1. Beginner/Pytorch5_CNN_Classifier(CIFAR10 dataset).ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### CIFAR10 Classification"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-09T12:45:37.037080Z",
     "start_time": "2019-12-09T12:45:33.352178Z"
    }
   },
   "outputs": [],
   "source": [
    "import torch\n",
    "import torchvision\n",
    "import torchvision.transforms as transforms\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-09T12:45:51.876545Z",
     "start_time": "2019-12-09T12:45:51.867553Z"
    }
   },
   "outputs": [],
   "source": [
    "# Hyper Parameters\n",
    "# torch.manual_seed(40)\n",
    "batch_size = 16\n",
    "EPOCH = 10\n",
    "lr = 0.01"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-09T12:45:40.096304Z",
     "start_time": "2019-12-09T12:45:37.182976Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Files already downloaded and verified\n",
      "Files already downloaded and verified\n",
      "Dataset CIFAR10\n",
      "    Number of datapoints: 50000\n",
      "    Root location: ../data\n",
      "    Split: Train\n",
      "    StandardTransform\n",
      "Transform: Compose(\n",
      "               ToTensor()\n",
      "               Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])\n",
      "           )\n",
      "****************************************************************************************************\n",
      "Dataset CIFAR10\n",
      "    Number of datapoints: 10000\n",
      "    Root location: ../data\n",
      "    Split: Test\n",
      "    StandardTransform\n",
      "Transform: Compose(\n",
      "               ToTensor()\n",
      "               Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])\n",
      "           )\n"
     ]
    }
   ],
   "source": [
    "# Dataset & DataLoader\n",
    "\n",
    "transform = transforms.Compose(\n",
    "[\n",
    "    transforms.ToTensor(),\n",
    "    transforms.Normalize(mean=[0.485, 0.456, 0.406],\n",
    "                                 std=[0.229, 0.224, 0.225])\n",
    "])\n",
    "\n",
    "trainset = torchvision.datasets.CIFAR10(root = '../data', train = True,\n",
    "                                      download = True, transform=transform)\n",
    "\n",
    "trainloader = torch.utils.data.DataLoader(dataset=trainset, batch_size = batch_size,\n",
    "                                          shuffle = True, num_workers=2)\n",
    "\n",
    "testset = torchvision.datasets.CIFAR10(root = '../data', train = False,\n",
    "                                     download = True, transform=transform)\n",
    "\n",
    "testloader = torch.utils.data.DataLoader(dataset=testset, batch_size = batch_size,\n",
    "                                         shuffle = True, num_workers=2)\n",
    "\n",
    "classes = ('plane', 'car', 'bird', 'cat',\n",
    "           'deer', 'dog', 'frog', 'horse', 'ship', 'truck')\n",
    "\n",
    "print(trainset)\n",
    "print('*'*100)\n",
    "print(testset)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-09T12:45:43.742211Z",
     "start_time": "2019-12-09T12:45:40.108297Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([16, 3, 32, 32])\n"
     ]
    },
    {
     "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": [
    "# Visualization Datasets\n",
    "\n",
    "def show(img):\n",
    "    print(img.size())\n",
    "    grid = torchvision.utils.make_grid(img, padding = 1) # make_grid 함수는 3채널로 만든다(모두 같은 format으로)\n",
    "    tranimg = grid.permute(1,2,0)\n",
    "    plt.imshow(tranimg)\n",
    "\n",
    "images, labels = iter(trainloader).next()\n",
    "\n",
    "show(images[:32])\n",
    "plt.savefig('aaaa.jpg')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Make Model\n",
    "\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "\n",
    "class Net(nn.Module):\n",
    "    \n",
    "    def __init__(self):\n",
    "        super(Net, self).__init__()\n",
    "        self.conv1 = nn.Conv2d(3, 10, 5)\n",
    "        self.conv2 = nn.Conv2d(10, 20, 5)\n",
    "        self.conv3 = nn.Conv2d(20, 20, 5)\n",
    "        self.conv4 = nn.Conv2d(20, 10, 5)\n",
    "        self.conv5 = nn.Conv2d(10, 10, 5)\n",
    "        self.pool = nn.MaxPool2d(2,2)\n",
    "        \n",
    "        self.fc1 = nn.Linear(10 * 6 * 6, 256)\n",
    "        self.fc2 = nn.Linear(256, 128)\n",
    "        self.fc3 = nn.Linear(128, 64)\n",
    "        self.fc4 = nn.Linear(64, 10)\n",
    "        \n",
    "    def forward(self, x): # 32 32\n",
    "        x = F.relu(self.conv1(x)) # 28 28\n",
    "        x = F.relu(self.conv2(x)) # 24 24\n",
    "        x = F.relu(self.conv3(x)) # 20 20\n",
    "        x = F.relu(self.conv4(x)) # 16 16\n",
    "        x = F.relu(self.conv5(x)) # 12 12\n",
    "        x = self.pool(x)          # 6 6\n",
    "        x = x.view(-1, 10 * 6 * 6)\n",
    "        x = F.relu(self.fc1(x))\n",
    "        x = F.relu(self.fc2(x))\n",
    "        x = F.relu(self.fc3(x))\n",
    "        x = self.fc4(x)\n",
    "        \n",
    "        return x   \n",
    "    \n",
    "model = Net()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-09T12:48:41.400245Z",
     "start_time": "2019-12-09T12:48:41.291310Z"
    }
   },
   "outputs": [],
   "source": [
    "# Make Model\n",
    "\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "\n",
    "class Net(nn.Module):\n",
    "    \n",
    "    def __init__(self):\n",
    "        super(Net, self).__init__()\n",
    "        self.conv1 = nn.Conv2d(3, 10, 7, padding = 3)\n",
    "        self.conv2 = nn.Conv2d(10, 50, 7, padding = 3)\n",
    "        self.conv3 = nn.Conv2d(50, 120, 7, padding = 3)\n",
    "        self.conv4 = nn.Conv2d(120, 100, 5)\n",
    "        self.conv5 = nn.Conv2d(100, 20, 5)\n",
    "        self.conv6 = nn.Conv2d(20, 10, 5)\n",
    "        self.conv7 = nn.Conv2d(10, 10, 3)\n",
    "        self.pool = nn.MaxPool2d(2,2)\n",
    "        self.fc1 = nn.Linear(10 * 10 * 10, 120)\n",
    "        self.fc2 = nn.Linear(120, 360)\n",
    "        self.fc3 = nn.Linear(360, 50)\n",
    "        self.fc4 = nn.Linear(50, 10)\n",
    "        \n",
    "    def forward(self, x):\n",
    "        x = F.relu(self.conv1(x)) # 32 32\n",
    "        x = F.relu(self.conv2(x)) # 32 32\n",
    "        x = F.relu(self.conv3(x)) # 32 32\n",
    "        x = F.relu(self.conv4(x)) # 28 28\n",
    "        x = F.relu(self.conv5(x)) # 24 24\n",
    "        x = self.pool(F.relu(self.conv6(x))) # 10 10\n",
    "        x = x.view(-1, 10 * 10 * 10)\n",
    "        x = F.relu(self.fc1(x))\n",
    "        x = F.relu(self.fc2(x))\n",
    "        x = F.relu(self.fc3(x))\n",
    "        x = self.fc4(x)\n",
    "        \n",
    "        return x   "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## GPU 사용여부 확인"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-09T12:48:41.954929Z",
     "start_time": "2019-12-09T12:48:41.942939Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cuda\n"
     ]
    }
   ],
   "source": [
    "if torch.cuda.is_available():\n",
    "    device = torch.device('cuda')\n",
    "else:\n",
    "    device = torch.device('cpu')\n",
    "print(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-09T12:48:42.189794Z",
     "start_time": "2019-12-09T12:48:42.122833Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Net(\n",
       "  (conv1): Conv2d(3, 10, kernel_size=(5, 5), stride=(1, 1))\n",
       "  (conv2): Conv2d(10, 20, kernel_size=(5, 5), stride=(1, 1))\n",
       "  (conv3): Conv2d(20, 20, kernel_size=(5, 5), stride=(1, 1))\n",
       "  (conv4): Conv2d(20, 10, kernel_size=(5, 5), stride=(1, 1))\n",
       "  (conv5): Conv2d(10, 10, kernel_size=(5, 5), stride=(1, 1))\n",
       "  (pool): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
       "  (fc1): Linear(in_features=360, out_features=256, bias=True)\n",
       "  (fc2): Linear(in_features=256, out_features=128, bias=True)\n",
       "  (fc3): Linear(in_features=128, out_features=64, bias=True)\n",
       "  (fc4): Linear(in_features=64, out_features=10, bias=True)\n",
       ")"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Weight 초기화\n",
    "def weight_init(m):\n",
    "    if ((type(m) == nn.Conv2d) or (type(m) == nn.Linear)):\n",
    "        torch.nn.init.xavier_uniform_(m.weight)\n",
    "        m.bias.data.fill_(0)\n",
    "        \n",
    "model.apply(weight_init)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-09T12:48:42.189794Z",
     "start_time": "2019-12-09T12:48:42.122833Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CrossEntropyLoss() SGD (\n",
      "Parameter Group 0\n",
      "    dampening: 0\n",
      "    initial_lr: 0.01\n",
      "    lr: 0.01\n",
      "    momentum: 0\n",
      "    nesterov: False\n",
      "    weight_decay: 0\n",
      ")\n"
     ]
    }
   ],
   "source": [
    "import torch.optim as optim\n",
    "\n",
    "loss_func = nn.CrossEntropyLoss()\n",
    "optimizer = optim.SGD(model.parameters(), lr = 0.01)\n",
    "scheduler = optim.lr_scheduler.StepLR(optimizer, step_size = 30, gamma = 0.1)\n",
    "print(loss_func, optimizer)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-09T12:48:49.964331Z",
     "start_time": "2019-12-09T12:48:42.718489Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([16, 10])\n",
      "Test :  tensor([[ 0.1032, -0.1221,  0.0394,  0.0119, -0.0315,  0.0116, -0.0928,  0.0652,\n",
      "         -0.0257,  0.0550],\n",
      "        [ 0.1041, -0.1226,  0.0386,  0.0122, -0.0324,  0.0123, -0.0932,  0.0645,\n",
      "         -0.0250,  0.0553],\n",
      "        [ 0.1044, -0.1221,  0.0395,  0.0121, -0.0315,  0.0119, -0.0930,  0.0643,\n",
      "         -0.0258,  0.0554],\n",
      "        [ 0.1038, -0.1220,  0.0390,  0.0120, -0.0316,  0.0119, -0.0928,  0.0646,\n",
      "         -0.0256,  0.0554],\n",
      "        [ 0.1037, -0.1223,  0.0391,  0.0122, -0.0312,  0.0117, -0.0931,  0.0646,\n",
      "         -0.0255,  0.0557],\n",
      "        [ 0.1036, -0.1219,  0.0386,  0.0118, -0.0318,  0.0119, -0.0929,  0.0645,\n",
      "         -0.0256,  0.0551],\n",
      "        [ 0.1037, -0.1215,  0.0386,  0.0117, -0.0317,  0.0118, -0.0926,  0.0648,\n",
      "         -0.0260,  0.0552],\n",
      "        [ 0.1037, -0.1220,  0.0393,  0.0116, -0.0317,  0.0116, -0.0926,  0.0644,\n",
      "         -0.0256,  0.0550],\n",
      "        [ 0.1038, -0.1220,  0.0384,  0.0117, -0.0321,  0.0119, -0.0931,  0.0653,\n",
      "         -0.0251,  0.0552],\n",
      "        [ 0.1041, -0.1219,  0.0391,  0.0121, -0.0315,  0.0124, -0.0931,  0.0644,\n",
      "         -0.0256,  0.0554],\n",
      "        [ 0.1037, -0.1217,  0.0391,  0.0121, -0.0318,  0.0122, -0.0927,  0.0647,\n",
      "         -0.0260,  0.0555],\n",
      "        [ 0.1037, -0.1213,  0.0389,  0.0121, -0.0323,  0.0120, -0.0925,  0.0644,\n",
      "         -0.0257,  0.0549],\n",
      "        [ 0.1044, -0.1221,  0.0394,  0.0123, -0.0311,  0.0121, -0.0929,  0.0642,\n",
      "         -0.0259,  0.0554],\n",
      "        [ 0.1033, -0.1220,  0.0389,  0.0122, -0.0322,  0.0121, -0.0927,  0.0646,\n",
      "         -0.0254,  0.0549],\n",
      "        [ 0.1038, -0.1222,  0.0392,  0.0121, -0.0316,  0.0118, -0.0931,  0.0647,\n",
      "         -0.0256,  0.0555],\n",
      "        [ 0.1038, -0.1221,  0.0389,  0.0118, -0.0324,  0.0116, -0.0927,  0.0643,\n",
      "         -0.0251,  0.0548]], device='cuda:0')\n"
     ]
    }
   ],
   "source": [
    "model.eval()\n",
    "model.to(device)\n",
    "\n",
    "with torch.no_grad():\n",
    "    images, labels = next(iter(trainloader))\n",
    "    images, labels = images.to(device), labels.to(device)\n",
    "    \n",
    "    example = model(images)\n",
    "    \n",
    "    print(example.size())\n",
    "    print('Test : ', example)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1/10]-----[3125/3125] LOSS : 4602.747------ Time : 12 \n",
      "\n",
      "[2/10]-----[3125/3125] LOSS : 4342.556------ Time : 12 \n",
      "\n",
      "[3/10]-----[3125/3125] LOSS : 4199.608------ Time : 12 \n",
      "\n",
      "[4/10]-----[3125/3125] LOSS : 4085.875------ Time : 12 \n",
      "\n",
      "[5/10]-----[3125/3125] LOSS : 3990.505------ Time : 12 \n",
      "\n",
      "[6/10]-----[3125/3125] LOSS : 3912.424------ Time : 12 \n",
      "\n",
      "[7/10]-----[3125/3125] LOSS : 3834.343------ Time : 12 \n",
      "\n",
      "[8/10]-----[3125/3125] LOSS : 3763.016------ Time : 12 \n",
      "\n",
      "[9/10]-----[3125/3125] LOSS : 3684.760------ Time : 12 \n",
      "\n",
      "[10/10]-----[3125/3125] LOSS : 3619.109------ Time : 12 \n",
      "\n"
     ]
    }
   ],
   "source": [
    "import time\n",
    "\n",
    "# Train\n",
    "\n",
    "model.train()\n",
    "for e in range(1, EPOCH+1):\n",
    "    start_time = time.time()\n",
    "    running_loss = 0\n",
    "    \n",
    "        \n",
    "    for i, data in enumerate(trainloader):\n",
    "        images, labels = data\n",
    "        images, labels = images.to(device), labels.to(device)\n",
    "        \n",
    "        optimizer.zero_grad()\n",
    "        outputs = model(images)\n",
    "        loss = loss_func(outputs, labels)\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "\n",
    "        running_loss += loss\n",
    "        now = time.time()\n",
    "\n",
    "        print('\\r[%d/%d]-----[%d/%d] LOSS : %.3f------ Time : %d ' \n",
    "              %(e, EPOCH, i+1, 50000/batch_size, running_loss, now - start_time), end = '')        \n",
    "    print('\\n')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-09T12:50:49.496727Z",
     "start_time": "2019-12-09T12:48:55.777997Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1/10]-----[6250/7500] LOSS : 14400.677------ Time : 20 ------- Current lr : 0.050000\n",
      "\n",
      "[2/10]-----[6250/7500] LOSS : 14400.625------ Time : 20 ------- Current lr : 0.050000\n",
      "\n",
      "[3/10]-----[6250/7500] LOSS : 14401.487------ Time : 21 ------- Current lr : 0.050000\n",
      "\n",
      "[4/10]-----[6250/7500] LOSS : 14399.734------ Time : 20 ------- Current lr : 0.050000\n",
      "\n",
      "[5/10]-----[6250/7500] LOSS : 14400.587------ Time : 20 ------- Current lr : 0.050000\n",
      "\n",
      "[6/10]-----[6250/7500] LOSS : 14399.898------ Time : 20 ------- Current lr : 0.050000\n",
      "\n",
      "[7/10]-----[6250/7500] LOSS : 14400.886------ Time : 20 ------- Current lr : 0.050000\n",
      "\n",
      "[8/10]-----[6250/7500] LOSS : 14401.681------ Time : 20 ------- Current lr : 0.050000\n",
      "\n",
      "[9/10]-----[6250/7500] LOSS : 14400.815------ Time : 20 ------- Current lr : 0.050000\n",
      "\n",
      "[10/10]-----[6250/7500] LOSS : 14401.087------ Time : 20 ------- Current lr : 0.050000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "import time\n",
    "\n",
    "# Train\n",
    "\n",
    "model.train()\n",
    "for e in range(1, EPOCH+1):\n",
    "    start_time = time.time()\n",
    "    running_loss = 0\n",
    "    \n",
    "    for param_group in optimizer.param_groups:\n",
    "        learning_rate = param_group['lr']\n",
    "        \n",
    "    for i, data in enumerate(trainloader):\n",
    "        images, labels = data\n",
    "        images, labels = images.to(device), labels.to(device)\n",
    "        \n",
    "        optimizer.zero_grad()\n",
    "        outputs = model(images)\n",
    "        loss = loss_func(outputs, labels)\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "\n",
    "        running_loss += loss\n",
    "        now = time.time()\n",
    "\n",
    "        print('\\r[%d/%d]-----[%d/%d] LOSS : %.3f------ Time : %d ------- Current lr : %f' \n",
    "              %(e, EPOCH, i+1, 60000/batch_size, running_loss, now - start_time, learning_rate), end = '')    \n",
    "    scheduler.step()        \n",
    "    print('\\n')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Net(\n",
       "  (conv1): Conv2d(3, 10, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3))\n",
       "  (conv2): Conv2d(10, 50, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3))\n",
       "  (conv3): Conv2d(50, 120, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3))\n",
       "  (conv4): Conv2d(120, 100, kernel_size=(5, 5), stride=(1, 1))\n",
       "  (conv5): Conv2d(100, 20, kernel_size=(5, 5), stride=(1, 1))\n",
       "  (conv6): Conv2d(20, 10, kernel_size=(5, 5), stride=(1, 1))\n",
       "  (conv7): Conv2d(10, 10, kernel_size=(3, 3), stride=(1, 1))\n",
       "  (pool): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
       "  (fc1): Linear(in_features=1000, out_features=120, bias=True)\n",
       "  (fc2): Linear(in_features=120, out_features=360, bias=True)\n",
       "  (fc3): Linear(in_features=360, out_features=50, bias=True)\n",
       "  (fc4): Linear(in_features=50, out_features=10, bias=True)\n",
       ")"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "torch.save(model, '../data/mnist_classifier.pth')\n",
    "test_model = torch.load('../data/mnist_classifier.pth')\n",
    "test_model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([8, 3, 32, 32])\n",
      "GroundTruth :  plane deer horse deer plane deer truck frog\n",
      "Predicted :  truck cat dog horse plane frog car frog\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXMAAABMCAYAAACf4VdeAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAACA4klEQVR4nOz9WaytW5bfCf1m97Wr2/1pbx9dRmRk2pnOdFbicpUsJBqJAslAgYRAQvITD0g8YPHEC6iekECUSkoEEkglVSGBBA8locJlsMtVTqcznRmRERn33rjd6c9uV/s1s+Vhrn3OiYh0RrjSShvrjKt991lrf6v71pzjG+M//uM/REqJt/bW3tpbe2v//23yX/QbeGtv7a29tbf257e3zvytvbW39tb+FbC3zvytvbW39tb+FbC3zvytvbW39tb+FbC3zvytvbW39tb+FbC3zvytvbW39tb+FbA/lzMXQvxXhBAfCyF+LIT42/+83tRbe2tv7a29tX82E/9FeeZCCAV8AvyXgSfA7wH/vZTSD//5vb239tbe2lt7a7+I/Xki898AfpxS+jylZIH/APi3/vm8rbf21t7aW3tr/yym/xyPvQ88fuP2E+A3/6wHTOcH6eTOfRC394ifOUYIke//iT+JfPNnD399POnVsa8ffvuARAJyEpJ4nYwk3sxLBBBC5OL5U5SSpBhJKVFVJZOmzo9IiRgjIUKMiZgSt9mN94EQIzEmpFT4EEjBkaIjRkcMMb/i/nghJEIqQJBSJAWPFHL/XhMIQVEW1HVFU9dUZUVM4KzFuZHge5RS+3OQXn2mGPc/IeFdoqoqhIBEJKWA937/eSVS5te+/SwgqKoaqSQxRrz3CCEQQiClRAqJlBJr3f785MdCIsV8bhMJbUx+XZPwakDESAoRSAgBQiqk1ghUfh8ioQxIpZBSIoBCK5SskBQIzBvfUnr13QnUq8+dTf7pC4UERGJy2LTDup6h79ldD4SQoJgiiik/L0/9mWf+M9YyIvHmE4qfWJtvvLP0+v2//hZ/4qH79QBy3CKDfeM131j3+98p5e9BiLzGbtf87bqXUrw6PgSPkHkdSCERQhBjIKZIjBElFUIKQsi3hRBIsX+N/fPGGHhjS+VfJq9jUkIiMFpjtEFrDQKcd/mMiHxeEq/3U0wJoxRSCLzz2NG+OpXpdr8JKMuSEPIadc7lTy8kZVkwm0ypq5pdt2McR0ZrX32eECMheGJK1FX2ISklUsjPq5REinx+tAQhIghICIRQkAQhhv1+joSQ8h4UAgiUpcKYvK6djwDEkI+LPp/DYn9+Qow4n/eEVBKpQIqEEAmpDEoZtCp5/PjqMqV08rOL7LX9eZz5P23H/ORBQvwt4G8BHJ3e5X/57/0HrxwDQiP2zuvV8XLv4GReWKT9iZISIUX++xvOReuf/Qi3zkAIgZD5i/IhEELCx5/aNCmRiPkOIdisVvy7/+v/FUKAcxYlEh++e5e/8ssf8c7dOyxmM1KSPLvasB08293Iar1ls9ux7jpW647VcocUhsE7+vUL+vUTYv+Ssdti7UCMDi0Fulowmd9FqYK+XzPcXCBSxAePDZ56OuPu/fv8+q/9ZX7rN36dD955l9/9R7/H8xfP2W6uKOMlVVURgieEQIwBSFRVzdAlrl52fPzDJ3z7O9+grgVJ9Ag5YG1kuw4MXUAgmEwl/TASEUxmc37tr/wG9aRmdJbVZo2UChEFVVEynbTcuXPKjz/5mGdPHnN1dYFzluQjKUpihBATx6en3HvnIdN3wXxwwXa9xl5agh+J2qFMybSaUhUTqmrC4mTC5DhfvIypMLrhUFfMinco9AJkTaXOKJkj0PsLaAKx2S/FAsEM8UayKX5ihUbA48MjluIP+erJH/P9f/KH/Cf/4fdYXgXEu38D/f6/SYq3bjS9WuWvoEjxKqwAsXdEIjuiVy+2X5tRxNsFtr+8ZCd4e6m5dUggEHsHFlIiu4dERBBTfteRhE8RhGD65T+gvHkE5CBAphwgxJgYXSQJmZ2+ElR1RdM0OOew1hO9IISU98fegfW7LVortDaYskAphQ8j3juC9yCgKipG6xhHyzjaVxd5RMJUiiQSzjlCCBAhuES6X1O2DdOy5XCywCTBbDajKCueXb7kxc0lBwcHHC4WLNoZXz1+CkqjTUlTNXz3o6+zuVry7KunPP3yCWeHp1xtVpxfXrDerDk6PObw6JCvHj3i6bPnrDcbjDG0bcuH73/AX//tv8b79x7w//47f4fPvvicy+srTFHyzsP3uLlZ8uzlc7Zdxy9/V2O0ggjJegQJZTQCiN4T7I66ipSVpKhL2uaY0kwQQhCiZ9d3PH1ySQg5sAlpx507hrt3DqirhmHUWO9YXXd0W8fQB0ojeHhkKHTERc+zK4syDaIQqALqVmCMQ8sGKVpgwv/+37366k/xtz9hfx5n/gR4+MbtB8Cznz4opfQ7wO8AvPf1b6fbq3tKCSF/Mky5dd4pxte3yYeIFElJIlPa7wFx+/w/88aygwaREqT9C9xGP2885xuvTN42+Z8ehXcOH0AJuFztuLha8u79B8znB8QIL1cjRVVQJcPoEzfbDo8iSEFSCecGXLDYMOKCR0RQWmOSBhJFoUmqoKwblC4JyRGrCm97pBCUpmS2WNBOptRNgzElznu+/PJLLq8uSHHkbCFzJpDEq58QIhNd07QGO63ouy8YB8dkOqFpK2La0VSGsduyGtZsNztMsUDrCbo0NO0MoTRSlbRVSzs7QCvN+maFElAUJXVd0zQ1RWXQWryKplAShUInQRKQUqTQisW0Au/oC80YJS5KiBHpAkoHtIqoUqCVQUmTL/ACdn1PdOdItSJqaOSSmb5LqeZo2ULSeHEFSCRTjDjI0fBPROm8+ndKEiVKJvIOTbpA+zkxOFIUiJQzFUTkZ6L/N64K6ZXPztFterVA31zIt9Hmm+vt1WXg1fq8/XN643lvj30zq8zH5chPSolScv/ZJOLW45Mj4BATSe5zFglaCVIURCnwcn+hiTE7n5gwUuaoN8YcBVtLjO5VGO+sQ0RJQiCFIvjsuH0I+VyriDIqZ5b7qDnEiBKJqtTMZxPOjo/xvaWoSmKKXN7csNztqCYTXEz4EOm6nuOzu5yd3OHB3ft8570P+OSHH9NfbwnHlu/80i/z6Pkz2Gezd+/eZbPZsNls2e52pBiRUhLCPvPcn9fRWgZrcSFA8K+3O6D3TltKgdqf1+Ddq4ujD4H1bsAFQRkkZYj4sKFtJE3TUjUtSRiM3iFlXi8uwDB4NptEigKlCpxzuKAJCFARG0a899RGUBSStlQEJRFaobREK0nwHh8GSBEhfrG65p/Hmf8e8DUhxPvAU+DfBv77P+9B8U1HHQGRXkXaQuwXvAQir6LvV1DKfrEk+XpjxJ92/EIQU3yVut0uyrzp5J+aTuRNu4c+SAQf8AlCSvjgeX5xxdPDlu+6gDaG7WbHcrMhypKEoChKrI9s+p5u7PD0xDSw3W0Y+ivcuEWOjtIoVGlQSlI3BTurkcYglUEogyoMzu2QWlI2Dcdnd1gcHKJ0wXbX88y/4MeffsJ6t6GuFcfzOSJKjCkwZUmMkb4fKKqaupjRlgql/pAYBYeLE9559w7b7gpnYehecn1pWS4vOTo+5s79U6bzKaYy7DqLMJ55M+Pk9IS6rPi0+yF26LB2pO92hOjQSlBWhoTFxYSUBiELQDBaSyJRaMm0UjitgEByGu8CqdqihUfHhEwR7weEL5DaIFJBjIFdt2XLkig9wXjKMOGkeZdF/ZBp+RAR5nTiJSDQDGh1jFBqDy3I/e/9N5wiAkliSsXXKf0Au8+x1pLS7TYQ3F4AhMhrTfzEKuG1884H/URw8Co4f+XkfxJmYR/EsI/mxRvrLnLr0N8IPn4CPUsgJNpIjMkZkBCS6HLkLoSgKgzWBSIRmRIyYxcoBEYKEAmlDeMwEPaRdKWKDBXGRPCebugheJSUSCVxg8WPHm0KEvlCEGPGbPJ6G9Fe4YPfR6cwDo5prFASjJE0dYUTEqE1vR1Z9zt65+m9ZzdYsCucC9w9ucMvf/u7fPeXvsvdyZSrxy/Zzda0uuE3/8pvMv3sU6QxtJMp77//Pn//7/09ttsd3nmKsqCqKsZxZBgGANpJSzud0LQN26Gj73s22w3DMJAS1HWNSDuUiBilKJVk7GzOjET+QmISuKiIY2S0PZuNpW1HDg8PWchDpDAoVVJXBabQjLbC+iVdr5FC0TQFfb8liRZVQBKe7WagGzyNkdRaYiT4GCh0SVmVSBHZ7UZC8BRFzcHi6E/1Wj9t/4WdeUrJCyH+J8D/i5zz/Z9SSj/4OY967XylhJSjICHfdNqvMcTbKAL2kImSeTNESGK/EeSr9wOA3KeZ7B3+LcadN6Ig/ixiSXbmYR/9ZFw5xUDylhQcwWiQmqvlDZ/8+BO+/OoxT693zOcnzGYLZk2L1oqXL8/ZbdcUMjCvNGnsCH1Pcg6lBFImyrKkKEFp2Awe5zyJHI12fYe3lqKqqNs59x6+R3CWzz7/ks8++4K2LPnqqyc5mltUbDo4Oz3j9PSEumnwLrJZb0kpUdSK48MDvvVL79F3A8ZM+OjDX0EXkT/6wz/gwQNDXc0Z+kCIFb/1W3+D7/zKL5GE5/PHX7DddSAM3ktiASGMhDDm70PAdrum6zusddjRI2XJ4dEdtK64vLpmvbnm8vKCdllxOkbKmxXnyyXXO8nGFdTJ8mAmOGwb6mZOYKQtaubNA2pzhuKI1G4Y7HPGsGR0W1y3ZO1+hAwrShmo5XcQviWkJY7HdGKJlncwnKHFfL8sb02AUAgWgMS5R/Q7iRtvETaRM4IU+NPx71eoSF41txDJTx/2Kqi4Pfi1N799nVcO/BZffgMvfzOvEPzUHSTKUjCZZxzV2sB6bQkxIZJkXreU05K6rqjamqquiTGilUZKifMeZx3L5ZLtdkvf94gksM7jUsSRo9sYRMaDfYZNdtstSUKSghBvM9iEEJGYEl3f72G+jKuP1lG4hm23QwRJt+oAqNsGVRTMFnNePnrMdtNhkmZMgof3HvKr3/lVvvHRN1i0U8Ztz9iNHMwPOf3ojMXBEU3znO9855f5tV//K9y9c4ff/8e/z/HhEWVV4YLnYLHg5uYG5z0vz1/yX/qrv8V3vvvL6Kog/ehHfP+P/5hPP/2UoigxZUHZNhRyQ7IdLoFUcDhtSTHlTFeUBG9RyhC8Z+h3hGAZuwuWV1dorSlMxW43cufOHebTI6Dlycsl2+1APzjK9Yayqfj6N79FOzkkBfj+9/8A3z1lsA4jEyFBiCCkRkmNGzZsrrc44OBoxr17D4B/8rNr8qfszxOZk1L6j4D/6Bc/nldFFBEjSImUt5E5yAyNvw5zxE8VhFIEIfeIYsYf5RuRT8rJYL4/yFep7u0FIaX8qPRTBUNEeiOxTQT0HiaAGDw+BB4/fsy4OqepFC4CZkpbNRRyxsG0pi4UttuxW10TpGdGxbyAstZYcuHH2R0QiRGiD6Sk6XYbhLREPxK8QymdMeOioWlnfPbJj1gul4zDiEyJy4sbqqqkrhRaS7bbFcYY+r4jJXA2v99+t+UyPuP0Tk3fCbQRXF8v6fstfe+5d+8eH3zYMD+Yc7PaMj+cMZ1NOT495Oz+XW6WS9abDevthsuLC5xPbLuei8tLXrx8kfH/0RF9JPmEKhQKhUgKoiC6RFXUHB7c4d7DA3bhUw6mX+flKvB8s0NOXzI9PqUqphRKYSMUqaGOR0y5T10+RFfg/ZoxXDC4x4TyEh9GCjUnoSFtmcQZUQqS6kmqJbEhpTmJCXl5v15fKb0OeKN32GFLsIHo5R4rF28U1MUb6e3e2e4zx1eQiLh9vp9Mg4UQJJFeu/pXBe/bh4lXfj69WcN54/B/WmKtdKIsoSw1MTmCcEQlmDQN3/7aN5momqqqKZuGdjpFKbWHZWAceoSAfhgY9z9dN3J+c8PL60ueXZ5j/UhZFBBypA5xn+2GfYHSo7R6BVsqISjrlmEYcC7XbnSRkEWJi7Da7RiER2kYkqesatqmQXlgDIgyBzjf/PDrlELTLdcsXUI7j5Kaw8MZDx6+Q9U0zOZzTFlwcHDAyckJX/va19FFwfnVBS/PXyClZL5Y0NQNN8sVwzjyznvvsdxu+MHHf5KL+Ua8uuB65/N59A5vHQFBCOyJCQKtBHeOJ7gx0G0dLiSkKsjxYkAEj/dbpIhst1eEOOzx9x4bAKsIpmJyvMBhCQwUxmC0ZLn1WDmyqxN9kjgBdAPWWhQDRikSimFwfPbZj/8pq+En7c/lzP/ZbR9m7IuNt3HybXIrhUCKlHFLkRBCkdIbK/w2Ir916mkftfOaCxBJiCReOfawj+Bf7cvb22/AKtmf74uqKafYvNpkghQTN9c3jBvPYlrz8P33We0cfb8luIGmNtRlgVYKYsC5HcmOVEDE4vGAyIVKmTE6ASilMzbnLW7sESndXorwAbrBsl5v2az3qaEPWOtyRd5HyrLeF7ccxhQURcnQ502aQkCEiI+eqp7i3I5nzx7x4sVLikIjzwSzecu779+lvLjCuh3XN+dMZg0H8wVSCJQQBG/ptwJjSpSuQAxcX18TbI+3I6SEUhKRIuPQIfAEayEmvPWoVHE0fZ/pvSniaMLBtuNoe4VvjmnbgjoJtA+YqKmEoZYtpZqiUsnYbRCqxPUFN5c9108f4dzA7OSI03c1sqhRYUKUiqRKlC+RxQpUTxIDJL2vz/ATmHZKCessu+2W4PcsHF5HyLdr6nbF3rrr20LnmxH67aOkuEVdXmPkr6A7cYuY/2SN6M0XynwQiSAh9/DLbRaZ9sufJFBKYoTCFDoHPjKgtKKqFYvZjIYKKRTCC4RLTOoWYwxSCryuKAqd60EuM0BW2x1l2aB0gQ2Rlzcv9+SE/FlDCujSoIQmErA+kQkyMV8gBUzaNgdRcsA5izACtMYjcC4QkkXFRJC5kFuomkoVtKZmXk84ns95eOc+07JGRYijwzrP4dExk8mM+eEhqig4PjlBG8NkOqVpWh6+8y4XV1dcXl9RVRVSSg5mM2bTGdPpFKU1bWEoypIQE5PJhKPFEUJIfIygJAK534OBSGK53iGlQhtFWRv03h8pmTILSGVcmyiIeAgghEcqhTJF3tMh4F3Ch8Q4jDSbgareMPYWPziuLy/oxxEnPS5BlBCkw4eE0YJSeYzWKF0ilKbvuj/bre7tL9iZZxhR7PNHkRJivyiyZbBckF4VjJJQ5AuAQMZEEJEocgHtliYlSKQ9BJ/2x5LyBgw631R79kCUeduIlB26TCByDpnfQQQtIiFm2EUqTYqBzXaDFSNtXfD+e+/zRz/8Mdvdmn7YUJYyR8t1gzYF0TtUGjAIbOogDDgvcKNFIvNikQqlCogCHwJj31OQaYLBB8bRcXW1ZLfdZozTWoLNdCrvMzOnKqa4cb13SIKyqFjHdY5AUkRL2HVLjg+njHbL02cDn37yJV/72gdYO5KS4/B4SpCOze6ar77ySKn41uRbGCmpCs2kKujrmm09wftIirC8vma72yGiR4lEYRQpBIbdmhAlbhwRKbDbbBh3jlYdMT1+H4Pk0C255xYMZWSkQ41b5LBl7CKNFFRaYZTE+4FHX/6IsmxZr8/54pOP+eyP/jNSsDz41n2oLHIhEPEeUWmEgio6qnlHKJckDEoYJOW+sL5ff3vobhgGVqs13mWoLr3J3bs9+A284yegjz+98PJn/eGNI372mATEvcMX+xpCeLWabx16TlmV1uhk0EajAK0SUkNdKaRIBBcYRouzgd26o5AlspYooymFoRQGoQ1RRILyRA9hLolC4lNkN6wYvQMt0cJgk6Mw5b5kFcEKYrT7WCfDMpOmRewLrj0BpTVJ5Aw2hIjzDp0CgYAPAZsCja44aGecLI54ePcO90/vUJsKIzUqCgbvOb1zxmQyo2lbQoST0zMSoJRCSsX9e/f54x/+gBgTBweHxBA4Ozvj9PiE0+Nj6rrCxYjSCm0Md+/d5/7JHYZhZL3dEEQiBIF34F2mBne7NUqbXAuSFaVIEEL+rFpQ1iVKCryXOCRCSpzzmLKibmYoFM4nnE04F7HWUVyt0KLGucjV5RXLm0uqIhG0wiWFkIkkLUpGgpYIE2gqg6pqkjAEF3/uuoJ/Ac78NgKS5OJCipliBRKnyKzjFJEp7DdXQAqNEhISyACCfVEnSZAalPqZXZYkBAHhjduCAMKhSBgfUT4iQsKLkqiLVzj7vPAs7UhSCkxNUJGuX+MAqUrunj3g97//KZc31ywOpgid0IVhdnjG4D2+Exy019TOoYJmGAQvL1fEYFFUaKmwIWKTQ2mPkgklErYfcmQfBF5tefrkGZvVhjDuiN5iR0sIHicEzgZEbNAi0m0Ddlgz9A4pJYcHhxilIcDy3HJ0dJfC1Ngx8M6792inJTEN9MOGejLh7OQu5+dXXF5cc3P1B9w9u4tzI7vthu16jVEy8+xjQKaAfvgOn3y8ZBg6ogg00wqA6CxhCFi/Y3SOFy9Gvvoi8dWPQRa/iWlvaGrPtFVMyilaHSOtJpaWtX6E0OdI/SUhOVbrI/793/nfIZOj3y65ePYIbEdZgvVXmMkS98El1nwXWbZIbVGbj7kznGCmc0x7QtN8iyo9QGBeF9fJTKNut2O13OF9QtXyNbPqDXzjpyGVN7A4fuqon73rT7n/TVYM5AuISKCTyhmgiCAi+b83oZdbfEagjESnzLrRSjCf1BRFzXwy5fMvvkCPBSKpzIcWAh8iBwcHTCctWuULrUIQfWDoOp5eXhC1wZFoy5aTwyM24xZd5IL88KzD+pxpkaAoNTHmLDOlsKdbSuqizhTLYAkSVtYzhIDrRrASlKOuDHXhqUkUsmTeTLl/dpfv/NJ3OD46Q/iIDDnIO9/sWBwsCCkyuAEhC4q6epUC2RQo65qTk1NcCNx9cI8Xz57yja9/xMN79zg9OmQ+rbHA3ft3+eVf+RUOpodU0vD8+VOePn+MTRa/eYntEikImqaiLPd4r4AUHUJVbHcjnfUkrflL33ifzc0NV+uem84hy4Zh5VjdbNguB0RSjGMgoZEYqqJAxcTq4go7BPrNSFuWCG1JKFwoMAJSsIzO4izoRhOMots4Ykpo9Yu56b9wZx73C/g25khSQJJEBN4nZPIcTCqmtcGPnt22g+SJQuKUIOiMlZuUF5HC5kJ/TMgIMiaED4joIUVqLfEAIaCcpxoG4jCSksCFwGgd6d4D4nRKVBIlPH/jX/+rXN5s6G2gHwPLVUd3OCG6LXXb8MVXz0CW6EISkuJ6ueXi8hqQNJMZahK5f9Rwp5ZcXK8xT6/YdI5xSEglM4wjEsFtiNEDirKAbohYG1AionwiBk/0FlJESfm6YSl4ppMp/7W/8V/n4vIlu92a0fY4P3B9fU7oHdZbghe8+87X+NVf+TWmkxkpJZarJWWpODs74fDokKJskaLl4vKaTz7+lL/7d/+/HC2OePe9ByQ8u82GwY6Z358SKUQKU/LRR99gvb5kt7tBCotSiagBEShswsaUi76mxcl7tK1kPnnIpKxoigqtRqRcMLBiE254sXnCsHtOW18wnbxkMvkuZw/O+PQHf8jq8pwQPdPWUBWB/mrLl//oGUX1NRbvNdTlEU2pKEvFyCqnv6nDsaQUx0Ag7aE3LRpG17PdLVltboAMbggh9ynjHjuHvT9O+SL/Jl3lZ+w15/yVwxZvNqe9ceQblNpb+EWg0SIhkwU8QUQsgpRULjwm9swvwR4YQCRPWUqUaShMQ6lLlrsdsRtIQYDSTGcTOjeSNmu2Q0dMkeA9RkhSiOy2Oy5urklCg1ZgYDFb0MSSgGMMI0XlSDGhkIgocaNnu3UoqTGqwChBqQqMyFlu8p4gEutIbiojYUdPwBFDAi/QsiDXMjRV1XJydIaQJboQqATEgDA1vXVEscVHj6knqFLnmldKhJgj9Pc/+Ih7777Pvfv3+ORPvs+Du6dURvHi0RfcvHzM6XsfcPfePf76X59zsjjDbnd89sUnlBPNcndFum5wvScEhVYBIeM+6xAoIfnw3gO6RY8k0TYF3zxdcBkGlmXJtVM86Ubi0HA4bVlMWoRQbDczrI301rOxA0o7tHLMa8OHRy19H7hYd4xEogoIIs5GRptRg0IrtEpsdj3dEPD/UkbmKZHcSNpTs0JSSKURSkHKeKEfHbKEptVU05pr22MHS9xjWkGQnVsAGRImeUgBkfZX9RDBe2L0pBRROqdCwTvS6Ai7kW60WKMZpaAjUXUTRrtlcJZgB37z177LphvZdpblpufLxxdcXVSIZCmMZLXt0brCe8F6O/LJZ1+xXm/3nZ8GJWvKGmYzCEmwGSOHFzvOzwekBCFBaYG0nhgsQhiUzqkjye9TfrDDkD9HCLDvtBR7PDPFQBpGaiHoncfverrdmnHVZVodCqUr7t9/yGw6paorYgxMY5NpeiJnQ4WuCF5he8dmueXy5SUf/+hjJm1F05YE59htVqgiw0ciCYbeY0zD0eFd5vM5m+0l69VFLuqKRNvWKFMRo0SW4IoeXW6pzAGlPMSkFu+/IIhrunDONj5nm3asu5HVpqNZO05OJ7z7jUMuz1uGzmCjpaigKnOTzW458vzJDQfvjZRmoCwLWjOjTyHzh4Wg755g05YkczehAVrzHtfXz7m+eUbXrTP7aY9l5wazWzrjvmL6KiK/Zam8dsa3t28LrLf/fo2V55oL6c3jXj9PZgblQAYRaFNCpoSLiTEmvFT5srKv2gohUEIhkyQFT1UYQshdidEGkk+EANZ6fBxJEpQxbPsOITKbJcVAZQpEItdXnMOUmU1FStRFTa01Y9iRxoHFvGL0HqIgeRiTIJWJup5QlxWlVjT1PBfoyRGmi4EYMv1PK82YXG5+igkpFCdHpxwtzrh39z7TdsbYWc7das93T6TgudoOzCclQil0CMRhpDQlWpt9B2surTXNhNYYjk9Osd37HM5bwtixfDmwvF7THB4zOTjhwf0HtMUUW9f0rsMyIs8l63UAE9Aq95QEcidoigFJZF5J7k8OmJSaWSmZRIuZGA6mhhNRYC82HDf3uHfUcrJo0IXh2fM1/RDY9SOXmxt8WjHRgtOJ4v68ZbkK6Bi4cZae3MySoiFhAIGUCmt9LsYiXxWwf579hTrzGAO238K+yUTKTJmSSqG0QinN2K2xYiTqyPToCIunG7f4fkC6iA1h33GYF68IIzLY3EyUEtF6fPD4FAhElBAkLbHe460j7UZW1tHVBV2h6Y1m0RR0g2W73UAK/Ld++29iI2x3lsvrDVJVKBEwKkGKLDc9CIMPjpvVjq7/Eh8NMd621hu0lpRlZDqFgyGymK+5vn6JlAKtBaZQ6AGcD4BCSP3KseQSaMQNPXHfRh2Dz0GiVAiR8M5y+eQRKXm2V1esV0s22zXj2KO1RhUlxigWixmIiHdjbkGOEe8d6/WOGCT9DoYu8uXnj3jy6Cmb1YZnT57x7jv3OYwzgndst2uKsoCkkAjs6CEZ5osDyvIIBJyfnxO9QwtJXVeUtWYYPWjHLl4w95LoDglpgfWJnms8HVt/zjZc4qRniBK36dmub/DxCw7PGmYHFcvLkjj2GJMwhSQFwTgEXjw655tujZQabQp0WWPUBCMEyQ1stk/ovSUZjRGCFoloLM9fPOHi4jFDvwVyg1OmUr0Gx7Mff12ABN5gr7xR6PxTQHRBZrOQcj3ntRO/Zcnk46QUr5BxJTwTEdEhZzWblHH0lER+WynTbrXUqKRxfqQwZWZguCzxEEMkJYVznm4cyOFP7oiOKTE4j0jQVBnq886htWZS1Uit8dFRatCVQTqHD4KjWcPgPMHmzs5KRKbVhNnskLZpKZRBJMN6vUGmRAoWESxx7EjELE9xm6IkgdYF9+7e5+H991lMZogkefb4OaNc41OmPnrn6aPnHebEGPDOIdWICIKmbijLAiEFMaTM/iprmsmU+w8eUBvJdpk/8263pe86ZgeSg/mc5BVSCY7OzggqMeLYfOYReISISJEpgjHEPU4eqIXj3aMFx5OCqQ5sL3fIRkFdMRY1F95TF3PePak4OyjQhUQHGGxi2w80ZiT0Ow4Lwd1ZwcOjgusaQrQ82yWuxsQYgahQRoOQSBlzIVmqXI9CAtuf61//Qp25HXpePPoxQgic8ywOF1g3IlKiUob7x6dcPXrEk68i60mDf+chqd9iNxtSP6CGwLDZcrVc4r2jEIopIJyjNLlqve12rO1ALxNOgPCJMXq8SozJs+o3rBPchMQOwaAUB5/8CYwjzo7Mpi3T5t8mImnLgtIYLq42XL58TgiOmCRF2bJ+dsOu22VNCBGZThdEWSGSQFKghKRpcuGqLvZaGgJMoajqkrou6PuAQBOCJLhACgEjI1J48D1xv4lDhOAzO6csNClF3DDw5cffxzlL3w845zKf3mdMPVYBpQuePX1MioG6rlFaY60nJcH11Qs26894+eKG9Wrk6vKK1WqFdxGjDc+fPWMYtkynFbvNCu8qjKmQoqQqWparNUJoZrOGtj1F8AXWDngcRVGiVKb29ds1j3/8MS/0D3n/5Bnz9gFFdchWfEkQI0SPiIEaA4sDeqaMa8v5589YPb/h2aMXbDcbRLSEKOhHQYoSHyLXnz3m8vET2ulA2RhurgXV4n36uKbvX3Bx+TFX/opCHlNTMk+ReDLwyQ8e8/izzxi3jugiqslfT2aN/CQv/E0O+K1zflUM/QnY5dbNsw/0b5krr0ueb/QY8erAlEA4GmU5FPlcdDFRioJA1ghJCKKWIBUFFTpW2BgQURJtxI0eO3iG3uFHxdAPDH1HdJY4WkKMWB8YQkBIxaRuKLVBAMenpxwen6K1ZrtdI7UnJUchC6ZlzSR5nISRRNSaxZ0Djo9PaJoWrUtCENxcdZSyRCOwQ4eXDr/zDM4j021rVNYrqsuGu3fuURU1L5695EeXN6wuVxR33ieWDTblrk1S4Fc/vE+VLG634vDwkNODO9w5vcPx8TG60IzeMThLKgqU1MwPDlHJEd3I4vAQCFRFlYUPYiLtIaxmPud+2yLqike/u6caI3NZwLPvQJcICdJb2rSj9jtktExmDSGN6CYwaSKn/YSSluNaMpeO7c0Ncbulqer896R4p73HgQfpPW7XcXCv5eDhAY8vCj57suHLl5YrkdBSgSwIyYM0VMYQg6Tr/C/kX/9CnbmWcFwptNZoXXB8qvF2JPQDwm5YiEhXrvGDJ9ws+XL5lLZpqKsGHzxfvnjOtV1xtVvivafE0KqGECMyarQs0POW1U5m52Yt9D3bfkQ0LRQar0vatib1A8UwMljLvLYk5fFFpDWBL7/4caZB6YZSQVsqvvX1D+iHkZvVmqsfPyH4hA/kFl2huLy+RIqscTGpSiRzykIRY0ld5fZ3rSVlVTCZNDRNyW4HcedJBJRILM5miOAJEWwUWClBTRi6gRR7jBHsc1GQiaI2FK2kbAVStbRtg9EFq9WGiMSUNeePP2d1dcnB8RkHR2dIXeFd4OWLa7768jF/8Hs/YLtxpBQpS8PxyYKzszNijGy3W2LM2PvQO4IXGANNbdBqilS5kNa2Uz569xtcXT6h2y6JNmBtnzONpNhRMKx7hu0XCPkJQVqaVHFwf0E9qajKkkpNOCpLxFHDWDle2K/4z//4c86fbUnW0hYJbwXCCKpKMS8V4zDww7/3H/Pk8zmH9+d8/fSEd799h8n8AQeL9zisPmRtB1ScoFJAyxua2T0qY5HuM9zGIcucEQopf4Lf/Wb35y1kzb5wn8lYt4JPIuf7/GTE/tpr/xmk8ZRQRKrkOcJixi0+gpcGVI1OkpD20XsiC0FFjaKiMlAoScCS3EC/GRn6QLeJOOsgBiazKYUpGZ3dd4UKgvWMaSDpgNEKsdc3qqsKJRKRAZsUyQ7EKLHbgYlsOJ0eUrULlNGsb67ZXr7ER7BBI2gwUjFtWpybU1cFaXlB3w8QQIsCoSRaGxKCx0+e8NnH/4Criyu6TUcMivf/tTlxlvBK47ynCZ5n59dsXzzi6Y9/xDe+/S0++lrEtBPagwOUDVgZ6YJl2G15eXHF8bymLiSqKCiblvdmU0zVYoTEDz1oQ5CCkCTIgsODOwghCTFkwa7g0UnhUUSRnXu0gTJZjFR4pbhxUBeStgaqAGGLHS6Rk4a61Lh+xXGwzEKk1YJ6AsIkihTxPuKiYHLYcu/dhm8OB/zai4Hf/eMNv/vFiqtNYNf39N4jTSaDRBfpNuMv5l//GXzxn9uUABEG+nHEujVlKjHCopxF+sjAmuBH7OBxneem37E4u8u7d++i0awfP2cZekKRsVk7OFzoSTIxq6bMjufMz+7RDCPr5RXry3M2yxu6zlGXFUZNwFRUTYEyCqEi/c0WgqZUEpUiftjxB3/w+5ycnnB4cEpZL1Aicef0mK7vEUIyaZcUxRIx9jkNlpKYIn23odCSWk0weoY2ArseWa2XbLZbtNL7QmamPjpnGccBgWA+a7h7NEdFzzAEVp3nqovEomQPmVMqgXdD1t9IgSF0TJuGsm1zQTV6SDIXihIkHItZzWq3Y71eInVF3Qq8i6+UEzMel92VUlBWBQeHcxCOmByjG7KCnmcftkiKQlDXRe6AKwqatkEc30EGzwrDbrNlPqlRUlI2jqTXjEPPpRcgPUl4uq7ARo+WChkFKeYsSMsW2yVefnXF+fmGsXeolIhGkHsL8jqqDLSVYXCO7cstdogUzjAUv8vB0QdMZw+YtHc4qFu01Ai2ufNAC4LoCSm3fsqUWVL7NgZe4edkZy33xbBXgIu4ZZ7f3p8fvO9xy3RXIYjiFQFkr/siEOmWa557kZWIqOBQtmN39RIRR6hbXF0RhcZHSYivBbiSADs6fPBZLTMVyBy/IwSMzrHadng7YhQIeQBCIZXBFFCpDNkVSmUtEgSVUpRKYBQEkdhsNhRNiZElnhrvFdIkaq2oBFxcPmfVX5GUIAlFTAZCjzYNSSZMXSKrsBe3ixAEWu87n4sSgeTmZsWzp8/YbjucCwhZMJA/oxcJLxNFiFjn2G62XF1dMYwjQQm8EngRM4vGCHz0DIPlydOnVPoeRjUIodCmoqoNwUGwFq8sRuvMGkoQI5k7H9PekYdX4mFRqNyKrw0HTUFdgCk0URv6bqDRuYNbFYkPzgpkN3JY7SgTTMqRs0NJGTzGJ7QPoAsQEakSRUqkzUDcZsVQbQqEVBzOW6RxlL0lrj0haayNBB/R5c+nvMJfsDNPKbLdXbPtN+y255SjoVYREwMyQliXrKyi3wXGnWXod4jFAbGuKaoJadpih4A0WWDI7qvFRifmk4qDsuSgadFVRXQ7uptI391gd5GiXqCKSNwrpN12Yjk30neSajpBS4H3no8//hOeP3/M2Z2HHB/fR+gFZ6enaKXoestiPqeqLpHdDhEjQsmMOQaLiJFgQSSL94LVZsuLiyuurpdAxvm89TgBzo4EN1AWhtm04PioxaRE1zuE7tneRrexQCSBlokYxxyYS9CVpGoNpjD44Fmv1vjRZvXCEIkhMF8cELceZweGsaeoJyCgKA3TacvxySGCDdYNVJVhMqmYTGtiUgxjYNy5fYt35gyDQylJVVXUdU1ZllRVBWFKV8/pt44+BeaTU5qiIjXXdPYcv+kYAkQFKIFdRbqtx3f5wm2dQE9LtDKEAW6e9ayXG6QPSJ2ISe4VAbNTQ0maWQO+YecT65vI48nAdfweJ8ue0+PA3TsHHN45zmqAUhHigI0r+uEG67pX2i2vI/L0Cil5hY3vf27bF16H2W9I7b5RAIUs2nUroXr7vJmAK7gFaiSRgkAaO+zNFatnX7HRCXN4ijyZ4IqUJVNfKSlCIhKcJQaLkZl/HH3Y11USw2jZ9TuCH6kLhVQSXRbIqJFBI7zPFxGZ32P0YIRAEZFEILDbbjCFzp8vFcS0124JjjB2rK/P6VgjSrPX0dEE2xOiIwhNkgm1Z+DEmCCALATGGLQxgKTbdazXG6zzJLLsayTmLtMEEPaNewFnLX3XEyIko4gyEVLY72FBjI7ttieMkQd3jgmxJsTEaB21Vnjrs4pjlMiyIAmzZwcJCJmdlSWj8+8gAyAplOawmXA0LakqhSg0Xpa4sEYUkERERM/dVqKFpBIjhoBsJEWpSNtI2kXiACQFJstHF1jccmB3abDKcLWFq61DKMVkpjBtSVCa0eULR2SkkP8SslnsOPDJ4x8y2B5hN1SrEmJuknE+EkLFEA2dd5lHWhQ0AW66kVk94eyb73K1/RI1BMbBstls+fTlMw6JnA2OVk/oLzu6ac24u2FYLrl+/pIUSmK6QG4HRiFwm4rprGXoLOubgZUYqaoJbVVT6MD15TlPH33KkyePOTp9hwcPvsVHH35ILyVaaY4ODphMWm52WxwJqRRFnajdlDRuGfst3W7Fs3PDZ4/O+fjz5zx9/oJpU2RaWYyE3pL6kVYLphPD4aygkImmLCiMQsjIbrPCCkenS/qiZBjG/YVMMJlUfO2b75Ji7i7tup7VOnF9s8KOHjdavLNYZxGqxpSKopRUtaY0NZO2Zjpp0FLz9Mk5FxfnKC25c+eIopBsdgPWDqTkaduGcXA4F/ba0VlzvCyzgqKUktFFdgNsBsHGas6KQ07u3KOrnvL0q0/xzxPUPb7UOF0yjpHNx5esnu3YXo8EYRDTIyQDwvfEcchCQ1KjhSYkRQCM0Til2VBSTO/TTh9i12s2u2ueuMD4+ILDq4rT88TleslH029ypN+hFQcYHrLaPebliyesVtcgcsE6C7dF1L6jOPecZbZP9u/pVbfjnqOYG9fYg+AivGK8SJHdIjF3ct5G5WrvwCUCmRKagAmW1fUlT3/0Q55//H2KqmT2zgfMbIE4KhHGIFWGQrK4sUbFERN7CmC1WjH0PV3v6XaR1WrH4DqUjKhCUZQFR0dH+OAZxp5dv8OYrFfurMe5rFs+DB1CJoQSjNYzdH3O9LxA6QYbVlzfXGdqorfoSubMI3iIHjtEfBpxKAZAh31nd8qqnkqrfX05F4KtdTgfCD43bImYSNYi+y1aSxKBsqwoZETGQLSekLKgWPSW0O2IBKYiEvuO3WYDrUZqQ0yC9a7jy8dPOJw0zMqGLJi14sgo1HSBQCMTeG9JMWe9mU2nSCJf7OZlxYcnp5zMCoqmpBeG616QZMoZgXe4ZU/qJGoI6EpjJgXNokDqiDM9Y0pse01IhubgEBEtbnnFbmPZPodnQ8ePr3Z878s1odCc3Dvh8M4Rs3sVztVcX75keXPBcnXxC/nXv1g2C4HKd7R9T3W9w1w9xxaSvlCstOJiEARq0BpVaIIoWF/t+PEffp+zgwUf3juhqI55dPGEZyvL1kcWWvHOaqD4k8949CePedEW2G+9S5oUDGFgEAnnR3aba8TYI6qC+cEp677nZjewGiKTSUs3OpRM1DprMizmU1wMPH32lMePrrh79y4+Kdbbka7rcONAIRWpqBDKUJgps3rC2C1J/Q0hwPWqY7XpGcasi65FZFIVHMxbZk3FO2d3MKXClIqqVIhgKXSuYGstGLcld977JhebwOdPr3n86BopE7IxVJOS2mgODg8pigIhJN/82kcsb3bcXK+5urzm5ctz1tsd8+MZk3bKfL5gPjvgYHGIQDAOlmmzQClD169JKXJ8ckRMkdVyhbWZ8reYH5Bmgr7v2e26V8MAhJTUdY0xhpcXV2ycZ+UiHZrn6y2puEZtexpfstoGjK/R4QG6eMiL5z/m8unnuNVIcg6hIsk6ojK5HV0qCDtcjMgQqKIGBbJRqMV76MNfws1bloWji56RQOwaprNjhPas+iXjJ2uuxiec3p0zO2qpFy1hteby8pqh9+hKITV77fHMGXgNi5CVB4mQHCEF8kwHhZQFUWRlvbQvYyjyzytxWmexuw3ddsX19QVlobhzesJsMqFQklJEjHeQRlYycEXE9QPbmzWuvsTvPNoopClQRUkxmVBN5xw0gjtK0TSaCsVL7+m2AT9IopPcuXNAoQRGGYiRfruEFEnRouTI+eVLun7EOxBUuM8CTiTu3L3H/OAA08zodjuEiCAjwUVEyJFz7qo24CG6PIylH0eiT4yhwwsIpaY6WKCNQam92mDXE1Og1DWyzHWluixRlaZpWg4PT3nv6JCoBEjQheab3/kWRz5QdRsuHn/FzWpNt97QL1dcbDb8/t/9T4jaEyaHzE4f8PVvf4f5dIZSgslkwte+/nVi3zHRJcNuYLlec3V5yXEzyzIIIUJw+JC1klISCKnRMrKop9yZzrjTlggBj1eSLy7XfPrsgl//qMKPmjj0+F2HGQWTxQmIDAs512FUZBciQZdUZ4cMtsOWGlW0qOmc7bMdv//jc857zyqW0N7BJkeXptTMOTi9y/mLSyIOJSOzdgqc/1z/+hfqzIWARYzo9YB6coO6WZLqRL1o4HBKPJiwsYJCGypdolCYzZY6CerRopZrZsstYtnj+kAQmmbwHKw7msGThCSYCSgHWpO7jxNBSbQCbQTVpMBUmu0wsrM9XkTq1gAW5yyEhFaKyaQm2MRuu2O1uub3f/93EbrCJY2LWfJyllp6Gxh8RKBRWlDVCVUammmLijcURjJtCszdI9pKMp/WzNqa2aSmMAVSy72GMeDzlJTbiSvTScPdO6cE0/PsZocyJnfHSkfdVDw4u8/iYLJnjuSfs8PA9qznxYsFhTY8e3lNP3guL27YDYLdLrFZ9fncjJbVzZrr6yXjaEkpcnl5hQ+WbrdBaUlZ1hRFtdcwb6mqHdvtlltNesga59N2wsFkiu8cykFylug9UkmIC2J9TKpGhKxIo2PcCaKZgx4QtkOEiPQ7bicHheDJYVvmaqcQaeqacnYPmiN6ofBB4QZHCBJlDKasEWbH6EZGF9hGwfbRS3aXmnrWYA5nNFvPdrnBB4fYNx9LkYvYUeiss07CBEeBo5IJKTwuWpwPOBeJSZNkpr8lqSiVQonMf08x4kbL7vwlq4sXrJdXdP2ae/fPSJUHJiSZmRIiBVrjeHD/CPwH3OwGypM7NPfvMlZzfAwEBElqEIYkFVoESixFUhy2hm5XsduMRO8xWnKwaCAGhq3lxfMn3Dk5ghRwfmSIuzydSkqSUjjruF5eMbucYaoKUzeYomS3uUEQEBJWN2tqOSBlps3ZMCJ1xPqYz0eMVFVJSh6tBaItaacttS4ooiTaAAHKpmAxmfPg/kO+ee8DfvPbf4m6bJhOphwdn3Bw9x5J5aYtIRLH8ymrFy9g934OAg+PmTUlVZVpqBuleXJ5zUxOeSg087bi8eefkrxDSUFRGkzRIk1FrRtEPcFpkesPe7leF0a8CwSfKZzIXGRudMGkKDAyMCbNDx5v+Px8w82upyoWCGcxQVDJkqKCMQRQJYgGJSW7oedmaxk3CZ02vHPPIA4MfdK8uLH8w0fX/MMfvWQ9RILURJ1oDmts7xh2jo1ec3P5nO1qiRsHlPiXsQNUgBkdejvCekBYj04eWUoKW6KmMkctUlNJg4yahXMs+o52HAijx/nAsB2wMYJRNC7SWEvlM+5Ua0EykmgkJE1d1aA0xtSUdU07nTApDc5bCg1lKZk0irLIkTMp0g+Ouq0ZR0/XDdhx4MsvPyOqgqQqyvoIVUwpq5IoPbbzWZpAGISpKaTCVBXSbmibkuODCXFWYmRk0hS0TZEdvDHcVs6UEpkrrQ0+RJz3VGWxF/CpKOoZzfQI7AYhOrRWhDGSXEDKPKaqKEqaEpQQDF3PfD5l3TnsasjdoyGQEgyjpd/1bFZbLs4vublZZineFDm/uKAftmgladoGyNOdiqJEa/2q2eW2PT6GrGDZNg0H0xm+s4ybDilgPm1JhWK1mRJNQzRbRNIkPxJspl8JpUAkFAktM7MnRnIX4B7GSCnjx2VdUUzPcOUJlinJVQR3RRI9SIcuOnza4pzHu0QSmvFmR1xpihuHuHG0u5S1omOGAmKMmL38LWShNZMiDY4Gy0SBFAEXPWNy9MEy2ohNMUMfImP5GTbIvQx9t2P9/BnbmyvGfocpoE0LCrdFjQml5b7xJaIVTGY1R3dOEdseNV9QTFp0PcEniU9ZIjVJlUe7ERHBEb2gkDUaDdESvKUsoK0L7GjpvGfsBpSYoGREJgfSU9UlPhpGq9hsE13Xc7O8pKxLmskUow3BB1J0KAXRJ8J+CEWSuVgYUsCH3IUplEGXBmMKRKkRbUlRV9RFTSW3JBSnsyM++to3ee+993nv3ff4pfe+RqsrmqqmrRtmsxmmrogx86uHbkd3dcnlbkPdlLz/9Y9oD46wmxW4QNSCwwfvcJEkqWgIZL32q/MXbJZLqrLi/sN3qGYTHCKrXRYlt4qlt1r1t70bMb7uSI9JUKgsnFfUBZed55MnK54tO3QJhSqQw4iOEi0KkoC1jeiyoixn6LpiXF0y2i297ZGxI5VzLIHl6HiyGXm88Xx107PeOEiCZprQTaLfblBCMQw9q+sLXL8jeZ8r/r+A/QV3gILb7FuOlcQcLVDDhmJMyJuethg4mh/kyR9EKqU5UYK57dHDwOZmxaMIVwlGXWB0wURItE5oISmqgpNJjSwKvC7YoegP76Ap0LKgKCua+RRpoJ5AIYDkOZoqJlUWCgo+8vTpDU07Ybvt2W56mqZhtblm01l8KpgfBQ7PSoQukEkjDQxjTkOVMCQJEYVWkoP5FLXXA7HDlrYuaKqCujIIAdb7LIMqFE1T5rZ559FKIqXm+csr1nFGMbvL4d05af0MwiUyJL73hz/i7G7DyekBB4cLDg8P8T6yWd/Q9xu0gbatSMpQTQ5ZHN3h5OwdrE08efSU1XrD559/SYiJmCDEwMuXL7B2ysHRAY1MpD1LxkeHlIqiKql8nWeBikwBtePItG7xc4vbDZzHFzRNwdc+fJ+d7Pn80xckY3B+RMVcxI1uQMQRgUPIiFGSplZY73MHoYiQ1CtaXggRNWvR01OiegfhTpHW4Tf/iFBsoIkIMTKOCec1AQ2FhpsZnZowekHsRq7Wkp0Ve5XCDCPIfT6gU8JES4NnLi1TFZgVuWknCs2oIxsfWXYbdsOA9QEfIptdz269Yeh7+q6n22xwuy1tWXA8n3Ln/h2Ojw4ptcIkTyVMHkgQPC5ExpAompp5UTMozdhvieUUM13kBqi0V2sREZH2YlphQClNv4sMncO7jnaiMFpg+0gMjkkruHNaUxmFIBDllKJp2A2S1TZyfjlymRzL6xfE4CjLkofvvJ97FZwlacF8uiD1N7lzWYGQit47hNYUxqCbPG1qcjBDNQWhULgomDYT5k1PExS//Wu/xX/3b/5NPnj/fY4ODymLrLGS93lCRM+w3THsdvTXN1w+eco/+Af/KdfB0RyfcPbeB9y9e4/H379m9fwCP5vzG7/x2zz4aMNnX31FDLkRzg4jz58+o6obTu8+wFQN11fXDMNACIG6baicz6yyJCAKlDJ5ncWQO0CjRCiNqWvMfMEf/Ogr/uTRFTYl3nuwQEeDieT6QVQsh8AyJA4OWg7nZxSHc9TG0sgOrTxJW25c4vz5ihcreHwVEe2CpGt2dmTsBqrWEOyO1Y1ls75AmoLgXa7VRLC2/4Xc619wZJ5LQcEU9PMF/cmE4/WGZr1GL3dIsUP5iuhWKBc5bBY0SrPaLNlsNlztOl62E86LiqFoUKNnsCMXIjIawaTQmAjpagtjoixKPrrzgEIY/GAJ3pO6HucTk9YgpiWkiolJ1KVBKoG1gWF0rDc7lssNNzdbrBvQg8InjakLptMJMQZ622GjRCiD1CpX4lNWFtxsdxw3gemkoFATrq8GkpYUhd5j3PmUOJuHQFMKmkmd6XJSYl3Aecnz856hOsLXd5mfCep5QeFbZrrn6nqNLHtmhy1Sa4qyQenEwXGknEyYH480L27YWSiqGfXkCKRgubzm5vqa9XpNjJHFYk7fdziXBVilBCkiKTm87+n6DdIIClPuNUygKHIWIZFEH1FSM1hLZwecjMiqyOyDtMX7z+AqMayvoJTEqsa7RKMilJoYS2II1E1F6oasRSFVpvztOyijUix1okg7Crmlrmq0/RR/2KP1PjsZ7tDUvwr1FhlXyGHNTl/TuZHgE1IFOqsJKETKU3okMg/k9SN6XMLqnBQGegJBRNYxIBI5MneWXbdlt+sYB8s4OoZhYLPJtFNSQvpAGQIIidKGiGK1GdiOFygFhVG0VYFSkpGIDZ5xtOAgCkNQBbGeUOsWM1mglUDLSFUoCq2RG9juPEJCCgObZSTYxGyiuPfeKcuVzTrdheL9h8fMp1DIkPF8qUg6kRSMAnQMHM9rugGMiQzbG4btIddX5zhrqeuadx7eY/AD9b547pSllwkrFckoTGsI0VHPZojSMEpBGD3zyYx3/+rX+NVv/wr/zf/qf4PDYoJJeY5B8hCloLcdm80NL59+yff+/j/k8fd/xM3zC3bbji8knH70AR+d3qedznj69Bn/+X/2e9ycX1GVFb/9b/4bfPu7v8pVWXDTddgo+Eu/8Zt89M1vMwwjPiTW2y0oSdnkIR1XV5f84I++z7RdcHJ4xrQpEElmeiwCLcAKzdPznuXqJT/47Ip//PGXSFNwMCmpYmC3HmD0jL1j3QcuXUAtato2UswscfyKSbmkPE64tqG3NS/6Od/77JqnFx2Xm8AuaKwfKcoMYx4cTphNDLowefLYvvDd9wO7Xc9q8/O7P+FfgJ65qDSpyUOIL31kKiSNkGiglpK07qisp/KBYhu5Gjquxh0rO7AJno2JUGcq2ND3KDugItgY2cSRUqwRLhK1ISlNWVYomSNIlyKu1KjjGfX0AFMUzFODMTm1yj0ukslkijYmD8kdMqOjiCXKtFQywxmFKairAhclW5vQxDzlPClKQEmXN58KiGgoTEHfbUn7cXimKElJUJQ6Nz1pMEVBoSRKKYYxEpIhiBlRzhCqoZCOg2ZKqzXHDXx0d85skfUudruOrh1o2yltO6GqG2ZzmMxO+OyrC27WlvOrF+y6xNXlkuura7puS1lKQugh2awLEwNSSJRUKKFIkVzwtDklVEqj1P6ClCCEwNDvVQhXS6J3HB5mtTuXslNu6xoparoiMPqRfgwk60muQ4TMxnHOY60j7KdCxT0M8up/UpHMhBgT1n6F50fU5Y5Gz7PuPQbLGUFsUfocWGUtmdpghEL6QHQ9u4tH+OWG0PkMEe2LdM4G4mBZXawR3QYVc5t3hnvyFHUffdaedy7DDQGSKCmOax7cvUOlFDjH+vqKp8+eM46WYbVkvV6ByhdBJcFIhdYFsqkRhSFpg5RFZnA0E6rJFC0lsVtTthXzquB4XjGrKsxzhR9jnseJIoSB0iQm05pJqVm6nkrBfF5y57Cm1Cn3d6QsLh18wPYj/c7S9zveffcMqSUog9QT6qbCGIO3A872XJxfk/ot1fGMqmooypZUeMqmQNQFptQMQ48wGk/CBUdIgq99+DXunr7Luw/f44++932GPmu3lNJw3ExRQ2B5/oKL58949OWPeXp1zs6N2GlNOpjxwdE9Tt59yOmDB1TtDOciiw/ew5ydYIxhnGhWdOziQB9GBucwhzMWRYN1nmG02LHDWZv1273Hp8im23J+fsOnn3xJYXLAUFf1Hn4JeCc4v9nR71aMo2dyNuOXvvEux7OaVkVuokU1Uyg8pg7Mkke1AudWXF2MVMIxbUpE25BUwm8lS1fz5bLnq4sr1puOdj7nwbuLV9ord46ntKVBaE0UMksETybs+h6h12x2/5yahoQQD4H/C3CH3FfxOyml/60Q4hD4D4H3gC+B/05K6ebPduUJXwCFRGhBFRKFCxgfKFKiCgFtLW1IublKBVbbLZeuZ4XHlQJmmuqgJHjButsioqUqDVlPTZFCxN2sskBPYVBtQx88AYiFJtQThFEkJVFa58HEJhekEhEloWnbXHQj4WNAx4wTaynRxmRnrguqZkIUimIMbLUlxBzplUJQFh6jO/B58iRkzeSQ50NRliUxQETn4bh4pNKYwhCjQOsBhCLJzJYxUtJqWMxmzNsDjmooypGm0YTo6PuRzaajqiZImbVujFC4ENhuOl4+X3G5HNls93WAoSelyGRS5mG0KXPgx9GSQktuKZGkkHDW4Vzmm0uZ2Qhaa1LIOhpjP7JaLgneZ90Mo7le5SlFQW8xIhKiRxgFyRBjILktfuyQ3r7i+lrr9nMkcyPOa/3CjHsiKtIoiGlHSpdEU6NSgwySFCVBBaJfokS3xz8rJDWliqQ45uHa6y1hZ4lj2Hdw7oWVkiTpEqopMSpS7uuGFHIDf8oKgGpPLY2vxLQEdaVpjw4wJOI4UKaEWq4I1mPtiM27NjNfRN7AShU0uqCsJpSTGUqVFFVNUdfopsnysn6klSWHdcGdacO8qdgZg0sKksL6uE/HPZXR6JQ7co2AaaVoCpnX3r7xaPBpXwcKDINltB1aQd1myDBIjS40R8fH7ErDOHRcXV0zNVlHXyhF2Ta40hMbjSg1SutMHxQpN5fFQIyStm6IPvLs6TMeffmYMUWU0tRFxf12AVdb3GbLdr3i5cWKTijM6R2mixmTw0MO53co5xPmhwfU7YQUI3c//BAXA1pLlBaMODrb0w0d/dDjU0IqjRIKIzQJTyQhtUI6TyMix3fvsF527LYjbnRMFwcQWlKEmDxutUWVCuEUKSoWi5b7Jy3TusAOPVeD5WAxpZaJIjhaPNJEJI7d1rONkkHUFFIjjERUmuubjuvdlq3ridqhzMi0LdGmpCgqjmczdJRYnxhCREVJVTckZfAB+m7gnxebxQP/s5TSHwghpsDvCyH+Y+B/BPydlNK/I4T428DfBv7nf6YzTwmrI6qIFMLzMGrmu4FmtFQxUY4DCxpaodBGsyoLrpdwIWBrFPqwZHp/gVkcMG56dpcXBGGZT2dIU6LRpNGxW95Q1Xu+dpkZKbKq0LOG4nhBsZjmSdhKUZiCpikYbI/3DimhaWp2vd1zW8npz96JFWVJWdVoUTCpG1RR0oSA2uzoBwdJUktN3SS03hBszIMlvMc6Twg5km7rGusCQoPzHuvy3EGpDEplrustbm60pCwShxPJ4eEhR4cLZtqzffkHzBcH+BAJ3nJ1ecNkMqMosu41SrDb7fj8s6/44tENN6vAMAqUlhgDVamYzyq0gvUKorcsb4Y8cDnrDJFCwg6WosyiTMZAWebuTx89wQdWqxWbzZbJZEI7abHes9yseXn+nGSuiX7HbrMlFJGwb5VOrscNW0QMmWec2A9XFpDVCohZWSxrkwgJoSQOEuUlEk2sjgi+JlpL9D1ptgS7Q4wtMCPGCsaWUnYIATIo0jYR+pCF2mTGo1MEoQ1FO+PwfoXwtwNTMq6eyBPSpQB1q1r5yjkLpIiEsccOO3wQ2PYAJteIwYOLSAKI+Iq/npBYpZnO5kzO7jI/PUOSMegkBUHkAb+NTBxUhtNpw+m0ZVYXUBRspCYGydi73KhGQARBtJHk8lCSykjkvmknpID1gW3vWa+2OC8Zx4B1I7tug64ERguENIQUePjOO2zWS16+fMYXXzxl+mBBlJEgAm3bEMrEoGJmhqQ8LyyQu0tTDHmIiR15+eQznj97weOvntKcnlDUNU1ZcV7P6Z+cczhf0EwmFO+8y0FtOHn3PqcP73H/4Ttop7jeLjGFoZlM0FLxwdcbZFkgJVw8/YqUAn23Zb1as9ms6boDkGbf9aoo6oqirvLkH++xfsLB2VmWp+0sl1dLhk+/Bz7vW+8dO/8lRaU5PFGEJDicWQ5Lhwgjy82GHT0ftUe0jaRBUxKIKbDuHMs+ctkpQhe4c1CwqEvKVrNcPWJ0W0wZmS1qmlpQFiPNpGQ2q5lXJaGH4aaj240krUkpD2qfzmYooYCP//zOPKX0HHi+//dGCPEnwH3g3wL+jf1h/2fg//PznDkkikLTikC1tbjlNabbUsnIvDAsdMFxUaObCV1peOY6flR4LkvNaLK64qwP6NTRd44uSZYiMKaemYSJEZgYCYcVHE8pz+aE+ZRxU+bmHlMym5dMpjVVUVBog9E6w7MiopQA4Qhx5Pp6RdcNSCmp6goXcmQmpMQUFXaIXF7fgJQZ/wuBbggQAl5GdlVkArjBMu52jGNPjLlmoKWmLgzOO6SSpAD94Pj8i5c8vHtKWSTKSjBpNB2CyUnD2b1THhyXOAt1NaFWluOTA4rC0K8GNqsd2+1zxtHx7vv3mB9M0BKGYeDHn37Jjz695HoZialEmcTZ6ZQHDw55+OCUu3eP+PSTL/DO020H3Oi5vlziRsvh4Yzp/HaWpKYoDIeHh2ipef7sOefn52xWG05Pz2jbFqM1kUjAcb1csolP2FZXjPUZSTt87Ih+hygSsRMQPFIEposS77OgWAhviF2FgChKVH2Ejw+pZIahZBSE/ghdL7Cbp7jdVZ4raWcM20gKN4Tpjq1+ydQotJeMfcT27rWYVkx76iMIkbt4ZbNXXxF5YERM6RVN71WhQ7wx1Wr/NKpaYGaRIgWaFCiP7rN++YLu5hrb97hxIMWIVJnpZOYTTt97yOHJCUcHh+gAhSmy5ngMSCm4czTjYNIwb0qmpaQpFBckrPPYMeGCoirrrPhHzqC00pRGU5jbrS2QQmO0oikNoYm4oPDCITcDnoRQEJJjt1ny+PENHzz8GhdXN3zy2ee8vL7gV375HtPTFrMwDGKkMFNiiIwu4LxFSbOnbeb5AEMfSOOGJ49e8qMffUJdNRzqBhE1wgrScY38+gP8fE5xfMLXP/iAb3z4EYvZlLosUEny6cefsFxeU7UNdVMzdCNlgIKIFgKtWrRz6G7Ebzd0uxUvXz7DpzzlSBcFRoVXrKuUEs574p5OmkrB9M6Cqx9Eos3MqRAEdTXHjY5xHBjHHadVzdcbSSsTNzHypAukcYfXGnQELHWpSRjWg+UHnz7n0U3ir/3lb/BL782ZTxz/+m/dRzUjj1/u2I2KO6fHeNfjQ2TYrpDjhlpP8a5n6EZGDJ17hlCCQivaqvx5bhr4Z8TMhRDvAX8J+F3gbO/oSSk9F0Kc/twnSFl0ZzqpObrX8OXuE0aVELqgrRvqqsIZzbhoWU5rHm8sV4cVGx+xMZHGxPb5CmRHCDCOES8Fl9HSo+i1oZ1oDusJxaRCaoNznqaqcaQsqIAjxp4UAh6LDzI71P1AX6U03kvWqzV26NAi0FQVq26g6wfkzZLV8oaj4zOsiwxDpNt5pCmoi5oUA0TLantNEx0ypFf0OlMWhCToBsdyvWMMsO1HLq/XfPXVU5Y3N/ylX/6Q99874fh4ipZLCtdx0irev3fE4azg8jLTBrWWtJOG9z54h/VyxYuXF2w/+4rvfe97rDZXPHh4l3sP7nGwWHB6csxXj1aMfR46Wzea4DXB91jXE6PbT41JmMJQFTWQaWspeaYHUw5PTphMZkzaCaUpefbsOS9fvGSz3XB0dMRsPsOYAm32mUStaNuSF8uezfIlTXsfW/QId0NynqoZiZuC5D1KREwB4xAIfi8kKHOnZB7VKJg0kqnZYu2KWJWY+oioDtjufkwMOzA1Sdxha3+IFAqlCzQ1tTpEhoEwjLhdvx+pl1738Ivb4dopy6qKLGwl9k0ygpSFtF719uepRCK9qamoeTXYQmiEgMn9h9QnJ0RnSbeib+TitjKaqiqYzaY0dUVdmCwL8Wp+KDSF4nhaM6k0jRFUWmCUYBwtu2FEyfLVjFCjJHWl2Ox6UogIvefq+3wetdYYpSjLTNVdbke2dgdKYYNjsAOFLDFGM/QrPvvsE15eXPD0/CVCKYIUeBlx2mMMdP1A2DONvEhEFbE21zy8y7O96rKhMAXstVBUkqikUNKwmB3w4JsfcXx2zOHRgpODBaftlEqqrMPjR0oVaCudG6ekQJkCyBzysjAs7t5jfX2Jd57N8pqL548Rw5ZN7xl92lM5X6sNJoCY0KpAaoVQkihhN1iiHUnRgghMZhV2yKP4YgrIKNmue2YTxQcnE96p55i2JMVAcIHOOYYxoEzNwWTCtz8qeflHL+m8ZmMlJwjuzlp+/evvc/9g4Po6cP/huzx9+YxnV1ecL2+4HtbIINluE5s+0QdJMSkp65LCGMZO/UL++Rd25kKICfB/A/6nKaX1T4/A+jMe97eAvwXQtkXWCVcCOdHsDisaGUgUyKIhKc1GCjqRuCRxoRK+MegAKkD0WXksBpf1SVRECwUSfEz0wSOFYCChvSeNYEgIo/ZDniPB9jgJ0pncJi1B6RItDFrnPr4QNbtdHuFWlxqtJCFErB3RpmcceibThq6PhOjQLlBXBdZHosy6FiFoxiCopMbsYQkhYfSe9a5DqoSLgpvNwMXliufnS1Y3a758fE47rTk8PqSqNIXfUaQNJT1GmozbK0VKiYvLG9778B0m8yl3pMAHzxdfeNabDS9eaJQyNO0hZ8czvvbBGbO2QRU1UnkWBw2HR1MmbVZZHK0DmXXWpcnaNVprTKEYR8/QWarS4bTn+uqG1XKFUorDwyNOT08pyxw9SJUZO1KX+DBhE6aYYYapAiEqUpqjVUEzkwxXA8kLtBAYLYnBEgMgRBaSUgpH5mXXraStepQBqTWmNCAHfL9DFQJJiywrpM0FXBkS0o5ZJ0QpXBD0gyOFuGfJ3O7yjCmLWwXEW5R+j6eLV+Ja6ZU/l2RufNZNycwfIbLY1W3LulAlqjCZepcCas/Hl3tK3qTQzKrcxVtqRUwZ3xUItJBMK0VbaupCUWqBErl/IMaI37fL596AgNaKpjZ4H1jvQnY0fq8Nk/YDrfddrUoqbqVJUJIkwAWHDAKpQOvEZruktzuEylOQQhKMPiCtxauekBJJGDzgyPMDbhtvcr0GClGgs0eEBCHrC6OlZDaZ8uGDdzk6WTCf1MybkpJEDGPWl1lv+erzT7H7WQdj1zMMI7tui1KSsjBMy4rtOGJTwDrLi2dP2V5f0o0RnyRSG3y0GaoDEJLkAzplTXMhwYvAYrBoISAlrO1y7U15tMlw5NVNz3OtMLKgrDXzWmNMIkaBF5oxFVmFMeThJw8OJxxNt9ysej5/DKKPHBxohl1CB8lcS+ZFwZU2aKlIJAZnMSk3ZSmVxy+mpEhR4H3A386+/Dn2CzlzIYQhO/J/P6X0f9/f/VIIcXcfld/ln4LQp5R+B/gdgJPjadr6wC4EzqWjOyw5SImd0+wo8CJxEz03Q8eF9FyngGlKqiTRKefD/TBm2l4I2Cggmv3sTon3ASsk665n9COV1zSxoQjmdp8So8UGR1B5JJeXAlO0lCqLRkkl8MHQdSOFhrYuIWW1OhslpcuMj7atCUR8NAgi7URyvdpk1cKiQtgKH3Vu9a8ryqpEDJF+tPTDyGgHxhC4XvZc3Wy5WXcMQ+TR02sWh4c8fEcymTRshh3SX+K2L/F11kk3WhOGwGefP+G9D9/h7OyAk7MjposWXWqePH7GcrmB9ILTEzg7aanLh4RvCRZHRwzjlrLW6LLAOvj0s+cMo0MoiSmzLkwzqajrhrKqkEKzW/eIpLG9p+s6UvIcHGRu+8HBAmst3lsgoRT4ICiKgrKp0bMCKZeoXUWMM5S6Q3voWD09JzkwItPD8oSXuC9e57mvIXl0KWkmiqq1yDQFmszQUC8ZCotSU7RsSOVIQYHcaWSXoN/sIzCJjZJd5/MkqhhfDREi3UbZt456Hx2L9Mqhv+nMJelVQTs/JiFkxtQF2ZFLcqEp7nVftJCYYl9SFrmfoTSSiTE0haY0kjElwr7IXkjFtDRURlNoiVYSJV/Pzogxl/aG0VGYiFaKulIkSq43Iyl4nMtyXmk/VDmLXgU8Ce99bppSClloIoEQHEJIqkqyWm8ROlJPK4YtuCgZbCB1I3gwtSZJSUgJGxwOlzVLpMZIhRABIwqk1LnIHxOOPKRaS5i2LfcOjphXJbUQNOOID4Ft8Kw2G14+fsrv/oP/lPsffcShz/M5t5uB59fn+P3Fa1rVmZmzr6u8fP6cJ0NPiAJlcgPUduhxPncSS6mwuwHps8pNIuCi5dfe09RtDSLR9wNDilSVQZs8MObxszW1UCQFqoJiIim1pDQaUZR00jAGRbAWFTzHVcW9wwmfPF/y/MUVNweKk4OKwTpkgtYYXL/FBwfELDImBLosQUqCiCgX9l3dkRQj/hf05r8Im0UA/0fgT1JK/5s3/vT/BP6HwL+z//3/+PnPJVntOq6Xa65XS7quZ2Yli6g5wFAZxVW/YXcpCW3J/OEJD85OKFSevJOiJ8SWvrcMo6MbLOOQOQa347eSjyx7Sxg8aRMxZkNb15TGUJeGw7YgavA2azwnCSJpkBC9A1kSmbMbIpQQk6TrRjbrHVWbYYaTwwU3NzfsOkFKmsmkAbkihBUhVYhkiBGGAG2lKMqSpm6IN3kUnbUj20ESk2M3BFzwuZVZSa6XI4+e7rj3tOdXvvYAKT8jpkuWFz+mrA6Q+hCUYnSBZ0+v+Me/932+852PuP/glNmi5bu/8kscHx9z/uKcy4sLhnFHXcusO6EN3/7lD2in+eIyusAPf/SYR09WlKUnhNdj0+q2ZrFYcHR0wv17D1guNzx+9JjHjx8hJfyV3/h1ptMpbdtQVRVKwTAEnHeMdtxroUf6ccmq+4ymMSj9gCQLkq4RZqScG7wvYBSItB+qTC44z87OSD2M/hxVSdo7DZ2GOC5yG3/c4PT3wRkq6xBxh4s3KH0EsiPJLRQd0zsPWXcD41VPd23RLuFvIZE9zJIx8+xoXw+nkG+oJN6qtmT9w5jEK12WJHJjlZBZ2+XWz+t98TzGLNKliJRIjBSUSrOoC6aFolSgZO48zuJ6gkpL6kJhlEDLPDZMCEESkn60bPuBujS4EGgbidYeH7YYrSmKRHCvtR198EQXCCEQbAAh8FEglUQViuliRl1GBJHgAiSHNB5RRPAJGxJdHym2gSQiMkp0k5DFvvC5S3gfKGqDLhVCZdG3qmjRutizk3L0L1LAeE0IjrHbcL67Ig479G6HKBtWk5JnF5d8/+//I37n3/s/8K/9tX+d7/7lX+Nr3/gWF1dL/smf/DE3qxXWWuqyZLFYUJiCs7MH3Fy85NMvPmXod0zalgcPHnJxdUU3WkAhheb6+SWu73ImJSHi+Wtf/01ODuaYQjKpHEoIyrJkve15tt3wxZMNCpMhLSXYDZYP7k04mgkmtaStDTUF4yYwDlvsZsf9ZsYPthv+5PElv7veMp82iGg5mBW8c2/B07GnC46b7Yr1ZsnoRoTQeCvpbaIbPFZYiiJgdEIr8/Nca15zv8Axvw38D4DvCyH+cH/f/4LsxP+vQoj/MfAI+G//vCeSUnG8OMBIQykNS9WxWW14MjqehJFkc5GxnTQsDiccnSwwRSKEkRQDIiVKXdFOWkhZlnW57fFxH5W7wDh4Cl8Qgt/reycIueMvqoSUJVKXBDdmnDwmNFlcS5ClME/uvE8zP6Fbn/PsxTV13eJcYFE2HB8d896777CzmZvsQ8L7jl2/hCSRQLAD0YfsdFzC4XJ1vaowg8clgU8erQoSI+GVA0mZVVHPmSwecHQyIdkbltuBbn3N9dU11cECJyzeBeazYx59+RzvLdc3N3zjmx9ydHzAnTunNHVFU1eslivmBxPUdsdms+WTT35A3dY0k5YQJU8ev+Ty8hrnsvRnVTWkCIWpMLpESc1yueTmZklZGb7+jQ+Zz2ekFDg/f8nV9QVNU++1VHKDj9Y6wyRCIqghnODj/6+9M4mVLEvv+u87wx1ielO+zMrMrqzq6urJGNQ2yBIyMgskwL0x7LzzAokNSLBgYeSNtyDBFgkEkoUQ3gDCSxBCsvACY5oeqrvpsaq6qnJ6cwx3OhOLc+NlVtPV3chVmUVW/KV4Ee++Ie49ce53vvMN/39NSo4YHpL8aY6LTvaYeE3tO6YLzdVyIIlDFwZbGdq2IQHOO5arc/ZrhUkXpGjxocdvDMpBshEqcH5K1wk6GXQ1Q6YLzq46NhdrNqcbfBOyZ52ehB9E6ZECdVsI+VT4ULZEuFw/v4+M9Ekh/LXg8uirPymPl5F4SwQrilorJkYyJ7Zk7nMlUG51SrWm0JpCgdlqfo6LSUxgtKaqSorpjNReUu8Z6imI9qSYCdpiVHit8BZInhgCIXm89gzNQB+yWtNsZiknCVtlKoM4gFsNmCnUZqx77gNoTYiKEBRVMSUpS9IWJYq6VoR1ppLNlTSKFIdMqTAaciWK4HIHcXCOtm148+23OTl9wMWj+/T3H/D6F36R4rOv8t6DR3ztf32dtml45+13eO31z3F84wZHt16ij57L5RLnHNZoEGE+n0NKvKsV3/r6N3B9oFeOrmnREjg+PqDvHJenl/SbJcFlz94YRfCeGK/YbHqkFVLqiVHoh5amdWir2bu1z6PNmvbdlsvNnM/c3mPTdRzPHTdmhsOZoppMCUMgBUF8Ar/GFBvKhaOqLIPyFEBXBB62ay4fDShNtk86z5d122FVTVFayiKx3nR5YTE5ef3z4OepZvlvT83oH8df+bneZYQSYTKd5LZwW1BXLRJhpTb0fYePYIxQzSvmi5rSKlLK6vQCGFNQVTVKzKiFqcEYUBrvcxKm7z1u8LlRwHm8D3nyFwXTumQxn1KWOakS4sjvrDXlyDyYpKQ8usW9Vz/Lo/c068sTNq1DKUMIkWbTcH52yvTgmKpU9EP2RmO0WF2SUAQiyujMm5Hkunba2hJtepIM+JDjlyFkXmnISiz1dMLRzZvc+/Tr7B8m2ot32bSJ0EXOzy+Y2QY6jfGOe/de5a23v8P52Wr0KIWX791hsTfH2pKbxy9BVFytrrLeo1XM51OKuiL4yGrdsFxuWC1X+JB3TtYU1zwlw+C5ulpydnqGLSzz+Zz9/QV7+3ukmGOVbnwMQ48e1WSUUsxmM0SEaphCqHHrQEgdKfakaFBMmM4XzEvDNC5JElApq6torTGVARWyKrxXdOvExq6oyxXoBZIqZNhHQgPak1ILPucrvESCOLTOMmb9eo3bdOATKMmq8aNBz1GUp2LoT4LpbKlvt99vU55P17+rkRtXnjLgSGT7H3PohTEPIBijsAa05Bg6KfcmiBK05JCMljG2Lk+WlhBSbkmvS6azGq+FIAHRBl1kbp4QCowd6MaQ15BcbhpSgihNUokiZimyiCIVlnpWYOwYixdhuj/FToWqjWzWEa0S1XxCMdEUlcYWNWItUUlW4tnWaAaI4nEkgnd0riX4QFGMHEQx5ZxVCGzWa84vLjg/P+fs9Jz29IzFySl2NuHR/Yc8fvSISTVhtVxyeXFJ2/cc37nBpz/zGv3QkwAzihxPp1Nc3xPdwGJvnxhS5i0Kia5pmMznKEnE4Ah+wJjcwV3XNZvNFZdXJ3SdyZ+UgGAYnGPT9lxctYixKG3oVeJx08JjTddZmrak7wvWm4BPHUZpjGhUqvnR2TnLvgXjmFX2yWerNdEqBlLuhE0xOxUolGQbVJgKkZJ4nnKuRKkxwfGz8Yzb+XNpotaaqqqZVj1hcBgjrBthSJGqNOztT1nMKrRKDD43b2ilKG1NVWbyJ6U0WmkiAT12a3oXcT4y9AE/+CwC4T2TesJkUlNXFbO6gOQxxoxVHBGlLNbmWnLRFcX+Ia+89nmUUjxUhsuzEwqTu+dWyyWPHj7kczfuUlWSDbN3aDVB6wmZLLRHS8U8TilTg4qZqKosdU6G9j0++LxlD0KMksWtS8tkPmf/xg1u3r7LdN5QVntY64Ce88sLhmpNUsJeEXjttde4Wp6x2Vxycb7Gubfpuo47d29zsHfAbDJlPj/k0ckJ/TBgreHo6JBqMuPicsn5RUPXDWw2Lc5HtLbUlWBtMaqDb1ivN7hh4N69T7G3N+fGjRvM5lOMUQxDz2az4fLyHO+zEbbWUhQF8/mclBLVqgQXce0StEPpkTxMTZjOSw7xzOPA+VWTDdvYBq+rCWJXGK2RpPGtZqM6ZL5BFzVJpqi4j0RPSj0peTQTlEm46IipJfpAcD2u2eDbPic+RV0bc+BaTUi2m6PtfTNSCbxv+vKkfmX7/VMm+0mM/cd+nuvTx9CGyUZdEZGUS1VFJC9icG3ItyV1pEQa1XBSihSVpZoUrIaANml0+zW6MOANtjCobsh0rMFRV2rs5DUkn8mhJEiu+KkL6mmFKMGHgFGahS1JWJpNwFae2dwwXcyprKIoFNqWiNVESYQYxnlMVs8KAR8cIQQul5e0bYsZ50SMIRNbNYmTx4/RWnNxccbyakm3bnjw+DGJwINHJ2zWG24d3+Sqabi4uuRyueRTr7/OnZfv5pCV0ddC0VVhadZrTh484PDoKFfwJDA2yyNu8xj5/Bzz2Zzj42MODvd5/FhxtX7MepOIIUscJjH0fU7ENl2gniyoJhYlkY33vHu5JrgSpTRlWTG4wHtnS7QpKIusk/ut+0su2g4kMq1yz0dOlOuRnz7vBtO4GMYolLpgUpVMqpq6UvTDgKiY92Tx/fPwg/BMjXkIgb5v0dpSVSWlMSh1m731jGWThR6qqmBWGyaFgrHbzmqLtSV1PWc6mY/ixTkOqFTu7tPKUNQGrSygSV5y4wt59a6rCmvy5fZ9O4oze2L0DIMfqxEyQ6Aoxc07r1HOjji6/Tpv/eC7DOsTXL8hxYQyFm1nVNagzJC3yp1H6wpthMJMOJhpbk0OcStLt1ZU033cD94hJMGagtVmncUflEZsQaUmJIRqcUDAcnq+4tZMYab7VAso3IbTty9Z2kuSGMxhzWc//zpFIZycPubs/ISTkwd89Svf5v57p+wt9phP59y9c4e6XtA0HVdXa1JSHB7eROkpm0aR0iPcuKMRCZAUwcN6dUpKCWMMr7/+Gi/dfon9/UXengY/eooOpRKTSTVqnNpR3zUvjn3fk9wK3f2I5uFjZHFAuX+PYnoHv3yXWi8pdEB6zdnZeiwdFcQu0MXLTI80ffsmYhKFtaBnrHwH/gyVBmZTg2sVUSxiYK9OeLkgtBtC09FtIvRCaAZ835NSxCeBqK85y+FJleL7ID/hGIz8QlxTDm899NH5vo7DZ27zfNyIUGtDqTVWSaaKfV91TIYawzL6On4f8T6fRYh5Mrs0EFSP1ppbN/cojCJ4oW1Tbu2vDYtUUQ+GAqFQo2eMgiEwjMZXG8N0Pqeq5yQiJoGtBGUKXNKU08RkD6p6QfADoXckH4hK5/g6Dp8cLoQcMkiB4F1eVG3BN7/5LR68e0nTNBwcHLBZr1mtVnRdx8nDh0zn8/H/dkjT8vbDU9Y6V71MpxO+8Gc+zx/+0R+xXq+5uLzAlgVJp9GbzeyKm82ajUC7XtM2S27duplDHaK4dXyDEB2HRzdZrZac6TMInuNbR3zxFz/Hy/c+xbe+ZSj0Ctf3DENi6CNnl1eYwlJWUw4Op5ydX4E1TKaGamppVi3nXvNyVbP30m0OphPeOPkuP/z+Ix6fr2kcXLqOxYFmMa+wUY+8O7kfxaVAVBGlIKLwMeL6xGK/YDotmU8KvIfVqmDV5GIPpT6GnnmmMs1t4SkFMMJsf0I5K9jzCxJQFRajMiVqItG17RjbNIiU9H3WzgzBg8S8vYyeFMKox5lLfpTVqCKHcwpr8ns7hyhDijmEEKPPVQgiY/dhQlnFQV3RDHnSG2uZTmeIu2RSW2aLBYc3X+Fs7VEqURhhNp8haoVWEWsMk9Ly6t0bHE46mktPv6m49+kvIPYAlEHpnLxq2hXf/d6bvHf/hNPzFU3fYeoZp5dr/vgr3yC1c+pCcHqKFEJkxeOTE5SyHFWHFPN9ju7co4uCTwZTzrmr4VP3XmY2maJQ+K5nUhvO6Li63PDWm2fs7b9C8BZJFtcF/ADN2jP0PWexzVvjQjOd18zmE4pS6Po1TScok0DnMs2+6zIHPFlLNEaP99kTE5Vou452aNCFop58Fjs/QhUF4leEKHRmxsrlJiVTHnBwPLAZElIYVDxnMduwNgqnLIOdUiwMZSzRaQ/FAcn0dPEhQ8jxcuEuqTlBGo0JM8pyznL4UVbdmegsJ7aMKBnN7GjF01bA8324NvXv+5mQf39rzLdeuVbZ+cjdomBEYVSuXNEiFEpTqCeVKYIfDfqTx7b8HXLTUkjZeMWQG4ms1gzJ04WervcYVUOaEbzg3YALPaWdcHCwh1VFVu1JHu9HVsxk8EpBmSXU7GSC8zC4LNcmymK1ZdMNJCzFrMZOytyh20BoYTN02FqIKpIkgAqjgU1jjsQg2vDmm2/x8N0LqrrmlV/9VVbLJRdn57z95luZAkBriIEUAsp5xFicJA6PjviFL36RVz/3Gb7yrTdAYL1e8+C9+7x3/202mzV93xGcZ/Ce48N9CqPYrK84urHHZrNC0BweH3LVHqOrkrgRgiTMrERNAk265HIomNwoqFpFaS1xoolOs25DJtQrZ9y6fYd1kxvyukaYzwq0NTSD5uJq4PJsxZ0bU77863+ZN/7323zjW2/yje+/y8QkppWlKg3aKoxVFOjr1Tomn0tEY9bXK8sSW9T4CJs+d2DP5zP6oGiWDauPJ9FW9l4SeXImrVBm1MYTgxKhLkusVtkziQmizvzKKJSyhOCJIRtga20u7Qnx+kaIetwKpzgKp+dGlBjyhE0q0A995iAhYLQikQWOU4LCJA4XM6JAUSgKKxhJzCd7HBzsY8uai3Vk3QW0eGIBpVbcOlowm06oCktVWG4eLqhNR20iYW+Pwxt3eP1zBpRBm8xj7lzP/sEdHj0+4+T8kk27YdkM9IOw3jRcLA1dpfA+4aKmrie0Xciescvsj4v9I0KEyWzBZrNGa+H41s3cTegDzvbM1AFtH9hsHBcXGzarDmst88mcO7du8/biDN+D61raTUtdzbBWkxuHNpxfnGFKRZLculwOZS5FdG7sVlTX8V0lCmc9ISa6oafvIyHM0PMKZRMpbAiuIzoQOWAYDH5IOAFtLyi1IJWgzECSBMUohJwiVmkKpZFuRWgvGKIjBE9SJa5XLNs1rVsSVY+UCj0dUKHFTiP4rMru1iG3uac0xit5Umq49ayfnqxPG/KnmnquX0u+ZqVy7FMSiOI6sSlbfdDr390mRdP1I//lNiyTxtLJmOPa79MATbliSymMMagx/CKSy+gSHpQmJEVyka6P+BhG6lrJTJTaoguLlAWiDINLtJ3PvEJa8HhikGsObZ8cjp6gPEEHQsrhIe+H3CIfPMF5jNFoq1HKElMuf3TeY52jbRr6rsf1A8F5UkigfCYSv37kpdANjt4PXDSbXGmmM9/S6uKS5fkFzg3jYpn5evzQ0697Hj18gC4K+iHz2Z+cK1bdBUZZnGqY36yYLm6zf2tCKjvW/gwzC3SrPgu+SAQT2T8s8MFgi8TQr1jMS7QOSArgAy44mmXPY6N5uKh4uT8kpAum85J7r77ERd/z4OKKvYMps1lJWSlEBcqyRmlDBNabNcYUCDl/YVQHyrFuutzI1zlKK6gkVNqwiYqfB89cnAKdb52QItuqXlGC1irfrCY3xihRJJWNq8cTE7ludWs8VK63TilXYUgab7CYy8tiCkAkiBBjVhaJCaIaPfuY602jyoml4HPrrxZhMZ0QVEIbUCoRfGSxv2C62CegWZ6c0bmIVZHKKqrCcPfWIQd7M6rSUmhFXZUYLKk0KALTxR53bE0ShSiNkHUEbTHj+Oaa1abB+Z4HJ2ecnq24umyxxSQbNInYwnDr2FB2FjcM7E0L/OCppzXzxT7GFlR1jVLCpJ6hdCae0spSlQVH3ZBDLZcXRA/KaiblhLsv3ebunVNiUASXaJuOqiiwIwd217VcXl5QTUqMySLBIYSROTA8KeXbNqWIwthM69s7Rz9EfCzQk0BU66zhOAwomUPUdIMi9govBYGAsoIqAqI8UWkosl6jUh6JA8RA7Brc5oo2RVDHpFDik9CuT3ByiZ6CrTV6Hij9QOojKgqDVyhJI4GWZB6YbfE2W1suT74ZYzG5vX978Kn6ltFQb1+PL66N9jbuveV5ERnr13li3J9+rbZvMyoWRXl/CCiRjZs2BSYl0jAeF4U2GkMCVRC9MAyRYYhEUWib47ZGDFibd0eFJYrgXcwLbspNeElFkujcDRsjxAFPh2ck01IJH2EIY5FB8Lml39ZZQNpaht4BTxqoHj96RNd2dG031tmrfP/FXL+ZK4vGcJIPbNqWk8vzTDUwJqxd0xH6zOwpKlf3+L7jstuwXi+5/947LI4OaLorhqFHrQI9a3xUSAX7L9XUZko5MSjr6VlRVZbODRjJRRCgmcwsMQiCxw9L5lOL0RXBDaTg6Fyg6wPnq4YHFxvOVh0b31CUMw6O9nnlVUeHUE+nVHVFPcnMlpPJDG0sPoLzBhGDYEgq5w9jaBj6SNsOuF5BjGgShYLafniliR8aRBRixg5JpSlMgU4jgVGhqcpyDPhnjwJyskPUtsUbtLa5O2qMncUgEDJlqzUWbQxhcJlCNSXS+OGPTXAEFOgsDQdASrjOjetMJviqCs1Cqqy3qRR9rwmx4v6jNZvOs2w9YhRlqTnYm/Dqp27whc/cpi5VbmYh4QNYXWL0AmMEDCyqKmugh0TX9fQbjzUVxzcm3HulZm9/ysXFJVdXG66WLaV0rNYbuj7HqGfTKdFMCK5noiLJOzYXPc7nsJNVgtKaoWtQOidbdKlBw8HRPlX1WbpNw/5iP7ewS+Lu7Vv8yp8X9mY/5O26pCgUtspjHkeFn+Vqw17TMUwdwWfudZCcdPaevu+JPqJQI39L5oT3KdAOHT0dmCVWhszPoaeU8wV+dUbXXOFWHUU9oXUdSMLGSHRTmMwwVQMqUE4dfXtGu/SksfNxszFMJkdgKlJoWJ+/Syw881ozqRSLmWdWRrrCsNFwfjV6hXlVz5//yJa55U2/xrWr/CTF+b65zNNxc64ZGHnyJ3khGGsTs3+9/ad5kdBKcqXJGCfXersIZEGQXNszns51DX5NNZljXCKZSJQ8501pQKYkb/BECB6XImJLTGnQhUJSRKwlaJW9/RAYBo8L5F2FzuRSRnSuTsFDCYEeFwZcyIIh9JkILMsbPhGzlnHH0PV93mXXuejg61/7Gr7PnDh1WWHLkqvLi5GtchywmO2Ad46Lq0senJ3g4igh4gOF1kzKgvWmZ7W84vLqgkcPHrJcXrJaX3Fxdcb+1QEu9OhC4bVw8JIlWrCFoipLKi3jDieQ8ARxtMOQ6SRsLqXVeKy1aDFolSgKS1UUBK/wg2DbiJtH2hj50dmauyeZrGyxiCz2D3n9s5/nqjEs1z1dqyiKmq4LOdGtIi54RCb0rWMYepzLWgDz2QJjKqp6L1f+uJb26jEqtBztfwTcLH9qiMLoImeXlcEojY5p3KZmKdyQAiluJbNhGPrr5M+2WkLE5FIq73OVgIQnrIZFwWYYiDFASiMtrs/ed0o4EmZsVYeclFVaX+8QctKpR4hUVUXSFeuu5MHDsxxXJ7E3KfjU3SNevnPIrYMZx4saIx7X97gx3uMRnM4hF7RGE8dJlBelqjIYWaBVQ9f3NOsVh/s1L93Y5+bBAYOLuK5j2XqaIeIjzKc1pS0Ivse4lklc59K7rsMNPUkSi4N9isLkRK5AWdd0TUNdVyxmc4w2uGHg6uKSVdNwfnbCzaMFN/7il/jiF17jjW9/jz/5ytcZhszxUZQlIXiWyzXWlhhjGXqP956u6+i6jr7PlSLWWMpxh6CtwcVszCkCyD49S1ARUQPD+h2s9yQdkVJBXaBWlug8ITnaeEboGmK3JJJYn0ZMvU/b55rnpCANjoELpHU5CZ5KXv70PtP9gXLmMJUwXdTEQljGgXS64tT3qDC23qsA1mAqhSoka4KOheRPkpujt35ty68D2+/DkwWB63pzrXQWV5ZEkSJWQvYCRWG1zvX4Y9natuJl68VDYuQ2uHbhE2CUxmqLSoIUmqRNThwrTfSabsj0vKmoKAuNC4kwhkwKDUGBJ1Pi+hjoXCBEk8MkhUKZRJ/6a7pe4xPO53b6YQikqDCmzAu3aLQtKEuN0YZEou02tN2Go/0FqdcMvWN9tSLGfD9NZjOObx1TFIquafDOIyI4n3fA0UdcM+CaQHCJ1WrNWz/6IQ8fv8d793+QqQa6hsH1aCPU04LysOCVezeoJiVia7QRjBWUgiC5SMLogB5zJTFFUgokApA7j93oHFrJyVytNLYIaKtpugbnHIRIVRYc1zWbjee87fjWD0555dWXuGoGLlb3sdWShw8fcnnV4ELEGMXV1QprilzJZMBag9ZQVpa6LihKi9Yu0xGXWdLOD4EYAjR+pMj+2Xi2nrnKdcgikJls8haLcUsaY6bR3PJlxLit4c53WAg5FKKUum5OSSk3+mRWv8xZkkhj9172IJwbcu9QSrTOZbY1Y67DNVVdE3wm5UmSORlwkWFwrBvH5XqFT47ZvGB/XvHSrT1effmYvVlNaRQqZk5upRSIEEm58y7liYrkrj4RPTpluWvQWsVkmhsF4pjgEiWEmHAxsukDp5c95+uOZTdQasXR/oK6sEwImPUJJ48ecnp6QtNsSASObx7z0u3bzOazzJmeEiqlXANuDUoZNFAv5gQlhDigyDzkWs/59Cu3uFy+zLsPTti0HW3XY6xivW7QOmuA1vUkJ0D7gWEYCGHMPVgPyYMETKoIEXwf8K0Q+ylonQmNvIPO47wmtZHkPdE1DCkRB49ygNUMfZebmYICUdj6AGUCMeTEoEggus1IECbceX3BjbsgKoEIKgjWF6BhOrUcHAUObzj8ICSBqBJu5Lwmd4VfhzSebh/K3pyMyXK4nlzXT6NXPoZlchQgV6QYlbs8SwLWgNFjUlSrJyWIsvXmt6/HgE+C7eIhUUBr/BAYOg9BUKXJCTWfywSDA+8EH7K+pUggppTDK6JxY9gmSt6YBEkEIcfSleRkJjDEkAWryR+n7wPRpxydiuR6aqXH/BdoZbKX7j0h9AjwpT/3Z3nnzVO+//0fEl3+v4UtODo65Jd+6UsslxdcXV6wWa1om4bHJxcslyt86OnaFa5bE3xW/PnB2wO21ARa9DSxt1eh9CQbxVIwpcLWCqXTKK+Xd52Z54aRYiE/0naRTmMYadvnIRBDQhdqpADwaOeJwOD77BwKFGhEOZSJYOB8vWK+rChMQkmibxyJBmNaYnJ4H7Ig+NARvKawBqOKzA8TPd41bDae4O31Yp6CJ/iepAZ0LaQnfGE/Fc/UmKtRYSXFTIgVUxonczaCPoScuEyZk8CNCiHwJCZ5LSb81ONpw7yNvW2/upCz3kqpnAAdkzPAGCcDpXVeSFLMsd6+Y7luuNr0XDUDzaalLixHeyU3b8x4+aUZNw8KrESid/RNm8/D2tw0FCPt0GW5MWNxZUFK9vomzdeTF7IYx/tIqdEwRpresdz0nJ9teO9kzemy5bLpIEaO9xcsZlOOSpiyYb284vz0hIvLC9quZbVc4gbH4eEB89mMuixJSqPLkrIO1JMZyloKYEIi9FOGzSWFFbQpuX37Bqv2VTrn6B+csrxaU8ZiHPccmgghjQtvbg/XoyhziB7nc6s02uB89jDESSaHKgAEArhNQEWdDXcMSBhwQOwze6OaJaIqwILSCWPB1hAi4LLYhdUReocSobSGvSONUZmWIASFb2FFzKIbTSSiKfdqtMsLuyfghjRqsKpx4R0/H8iJV64d7Sfz8Np7fhLzvnagt4+UQyNaKawGK+QqLTUal3HOP/kbed/z9n8/vQtIKRF9wjsgCmJyHYyPiX7wRBcBS7hOjObPR6JC4hhqVAnSSFWgZczJMNISBEKCIWSCKo2QoiZ4II613epJh+/2/hJRRJ89SO89WhteuXcPfMXJ49M8UmOZa13XWXYxDLihxbuOvhfERFSRjXGkw4Ul5SQh2tGGS1xUVBNLOclsgrkvREgqs/AlkwgSxh19Jv26NtzIeO8zcrDnjl8fImxTKCnhfUJLyothHDVXdUeW7s5NXoGA92OICUUzdGy6BmqNVom+byF1WOsQcXl8JuBcDiMVVjOpFVVtM6ujOJLkPokcVhNiGhAJlLUFynyePweeqTHP4gYFrvNIlsTOsV2VjYT3Hi2MsfAci02jwVdKjfWyvM+QA2Pr+LglTplznJiTcM498biVVlRabWmsr//GeUdII7m+H3j46AH3H55ytW5pBo8tK27du8vt45obByWLKjCsz+l8wPWevhtyCMkUJMkdn6vNBsililVRMJlUwJNrydUPKfsLkmOn61WiaXuu1i2nl2vu37/gwemKZevYDAHvPO8YxWw65ZWjKV/8/JTFfMpJaRm85+HJKZerNeeXlxzs7XF8cMCNg0N8AlvVTPf3uPeZz1BPJkQB7TVVZVFOY7RQ2II79UsUkxlX64bLZcM77z2i7z11ba8Tn3ZsCioKm3dDMdG2a7xv8dGjgkKHRNdmStQCYTOcoaRFm9wo4zoHGkwI6JizGVFp/BBRMVIkjZ29RFEFjN4wmV5S3DgDUeguEJNjNlH4q7zI2CoQUsvZOxGlc0KtbRs2G0czJHyfSJtEmui8iPqRqyTmpiRR//etMK5dT+c+x/m3nTtPG94tXe5IuKWyMVc6x8I1Cq0TorjmoCGlsSzy/aWOSMoLyfvD9BDjqIJU5FBjMLgg9IOnaQaiF6rKMOp75HtDVGYsDBE/EpnpkcRMG02SAcSP1L8BEgx+yFKGSqCoIJiRQzy768WoYRqjx7Xxmsgrc9GDNZr53oKbN+H2ndvor71B8B6tc7PMW2++xYMH73B1dcF6s6Jp13gCulKUpcHWkYEzjl+eU9QFpjQoK8jYcCZKEUj03mWOpeRzg40WytJgTCbWSjp/pjGl613/4MZcW8phjO2Yx0guAAgDohUJjXcRUQ5T5LpwpRRDdAxDoO8S3ilEMv9MiDlEfLG+IviOwkBZKGIaw7lw3fk+m86YTGpidDjfIqqisDnsJiIMfY9WMmoFq5Fq4v7PtK+SfkL876OCiJwAG+D0mb3pxxM32I3BbgwyduOwG4Mtfto4vJJSOv5pf/xMjTmAiPxJSukvPNM3/ZhhNwa7MdhiNw67MdjiTzsOP181+g477LDDDh9r7Iz5DjvssMMLgOdhzP/5c3jPjxt2Y7Abgy1247Abgy3+VOPwzGPmO+ywww47fPjYhVl22GGHHV4APDNjLiJ/XUS+IyLfF5Hfflbv+7whIm+JyDdE5Ksi8ifjsUMR+c8i8r3x+eB5n+eHDRH5VyLyWETeeOrYB163iPzDcW58R0T+2vM56w8XHzAGvysi743z4asi8uWnfvYijsHLIvJfReTbIvJNEfl74/FP2lz4oHH48OZDSukjf5C5+n8AvAYUwNeAX3gW7/28H8BbwI0fO/aPgd8eX/828I+e93l+BNf9a8AvA2/8rOsGfmGcEyXw6XGu6Od9DR/RGPwu8A9+wu++qGNwG/jl8fUc+O54rZ+0ufBB4/ChzYdn5Zn/CvD9lNIPU0oD8PvAbzyj9/444jeA3xtf/x7wN57fqXw0SCn9IXD+Y4c/6Lp/A/j9lFKfUnoT+D55zvx/jQ8Ygw/CizoGD1JKXxlfr4BvA3f55M2FDxqHD8L/8zg8K2N+F3jnqe/f5adfyIuEBPwnEfmfIvK3x2O3UkoPIH/IwM3ndnbPFh903Z+0+fF3ReTrYxhmG1544cdARF4Ffgn473yC58KPjQN8SPPhWRnzH2eZgA+QWHwB8asppV8Gfh34OyLya8/7hD6G+CTNj38GfAb4EvAA+Cfj8Rd6DERkBvw74O+nlJY/7Vd/wrEXeRw+tPnwrIz5u8DLT33/KX4e5pgXACml++PzY+A/kLdKj0TkNsD4/Pj5neEzxQdd9ydmfqSUHqWUQkopAv+CJ1vnF3YMRMSSDdi/SSn9+/HwJ24u/KRx+DDnw7My5v8D+KyIfFpECuA3gT94Ru/93CAiUxGZb18DfxV4g3ztvzX+2m8B//H5nOEzxwdd9x8AvykipYh8Gvgs8MfP4fw+cmwN2Ii/SZ4P8IKOgWRawn8JfDul9E+f+tEnai580Dh8qPPhGWZzv0zO4P4A+J3nnV1+Rtf8Gjkj/TXgm9vrBo6A/wJ8b3w+fN7n+hFc+78lbxsd2cv4Wz/tuoHfGefGd4Bff97n/xGOwb8GvgF8fbxhb7/gY/CXyOGBrwNfHR9f/gTOhQ8ahw9tPuw6QHfYYYcdXgDsOkB32GGHHV4A7Iz5DjvssMMLgJ0x32GHHXZ4AbAz5jvssMMOLwB2xnyHHXbY4QXAzpjvsMMOO7wA2BnzHXbYYYcXADtjvsMOO+zwAuD/AFGOVlTrhIoHAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "model.eval()\n",
    "with torch.no_grad():\n",
    "    classes = ('plane', 'car', 'bird', 'cat',\n",
    "           'deer', 'dog', 'frog', 'horse', 'ship', 'truck')\n",
    "    \n",
    "    test_data = iter(testloader)\n",
    "    test_images, test_labels = iter(testloader).next()\n",
    "    show(test_images[:8])\n",
    "    test_images, test_labels = test_images[:8].to(device), test_labels[:8].to(device)\n",
    "    \n",
    "    test_outputs = test_model(test_images)\n",
    "    predicted = test_outputs.argmax(dim=1)\n",
    "    \n",
    "    \n",
    "    print('GroundTruth : ', ' '.join(classes[test_labels[j]] for j in range(8)))\n",
    "    print('Predicted : ', ' '.join(classes[predicted[i]] for i in range(8)))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy of the network on the 10000 test images : 46.370 %\n"
     ]
    }
   ],
   "source": [
    "correct = 0\n",
    "test_loss = 0\n",
    "model.eval()\n",
    "\n",
    "with torch.no_grad():\n",
    "    for data in testloader:\n",
    "        val_images, val_labels = data\n",
    "        val_images, val_labels = val_images.to(device), val_labels.to(device)\n",
    "        \n",
    "        val_outputs = model(val_images)\n",
    "        pred = val_outputs.argmax(dim=1, keepdim=True)\n",
    "        correct += pred.eq(val_labels.view_as(pred)).sum().item()\n",
    "        \n",
    "        \n",
    "print('Accuracy of the network on the 10000 test images : %.3f %%' \n",
    "      %(100 * correct / len(testloader.dataset)))"
   ]
  }
 ],
 "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": "3.8.10"
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}