufRfwdSUlJkxhw7dkyir6/f7J/PokWLJPX19TLPf/vttx/752tiYtLkvKSkJIm1tfVjz/3555/b/Pyt0RGfQ1VbVe0gImqGnbE2lk/ohzdCXLEtJg1bYlJRWN50Kkl6YQU+OnAdP5y8gzk+jpjt4wAjHfVOyJiIiIiUxYsvvog33ngDlZWV2LZtG9588025sQ3bR7S1tTF58mTp+3V1dTAyMsKECRMQGBgIV1dX6OjoIDs7G5cvX8ZPP/2EgoICvPbaa+jTpw9GjhzZ7s/x9ttvN2ou+u6776J///4oKSnB77//jnXr1mHKlCnNXqOurg46OjoYN24cRo4ciT59+kBfXx95eXm4fv06fvrpJ6SlpWHFihVwc3NrtNpi6NChuHbtGg4cOIAPP/wQAHDs2DFYW1s3uoeTk1OLn+nw4cOYO3cuJBIJdHV1sWTJEowaNQqqqqo4d+4cvvrqKxQUFGDZsmUwNDTEq6++Kvdax44dw4ULF9C/f3+8+eab8PT0RGVlJfbt24cff/wRFRUVmDVrFu7cuQN19fb/+TMhIQHjx49HTU0N1NTUsGjRIkycOBE6OjqIj4/HihUrkJKSglWrVkFHRwdff/11o/P/+usv/PDDDwCA/v3749VXX4W7uzsMDAxQUlKCmzdv4sSJE4iJiWly71mzZiE7OxtqampYsGABnn76aVhaWkIsFiM7OxuxsbHSf3e6rXYpg1CXxhUY1FkqquskW2NSJf5fn252RUafD49KPjmQKEm/X97ZKRMREXUt9fUSyYN85fol51vp9jB16lTpz8VJSUkyY65fvy6NmT59eqNjmZmZkvJy+T+vFBcXS/r37y8BIPHz85MZ8yQrMK5cuSIRCoUSABIvLy9JWVlZk5gtW7Y0+rZd1gqM/Px8SVFRkdznqK6ulowePVoCQOLg4CCpq2u6arYlKxEaNPfMNTU1EhsbGwkAia6uriQ+Pr5JTGpqqsTKykq6giI/P79JzKPP/H//93+S6urqJjGff/65NGbv3r3N5izP45576NCh0hUjx44da3K8sLBQ0rdvXwkAiVAolCQmJjY6PmvWLOmfu6x/vg3u37/f6PfJycktWmEhFoslhYWFj3nK9sEVGETUrWipq2DWcAdM97bH34n3sCYiGdeySprEVdbWY/O5VGw7n4ZxnlZYGOAMDxuDTsiYiIioi6ksBL516ewsFOudZEBHdpPwJzV79mzs3r0bwMNVFl9++WWTmEebfDZML2lgY2PT7PUNDAzw6aef4tlnn0V0dDTu378PExOTdsj8oTVr1kAsfthLLCwsDLq6uk1iGp7x6NGjcq9jatr8n6+6ujq+/fZbDBw4EGlpaUhISMDgwYOfLHk59u3bh6ysLADAv//9bwwcOLBJjIODA7799lvMnDkTFRUV2LRpE9555x2Z19PU1MSmTZtkrq5444038Omnn6KmpgZRUVGYNGlSuz5LbGwsLl68CAB4+eWXMWbMmCYxRkZGCAsLg5+fH8RiMVavXo1Vq1ZJj9+7dw8A4OXlJfOfbwNjY+NGv284DwACAgLknicQCGBkZNSyB+qC2MSTiDqcilCAcf2tcPA1X+xcMAyBbmYy4+rFEhy8ko1nfo7GzPUXEHUnn5NLiIiIqN2MGTMGVlZWAB42wfznzxkSiQQ7duwAAFhZWWHUqFHNXq+8vBypqam4fv06EhMTkZiYCDU1NenxK1eutGv+J0+eBAB4eno2W1CYP39+q65bXV2N9PR0JCUlSZ/j0T+b9n6ORzU8k0AgaDbvF154AQYGBo3OkWX06NEwNzeXeUxPTw+urq4AAJFI1NaU5Xo0r5deeklunK+vL9zd3ZucA0D672dkZCSSk5NbfO+G8wBg8+bNLT6vu2EBg4gURiAQYISLKbbM98bRN/0xaZANVOVMLom+W4BZG2Ix7qdoHEjI4uQSIiIiemIqKiqYPn06ACA9PR0RERGNjoeHhyMjIwMAMH36dKioqDS5RkMvht69e0NPTw9OTk7w8PCAp6cnPD09MW7cuEax7aWqqgp3794F8LAPRXO8vb0fe73y8nJ89dVXGDBgAHR0dODg4IB+/fpJn2PQoEHS2PZ8jn9KTEwEADg6OsotPAAPV4U05NRwjix9+vRp9n4NKxfKyspam+pjNeT1aK7yDBs2DABw584d1NT8r2dcw6qf+/fvw8PDA1OnTsWmTZuk/+zlcXJygr+/PwDghx9+QL9+/fDxxx/j9OnTqKioaPMzdTUsYBBRp3C30scPUwYi4t1gzPd1grZ60x8QACAppxRv7k5A0MpwbD6bgoqaOgVnSkRERD3JnDlzpK8f3S7yz9//c/sIAMTFxaFPnz746quvcPv27ceuFK2srHzCbP+nuLhYer/mPugDgIWFRbPHU1NT4enpiWXLluHq1auor69vNr49n+OfCgsLATw+ZwCwtLRsdI4s2trazV5DKHz4Efhxz9wWDXkZGxtDVbX5bg0NzyKRSFBUVCR9PyQkBL/88gu0tLRQVVWFPXv2YP78+XB1dYWtrS1eeeUVuStidu3aBR8fHwBAUlISPvvsM4SEhMDQ0BCBgYFYs2YNqqqq2uNROw17YBBRp7Ix1MLH4/vijZBe2H4+DZvPpaLgQdPJJZlFlVh+KAn/PXUHs30cMcfHASa6Gp2QMRERkQJpGT/sCaFMtIwfH/MEPD09MWDAAFy5cgV//PGH9MNiZWWldELDgAED0L9//0bn1dTU4MUXX8T9+/ehpqaG119/HRMnToSbmxuMjIygofHw5xKRSAQXl4d9S9pzK+yj1xIIZK9gbalZs2YhJSUFAoEA8+bNw9SpU+Hu7g4zMzPpc4jFYukKFEVs6W3JM3WXrcVP+iyLFi3CCy+8gJ07d+LEiRM4e/YsSkpKkJWVhbVr1yIsLAzLli3D559/3ug8GxsbnDt3DqdOncLevXsRERGBpKQk1NbWIjIyEpGRkVi5ciWOHDkCNze3J37OzsACBhF1CYba6nhtpCte9nfG3stZWBclQkpBeZO44opa/HTqDtZGJOPFIXZ42d8JDiY6nZAxERGRAgiFHdbQUpnNmTMHixcvRmlpKQ4ePIgpU6bgwIEDKC0tBSB79cXp06elfRNWrVqFBQsWyLz2o9+mt6dHGy/m5uY2G9vc8Zs3byI6OhoA8MEHH+CLL76QGddRz/FPDVs6Hm1CKU/Dc/2zgWVX0ZDX/fv3UVdX1+wqjIZnkddU09zcHG+99RbeeustiMViJCQkYO/evVi1ahWKi4vxxRdfYOjQoZg4cWKTc0NCQhASEiLN5eTJkwgLC8Pp06eRnJyMKVOmID4+vj0eWeG4hYSIuhRNNRVMH2aPk4sDsWamFwbYGcqMq64TY9v5NASvDMeiHZdxNbNYoXkSERFR9zV9+nTph8uGbSMN//ton4xHXb9+Xfp66tSpcq996dKl9kxVSlNTU9qAsmHShTzNHW+v53jSVSANPDw8ADzc1pKXlyc3rra2Vvqhu+GcrqYhr5qamscWCGJjYwEArq6uMiemPEooFMLLywuff/45Tp06JX3/t99+e2xOJiYmmDJlCk6dOoUJEyYAABISEnDnzp3HntsVsYBBRF2SilCApzyssP9fI7Bn4XCM7CN7r6dYAhy+loMJv5zFtLDzCL+V122WFxIREVHnsLCwkI64PHbsGBITE3H8+HEADyeVNPQneFRd3f/6cMlriigWixEWFtYBGT/UMBXl2rVrzX5A3rhxo9xjLXkO4OHI1uZoampKX1dXVzcb25yGZ5JIJM3m/ccff6CkpKTROV3No3lt2LBBblxMTAySkpKanNMSXl5e0hUbrW2u2rAqoy3ndhUsYBBRlyYQCDDM2QQb5w7F8bcDMHmwLdRUZFf8Y0T3MXfTRTz9YxT2xWeilpNLiIiISI6GZp51dXWYOnWq9IO9rO0jAKSrHwBgy5YtMmM++OADXL58uZ0z/Z/Q0FDpyoeFCxeivLzpdtsdO3bgyJEjcq/Rkuf49ddfsX///mZzeXRsZ2vGff7TpEmTYG1tDQD48ssvZTaozMjIwNKlSwE8bNI5b968Nt+vI3l7e0snxKxfvx4nTpxoElNSUoLQ0FAAD1dWvPrqq42O79mzp9mmqZcuXZJu73FycpK+n5CQgISEBLnnSSSSRiNrHR0dW/RMXQ17YBBRt+FmoYeVLwzAkjFu2HQ2FTsvpONBddOpJDfvleHtPVfw7d+38JK/M6YOtYOOBv9zR0RERP8zYcIEGBoaori4WLqtQl9fX2ZPAQAYO3YszM3NkZeXh3//+99IS0vDhAkTYGpqirt372LdunU4deoUfH19cfbs2Q7JecCAAVi0aBF++eUXXLp0CUOGDMF7770HT09PlJSU4Pfff0dYWBiGDBkidwvIoEGD4OHhgcTERPz6668oLi7GjBkzYGVlhYyMDGzfvh1//PHHY59j0KBB0NTURFVVFT766COoqqrC0dFROuXDxsYGWlpaj30mNTU1hIWFYfz48SgrK4Ofnx/eeecdhISEQFVVFefOncOKFSuk20tWrlwJU9Ou2xcmLCwMw4YNQ01NDcaNG4fXX38d48ePh66uLuLj47FixQppL5WlS5c22Q7z3nvv4ZVXXsHEiRMREBAANzc36Ojo4P79+4iOjsbPP/8M4OFWp0f7sCQkJGDevHkYOnQoxo8fDy8vL1haWqK2thYpKSnYtGmTtKAyceLERgWobkVCPV5GRoYEgASAJCMjo7PTIWo3xRU1ktVn7kqGfH5C4vDeX3J/eX7yt+Tbv29K8kqrOjtlIiIi6kIWLFgg/TkZgOSll15qNv7vv/+WaGpqNjrn0V9BQUGSxMRE6e83bdrU5BqffPKJ9LgsgYGBEgCSwMBAmcdramokzz33nNwcnJycJCKRSPr7Tz75pMk14uPjJUZGRnKv4enpKcnOzm72GhKJRPLuu+/KvcaZM2da/MwSiUSyefNmiYaGhtzrqaioSL788ku55z8u1waP+/N9nE2bNknvlZKSIjPm2LFjEn19fbnPAkCyaNEiSX19fZNzHRwcmj0PgERTU1OyZcsWuXk198vPz09y//79Nj17a3XE51BuISGibstASw2vBrkg+r1gfP28J5zNZE8jKa2qwy9n7sL369P4YO81iPIfKDhTIiIi6ooatpE0kLd9pMHYsWNx6dIlzJw5E9bW1lBTU4OZmRkCAwMRFhaGU6dOQUenY6ejqamp4c8//8S2bdvg7+8PAwMDaGtrw93dHcuWLUNcXFyjrQWyDBw4EAkJCXjllVfg4OAANTU1GBsbw9vbGytXrkRsbGyLvqFfsWIF1q1bB39/fxgbG0vHrrbFnDlzcPPmTbz55ptwd3eHjo4OtLS04OLiggULFiA+Ph4ffPBBm6+vSGPGjMHdu3exbNkyDBw4EPr6+tDQ0IC9vT1mzJiBqKgo/PLLL9LVKo+KjIzE+vXrMWXKFHh6esLMzAyqqqrQ19eHl5cX3nnnHSQlJTX5d3X69Ok4c+YMli1bBn9/fzg5OUFbWxvq6uqwtbXFhAkTsHPnTkRERHTZKS4tIZBI2O2up8vMzISdnR2Ah/vHbG1tOzkjoo4hFktw8kYu1kaKEJcmf/SXQACM7WuJ0EBnDLJvOraKiIiIiIieTEd8DuWmcCLqMYRCAcb0s8SYfpa4lFqItZEinEhqOgddIgH+vn4Pf1+/B28nY7wS6IwgN3MIhe0zDoyIiIiIiNofCxhE1CMNcTTGEEdj3M0rw7rIFOyLz0KNjKkksSmFiE0phJuFLhYGuGDCAGuoq3J3HRERERFRV8MtJEqAW0iIgNzSKmw6m4od59NQJmNySQNLfU3M93PENG976GmqKTBDIiIiIqKeoyM+h7KAoQRYwCD6n7KqWuyKTcfG6FTcK62SG6enoYoZwx0w39cR5vqaCsyQiIiIiKj7YwGD2oQFDKKmaurEOHglG2sjknEnT/5UEnUVISYNssGCAGf0MtdVYIZERERERN1XR3wO5UZvIlJK6qpCTB5si2NvBWDDnCHwdpI9TqqmXow9lzIw6vsILNh6CXFphQrOlIiIiIiIADbxJCIlJxQKEOJugRB3C1xOL0JYhAjHku5B1tq0E0m5OJGUiyEORggNdEFIH04uISIiIiJSFBYwiIj+Py97I6yZNRii/AdYF5WCPy9noqau6eSSS2lFuLT1ElzMdLAwwBnPDrKBhqpKJ2RMRERERKQ82ANDCbAHBlHb5JVVYcu5VGyLSUNplfzJJeZ6Gpjn64Tpw+xhoMXJJURERERE7IHRQdLT07F06VK4u7tDR0cHxsbG8Pb2xsqVK1FRUdEu90hKSsLrr78OT09P6OvrQ11dHWZmZggODsYPP/yAsrKydrkPEbUfcz1NvDO2D859EIKPnukLawPZ00jyyqrx9d834bviNL48cgM5JZUKzpSIiIiIqOdT+hUYhw8fxowZM1BSUiLzeO/evXHkyBE4Ozu3+R7fffcd3n//fdTVyf8G18HBAQcPHkT//v3bfB95uAKDqH3U1ovx19VsrI0Q4eY9+UVHNRUBJg60wcIAZ7hZ6CkwQyIiIiKiroFjVNvZlStXMGLECFRUVEBXVxcffPABgoODUVlZid27d2PdunUAgD59+uDixYvQ1W39CMXffvsNU6ZMAQCoq6tj0aJFGDVqFExNTZGcnIzVq1cjOjoaAGBlZYUbN27AwMCg/R4SLGAQtTeJRIKI2/lYGyFCjOh+s7Ej+5jjlUAXDHU0gkDAhp9EREREpBxYwGhnwcHBCA8Ph6qqKiIjI+Hj49Po+Lfffot3330XAPCf//wHH3/8cavv4enpicTERADAX3/9hXHjxjWJef7557F3714AD1drLF68uNX3aQ4LGEQd50pGMcIiRTiamANxM/81HWRviNAAZ4zuawkVTi4hIiIioh6OBYx2dPHiRXh7ewMAQkNDsWbNmiYxYrEYHh4euHHjBoyMjJCbmws1tZY36CstLZWupvDy8kJcXJzMuKtXr2LAgAEAHhYz/vjjj9Y+TrNYwCDqeKkF5VgfLcLvlzJRLWNySQMnUx0s8HfGc1420FTj5BIiIiIi6pnYxLMd7d+/X/p63rx5MmOEQiFmz54NACgqKkJ4eHir7lFTUyN93VwPDRcXF+nr6urqVt2DiLoGR1MdfP6sJ86+PxJvhLjCUFt2sTOloBzL9l2D39dnsOrMXZRU1Co4UyIiIiKi7klpCxhRUVEAAB0dHQwePFhuXGBgoPR1Q6+KljI1NYWxsTEAQCQSyY1LTk6WvnZzc2vVPYioazHV1cDi0W449/5ILB/fFzaGWjLjCh5U49tjt+Cz4hQ++ysJWcWcXEJERERE1BylLWDcuHEDANCrVy+oqqrKjevTp0+Tc1pj4cKFAIDLly/j6NGjMmM+++wzAICKigpefvnlVt+DiLoebXVVzPV1QsQ7Qfhx6kD0tdKXGVdRU48N0SkI/OYM3t6TgBs5pQrOlIiIiIioe5D/yb0Hq6qqQkFBAQA8dh+OkZERdHR0UF5ejoyMjFbf69///jcuXbqEkydPYtKkSXjttdcQEhICU1NTiEQi/Prrr4iIiICKigp++uknuLu7t/oemZmZzR7Pyclp9TWJqH2oqggxcaANJgywRvTdAqyNECH6bkGTuDqxBPvis7AvPguBbmYIDXSGj7MJJ5cQEREREf1/StnEMz8/H+bm5gCAKVOmYPfu3c3GW1hYIC8vDx4eHrh27Vqr71dXV4fNmzdjxYoVjbaLNHjuuefw7rvvYtiwYa2+NoBWfcBhE0+izpeYVYK1kSIcvprd7OSS/rYGCA1wwVMenFxCRERERN0Lm3i2k6qqKulrdXX1x8ZraGgAACor27ZH/dKlS9i1a5fcPhgnT57Eli1bUFrKpeNEysDDxgA/TxuEiHeCMXeEIzTVZP+n+GpmCRbtvIyR34Vj2/k0VNXWKzhTIiIiIqKuQykLGJqamtLXj04KkadhMoiWluxmfM35448/EBQUhNOnT8PT0xP79u3D/fv3UVNTg+TkZHz55Zeora3Fr7/+ihEjRuDevXutvkdGRkazv2JjY1t9TSLqeHbG2lg+oR/OvR+Ct0e5wVhHdkE17X4FPtqfCN8Vp/HTqTsoKn/8f7eIiIiIiHoapdxCUlVVJS1GjBs3Dn/99Vez8bq6uigvL8fw4cMRExPT4vvk5ubCxcUF5eXl6NevHy5cuAAdHZ0mcadOncLo0aMhkUgwefJk/P777617oMfoiKU7RNT+Kmvq8UdcBtZFpSC9sEJunJaaCqYMtcNLfk6wM9ZWYIZERERERC3DLSTtRFNTE6ampgAe3wCzqKgI5eXlACD9w2+p3bt3S89dtmyZzOIFAISEhCAkJAQAsHfvXhQVFbXqPkTUM2ipq2CWjyNOLwnEL9MHwdPGQGZcZW09Np9LRdDKcLyxKx6JWSUKzpSIiIiISPGUsoABQDrt4+7du6irq5Mbd/PmzSbntNSjY1e9vLyajR08eDAAQCwW4/bt2626DxH1LKoqQjzT3xoHX/PFzgXDEOhmJjOuXizBwSvZeObnaMzacAHRdwqghIvqiIiIiEhJKG0Bw8/PDwBQXl6OuLg4uXERERHS176+vq26h6rq/6bUNlckAYDa2lqZ5xGR8hIIBBjhYoot871x5A1/TBpkI3caSdSdAszccAHP/ByNAwlZqKsXKzhbIiIiIqKOpbQFjGeffVb6etOmTTJjxGIxtm7dCgAwNDREcHBwq+7h5OQkfR0VFdVsbGRkJICHH1gcHR1bdR8i6vn6WuvjhykDEfFOEOb7OkFbXUVm3PXsUry5OwFBK8Ox5VwqKmqaL54SEREREXUXSlvA8Pb2hr+/PwBgw4YNMptzfvfdd9JtIG+++SbU1NQaHd+8eTMEAgEEAgGWL1/e5Pxx48ZBIHj4bekXX3yBrKwsmbmEhYXh0qVLAIDhw4fDxMSkzc9FRD2brZE2Ph7fF+feH4mlY9xgqit7cklmUSU+OXgdvitO4/sTt3H/QbWCMyUiIiIial9KW8AAgB9//BFaWlqoq6vDmDFj8NVXX+H8+fM4c+YMQkND8e677wIA3NzcsGTJklZfv0+fPpg3bx4AICsrC4MGDcKXX36JqKgoJCQk4NChQ5gxYwZCQ0MBACoqKvjyyy/b7wGJqMcy1FbHayNdEf3eSHwxyQOOJrKnkRRV1OKnU3cwYsVpfLQ/EWn3yxWcKRERERFR+1DqZguDBg3Cnj17MHPmTJSWlmLZsmVNYtzc3HD48GHo6em16R6rV69GeXk59uzZg/z8fPz73/+WGaejo4OwsDAEBQW16T5EpJw01VQwY5gDpg61x4mke/g1QoQrGcVN4qrrxNh2Pg07LqThaU8rhAY4o7+tocLzJSIiIiJqK6UuYADA+PHjcfXqVfz44484fPgwMjMzoa6ujl69euGFF17Aa6+9Bm1t2d9stoSGhgZ2796N0NBQbN68GefPn0dWVhaqq6uhr6+P3r17Y9SoUVi4cGG7zMUlIuWkIhTgKQ8rjO1nidiUQqyNFOH0zbwmcWIJcPhqDg5fzcEIFxOEBrogwNVUut2NiIiIiKirEkg4c6/Hy8zMhJ2dHQAgIyODhRIiJXHrXhnCIkUPp5KI5f+nvo+lHkIDnfFMf2uoqSj1zkIiIiIiaicd8TmUBQwlwAIGkXLLKanExugU7IrNwINq+VNJrA008ZK/M6YOtYOOhtIv0CMiIiKiJ8ACBrUJCxhEBAAllbXYeSEdG8+mIL9M/lQSAy01zBrugDkjHGGmp6HADImIiIiop2ABg9qEBQwielR1XT32x2dhbaQIonz5U0nUVYWYPNgWC/yd4WSqo8AMiYiIiKi764jPodzsTESkZDRUVTBlqD1Ovh2IsFmD4WVvKDOupk6MnRfSMfK7cLy6PQ7x6UWKTZSIiIiI6BHc5ExEpKSEQgHG9LPEmH6WuJRaiDURIpy8kdskTiIBjibew9HEe/B2MsYrgc4IcjOHUMjJJURERESkOCxgEBERhjgaY72jMe7mPZxcsi8+C7X1TXcYxqYUIjalEG4WulgY4IIJA6yhrsrFfERERETU8dgDQwmwBwYRtVZuaRU2nU3FjvNpKGtmcomlviZe8nPCVG876GmqKTBDIiIiIurK2MST2oQFDCJqq7KqWuyKTceG6BTklsqfXKKnqYqZwx0wb4QjzPU1FZghEREREXVFLGBQm7CAQURPqqZOjAMJWQiLFOFO3gO5ceoqQkwaZIOFgc5wMdNVYIZERERE1JVwCgkREXUKdVUhXhhih2NvBWDDnCHwdjSWGVdTL8aeSxkY9X0EFmy9hLi0QgVnSkREREQ9FZt4EhFRiwmFAoS4WyDE3QJxaUUIi0zG8aRc/HMtn0QCnEjKxYmkXAxxMEJooAtC+nByCRERERG1HQsYRETUJoMdjLB21hCI8h9gXVQK/ryciZo6cZO4S2lFuLT1ElzMdBAa4IKJg6yhoarSCRkTERERUXfGHhhKgD0wiEgR8sqqsOVcKrbFpKG0Sv7kEnM9Dcz3c8L0YfbQ5+QSIiIioh6JTTypTVjAICJFelBdh93/f3JJTkmV3DhdDVVMH2aP+b5OsDTg5BIiIiKinoQFDGoTFjCIqDPU1otx6Eo2wiJFuHmvTG6cmooAEwfaYGGAM9ws9BSYIRERERF1lI74HMoeGERE1CHUVIR4zssWkwbZIOJ2PtZEJOO8qOlUktp6Cf6Iy8QfcZkI6WOO0EAXDHU0gkDAhp9ERERE9D8sYBARUYcSCAQI6m2OoN7muJJRjLBIEY4m5kAsY/3fqZt5OHUzD4PsDREa4ILRfS2gwsklRERERAQWMIiISIEG2Bli1QwvpBaUY320CL9fykS1jMkl8enFeGV7HJxNdbAgwBmTBtlAU42TS4iIiIiUGXtgKAH2wCCirqrgQTW2nkvFlpg0lFTWyo0z1dXAPF9HzBzmAANtTi4hIiIi6urYxJPahAUMIurqyqvr8NulDKyPSkFWcaXcOG11FUzztsd8PyfYGGopMEMiIiIiag0WMKhNWMAgou6irl6Mw9dysDZChKScUrlxqkIBJgywxsJAZ/Sx1FdghkRERETUEpxCQkREPZqqihATB9pgwgBrRN8twNoIEaLvFjSJqxNLsDc+C3vjsxDU2wwLA5zh42zCySVEREREPRgLGERE1OUIBAL4u5rB39UMiVklWBspwuGr2TInl4Tfykf4rXwMsDVAaKALxvaz5OQSIiIioh5I2NkJEBERNcfDxgA/TxuEiHeCMcfHAZpqsv/qupJZgn/tuIyR34Vj2/k0VNXWKzhTIiIiIupI7IGhBNgDg4h6ksLyGmyNScWWc6koqpA/ucRERx1zRjhi1nAHGOmoKzBDIiIiImITT2oTFjCIqCeqrKnHH3EZCIsSIaNQ/uQSLTUVTBlqh5f8nGBnrK3ADImIiIiUV0d8DuUWEiIi6pa01FUwy8cRZ5YE4Zfpg+BpYyAzrrK2HpvPpSJoZTje3B2P69klCs6UiIiIiNoDm3gSEVG3pqoixDP9rTHO0woxyfexJlKEyNv5TeLqxRIcSMjGgYRs+LuaIjTABb69OLmEiIiIqLtgAYOIiHoEgUCAEb1MMaKXKZKyS7EuSoSDV7JRL2N0SdSdAkTdKUA/a32EBrrg/zwsoarCRYlEREREXRl7YCgB9sAgImWVWVSBjdGp2H0xHRU18qeS2BppYYG/M14YYgttddb2iYiIiJ4Um3hSm7CAQUTKrriiBtvPp2HT2VTcL6+RG2ekrYbZPo6Y7eMAE10NBWZIRERE1LOwgEFtwgIGEdFDVbX1+PNyJtZFipB6v0JunKaaEC8OscPLfs6wN+HkEiIiIqLW4hQSIiKiJ6CppoIZwxxwakkQfp3hhQG2sieXVNWKsTUmDUErz2DRzsu4lsnJJURERESdjRt9iYhI6agIBXja0wpPeVjiQkoh1kYk48ytppNLxBLg8NUcHL6agxEuJggNdEGAqyknlxARERF1AhYwiIhIaQkEAgx3NsFwZxPculeGsEgRDiRkoU7G5JJzyfdxLvk++ljq4ZVAF4zrbwU1Ti4hIiIiUhj2wFAC7IFBRNRy2cWV2HQ2BTsvpKO8mcklNoZamO/nhKlD7aCjwe8DiIiIiB7FJp7UJixgEBG1XkllLXZcSMPG6FQUPKiWG2egpYZZwx0w19cRppxcQkRERASABQxqIxYwiIjarqq2HvvjsxAWKYKooFxunLqqEJMH22KBvzOcTHUUmCERERFR18MpJERERAqmqaaCqd72OLk4EGtnDYaXvaHMuJo6MXZeSMfI78Lx6vY4JGQUKzRPIiIiop6Om3aJiIhaQCgUYGw/S4ztZ4lLqYVYEyHCyRu5TeIkEuBo4j0cTbyHYU7GeCXQBUG9zTi5hIiIiOgJsYBBRETUSkMcjbHe0Rh38x5OLtkXn4Xa+qY7Mi+kFOJCSiF6W+hhYYAzxg+whroqFz8SERERtQV7YCgB9sAgIupYuaVV2Hg2BTvPp6Osuk5unKW+Jl7yc8JUbzvoaaopMEMiIiIixWITT2oTFjCIiBSjtKoWuy6kY+PZFOSWyp9coqepipnDHTBvhCPM9TUVmCERERGRYrCAQW3CAgYRkWLV1IlxICELayNFuJv3QG6cuooQz3nZYEGAM1zMdBWYIREREVHH4hQSIiKibkBdVYgXhtjh+FsB2DBnCLwdjWXG1dSLsftiBkZ9H4GFWy8hLq1IwZkSERERdR9s4klERNRBhEIBQtwtEOJugbi0IoRFJuN4Ui7+ufZRIgGOJ+XieFIuhjoaITTABSP7mEMo5OQSIiIiogYsYBARESnAYAcjrJ01BMn5D7A+SoQ/47JQUy9uEncxtQgXUy/BxUwHoQEumDjIGhqqKp2QMREREVHXwh4YSoA9MIiIup68sipsPpuK7efTUFolf3KJuZ4G5vs5Yfowe+hzcgkRERF1E2ziSW3CAgYRUdf1oLoOu2PTsSE6BTklVXLjdDVUMWOYPeb5OsHSgJNLiIiIqGtjAYPahAUMIqKur7ZejENXsrE2QoRbuWVy49RUBHh2oA0WBjjD1UJPgRkSERERtVxHfA5lDwwiIqIuQE1FiOe8bDFpkA3Cb+djbUQyzosKm8TV1kvwe1wmfo/LREgfc4QGumCooxEEAjb8JCIiop6NBQwiIqIuRCAQILi3OYJ7myMhoxhhkck4mnivyeQSADh1Mw+nbuZhkL0hQgNcMKavBSeXEBERUY8l7OwEiIiISLaBdoZYPWMwziwJwoxh9tBQlf3Xdnx6MV7ZHodR30dgV2w6qmrrFZwpERERUcdjDwwlwB4YREQ9Q8GDamw9l4otMWkoqayVG2eqq4F5vo6YOcwBBtqcXEJERESKxyae1CYsYBAR9Szl1XX47VIG1kelIKu4Um6cjroKpnnbY76fE6wNtRSYIRERESk7FjCoTVjAICLqmWrrxThyLQdrIkS4kVMqN05VKMCEAdZYGOiMPpb6CsyQiIiIlBWnkBAREZGUmooQEwfaYMIAa0TdKUBYpAjRdwuaxNWJJdgbn4W98VkI6m2G0AAXDHc25uQSIiIi6lZYwCAiIurmBAIBAtzMEOBmhsSsEqyNFOHw1WyIZayxDL+Vj/Bb+Rhga4DQQBeM7WcJFU4uISIiom6AU0iIiIh6EA8bA/w8bRDClwZjjo8DNNVk/1V/JbME/9pxGSO/C8f282mcXEJERERdHntgKAH2wCAiUl6F5TXYGpOKLedSUVQhf3KJiY465o5wxCwfBxhqqyswQyIiIuqJ2MST2oQFDCIiqqypx+9xGVgXJUJGofzJJVpqKpgy1A4v+TnBzlhbgRkSERFRT9IRn0O5hYSIiEgJaKmrYLaPI84sCcLP0wbBw0b2NJLK2npsPpeKoJXheHN3PK5nlyg4UyIiIiLZ2MSTiIhIiaiqCDF+gDWe6W+FmOT7WBMpQuTt/CZx9WIJDiRk40BCNvxdTfFKoAtGuJhwcgkRERF1GhYwiIiIlJBAIMCIXqYY0csUSdmlCItMxqGrOaiXMbok6k4Bou4UwMNGHwsDXPB/HpZQVeEiTiIiIlIs9sBQAuyBQURELZFZVIEN0SnYHZuBymamktgaaWGBvzNeHGIHLXUVBWZIRERE3QWbeFKbsIBBREStUVReg+3n07D5XCrul9fIjTPSVsNsH0fMGeEIYx1OLiEiIqL/YQGD2oQFDCIiaouq2nr8EZeJ9VEipN6vkBunqSbEi0Ps8LKfM+xNOLmEiIiIOIWkw6Snp2Pp0qVwd3eHjo4OjI2N4e3tjZUrV6KiQv4PbG1x8uRJzJ07F7169YKOjg4MDAzg5uaGyZMn49dff8WDBw/a9X5ERERtpammgpnDHXBqSRB+neGFAbYGMuOqasXYGpOGoJVn8NrOy7iWycklRERE1P6UfgXG4cOHMWPGDJSUyP5hq3fv3jhy5AicnZ2f6D5FRUWYN28eDhw40GxcfHw8Bg4c+ET3+ieuwCAiovYgkUhwIaUQayOSceZW08klj/LtZYKFAS4IcDXl5BIiIiIl1BGfQ5V6CsmVK1fw4osvoqKiArq6uvjggw8QHByMyspK7N69G+vWrcOtW7cwbtw4XLx4Ebq6um26T0lJCUaPHo24uDgAwLhx4zB16lT06tUL9fX1SEtLw8WLF/HHH3+05+MRERG1K4FAgOHOJhjubIKb90oRFinCwYRs1MmYXHL27n2cvXsf7lb6CA1wxrj+VlDj5BIiIiJ6Akq9AiM4OBjh4eFQVVVFZGQkfHx8Gh3/9ttv8e677wIA/vOf/+Djjz9u031mz56Nbdu2QVVVFdu3b8eUKVNkxkkkEtTX10NVtX3rSlyBQUREHSW7uBKbzqZg54V0lNfIn1xiY6iFl/ycMGWoHXQ0lPr7EyIiIqXAJp7t6OLFi/D29gYAhIaGYs2aNU1ixGIxPDw8cOPGDRgZGSE3Nxdqamqtuk90dDT8/f0BAMuXL8cnn3zy5Mm3EgsYRETU0Uoqa7HjQho2Rqei4EG13DgDLTXM9nHAnBGOMNXVUGCGREREpEhs4tmO9u/fL309b948mTFCoRCzZ88G8LCHRXh4eKvv88svvwAAdHV1sWTJklafT0RE1B0YaKnhX0G9EP1eMFY85wlnUx2ZcSWVtfj59F34rjiNf++7htSCcgVnSkRERN2V0hYwoqKiAAA6OjoYPHiw3LjAwEDp6+jo6Fbdo6amRtq08+mnn5b20Kirq0NaWhrS09NRU1PT2tSJiIi6LE01FUz1tsfJxYFYO2swBtkbyoyrrhNjx4V0BH8Xjn/tiENCRrFC8yQiIqLuR2k3od64cQMA0KtXr2Z7TvTp06fJOS115coVVFVVAQB8fHxw7949fPDBB/j9999RXv7wGydNTU0EBwfjww8/xIgRI1r7GAAeLs1pTk5OTpuuS0RE1FZCoQBj+1liTF8LXEorwtqIZJy8kdckTiIBjly7hyPX7mGYkzFeCXRBUG8zTi4hIiKiJpSygFFVVYWCggIAeOw+HCMjI+jo6KC8vBwZGRmtuk9SUlKje3p6ekrv++j7R48exbFjx/Ddd9/hrbfeatU9AEj3FREREXU1AoEAQx2NMdTRGHdyy7AuSoR98VmorW/agutCSiEupBSit4UeFgY4Y/wAa6irKu1iUSIiIvoHpfypoKysTPq6JaNRdXQe7uN98OBBq+5TWFgoff2f//wHBQUFeOaZZ3Dp0iVUVVUhNzcXq1evhr6+PsRiMRYvXoyjR4+26h5ERETdhauFHr6ZPABR745EaKAz9ORMI7mVW4Ylv19B4LdnsD5KhAfVdQrOlIiIiLoipV2B0UBdXf2x8RoaD7ukV1ZWtuo+DdtEAKC6uhrjx4/H/v37IRQ+rBuZm5vj1VdfhaenJwIDAyEWi/Huu+/iqaeeatXS2cetDMnJyZFOXCEiIupslgaa+OBpdywK7oVdF9Kx8WwKckubTi7JKanC54dv4MdTdzBzuAPmjXCEub5mJ2RMREREXYFSFjA0Nf/3w09LmmhWVz/8oUpLS6vN9wGAb7/9Vlq8eJSfnx+ee+45/PHHH0hMTERiYiI8PT1bfB+ORSUiou5IX1MNoYEumOvriAMJ2QiLFOFuXtPVjmVVdfg1PBkbolLwnJcNFgQ4w8Xs8SsoiYiIqGdRyi0kenp60tct2RbSsJKiJdtN5N3HyckJvXv3lhs7duxY6euLFy+26j5ERETdmYaqCl4cYofjbwVg/ewhGOpoJDOupl6M3RczMOr7CCzceglxaUUKzpSIiIg6k9KuwDA1NUVBQcFjJ3gUFRVJCxitbZb5aPzjVkk8GpuX17RLOxERUU8nFAowqq8FRvW1QFxaIdZGiHDiRi4k/+j3KZEAx5NycTwpF0MdjRAa4IKRfcwhFHJyCRERUU+mlCswAMDd3R0AcPfuXdTVyW8OdvPmzSbntFS/fv2kr+vr65uNffR4c2NdiYiIlMFgB2OEzR6Ck4sDMc3bDuoqsn9kuZhahJe3XsKY/0bit0sZqK5r/u9bIiIi6r6UtoDh5+cH4OH2kLi4OLlxERER0te+vr6tuoeDgwPs7e0BAMnJyc3GPnrcxsamVfchIiLqqVzMdPHVc/0R/V4w/hXkAj1N2UX+u3kP8O4fVxHwzRmsiUhGaVWtgjMlIiKijqa0BYxnn31W+nrTpk0yY8RiMbZu3QoAMDQ0RHBwcKvv8/zzzwMAcnNzce7cOblxe/fulb729/dv9X2IiIh6MnN9Tbz7VB/EfBCCD8e5w8pA9jSS3NJqrDh6EyO+Oo2vjtzAvZIqmXFERETU/XTJAkZubi42bNiAr7/+Gr/99lurx5e2hLe3t7RQsGHDBsTExDSJ+e6773Djxg0AwJtvvgk1NbVGxzdv3gyBQACBQIDly5fLvM9bb70lnUbyxhtvNBqt2mD79u0IDw8HAIwbN45TRYiIiOTQ1VDFy/7OiHgnGN+/OAC9LfRkxj2orsPaSBH8vzmNd36/gju5ZQrOlIiIiNqbwpst3LhxA5988gkEAgHWrl0LQ0PDRscPHjyI6dOnNypa2NnZ4eDBg+jfv3+75vLjjz/C19cXlZWVGDNmDJYtW4bg4GBUVlZi9+7dCAsLAwC4ublhyZIlbbqHvb09Pv30U7z77ruIi4uDt7c33n33XXh4eKCkpAR79+7FmjVrAAD6+vr44Ycf2u35iIiIeip1VSGe87LFpEE2CL+dj7URyTgvKmwSV1svwe9xmfg9LhOj3M0RGuiCIQ5GEAjY8JOIiKi7UXgBY//+/fjjjz8QEBDQpHiRl5eHmTNnoqKiotH76enpGD9+PJKSkqCjo9NuuQwaNAh79uzBzJkzUVpaimXLljWJcXNzw+HDhxuNRG2td955B4WFhfj666+RlJSEuXPnNokxNzfH/v374erq2ub7EBERKRuBQIDg3uYI7m2OhIxihEUm42jivSaTSwDg5I08nLyRBy97Q4QGumC0uwUnlxAREXUjCt9CcurUKQgEAjzzzDNNjq1evRoPHjyAqqoqvv/+e1y5cgXffPMNhEIhMjMzsW7dunbPZ/z48bh69SrefvttuLm5QVtbG4aGhhgyZAi+/vprxMfHo1evXk98n6+++gpnz57FrFmz4OjoCA0NDRgYGGDo0KH47LPPcPv2bfj4+LTDExERESmngXaGWD1jME4vCcKMYfZQV5X9Y87l9GKEbovDqO8jsDs2HVW1nFxCRETUHQgkElnfUXQcNzc3JCcn4++//8bo0aMbHevfvz+uX7+OefPmYf369dL3Q0NDsW7dOgQFBeH06dOKTLdHyMzMhJ2dHQAgIyODPTaIiEgp5JdVY2tMKrbGpKGkUv5UElNdDczzdcTMYQ4w0FaTG0dEREQt1xGfQxW+AiM/Px8AYGZm1uj9goICXL9+HQAwffr0RscmTJgAANLjRERERI9jpqeBJWN649z7I/HxM31hY6glM67gQTW+PXYLI1acwud/JSG7uP2bhxMREdGTU3gBo6G/RVVV47Fm0dHRkEgkUFdXh6+vb6NjVlZWAIDi4mKF5EhEREQ9h46GKub7OSH8nSD8OHUg3K30ZcaV19RjfXQKAr45g8W/JeDmvVIFZ0pERETNUXgBw9jYGMDDxpyPOnXqFABgyJAh0NDQaHSsrq4OAKCrq6uADImIiKgnUlMRYuJAGxx5ww9b53vDt5eJzLg6sQR7L2fhqf9GYe6mWMQk34eCd9wSERGRDAovYAwYMAAAsHPnTul7lZWV+P333yEQCDBy5Mgm56SlpQEALCwsFJMkERER9VgCgQABbmbY8fJwHHrND8/0t4K8YSTht/Ixbd15PLvqLI5cy0G9mIUMIiKizqLwAsbUqVMhkUhw6NAhTJ06Fb/88gvGjBmDvLw8CAQCTJs2rck5Fy5cAAA4OzsrOl0iIiLqwTxtDfDLdC+ELw3GbB8HaKrJ/tHoSmYJ/rXjMkK+C8f282mcXEJERNQJFD6FRCwWIygoCNHR0RAI/vd1h0QiwUsvvSRzVKqzszPS0tLw7bffYvHixYpMt0fgFBIiIqKWuf+gGltj0rA1JhVFFfInl5joqGPuCEfM8nGAoba6AjMkIiLqHnrEFBKhUIijR49i8eLFsLW1haqqKuzs7PDRRx/h119/bRJ/6NAhpKamAkCTsatERERE7clEVwNvj3bDufdD8OnEfrAzlj255H55Db47cRsjVpzGfw5dR2ZRhYIzJSIiUj4KX4HRWkVFRSgtfdgF3MHBoZOz6Z64AoOIiKht6urFOJp4D2sjk5GYJX8qiYpQgPH9rbAwwAV9rWVPOSEiIlImHfE5tMsXMOjJsYBBRET0ZCQSCc4l38eaiGRE3SloNtbf1RSvBLpghItJo+2yREREyqQjPoeqPvEViIiIiHo4gUAA316m8O1liuvZJVgXKcKhq7KnkkTdKUDUnQJ42OgjNMAFT3tYQlVF4bt2iYiIehyF/21aW1uLpKQkJCUlobq6usnxqqoqLFmyBHZ2dtDS0kLfvn3xyy+/KDpNIiIiIpn6WRvgv1MHIeKdIMzzdYSWmorMuMSsUry+Kx7B34Vja0wqKms4uYSIiOhJKHwLyW+//YZp06bB1NQUGRkZUFdv3Ln76aefxvHjx/FoWgKBAK+99hp+/PFHRabaY3ALCRERUccpKq/B9vNp2HwuFffLa+TGGWmrYc4IR8z2cYSxDieXEBFRz9YjppAcO3YMEokEzz33XJPixeHDh3Hs2DEAgK2tLSZNmgQbGxtIJBL88ssviImJUXS6RERERM0y0lHH6yGuOPv+SHz+rAccTLRlxhVV1OK/J+9gxIpT+ORAIjIKObmEiIioNRRewLh8+TIEAgECAgKaHNu0aRMAwM3NDdevX8eff/6JxMREuLu7AwDWr1+v0FyJiIiIWkpTTQUzhzvg9JIgrJ7hhQG2BjLjqmrF2BKThsBvz+C1nZdxLbNEwZkSERF1TwovYOTl5QEAnJ2dG70vFotx8uRJ6XYRPT09AICBgQFee+21h92/z51TdLpEREREraIiFOD/PK2wf5Evdi0YjuDeZjLjxBLgr6s5GP9LNGasP4/I2/ngcDgiIiL5FD6FpKDg4egxTU3NRu8nJCSgtLQUAoEA48aNa3TMw8MDwMN9M0RERETdgUAggI+LCXxcTHDzXinCIkU4mJCNOhmTS87evY+zd+/D3UofrwQ64/88raDGySVERESNKPxvxoa+Fw2FjAaRkZEAHva+cHR0bHSsYTVGfT27dxMREVH308dSH9+/OBCR7wbjZT8n6KjLnlxyI6cUb+5OQNC34dgYnYLy6joFZ0pERNR1KbyA0VCcuHDhQqP3Dx06JLc3RmFhIQDAzEz2EkwiIiKi7sDaUAsfPtMX594PwTtje8NUV0NmXFZxJT79KwkjVpzGd8dvoeBB09HzREREykbhBYzg4GBIJBL8/PPPuHHjBgDg4MGDCA8PBwD83//9X5NzEhMTAQBWVlYKy5OIiIiooxhoq2FRcC9EvxeMr57zhLOpjsy4kspa/Hz6LnxXnMa/911DakG5gjMlIiLqOhRewHj99dehrq6OvLw8eHh4wNTUFJMmTYJEIoGNjQ2ef/75JuccP34cAoEAQ4YMUXS6RERERB1GU00F07ztcWJxINbMHIxB9oYy46rrxNhxIR3B34XjXzvicCWjWKF5EhERdQUKL2C4urpi27Zt0NbWhkQiQWFhISQSCQwNDbFr1y5pj4wG9+7dw4kTJwAAo0ePVnS6RERERB1ORSjAUx6W2PvqCPz+ig9GuZvLjJNIgCPX7mHiqrOYGhaDM7fyOLmEiIiUhsKnkADACy+8gMDAQBw+fBj37t2DlZUVJkyYAGNj4yaxV69exfTp0wEAI0eOVHSqRERERAojEAgw1NEYQx2NcSe3DGGRIuxPyEJtfdMixXlRIc6LCtHbQg8LA5wxfoA11FU5uYSIiHougYRl+x4vMzMTdnZ2AB6OorW1te3kjIiIiKil7pVUYdPZFOy4kI4HzUwlsTLQxEt+TpjqbQ9djU75joqIiEiqIz6HsoChBFjAICIi6v5Kq2qx60I6NkSnIK9M/lQSPU1VzBrugLm+jjDX01RghkRERP/TYwsYubm5SExMlI5LNTY2hoeHBywsLDo5s56BBQwiIqKeo7quHgcSsrE2IhnJ+fKnkqirCPH8YBu87O8MFzNdBWZIRETUwwoYEokEYWFh+OWXX5CUlCQzpm/fvnj99dexYMECCAQCBWfYc7CAQURE1POIxRKcvpmHtZHJuJhaJDdOIADG9LVAaKALvOyNFJghEREpsx5TwCgqKsL48eMRExMDAHK7ZzcULUaMGIFDhw7B0NBQUSn2KCxgEBER9WxxaYVYGyHCiRu5aO4nO29HYywMcMbIPuYQCvnlEBERdZweUcCQSCQIDAxEdHQ0AMDExAQvvvgihg0bBktLS0gkEuTm5iI2Nha//fYbCgoKIBAI4Ofnh4iICEWm2mOwgEFERKQc7uY9wPooEfZezkJNvVhuXC9zXSwMcMbEgdbQUFVRYIZERKQsekQBY8eOHZg1axYEAgGmT5+O1atXQ09PT2bsgwcPsGjRImzbtg0CgQDbt2/HtGnTFJluj8ACBhERkXLJK63C5nOp2HY+DWVV8ieXWOhrYL6vE6YNs4e+ppoCMyQiop6uIz6HKnxY+M6dOwEAgYGB2LZtm9ziBQDo6upiy5YtCAwMhEQiwfbt2xWVJhEREVG3Za6viXef6oOYD0Lw4Th3WBnInkaSW1qNr47ehO9Xp/HV0RvILa1ScKZEREQtp/ACxuXLlyEQCPDaa6+1+JzXX38dABAfH99RaRERERH1OLoaqnjZ3xkR7wTjuxcGwM1C9jSSsuo6rI0Qwe/r03jn9yu4k1um4EyJiIgeT1XRN2wYlerk5NTicxpiG84lIiIiopZTVxXi+cG2eM7LBuG38rEmIhkXUpr+XFVbL8HvcZn4PS4To9zNERrogqGOxp2QMRERUVMKX4FhYGAAAMjOzm7xOQ2x+vr6HZITERERkTIQCAQI7mOOPaE+2PevEXjawxLyJtWfvJGHF9bE4LnVZ3Hs+j2IxQofXEdERNSIwgsYHh4eAIBNmza1+JyNGzc2OpeIiIiInswgeyP8OnMwTi8JwvRh9lBXlf1j4eX0YoRui8OoHyKwOzYdVbX1Cs6UiIjoIYUXMCZPngyJRIJ9+/Zh+fLlaG4IikQiwfLly7Fv3z4IBAK88MILCsyUiIiIqOdzMtXBl5M8cfa9kXh9ZC8YaMmeRiLKL8f7e6/B/5szWB1+FyWVtQrOlIiIlJ3Cx6jW1tZiwIABuHnzJgQCAfr27Yu5c+di2LBhsLCwgEAgwL1793DhwgVs2bIF169fh0Qigbu7O65cuQJVVYW37ej2OEaViIiIWqq8ug57LmZgQ3QKsoor5cbpqKtgmrc95vs5wdpQS4EZEhFRd9ARn0MVXsAAgLS0NIwcORIpKSkQyNt4+f9JJBI4Ozvj1KlTcHBwUFCGPQsLGERERNRatfViHL6agzURybh5T/5UElWhABMGWiM0wAW9LfUUmCEREXVlHfE5VOFbSADAwcEBV69exZIlS2BgYACJRCLzl4GBAZYuXYqEhAQWL4iIiIgUSE1FiGcH2eDom/7YOt8bvr1MZMbViSXYezkLY/8biXmbYnFedL/ZLcJERERt1SkrMB5VU1ODuLg4JCYmSsekGhsbw8PDA4MHD4a6ujpEIhGio6MBALNnz+7MdLslrsAgIiKi9nAtswRrI5Nx5FoOmhtKMsDOEK8EOGNMP0uoCJtfbUtERD1Tj9lC0lpbtmzBvHnzIBQKUVdX19npdDssYBAREVF7Sr9fgfXRIvx2KQNVtWK5cY4m2lgQ4IznvWyhqaaiwAyJiKiz9ZgtJG3VDWotRERERD2evYk2Pp3ogbPvjcSbIa4w0pY9uST1fgX+vS8Rfl+fxi+n76C4okbBmRIRUU/SrQoYRERERNR1mOhq4O3Rbjj7/kj8Z0I/2BrJnkZS8KAGK4/fxogVp/GfQ9eRWVSh4EyJiKgnYAGDiIiIiJ6Itroq5oxwRPjSIPw0bRD6WevLjKuoqcems6kI/DYcb+2OR1J2qYIzJSKi7ky1sxMgIiIiop5BVUWICQOsMb6/Fc4l38eaiGRE3SloElcvlmB/Qjb2J2QjwM0MrwQ4w8fFBAIBG34SEZF8LGAQERERUbsSCATw7WUK316muJ5dgrBIEf66moN6GaNLIm/nI/J2Pjxs9BEa4IKnPSyhqsJFwkRE1BT/diAiIiKiDtPP2gA/Th2E8KVBmDvCEVpyppEkZpXi9V3xCP4uHFtjUlFZU6/gTImIqKtjAYOIiIiIOpydsTaWT+iHc++PxOLRbjDWUZcZl1FYiY8PXIfv16fx35O3UVjOySVERPQQCxhEREREpDBGOup4I8QV594fic+e9YCDibbMuMLyGvz35B2MWHEKnxxIREYhJ5cQESm7Du2B8emnn7bLdRISEtrlOkRERETUNWiqqWDWcAdM97bHsev3sCYiGVczS5rEVdWKsSUmDdvOp2Fcf2uEBjjDw8agEzImIqLOJpBIJE27KbUToVDYbt2kJRIJBAIB6uu5H7K1MjMzYWdnBwDIyMiAra1tJ2dERERE1JhEIsF5USHWRiYj/FZ+s7F+vUyxMMAZ/q6mnFxCRNRFdcTn0A6fQtKB9REiIiIi6iEEAgF8XEzg42KCGzmlWBcpwsEr2aiTMbkk+m4Bou8WwN1KH68EOmOcpxUnlxARKYEOXYERERHR7tcMDAxs92v2dFyBQURERN1RdnElNkanYFdsOsqbmUpiY6iFl/2dMGWoHbTVO/z7OSIiaoGO+BzaoQUM6hpYwCAiIqLurKSiFtsvpGHT2VQUPKiWG2eorYbZwx0we4QjTHU1FJghERH9EwsY1CYsYBAREVFPUFVbj33xWQiLFCGloFxunIaqEC8MscXLfs5wNNVRYIZERNSgIz6HcrMgEREREXULmmoqmOZtj5OLA7Fm5mAMtDOUGVddJ8b28+kY+V04Fu24jCsZxQrNk4iIOgY3CRIRERFRt6IiFOApD0uM7WeBi6lFWBuRjFM385rEiSXA4Ws5OHwtB8OdjREa6IIgNzNOLiEi6qZYwCAiIiKibkkgEMDbyRjeTsa4k1uGsEgR9idkoba+6Q7p86JCnBcVoo+lHhYGOGP8AGuocXIJEVG3wh4YSoA9MIiIiEhZ3CupwqazKdhxIR0PquvkxlkbaGK+nxOmettDV4Pf6RERtTc28aQ2YQGDiIiIlE1pVS12XkjHxugU5JXJn1yir6mKmcMdMNfXEeZ6mgrMkIioZ2MBg9qEBQwiIiJSVtV19TgQn421kclIzpc/uURdRYjnB9tggb8znM10FZghEVHPxCkkREREREStoKGqgheH2uHE24FYN3sIhjgYyYyrqRdjV2wGQr6PQOi2S7icXqTgTImI6HG44Y+IiIiIejyhUIDRfS0wuq8F4tIKsTZChONJuU3iJBLg2PVcHLueC29HY4QGOiO4tzmEQk4uISLqbCxgEBEREZFSGexgjLDZxrib9wDro0TYezkLNfXiJnGxqYWITS2Eq7kuFgY4Y+JAG6ircgEzEVFnYQ8MJcAeGERERETy5ZVWYdO5VGw/n4ayKvmTSyz0NTDf1wnThtlDX1NNgRkSEXU/bOJJbcICBhEREdHjlVXVYndsBjZEp+BeaZXcOD0NVUwfbo/5vk6w0OfkEiIiWVjAoDZhAYOIiIio5WrqxDh05eHkktu5D+TGqakIMGmQDRYGOKOXuZ4CMyQi6vo4hYSIiIiIqIOpqwrx/GBbHHsrAJvmDsUwJ2OZcbX1Evx2KROjvo/Ey1su4VJqoYIzJSJSLixgAEhPT8fSpUvh7u4OHR0dGBsbw9vbGytXrkRFRUWH3DMnJweGhoYQCAQQCAQICgrqkPsQERERUdsIBAIE9zHHnlAf7PvXCDztYQmBnGEkJ2/kYvKaGDz/6zkcv34PYjEXORMRtTel30Jy+PBhzJgxAyUlJTKP9+7dG0eOHIGzs3O73nfy5Mn4888/pb8PDAxEeHh4u96jAbeQEBEREbWPlIJyrIsS4Y+4TNTUNZ1c0sDZTAcL/Z0xycsGGqoqCsyQiKhr4BaSdnblyhW8+OKLKCkpga6uLr744gucO3cOp06dwoIFCwAAt27dwrhx4/Dggfz9j6116NAh/PnnnzA3N2+3axIRERFRx3My1cGXkzxx9r2ReC24F/Q1VWXGifLL8f7ea/D7+gxWh99FSWWtgjMlIup5lLqA8dZbb6GiogKqqqo4fvw4li1bBh8fH4wcORJhYWH45ptvAAA3b97E999/3y73fPDgARYtWgQAWLlyZbtck4iIiIgUy0xPA0vH9kbMByH4+Jm+sDHUkhmXX1aNb/6+hRFfncIXh5OQU1Kp4EyJiHoOpS1gXLx4Ubpl46WXXoKPj0+TmCVLlsDd3R0A8N///he1tU9eOV+2bBkyMjIQHByMWbNmPfH1iIiIiKjz6GioYr6fE8LfCcJ/pwxEH0vZ00jKa+qxLioF/l+fwZLfruDWvTIFZ0pE1P0pbQFj//790tfz5s2TGSMUCjF79mwAQFFR0RP3qIiNjcWqVaugrq6OX3/99YmuRURERERdh5qKEM8OssHRN/2xZb43RriYyIyrE0vw5+VMjP1vJOZvvogLovtQ8pZ0REQtprQFjKioKACAjo4OBg8eLDcuMDBQ+jo6OrrN96urq8PChQshFovx3nvvoXfv3m2+FhERERF1TQKBAIFuZti5YDgOvuaLcf2tIJQzueT0zTxMCTuPSavP4ei1HNRzcgkRUbNkdx1SAjdu3AAA9OrVC6qq8v8Y+vTp0+Sctli5ciWuXLkCFxcXLFu2rM3XkSUzM7PZ4zk5Oe16PyIiIiJ6vP62hlg13Qtp98uxPioFv13KQLWMySUJGcV4dcdlOJpoY0GAM573soWmGieXEBH9k1IWMKqqqlBQUAAAjx3lYmRkBB0dHZSXlyMjI6NN9xOJRPj0008BAKtXr4ampmabriNPw2gaIiIiIup6HEx08NmzHnhrlCu2xqRhS0wqiiua9lZLvV+Bf+9LxA8nbmPuCEfMHO4AQ231TsiYiKhrUsotJGVl/2uapKur+9h4HR0dAGjzKNXQ0FBUVlZiypQpGDNmTJuuQURERETdm4muBt4e7YZz74/Efyb0g62R7MklBQ9qsPL4bYxYcRqfHkpCVjEnlxARAUq8AqOBuvrjq9oaGhoAgMrK1v/lsXXrVpw8eRL6+vr44YcfWn1+SzxuZUhOTg68vb075N5ERERE1Dra6qqYM8IRM4bZ40jiPayNSMb17NImcRU19dh4NgVbYlIxvr8VFga4oK+1fidkTETUNShlAePRLRw1NTWPja+urgYAaGnJrpLLU1BQgCVLlgAAvvjiC1hZWbXq/JZ63DYYIiIiIup6VFWEmDDAGuP7W+Hs3ftYG5mMqDsFTeLqxRLsT8jG/oRsBLiZ4ZUAZ/i4mEAgkNMdlIioh1LKAoae3v/mc7dkW0h5eTmAlm03edTixYtRUFCAIUOG4F//+lfrkiQiIiIipSAQCODnago/V1MkZpUgLFKEw3KmkkTezkfk7Xx42hggNNAZT/WzhKqKUu4KJyIlpJQFDE1NTZiamqKgoOCxEzyKioqkBYzWNMvMzs7Gtm3bAAAjR47Eb7/91mx8Xl4edu/eDQBwcnLCsGHDWnwvIiIiIuoZPGwM8NO0QXhnbG9siE7BnosZqKytbxJ3LasEr+2Mh72xNhb4O2HyYDtoqXNyCRH1bEpZwAAAd3d3REVF4e7du6irq5M7SvXmzZuNzmmpR7emfPPNN4+Nv3HjBqZNmwYAmDNnDgsYRERERErMzlgbyyf0w5shrth2Pg2bz6WisLzp1uf0wgp8dOA6fjh5B3N8HDHbxwFGOpxcQkQ9k9KuN/Pz8wPwcHtIXFyc3LiIiAjpa19f3w7Pi4iIiIiogZGOOt4IccXZ90bis2c9YG+sLTOusLwGP5x8OLnkkwOJyCisUHCmREQdT2kLGM8++6z09aZNm2TGiMVibN26FQBgaGiI4ODgFl/f0dEREonksb8aBAYGSt/bvHlzm56JiIiIiHomLXUVzBrugDNLg7Bquhf62xrIjKusrceWmDQEfnsGr++KR2JWiYIzJSLqOEpbwPD29oa/vz8AYMOGDYiJiWkS89133+HGjRsAgDfffBNqamqNjm/evBkCgQACgQDLly/v8JyJiIiISLmpCAUY198KBxb5YteC4QjqbSYzTiwBDl3JxjM/R2Pm+guIupPf6MszIqLuSGl7YADAjz/+CF9fX1RWVmLMmDFYtmwZgoODUVlZid27dyMsLAwA4ObmJh2HSkRERETU2QQCAXxcTODjYoIbOaVYFynCwSvZqJMxuST6bgGi7xagr5U+QgOdMc7TipNLiKhbUuoCxqBBg7Bnzx7MnDkTpaWlWLZsWZMYNzc3HD58uNHoVSIiIiKirsLdSh/fTxmIJWN7Y2N0CnbFpqOipunkkqScUry5OwHf/H0LC/yd8OJQO2irK/XHASLqZpS+9Dp+/HhcvXoVb7/9Ntzc3KCtrQ1DQ0MMGTIEX3/9NeLj49GrV6/OTpOIiIiIqFk2hlr46Jm+iHk/BO+M7Q1TXdnTSLKKK7H8UBJGrDiN70/cxv0H1QrOlIiobQQSbobr8TIzM2FnZwcAyMjIgK2tbSdnREREREQdraq2HnsvZ2FdlAgpBeVy4zRUhXhhiC0W+DvDwURHgRkSUU/WEZ9DlX4FBhERERFRT6SppoLpw+xxcnEg1sz0wkA7Q5lx1XVibD+fjuCV4Vi04zKuZBQrNE8iopbipjciIiIioh5MRSjAUx5WGNvPEhdTi7A2IhmnbuY1iRNLgMPXcnD4Wg58nE0QGuiMQDczCASCTsiaiKgpFjCIiIiIiJSAQCCAt5MxvJ2McTu3DGGRIhxIyEJtfdMd5TGi+4gR3UcfSz0sDHDG+AHWUOPkEiLqZOyBoQTYA4OIiIiIZMkpqcSms6nYeSEdD6rr5MZZG2hivp8TpnrbQ1eD34ES0eN1xOdQFjCUAAsYRERERNSc0qpa7Difjo1nU5BfJn8qib6mKmb5OGDuCCeY6WkoMEMi6m5YwKA2YQGDiIiIiFqiuq4e++OzEBYpQnK+/Mkl6qpCPO9liwX+TnA201VghkTUXXAKCRERERERdRgNVRVMGWqPE28HYt3sIRjiYCQzrqZOjF2x6Qj5PgKvbItDfHqRgjMlImXEDWxERERERNSIUCjA6L4WGN3XApdSC7E2UoQTSblN4iQS4O/r9/D39XvwdjJGaIAzgnubQyjk5BIian8sYBARERERkVxDHI0xxNEYd/MeYF2kCPvis1BTL24SF5tSiNiUQria62JhgDMmDrSBuioXfBNR+2EPDCXAHhhERERE1F7ySquw6Vwqtp9PQ1mV/MkllvqamO/niGne9tDTVFNghkTUFbCJJ7UJCxhERERE1N7KqmqxOzYDG6JTcK+0Sm6cnoYqZgx3wDxfR1joayowQyLqTCxgUJuwgEFEREREHaWmToyDV7KxNiIZd/IeyI1TVxFi0iAbLAhwRi9zTi4h6uk4hYSIiIiIiLoUdVUhJg+2xbG3ArBx7hB4OxnLjKupF2PPpQyM+j4CC7ZewqXUQgVnSkTdHZt4EhERERHRExMKBRjZxwIj+1jgcnoRwiJEOJZ0D7LWe59IysWJpFwMdjBCaIAzRrlbcHIJET0WV2AQEREREVG78rI3wppZg3FqcSCmD7OXO40kLq0IC7fFYdQPEdhzMR3VdfUKzpSIuhP2wFAC7IFBRERERJ0pv6waW86lYmtMKkqbmVxipqeB+b5OmD7MHgZanFxC1J2xiSe1CQsYRERERNQVPKiuw56LGdgQJUJ2ifzJJboaqpjmbYf5fk6wMtBSYIZE1F5YwKA2YQGDiIiIiLqS2nox/rqajbURIty8VyY3TlUowMSBNlgY4IzelnoKzJCInlRHfA5lE08iIiIiIlIoNRUhJg2yxbMDbRB5pwBrwpMRI7rfJK5OLMGflzPx5+VMjOxjjtAAZ3g7GUMgYMNPImXEAgYREREREXUKgUCAQDczBLqZ4WpmMdZGinD0Wg7EMtaIn76Zh9M38zDQzhCvBDpjdF9LqHByCZFSYQGDiIiIiIg6XX9bQ6ya7oW0++VYH5WC3y5loLpO3CQuIaMYr2y/DCdTHSzwd8ZzXjbQVFPphIyJSNHYA0MJsAcGEREREXU39x9UY0tMGrbGpKK4olZunKmuOuaOcMTM4Q4w1FZXYIZE1Bw28aQ2YQGDiIiIiLqripo6/HYxA+uiUpBVXCk3TltdBVOH2uMlfyfYGHJyCVFnYwGD2oQFDCIiIiLq7urqxTiSeA9rI5JxPbtUbpyKUIAJA6yxMMAZ7lb6CsyQiB7FKSRERERERKSUVFWEmDDAGuP7W+Hs3ftYG5mMqDsFTeLqxRLsi8/CvvgsBLqZITTQGT7OJpxcQtQDsIBBRERERETdhkAggJ+rKfxcTZGYVYKwSBH+upotc3JJxO18RNzOR39bA4QGuOApD04uIerOuIVECXALCRERERH1ZBmFFdgQnYLdF9NRVdt0ckkDe2NtLPB3wuTBdtBS5+QSoo7EHhjUJixgEBEREZEyKCyvwbaYNGyJSUVheY3cOGMddczxccRsHwcY6XByCVFHYAGD2oQFDCIiIiJSJpU19fjjcibWRYqQXlghN05LTQVThtrhJT8n2BlrKzBDop6vIz6HCp/4CkRERERERF2IlroKZg13wJmlQVg13QueNgYy4ypr67H5XCqCVobjjV3xSMwqUXCmRNQabOJJREREREQ9kopQgHH9rfB/npaIEd3H2ggRIm7nN4mrF0tw8Eo2Dl7Jhl8vU4QGOsOvlyknlxB1MSxgEBERERFRjyYQCDDCxRQjXExxI6cUYZEiHLqSjToZo0ui7xYg+m4B+lrpIzTQGeM8raCqwoXrRF0Be2AoAfbAICIiIiJqLKu4EhuiHk4uqaiplxtna6SFl/2c8OJQO2ir8/tfopZiE09qExYwiIiIiIhkK66owfbzadh8LhUFD+RPLjHUVsNsH0fM8XGAia6GAjMk6p5YwKA2YQGDiIiIiKh5VbX12Hs5C+uiREgpKJcbp6EqxItD7PCyvxMcTHQUmCFR98IpJERERERERB1AU00F04fZ4+TiQKyZ6YUBdoYy46rrxNh2Pg3BK8OxaMdlXM0sVmieRMqMm7iIiIiIiIj+PxWhAE95WGFsP0vEphRibaQIp2/mNYkTS4DD13Jw+FoOfJxNEBrojEA3M04uIepALGAQERERERH9g0AgwDBnEwxzNsHt3DKERYpwICELtfVNd+DHiO4jRnQffSz1EBrojGf6W0ONk0uI2h17YCgB9sAgIiIiInpyOSWV2HQ2FTsvpONBdZ3cOGsDTbzk74ypQ+2go8HvjEk5sYkntQkLGERERERE7aekshY7L6Rj49kU5JdVy43T11R9OLlkhCPM9Di5hJQLCxjUJixgEBERERG1v+q6euyPz8LaSBFE+fInl6irCvG8ly0W+DvB2UxXgRkSdR5OISEiIiIiIuoiNFRVMGWoPU6+HYiwWYMx2MFIZlxNnRi7YtMR8n0EXtkWh/j0IgVnStQzcEMWERERERHRExAKBRjTzxJj+lniUurDySUnknKbxEkkwN/X7+Hv6/fg7WSMVwKdEeRmDqGQk0uIWoIFDCIiIiIionYyxNEYQxyNcTevDOsiU7AvPgs19eImcbEphYhNKYSbhS4WBrhgwgBrqKtygTxRc9gDQwmwBwYRERERUefILa3CprOp2HE+DWXNTC6x1NfEfD9HTPO2h56mmgIzJOoYbOJJbcICBhERERFR5yqrqsWu2HRsjE7FvdIquXF6GqqYMdwB830dYa6vqcAMidoXCxjUJixgEBERERF1DTV1Yhy8ko21Ecm4k/dAbpy6ihCTBtlgQYAzeplzcgl1P5xCQkRERERE1I2pqwoxebAtjr0VgA1zhsDbyVhmXE29GHsuZWDU9xFYsPUS4tIKFZwpUdfDJp5EREREREQKJhQKEOJugRB3C1xOL0JYhAjHku5B1vr4E0m5OJGUiyEORggNdEFIH04uIeXEAgYREREREVEn8rI3wppZgyHKf4B1USn483ImauqaTi65lFaES1svwcVMBwsDnPHsIBtoqKp0QsZEnYM9MJQAe2AQEREREXUfeWVV2HIuFdti0lBaJX9yibmeBub5OmH6MHsYaHFyCXUtbOJJbcICBhERERFR9/Ogug57LmZgQ5QI2SXyJ5foaqhi+jB7zPN1hJWBlgIzJJKPBQxqExYwiIiIiIi6r9p6Mf66mo21ESLcvFcmN05NRYCJA22wMMAZbhZ6CsyQqKmO+BzKHhhERERERERdmJqKEJMG2eLZgTaIuJ2PtREixIjuN4mrrZfgj7hM/BGXiZF9zPFKoAuGOhpBIGDDT+oZWMAgIiIiIiLqBgQCAYJ6myOotzmuZBQjLFKEo4k5EMtYU3/6Zh5O38zDIHtDhAY4Y3RfS6hwcgl1cyxgEBERERERdTMD7AyxaoYXUgvKsT5ahN8vZaJaxuSS+PRivLL9MpxMdbDA3xnPedlAU42TS6h7Yg8MJcAeGEREREREPVvBg2psjUnD1phUFFfUyo0z1dXAPF9HzBzmAANtTi6hjsMmntQmLGAQERERESmHipo6/HYxA+uiUpBVXCk3TltdBdO87THfzwk2hpxcQu2PBQxqExYwiIiIiIiUS129GIev5WBthAhJOaVy41SFAowfYI2FAc5wt9JXYIbU03EKCRERERERET2WqooQEwfaYMIAa0TfLcDaCBGi7xY0iasTS7AvPgv74rMQ6GaG0EBn+DibcHIJdUksYBAREREREfVQAoEA/q5m8Hc1Q2JWCdZGinD4arbMySURt/MRcTsf/W0NEBrggqc8OLmEuhZhZydAREREREREHc/DxgA/TxuEiHeCMXeEIzTVZH8cvJpZgkU7L2Pkd+HYdj4NVbX1Cs6USDb2wFAC7IFBRERERET/VFheg20xadgSk4rC8hq5cSY66pgzwhGzhjvASEddgRlSd8YmntQmLGAQEREREZE8lTX1+CPu4eSS9MIKuXFaaiqYMtQOL/k5wc5YW4EZUnfEAga1CQsYRERERET0OHX1Yvx9/R7WRohwLatEbpyKUIBxnlZYGOAMDxsDBWZI3QmnkFDPlhYDaBsDOmaApiEgZIsWIiIiIiJFUVUR4pn+1hjnaYUY0X2sjRAh4nZ+k7h6sQQHr2Tj4JVs+LuaIjTABb69OLmEOh4LGNQ11NcCm5763++FqoC26cNihq7Zw//VMQN0TAEd80de///31TQ7L3ciIiIioh5EIBBghIspRriYIim7FOuiRDh4JRv1MkaXRN0pQNSdAvSz1sfCAGeM87SCqgq/iKSOwS0kANLT0/HTTz/h8OHDSE9Ph4aGBnr16oUXX3wR//rXv6Ct3fb9XaWlpThy5AhOnTqFuLg4iEQiVFRUwMDAAP369cMzzzyDl19+GYaGhu33QP/QLbaQlN0Dvuvd9vM19BsXNB79pfuP33N1BxERERFRq2QWVWBjdCp2X0xHRY38qSS2RlpY4O+MF4bYQlud35crM/bA6ACHDx/GjBkzUFIie49X7969ceTIETg7O7f62kePHsWkSZNQXV3dbJyFhQV27dqF4ODgVt+jJbpFAePeNWCNn2Lu9ejqDh1TQPefKzrMubrj/7V35/FRVff/x9/ZV7JAElkChIABFFR22UEKVJBWxd0iS5V8VezDar8ofOv+sK79UdsqmwpCxaAWF0BcAAlBQEAFpYIKYYssSQhbSEIyyf39EWc6SWbJJJPJzeT1fDzuw0nmzD33+jgkmc+8zzkAAACAA6eLSvWvrYe0ePNB5Rc637kkPjJEkwamaPLAjmoVHebDK4RZUMDwsl27dmnQoEEqKipSdHS0Zs2apZEjR6q4uFgZGRlauHChJKlbt27avn27oqOjPTr/v/71L02aNEmBgYEaPXq0fv3rX+vyyy9XXFyccnJy9Oabb2r58uWSpMjISH3xxRe64oorvH2bTaOAcWiL9NbNUonzxYIajat0R5UCCOkOAAAANA8lZeX699c5WrgxWwdPOt+5JCw4UDf1ba87h3ZSx1ZRPrxCNDYKGF42cuRIbdiwQcHBwdq4caMGDhxY5fkXXnhBM2fOlCQ98cQTevTRRz06//Lly/X5559r9uzZ6tChg8M2//jHP/SHP/xBknTVVVdp3bp1dbgT15pEAcPKckE6ny+dz/vlv7m/PP7l68Jcu+fypIqyxr7iqqqnO6ISfylwOCmAkO4AAABAE1ZeYeiz749rbma2dh057bRdYIB0dc82Sh+WqsuS43x2fWg8FDC8aPv27erfv78kKT09XfPmzavRpqKiQj169NCePXsUHx+vEydOKCQkxOvX0q9fP+3YsUOBgYHKzc1Vq1atvHr+JlXA8IRhSCWn7Qoeeb8UOOy+tj/MmO4IbeF+kVJrwoN0BwAAAEzKMAxtO1Cg+RuztX5vrsu2gzq3Uvrwzhp2cQI7l/gxtlH1ovfff9/2eOrUqQ7bBAYG6o477tCsWbN06tQpbdiwQaNHj/b6tYwYMUI7duxQRUWFDhw44PUCht8KCJAi4iuPhIvdt6+R7sizS3jk1yyA+CLdUXpOKjgnFWS7bxsQZFfgsEt01CiAWNMdEQ1//QAAAIAqdy4ZkNpKA1Jb6Yfj57RgY7Y+2PmzLA52Ltm8/6Q27z+pbq1bKH14qq65rK1C2LkEtdBsCxhZWVmSpKioKPXp08dpu+HDh9seb9q0qUEKGPaLfAbyCXvDCQ6TYttVHu4YRmVio3qKozCv6pQWawHEF+kOo1wqPFF51EZoi9pNY4lKrCwCMfYAAADgBV1bt9Bfb7pcfxqbptc3HdBb246o8IKlRru9x8/pj8t36YWPf9Dvh6bqln7tFRXWbN+iohaa7ejYs2ePJKlLly4KDnb+v6Fbt241XuNtmZmZkqTg4GB16dLF49fn5OS4fP7YsWN1uq5mLSBAioirPGqV7iiVivIdTGGp9rW1AOKrdEfpOenUAfdtbekOV9NYSHcAAACg9trERuj/xl+iGVddrGVfHtbrXxxQ3rmaOzQePVOip1Z9r7+v+0mTruyoyYNSlNiCnUtQU7MsYJSUlCg/P1+S3M7DiY+PV1RUlM6fP68jR454/VpWr16tb7/9VpI0duxYxcTEeHwO67wiNKLgUCmmbeXhji3dUYtFSs/nmj/d4XAaSxLpDgAAAEiSYiNCdPeIzpo2JEXvf/Oz5m/MVnbe+RrtzhSX6Z+f79OCrGzd0CdZdw1NVacEdi7BfzXLAsa5c+dsj2uzNaq1gFFYWOjV6ygoKNC9994rSQoKCtJTTz3l1fPDpKqkO2qRuLGmO9xNY7EWQEyf7kh0vY4H6Q4AAAC/FBYcpJv7ddCNfdpr7Z4Tmpe5X18fPl2jXamlQsu+PKy3th3Wry9trenDUtWrQ7zvLxim0ywLGCUlJbbHoaGhbtuHhVXGl4qLi712DeXl5br99tt16NAhSdKf//xn9erVq07ncpcMOXbsmG3HFTRBdU53uJnGcj6vcheXhuZxuiO62hodCf/diaVKAYR0BwAAQFMUGBigMZe21phLW2vHwQLNy8zW2j01/1Y0DGnN7uNas/u4+ndqqf8ZnqoRaUkKDGTnkuaqWRYwwsPDbY9LS0vdtrcushkR4b1Phu+55x59/PHHkqTx48frkUceqfO5/GZbVNRffdMdzqaxWB+Xu//3Um+lhZVHbdMdka3cL1IaTboDAADAjPqmtNSrKS21L7dy55L3vvlZZeU1dy7ZdqBA2w4UKO2iaE0f1lm/ubytQoP5IKu5aZYFjBYtWtge12ZayPnzlfOzajPdpDZmzZqlBQsWSJKGDBmid955R0FBQV45N+ART9MdF85Wm8JS/bArgPgq3XE+t/KojdBox4uU1ljHI1GKaEm6AwAAwEe6JLXQ8zdcrgfHdNWiLw7qza2HdM7BziU/nijUn97ZpRc/+UG/H9JJt/RvrxbhIY1wxWgMzbKAER4eroSEBOXn57vdwePUqVO2AoY3Fst87rnn9Oyzz0qSevfurVWrVnk12QE0mIAAKTy28qh1uuNk1TU6zuc52KklrxHSHQfdtw0IlCIT3E9jsX4dGtnglw8AAODvLooJ18NXd9O9IzvrrW2H9dqmAzpxtubOJcfPlujpj/bo7+t/0u+u7Kipg1KUFBPu4IzwJ82ygCFJ3bt3V1ZWlvbt2yeLxeJ0K9W9e/dWeU19vPLKK3r44Ydt5/rkk08UGxtbr3MCphUcKsW0qTzcsaY7qkxhcbBIqbUA4pN0R0Ud0x2OprHYF0Csa3eQugIAAHCmRXiIpg/rrCmDOumDnT9rwcZs/ZRbMz1/rsSiuRv267WsA7quVztNH56qzoneSc7DfJptAWPIkCHKysrS+fPn9dVXX2nAgAEO22VmZtoeDx48uM79LV26VDNmzJAkpaamau3atUpISKjz+QC/Yp/uaNXZfXtbusPNIqXWAojZ0x1RiQ7W8Ugi3QEAAJq90OBA3di3vSb2TtbnP+Rqfma2th0sqNGutLxCy3cc0dtfHdGvul+k/xmeqj4dWzbCFaMhBRiGUXOFlGZg27ZttqJFenq65s2bV6NNRUWFevTooT179iguLk65ubkKCfF8ftWKFSt00003qby8XMnJycrKylJKSkp9b6HWcnJybNNfjhw5wqKfaF7s0x22IoeTaSzn86TiU419xTWFRFVbo8PBOh7RSaQ7AABAs/DVoVNasHG/Pv3+hFy9m+3bMV7pwztrVDd2LmkMDfE+tNkWMCRp2LBhysrKUnBwsDZu3KiBAwdWef6FF17QzJkzJUmPPfaYHn/88SrPL168WFOnTnX6vCR9+umnmjBhgkpLS5WUlKSNGzeqa9euDXI/zlDAADxQXlatuOFkGosv0x2eCAis3Jklqlqio/oipdaDdAcAAGiisvMKtTDrgP79dY5KLRVO23VOjFL6sM76ba+2Cgvmgx5faYj3oc12CokkvfTSSxo8eLCKi4s1ZswYzZ49WyNHjlRxcbEyMjJsO4WkpaXpwQcf9Pj8W7du1XXXXafS0lKFhIRozpw5Kisr0+7du52+Jjk5WXFxcXW9JQD1FRTi4dod52qmOBxNY/FVusOo+G/ftRES5WKRUvsCSBLpDgAAYCqpidF65vqe+uPoi/XG5oNauuWQzpbU3Llkf955zfz3t3rx0x80bUgn3Tagg2LYuaRJatYFjF69emn58uX63e9+p7Nnz2r27Nk12qSlpWn16tVVtl6trY8//lhFRUWSpLKyMt1+++1uX7No0SJNmTLF474ANIKAACk8pvKozdod5WWVa3cUVkt0OFvHo7zmitteV3ZeOn1eOn3IfVtbusPBIqX201hsa3dENfz1AwCAZi+pRbj+d2w33T2iizJ+2bnk2JmSGu1yz13Qs2v26p/r9+m2AR00bXAntY5l55KmpFkXMCRpwoQJ+vbbb/XSSy9p9erVysnJUWhoqLp06aIbb7xRM2bMUGQkEWsAXhAUIrVoXXm44yzdUWWnFrsCiJnTHe6msUQlSpEtSXcAAIB6iQ4L1p1DUzV5UIpW7jqqBRuztff4uRrtCi9YtGBjthZ9cUC/vaKdpg9LVdpFnn9gDd9r1mtgNBesgQE0A9Z0h7tFSq0FEF+kOzxRJd3hYhoL6Q4AAFBLhmEo88c8zcvcr63ZNXcusTeqW5LSh3dWv5R4BQSw4Kc3sAYGAMCxOqc7XC1Sat2ZxfUvfK/wON0RWXMai20KS7WdWkh3AADQLAUEBGhE1ySN6JqkXUdOa8HGbK3ZfUwVDj7CX7c3V+v25qpXhzilD+us0ZdcpCB2LjEdEhjNAAkMAPVin+5wNY3FrOkOBVSmO6KrJTqqH9YpLqQ7AADwWwfzz+vVTdl6Z0eOLrjYuSQ1IUp3DUvVdb3aKTyED0Lqgm1UUScUMAD4jGFIpYW1m8biq3SHp0IiqyY4nC1SGpVEugMAgCYqv/CClmw+qDe2HNKZ4jKn7RKiwzR1cIp+N6CjYiPZucQTFDBQJxQwAJhWueWXdIebaSzW75k13VFjkVL7AojdFBfSHQAAmMr5Cxa9veOIXs06oJ9PFzttFxkapFv7d9C0IZ3ULi7Ch1fYdFHAQJ1QwADgF6zpDvutZl2t42HqdIebaSxRiZWFEdIdAAD4hKW8Qqu/O6b5mdn6/thZp+2CAwP0m8vbavrwVHVrHePDK2x6KGCgTihgAGiWbOkOuzU6bGkOBwUQS8394htXQM2dWWqs41FtZxZWTQcAoF4Mw9Cmffman5mtTfvyXbYd0TVR04elamBqK3YucYBdSAAAqK2gYKnFRZWHO/bpDleLlFqLHkUFkhq6/m9IRfmVR202ZwmOqN00lqhEKaJl5f8fAABQRUBAgIZenKihFydq989nNH9jtlZ/e9ThziUbfsjThh/ydHlyrNKHd9bYS1uzc0kDI4HRDJDAAAAvq5LucHAUNoV0R8uqCY4a63jYHaQ7AADN2JGCIr2ala3lO46opMz5ziUdW0XqzqGpurFPMjuXiCkkqCMKGADQiAxDKj3vZhpLno/THR4KjqhW4LDbiaX6FBfSHQAAP1VwvlRLthzUG5sP6lSR851LWkWFavKgFE26sqPio0J9eIXmQgEDdUIBAwCakHJL5QKkDqex2C9UmmfydIebRUqtU1xIdwAAmpji0nK9+9URLcjK1pEC5zuXRIQE6eZ+7fX7IZ3UvmWkD6/QHChgoE4oYACAn7KlO1xNY7GbymLmdIdtjQ4ni5Rad2Yh3QEAMAlLeYU+/s9xzc/M1nc/n3HaLigwQNdc1kbTh6Xq0raxPrzCxkUBA3VCAQMAIOm/6Q7rVrP2U1ocTXGxOP9UqXFUT3ckOJ7GYtuZJZp0BwCgwRmGoS37T2rexmxt/NH1yttDL05Q+rDOGtzF/3cuoYCBOqGAAQDwWJV0R77dFBYnO7UUnZSp0x2uprGQ7gAAeMn3R89qYVa2Ptx1VOWOti75xaVtY5Q+vLPG9Wit4KBAH16h71DAQJ1QwAAANDj7dIerRUrNnO6IiP/vVrNVFip1UAAh3QEAcCHnVJFe33RQGdsPq6i03Gm75PgI3TU0VTf2TVZkqH8V0ilgoE4oYAAATOdCoetFSu2LIKZMd4Q7WKg0wXEBhHQHADRbp4tK9a+th7Toi4M6eb7Uabv4yBDdMTBFdwzsqFbRYT68woZDAQN1QgEDANCkVZRXLkBaYycWBzu1mDLdocrtZauv0eGsAEK6AwD8TklZuf79dY4WbszWwZNFTtuFhwTqpr7tdeeQVHVo1bR3LqGAgTqhgAEAaFasa3e4msZiXcfD1OkON9NYbGt3hDT2FQMAaqm8wtCn/zmueZn7tSvH+c4lgQHS1T3b6H+GdVbP5Ka5cwkFDNQJBQwAAJywpTsc7cRSfaeWPKnM+admjSaipftpLNbHYS1IdwCACRiGoS8PFGh+5n59/oPrnUsGdW6l9OGdNezihCa1c0lDvA9lQiYAAGi+AoMqEw3RiZIucd/efmeWwmo7s1Sf4lJ0UjIqGvwWVFxQeeT/4L5tjXSHi3U8SHcAQIMJCAjQlamtdGVqK/1w/JwWbMzWBzt/lsXBziWb95/U5v0n1a11C/3P8M4af1kbhfjpziXukMBoBkhgAADQCKqkO5zsxmJfACk739hXXFNE/H+3mrUWPZyt40G6AwDq5ejpYi364oCWfXlY513sXNIuLkLThnTSLf3aKyrMvJkEppCgTihgAADQBNinO5wtUurrdIcngsJqrtERleC4AEK6AwCcOlNcpje/PKTXNx1UfuEFp+1iI0I06cqOmnFVF4WHBPnwCmuHKSQAAAD+KjSq8ohPcd/WXbqjShHER+mO8gvS2ZzKozYi4t0vUkq6A0AzFBsRontGdNG0wZ30/jc/a8HGbGXn1/w5fqa4TB/tPqYHRqc1wlU2DgoYAAAATU2VtTtqofR8zQVJHU1jOZ/rw7U7TlUe+T+6b2tNd7hbpNT6fdIdAPxAeEiQbunfQTf1ba/P9pzQ/Mz9+vrw6Spt0oelKjCw+RR4KWAAAAD4O1u6o6P7thXllYUFd9NYrAUQU6c7XExjsX4dFkO6A4CpBQYGaOylrTX20tbacbBA8zKztXbPCSW2CNO1vdo19uX5FAUMAAAA/Fdg0C9v9BOkpO7u29vSHfY7sVTfqcW6dke+udMdDqexVCuAkO4A0Ij6prTUqykttS/3nI6cKlZYsPnWvmhIFDAAAABQd3VNd9iv0eFsHY/Swoa/fk/THeFx7qexWHdqId0BoIF0SWqhLkktGvsyfI4CBgAAAHzDPt2h2qQ7ihxMYXGwSKkv0x0lpyuPWqU7QmsuSOpqHQ/SHQDgEgUMAAAAmFNopBTasZbpjgqpuNrOLI4WKfVpuqNUOvtz5VEb4XFVExw1ih92U1xIdwBohihgAAAAoOkLDKxnusPJNJbCXN+nO07+5L6tLd3hbJFSuwJIZIIUHNrQVw8ADY4CBgAAAJofj9Mdp9xPY7EeZk53uFyk9Jevw2NJdwAwJQoYAAAAgCuBgVJUq8pD3dy3Ly2qTG24m8Zi/a9R3uC3UPd0h4tpLKQ7APgYBQwAAADAm0IjpdAOUlwH922rpDusCQ+7REf1AkjpuYa/fo/THbHup7FYFyol3QGgHihgAAAAAI2lPukO6xodNXZq8XW640zlUZt0R2BI7aaxRCeR7gBQAwUMAAAAoKmoV7rDySKlvkx3VJRJ545WHrURHls1weFwHY9fDtIdgN+jgAEAAAD4I0/THWXF1dbrcLJIaaOkO/a5b2tNd1gTHFXW8aj+dSLpDqAJooABAAAAQAqJqEx21DbdUXLaLsHhbBrLL+t4mDrdUf2wL4BYd2aJI90BmAAFDAAAAACeCQyUIltWHrVOdzhJctRYxyPP3OkOV4uUWtsEhzX89QPNEAUMAAAAAA0rJEKKa195uGNNd9gSHM6msfxS9LhwtsEv3+N0R1is3TodCQ6KHHZFENIdQK1RwAAAAABgHvbpjsSu7ttXSXfk/zKFxcU6Hr5Id1w4U3nUOt2R4H4ai7UAQroDzRgFDAAAAABNV33SHdY1Oho93XGs8qiNsFi7Akf1wofdEZ1IugN+hwIGAAAAgObB43RHSe0WKW2MdEfBfvdtbekOa4Kj+joe1XdmId0Bc6OAAQAAAACOhITXL91RZZFS+ykuJk93WBMcztIdUQlSRDzpDvgcBQwAAAAAqK+6pDuK8t0sUmqX/KiwNPw9eJTuCK7dIqWkO+BFFDAAAAAAwNdCwqXY5MrDHVu6w9k0Fvuv8yuLEA2twlL3dIfTRUpJd8A1ChgAAAAAYGZV0h1p7ttb0x0up7GYPN0RmeB+Gkt0UmW7kPCGv36YAgUMAAAAAPAnnqQ7DEMqPuV6Gkthnu/THYXHK4/aCIuputWss2ksUb/szBIY2KCXj4ZDAQMAAAAAmquAAM/SHZYLNdfnKKyW6LCf4uKTdMfZyqMg231ba7qjxkKlDnZqiUok3WEyFDAAAAAAALUTHOZZuqPkdNWtZp1NYynM832640Qt2tvSHQ6msVSf4kK6o8FRwAAAAAAAeF9AQOWCnBHxHqQ7nExlqVEEMXm6o8pCpU52aiHd4TEKGAAAAACAxhccJsW2qzzcsaY7ajON5XyeVGLmtTtcLFJKuqMKChgAAAAAgKbFPt2RcLH79lXSHfZTWKrv1GJdu6Os4e+hPukOa2Gj4yCp+zUNf60mQQEDAAAAAODf6pPusCU8nExvacx0R/kFChgAAAAAADRL9U535KnmIqW5DZPuiEry3rmaAAoYAAAAAADUlcfpjjNuFim1m+LiLt0RleCde2giKGAAAAAAAOALAQFSRFzlUat0R6lUlO+gyPHLcdGlDX3FpkIBAwAAAAAAMwoOlWLaVh4Q+7AAAAAAAADTo4ABAAAAAABMjwIGAAAAAAAwPQoYAAAAAADA9ChgAAAAAAAA06OAAQAAAAAATI8CBgAAAAAAMD0KGAAAAAAAwPQoYAAAAAAAANOjgAEAAAAAAEyPAgYAAAAAADA9ChgAAAAAAMD0KGAAAAAAAADTo4ABAAAAAABMjwIGAAAAAAAwveDGvgA0PIvFYnt87NixRrwSAAAAAEBzYP/e0/49aX1QwGgG8vLybI/79+/fiFcCAAAAAGhu8vLylJKSUu/zMIUEAAAAAACYXoBhGEZjXwQaVklJib777jtJUmJiooKDzRm8OXbsmC0hsm3bNrVp06aRrwioijEKs2OMwuwYozA7xijMrimNUYvFYpsN0LNnT4WHh9f7nOZ8JwuvCg8PV79+/Rr7MjzSpk0bJScnN/ZlAE4xRmF2jFGYHWMUZscYhdk1hTHqjWkj9phCAgAAAAAATI8CBgAAAAAAMD0KGAAAAAAAwPQoYAAAAAAAANOjgAEAAAAAAEyPAgYAAAAAADA9ChgAAAAAAMD0AgzDMBr7IgAAAAAAAFwhgQEAAAAAAEyPAgYAAAAAADA9ChgAAAAAAMD0KGAAAAAAAADTo4ABAAAAAABMjwIGAAAAAAAwPQoYAAAAAADA9ChgAAAAAAAA06OAAQAAAAAATI8CBgAAAAAAMD0KGPC6w4cP609/+pO6d++uqKgotWzZUv3799eLL76ooqIir/WTkZGhsWPHqk2bNgoPD1dKSoomTZqkrVu3eq0P+KeGHKNnz55VRkaG7rrrLvXu3VtxcXEKDQ1VYmKiRowYoRdffFGnT5/2zo3Ab/nq56i9Y8eOKS4uTgEBAQoICNCIESMapB/4B1+O0bVr12rKlCnq0qWLoqKiFBsbq7S0NN1www2aO3euCgsLvdof/IMvxuj333+v++67Tz179lRMTIzt9/3IkSM1Z84cnTt3ziv9wH/k5uZq1apVevTRR3X11VcrISHB9nt3ypQpDdKn371nMgAvWrVqlREbG2tIcnh07drV2L9/f736KC4uNq655hqnfQQGBhpPPvmkl+4I/qYhx+hHH31khIWFOT239bjooouM9evXe/nO4C988XPUkYkTJ1bpZ/jw4V7vA/7BV2O0oKDA+O1vf+v2Z+o333xT/5uCX/HFGH3xxReN4OBgl2OzY8eOxq5du7x0V/AHrsbL5MmTvdqXv75nooABr9m5c6cRGRlpSDKio6ONp59+2ti8ebOxbt0646677rL9Y+nWrZtx7ty5Ovdz22232c41cuRI4/333ze2bdtmvPbaa0bnzp1tzy1cuNCLdwd/0NBjdOnSpbZfCGPHjjXmzJljrF+/3vj666+NDz/80Lj55pttfURGRvJHN2rw1c/R6j788ENDkpGUlEQBAy75aoyePn3a6NOnj+1848ePN5YuXWps2bLF2LRpk/Hmm28a999/v5GcnMzPUlThizG6fPly23lCQ0ONP/7xj8bq1auNL7/80li2bJkxZMgQ2/Nt2rQxTp8+7eW7RFNlX0Bo3769MWbMmAYrYPjreyYKGPCaESNGGJKM4OBgY/PmzTWef/75523/UJ544ok69bFhwwbbOSZMmGBYLJYqz+fl5RkdOnQwJBnx8fHGqVOn6tQP/FNDj9GMjAwjPT3dOHTokNM2f//73219XHXVVR73Af/mi5+j1Z07d85o3769IclYsmQJBQy45KsxOmnSJFs/GRkZTttVVFQYZWVlde4H/scXY7RHjx62c6xatcphm+uvv97W5q9//Wud+oH/efTRR42VK1cax48fNwzDMA4cONAgBQx/fs9EAQNesW3bNts/kvT0dIdtysvLje7du9v+oZSWlnrcz7hx4wxJRlBQkHHkyBGHbd566y3btbz44ose9wH/5KsxWht9+/a1JTXy8/MbpA80PY01Ru+77z7bpzOGYVDAgFO+GqNZWVm2fh5//PH6XjaaEV+M0TNnztj66N27t9N2u3btsrWbOHGiR32g+WioAoY/v2diEU94xfvvv297PHXqVIdtAgMDdccdd0iSTp06pQ0bNnjUR2FhodatWydJGj16tJKTkx22u/766xUTEyNJWrFihUd9wH/5YozWlnVxxIqKCh04cKBB+kDT0xhjdNu2bXr55ZcVGhqquXPn1utc8H++GqP//Oc/JUnR0dF68MEHPX49mi9fjNHS0lLb49TUVKftOnfubHt84cIFj/oA6sPf3zNRwIBXZGVlSZKioqLUp08fp+2GDx9ue7xp0yaP+ti2bZvtF4D9eaoLDQ3VlVdeaXtNWVmZR/3AP/lijNaW/R8ygYH8GEYlX49Ri8Wi6dOnq6KiQg899JC6du1a53OhefDFGC0tLdUHH3wgSbr66qsVHR0tqXK8Hjp0SIcPH67yBhKw54sxmpCQoJYtW0qSsrOznbbbv3+/7XFaWppHfQD14e/vmfjLGV6xZ88eSVKXLl0UHBzstF23bt1qvMbTPqqfx1U/FotFP/30k0f9wD/5YozWVmZmpiQpODhYXbp0aZA+0PT4eoy++OKL2rVrlzp37qzZs2fX+TxoPnwxRnft2qWSkhJJ0sCBA3X8+HFNnTpVcXFxSklJUceOHRUbG6tx48Zp8+bNdbgL+DNf/RydPn26JOnrr7/WmjVrHLZ56qmnJElBQUG68847Pe4DqCt/f89EAQP1VlJSovz8fElyGlGyio+PV1RUlCTpyJEjHvVj395dP+3bt3f4OjRPvhqjtbF69Wp9++23kqSxY8faonto3nw9RrOzs/Xkk09Kkl555RWFh4fX6TxoPnw1Rr///vsqffbs2VOLFy/W+fPnq3x/zZo1Gjp0qP72t795dH74L1/+HP2///s//epXv5IkXXfddfrTn/6kNWvWaPv27Vq+fLlGjBihd999V0FBQfr73/+u7t27e9wHUFf+/p6JAgbq7dy5c7bH1qinK9ZfGIWFhQ3Wj7WPuvQD/+OrMepOQUGB7r33XkmVn8hYP50BfD1G09PTVVxcrJtvvlljxoyp0znQvPhqjBYUFNgeP/HEE8rPz9c111yjHTt2qKSkRCdOnNArr7yimJgYVVRU6IEHHnD6CTiaF1/+HI2OjtaaNWu0cOFCJScn669//avGjRun/v3765ZbblFmZqauv/56ffHFF7rnnns8Pj9QH/7+nokCBurNGvWUKudSuRMWFiZJKi4ubrB+rH3UpR/4H1+NUVfKy8t1++2369ChQ5KkP//5z+rVq5fXzo+mzZdjdMmSJVq7dq1iYmI0Z84cj1+P5slXY9Q+aXHhwgVNmDBBH3zwgfr06aOwsDAlJSXp7rvv1urVqxUYGCjDMDRz5kwZhuFRP/A/vv5dv2PHDr311ltO18FYu3at3njjDZ09e7ZO5wfqyt/fM1HAQL3ZR49rs7CWdVGZiIiIBuvHfpFET/uB//HVGHXlnnvu0ccffyxJGj9+vB555BGvnRtNn6/GaH5+vm1Xh6efflpt2rTx6PVovhrjd70kvfDCCw4XOx4yZIiuv/56SdLu3bu1e/duj/qB//Hl7/p3331XI0aM0Pr169WzZ0+99957OnnypEpLS7V//3795S9/UVlZmebOnatBgwbp+PHjHvcB1JW/v2eigIF6a9Gihe1xbaJH1k9XahPvq2s/9p/geNoP/I+vxqgzs2bN0oIFCyRV/tH9zjvvKCgoyCvnhn/w1Rh94IEHlJ+fr759+xJrhkca43d9p06dXO6OM3bsWNvj7du3e9QP/I+vxuiJEyc0ZcoUXbhwQZdeeqk2b96sa6+9Vi1btlRISIhSU1M1a9YsrVy5UgEBAfrPf/6j++67z7ObAerB398zOV+eF6il8PBwJSQkKD8/Xzk5OS7bnjp1yvYPxX7RmNqwX4QmJydHffv2ddrWfhEaT/uB//HVGHXkueee07PPPitJ6t27t1atWtVkKtzwHV+M0aNHj2rp0qWSpKuuukpvv/22y/a5ubnKyMiQVPlGcsCAAbXuC/7HVz9H7dt7svhcbm6uR/3A//hqjGZkZNheO3v27CprCNgbNWqURo0apbVr12rFihU6deqU4uPjPeoLqAt/f89EAQNe0b17d2VlZWnfvn2yWCxOt67au3dvldd44pJLLnF4Hlf9sE0lrHwxRqt75ZVX9PDDD9vO9cknnyg2NrZe54T/augxah8jff75592237Nnj2699VZJ0uTJkylgwCc/Ry+99FLb4/Lycpdt7Z93tWUmmg9fjFH7LSp79+7tsm2fPn20du1aVVRU6Mcff+TnKHzC398zMYUEXjFkyBBJlTGkr776ymm7zMxM2+PBgwd71Ee/fv1sC9HYn6e60tJSbd26tcZr0Lz5YozaW7p0qWbMmCFJSk1N1dq1a5WQkFDn88H/+XqMAp7yxRjt2LGjOnToIEnav3+/y7b2z7dr186jfuCffDFG7YsiFovFZduysjKHrwMakr+/Z6KAAa+49tprbY8XLVrksE1FRYWWLFkiSYqLi9PIkSM96qNFixYaNWqUpMqVnZ3FA1esWGFb8fm6667zqA/4L1+MUasVK1Zo6tSpMgxDycnJWrdundq2bVunc6H5aOgxmpKSIsMw3B5Ww4cPt31v8eLFdbon+Bdf/RydOHGipMq1BjZv3uy03YoVK2yPhw4d6nE/8D++GKOdOnWyPc7KynLZduPGjZKkgIAApaSkeNQPUFd+/57JALxk6NChhiQjODjY2Lx5c43nn3/+eUOSIcl47LHHajy/aNEil88bhmGsW7fO1uY3v/mNYbFYqjyfl5dndOjQwZBkxMXFGQUFBd64NfgJX4zRTz75xAgNDTUkGUlJScbevXu9fBfwZ74Yo+5YXz98+PA6vR7+zRdj9NChQ0Z4eLghyejTp49RWFhYo83SpUtt5xk/fnx9bwt+pKHH6J49e4yAgABDktGuXTsjJyfH4XXMnz/fdp6BAwfW97bgpw4cOGAbJ5MnT67Va5r7eyayTPCal156SYMHD1ZxcbHGjBmj2bNna+TIkSouLlZGRoZtF4a0tDTbNn6euuqqq3TLLbcoIyNDH374oUaPHq37779fbdu21Xfffaenn35ahw8fliQ9++yzLJaEKhp6jG7dulXXXXedSktLFRISojlz5qisrMzl9n7JycmKi4ur6y3Bz/ji5yhQH74Yox06dNCTTz6pmTNn6quvvlL//v01c+ZM9ejRQ2fOnNGKFSs0b948SVJMTIzmzJnjtftD09fQY7Rbt26aOnWqXn/9df3888/q1auX7r//fg0dOlQtWrTQkSNHlJGRoWXLlkmSgoKC9Je//MWr94ima9OmTdq3b5/t6/z8fNvjffv21Ug8TpkypU79+PV7psauoMC/fPjhh0ZMTIyt4lf9SEtLM3766SeHr63tJ4dFRUXGuHHjnPYRGBhY508e4f8acow+9thjTs/r7Fi0aFHD3jCaHF/8HHXF+noSGHDGV2P04Ycftn3S7ehISkpy+Ak70NBjtKSkxLj55pvd/o6Piooy3nzzzQa8UzQ1kydP9ujvREea+3sm1sCAV02YMEHffvut/vjHPyotLU2RkZGKi4tT37599dxzz+mbb76p9wq3ERERWr16td58802NHj1aSUlJCg0NVfv27XXbbbdp06ZNevzxx71zQ/A7vhijQH0wRmF2vhqjzzzzjL744gtNmjRJKSkpCgsLU2xsrPr166ennnpKP/74owYOHOiFO4K/aegxGhYWpoyMDK1fv1533HGH0tLSFBUVpeDgYLVs2VIDBw7UI488or179+q2227z4p0Bteev75kCDMNuxS4AAAAAAAATIoEBAAAAAABMjwIGAAAAAAAwPQoYAAAAAADA9ChgAAAAAAAA06OAAQAAAAAATI8CBgAAAAAAMD0KGAAAAAAAwPQoYAAAAAAAANOjgAEAAAAAAEyPAgYAAAAAADA9ChgAAAAAAMD0KGAAAAAAAADTo4ABAAAAAABMjwIGAAAAAAAwPQoYAAAAAADA9ChgAAAAAAAA06OAAQAAAAAATI8CBgAAgA8dPHhQAQEBCggI0OLFixv7cgAAaDIoYAAAAJ/YsGGD7Y17bY/777+/sS8bAACYBAUMAAAAAABgesGNfQEAAKD5ufvuu3XPPfe4bZeQkOCDqwEAAE0BBQwAAOBzSUlJ6tGjR2NfBgAAaEKYQgIAAAAAAEyPAgYAAGgyUlJSFBAQoClTpkiStm/frltvvVXt27dXeHi42rdvrylTpmjPnj21Ot/KlSt1ww03KDk5WWFhYWrVqpUGDhyoZ599VoWFhbU6x+7du3XfffepZ8+eio+PV2RkpLp06aJf//rXmjt3rvLy8tye47PPPtOECRPUunVrhYWFqVOnTrr77ruVk5Pj8nVHjx7Vww8/rN69eys2NlahoaFq3bq1evbsqVtvvVWLFy/W2bNna3UfAACYXYBhGEZjXwQAAPB/GzZs0MiRIyVJjz32mB5//HGPz5GSkqJDhw5p8uTJGjZsmNLT02WxWGq0CwsL0xtvvKGbb77Z4XlKSkp022236b333nPaV9u2bbV69WpdccUVDp8vLy/X//7v/+qll15SRUWF0/NMnjy5ynapBw8eVKdOnSRJixYt0t69e/Xcc885fG1iYqIyMzPVvXv3Gs9lZWXpmmuucVugWLlypa655hqXbQAAaApYAwMAADQ5O3fu1LJly5SUlKRZs2apf//+Kikp0UcffaS//e1vunDhgn73u9+pU6dO6t+/f43XT5482Va8uPzyy/Xggw+qe/fuKigoUEZGhhYvXqyjR49q1KhR+vbbb9WuXbsa55g+fbpef/11SVKbNm00Y8YMDRo0SLGxscrLy9O2bdv07rvvuryPhQsXavPmzRo+fLjS09OVlpam06dPa8mSJVqyZIny8vI0bdo0bdmypcrrLly4oFtuuUVnz55VixYtdPfdd2vkyJFKSkpSWVmZDh06pC1btujf//53Xf8XAwBgOiQwAACAT9gnMGq7C0nXrl0VEhJi+9qawJCkjh07auvWrWrdunWV13z++ecaM2aMLBaL+vbtq+3bt1d5fvXq1bZEwqhRo/TRRx8pNDS0SpuFCxdq+vTpkqSbbrpJy5cvr/L8Bx98oGuvvVaSNHDgQH300UeKi4tzeA85OTlKTk62fW2fwJCku+66S/Pnz1dAQECV191111169dVXJUlff/21evXqZXtu/fr1GjVqlCTXCQuLxaKioiLFxMQ4fB4AgKaEAgYAAPAJ+wJGbR04cEApKSm2r+0LGO+++64mTpzo8HX33HOP5s6dK0natm2b+vXrZ3tu3LhxWrNmjUJCQrR//361b9/e4TlGjx6ttWvXKjg4WIcPH1abNm1szw0cOFBbt25VZGSkfvzxR4cJDWfsCxht2rTRgQMHFBYWVqPdDz/8oG7dukmSXnrpJf3hD3+wPbds2TLdfvvtkqQzZ85QoAAANAss4gkAAJqc+Ph4/fa3v3X6/LRp02yP165da3tssViUmZkpqbJA4ax4IVUmIKyv2bBhg+37J0+e1JdffimpMp3hSfGiuhtuuMFh8UKqTJ9ER0dLkrKzs6s8Z19MWbRoUZ37BwCgKaGAAQAAfO6xxx6TYRhuD/v0hb1evXopONj5Ul5XXHGFbVrI7t27bd/Pzs5WUVGRJGnAgAEur9H+eftz7Ny5U9YA67Bhw1zfqBvWhIUz8fHxkqRz585V+f6QIUOUmpoqSbr//vvVv39/PfPMM9q8ebNKS0vrdU0AAJgVBQwAANDkJCUluXw+ODhYLVu2lCQVFBTYvm//+KKLLnJ5Dvu1Nexfl5+fb3tsn4Soi8jISJfPBwZW/qlWXl5e5fshISFauXKlbXeS7du3a/bs2Ro8eLDi4uJ09dVXa9myZTVeBwBAU0YBAwAANDnVF7x0xN0yX7U5hzeuo6Fccskl+u677/Tee+9p2rRp6ty5sySpuLhYH3/8sW6//XYNGDBAubm5jXaNAAB4EwUMAADQ5Jw4ccLl8xaLRadOnZIkWxKj+uPjx4+7PIf98/avS0hIsD0+evRo7S64gQQFBenaa6/Va6+9pn379uno0aN67bXX1KdPH0nSV199pfT09Ea9RgAAvIUCBgAAaHJ27twpi8Xi9Pldu3bZ1oLo0aOH7fupqam2aRvWhTid2bZtm+2x/Tl69eplS15s3LjR84tvQG3atNG0adO0ZcsW9e7dW5K0atUqFRcXN/KVAQBQfxQwAABAk1NQUKCVK1c6ff7111+3Pf7Vr35lexwcHKzhw4dLkj777DMdOXLE6TleffVVSZUphxEjRti+37JlSw0aNEiS9Pbbbzd6CsORkJAQ231aLBadPn26cS8IAAAvoIABAACapAceeMDhVJLMzEwtWLBAktSnTx/169evyvP33nuvJKmsrEzTpk1zuGvH66+/rk8//VSSNHHixBqLdT700EOSpKKiIt144406c+aM0+vMycnx4K5qJysrS/v27XP6fGlpqW272OjoaCUmJnr9GgAA8DXn+48BAAA0kNzc3CpbkzoTERFhW5zS3uWXX67vv/9effr00axZs9S/f39duHBBH330kebMmSOLxaLg4GC9/PLLNV47fvx43XjjjXrnnXe0du1aDRgwQA8++KC6d++uU6dOKSMjw5bgaNmypf7f//t/Nc4xYcIE/f73v9drr72mzZs365JLLtGMGTM0ePBgxcTEKD8/Xzt27NDbb7+tyy67TIsXL/b8f5IL69at01NPPaWhQ4dq/Pjxuuyyy5SYmKji4mL9+OOPmjdvnr7++mtJ0p133ulyy1kAAJoKfpsBAACfmzt3rubOneu23eWXX66dO3fW+P4VV1yhGTNm6O6779aMGTNqPB8aGqo33nhDAwYMcHjeJUuWyGKx6L333tPOnTs1adKkGm3atm2r1atXq127dg7PMX/+fEVEROjll1/W0aNHNXv2bIftLrvsMhd3WHcVFRXKzMy0JS0cuf766/XMM880SP8AAPgaBQwAANAk3XnnnerRo4fmzJmjTZs2KT8/X4mJiRo1apQeeughXXLJJU5fGx4erhUrVmjlypVavHixtm7dqvz8fEVFRSktLU3XXnutZsyYoejoaKfnCAoK0j/+8Q9NnTpV8+fP14YNG/Tzzz/LMAy1a9dOF198sa677jpNnDjR6/c+c+ZMDRgwQJ999pm2bNmio0eP2rZLbd26tQYMGKA77rhD48aN83rfAAA0lgDD3SbpAAAAJpGSkqJDhw5p8uTJXp+WAQAAzI1FPAEAAAAAgOlRwAAAAAAAAKZHAQMAAAAAAJgeBQwAAAAAAGB6FDAAAAAAAIDpsQsJAAAAAAAwPRIYAAAAAADA9ChgAAAAAAAA06OAAQAAAAAATI8CBgAAAAAAMD0KGAAAAAAAwPQoYAAAAAAAANOjgAEAAAAAAEyPAgYAAAAAADA9ChgAAAAAAMD0KGAAAAAAAADTo4ABAAAAAABMjwIGAAAAAAAwPQoYAAAAAADA9ChgAAAAAAAA06OAAQAAAAAATI8CBgAAAAAAMD0KGAAAAAAAwPQoYAAAAAAAANP7/w+8+nS+2fJ6AAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 600x300 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 294,
       "width": 536
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "%config InlineBackend.figure_format = 'retina'\n",
    "plt.figure(figsize=(6,3))\n",
    "plt.plot(train_losses, label='Training loss')\n",
    "plt.plot(valid_losses, label='Validation loss')\n",
    "plt.xlabel(\"Epochs\")\n",
    "plt.ylabel(\"Loss\")\n",
    "plt.legend(frameon=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### model performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Correct: [19772 / 20000] with accuracy 98.86 %\n",
      "Correct: [7356 / 7455] with accuracy 98.67 %\n",
      "Correct: [6495 / 7172] with accuracy 90.56 %\n"
     ]
    }
   ],
   "source": [
    "cnn.get_accuracy(loader=train_loader, model=trained_model, device=device)\n",
    "cnn.get_accuracy(loader=valid_loader, model=trained_model, device=device)\n",
    "cnn.get_accuracy(loader=test_loader, model=trained_model, device=device)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Save Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# save model\n",
    "PATH = 'output/cnn/model.pth'\n",
    "torch.save(trained_model.state_dict(), PATH)\n",
    "# reload model\n",
    "#net = cnn.CNN().to(device)\n",
    "#net.load_state_dict(torch.load(PATH))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "bnn",
   "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": "3.9.13 | packaged by conda-forge | (main, May 27 2022, 16:56:21) \n[GCC 10.3.0]"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "d09e971681001a3bfdc1839b66ef386ac64c285d9e0d72fab495a2fdc017d205"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}