Boubaker.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# **Define of Model**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.optim as optim\n",
    "from torchvision import datasets, transforms\n",
    "from torch.utils.data import DataLoader, Subset\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def boubaker(n, x):\n",
    "    if n == 0:\n",
    "        return torch.ones_like(x)\n",
    "    elif n == 1:\n",
    "        return x\n",
    "    else:\n",
    "        return (2 * n - 1) * x * boubaker(n - 1, x) - (n - 1) * boubaker(n - 2, x)\n",
    "\n",
    "class BoubakerKANLayer(nn.Module):\n",
    "    def __init__(self, input_dim, output_dim, degree):\n",
    "        super(BoubakerKANLayer, self).__init__()\n",
    "        self.input_dim = input_dim\n",
    "        self.output_dim = output_dim\n",
    "        self.degree = degree\n",
    "        self.boubaker_coeffs = nn.Parameter(torch.empty(input_dim, output_dim, degree + 1))\n",
    "        nn.init.normal_(self.boubaker_coeffs, mean=0.0, std=1 / (input_dim * (degree + 1)))\n",
    "\n",
    "    def forward(self, x):\n",
    "        # Normalize x to [0, 1] using sigmoid\n",
    "        x = torch.sigmoid(x)\n",
    "\n",
    "        # Compute the Boubaker basis functions\n",
    "        boubaker_basis = []\n",
    "        for n in range(self.degree + 1):\n",
    "            boubaker_basis.append(boubaker(n, x))\n",
    "        boubaker_basis = torch.stack(boubaker_basis, dim=-1)  # shape = (batch_size, input_dim, degree + 1)\n",
    "\n",
    "        # Compute the Boubaker interpolation\n",
    "        y = torch.einsum(\"bid,iod->bo\", boubaker_basis, self.boubaker_coeffs)  # shape = (batch_size, output_dim)\n",
    "        y = y.view(-1, self.output_dim)\n",
    "\n",
    "        return y\n",
    "\n",
    "class MNISTBoubakerKAN(nn.Module):\n",
    "    def __init__(self):\n",
    "        super(MNISTBoubakerKAN, self).__init__()\n",
    "        self.trigkan1 = BoubakerKANLayer(784, 32, 3)\n",
    "        self.bn1 = nn.LayerNorm(32)\n",
    "        self.trigkan2 = BoubakerKANLayer(32, 32, 3)\n",
    "        self.bn2 = nn.LayerNorm(32)\n",
    "        self.trigkan3 = BoubakerKANLayer(32, 10, 3)\n",
    "\n",
    "    def forward(self, x):\n",
    "        x = x.view(-1, 28*28)\n",
    "        #x=x.tanh()\n",
    "        x = self.trigkan1(x)\n",
    "        x = self.bn1(x)\n",
    "        x = self.trigkan2(x)\n",
    "        x = self.bn2(x)\n",
    "        x = self.trigkan3(x)\n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.optim as optim\n",
    "from torchvision import datasets, transforms\n",
    "from torch.utils.data import DataLoader, Subset\n",
    "import numpy as n\n",
    "transform = transforms.Compose([\n",
    "    transforms.ToTensor(), \n",
    "    transforms.Normalize((0.1307,), (0.3081,))\n",
    "])\n",
    "\n",
    "train_dataset = datasets.MNIST(root='./data', train=True, download=True, transform=transform)\n",
    "test_dataset = datasets.MNIST(root='./data', train=False, download=True, transform=transform)\n",
    "\n",
    "train_loader = DataLoader(train_dataset, batch_size=64, shuffle=True)\n",
    "test_loader = DataLoader(test_dataset, batch_size=1024, shuffle=False)\n",
    "\n",
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
    "num_classes = 10\n",
    "\n",
    "\n",
    "\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "\n",
    "\n",
    "def train(model, train_loader, criterion, optimizer, device):\n",
    "    model.train()\n",
    "    total_loss = 0\n",
    "    correct = 0\n",
    "    \n",
    "    for idx, (data, target) in enumerate(train_loader):\n",
    "        data, target = data.to(device), target.to(device)\n",
    "\n",
    "        if isinstance(optimizer, optim.LBFGS):\n",
    "            def closure():\n",
    "                optimizer.zero_grad()\n",
    "                output = model(data)\n",
    "                loss = criterion(output, target)\n",
    "                loss.backward()\n",
    "                return loss\n",
    "            loss = optimizer.step(closure)\n",
    "        else:\n",
    "            optimizer.zero_grad()\n",
    "            output = model(data)\n",
    "            loss = criterion(output, target)\n",
    "            loss.backward()\n",
    "            optimizer.step()\n",
    "            loss = loss.item()\n",
    "            \n",
    "        total_loss += loss\n",
    "        \n",
    "    return total_loss / len(train_loader)\n",
    "\n",
    "def validate(model, test_loader, criterion, device):\n",
    "    model.eval()\n",
    "    total_loss = 0\n",
    "    correct = 0\n",
    "    with torch.no_grad():\n",
    "        for data, target in test_loader:\n",
    "            data, target = data.to(device), target.to(device)\n",
    "            output = model(data)\n",
    "            loss = criterion(output, target)\n",
    "            total_loss += loss.item()\n",
    "            pred = output.argmax(dim=1, keepdim=True)\n",
    "            correct += pred.eq(target.view_as(pred)).sum().item()\n",
    "\n",
    "    return total_loss / len(test_loader), correct / len(test_loader.dataset)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total trainable parameters of Boubaker: 105856\n"
     ]
    }
   ],
   "source": [
    "Model_Names=['Boubaker']  #Add names of other models\n",
    "model0 = MNISTBoubakerKAN().to(device)\n",
    "models=[model0]\n",
    "total_params = sum(p.numel() for p in model0.parameters() if p.requires_grad)\n",
    "print(f\"Total trainable parameters of {Model_Names[0]}: {total_params}\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# **Training of Model**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Boubaker, Epoch 1, Train Loss: 0.6494, Test Loss: 0.2118, Test Acc: 0.94\n",
      "Boubaker, Epoch 2, Train Loss: 0.1872, Test Loss: 0.1504, Test Acc: 0.96\n",
      "Boubaker, Epoch 3, Train Loss: 0.1376, Test Loss: 0.1293, Test Acc: 0.96\n",
      "Boubaker, Epoch 4, Train Loss: 0.1147, Test Loss: 0.1233, Test Acc: 0.97\n",
      "Boubaker, Epoch 5, Train Loss: 0.0996, Test Loss: 0.1127, Test Acc: 0.97\n",
      "Boubaker, Epoch 6, Train Loss: 0.0870, Test Loss: 0.1103, Test Acc: 0.97\n",
      "Boubaker, Epoch 7, Train Loss: 0.0810, Test Loss: 0.1026, Test Acc: 0.97\n",
      "Boubaker, Epoch 8, Train Loss: 0.0750, Test Loss: 0.0999, Test Acc: 0.97\n",
      "Boubaker, Epoch 9, Train Loss: 0.0687, Test Loss: 0.0954, Test Acc: 0.97\n",
      "Boubaker, Epoch 10, Train Loss: 0.0635, Test Loss: 0.0997, Test Acc: 0.97\n",
      "Boubaker, Epoch 11, Train Loss: 0.0599, Test Loss: 0.1061, Test Acc: 0.97\n",
      "Boubaker, Epoch 12, Train Loss: 0.0568, Test Loss: 0.1189, Test Acc: 0.96\n",
      "Boubaker, Epoch 13, Train Loss: 0.0528, Test Loss: 0.1115, Test Acc: 0.97\n",
      "Boubaker, Epoch 14, Train Loss: 0.0514, Test Loss: 0.0998, Test Acc: 0.97\n",
      "Boubaker, Epoch 15, Train Loss: 0.0473, Test Loss: 0.0989, Test Acc: 0.97\n",
      "Boubaker, Epoch 16, Train Loss: 0.0467, Test Loss: 0.0872, Test Acc: 0.97\n",
      "Boubaker, Epoch 17, Train Loss: 0.0433, Test Loss: 0.0928, Test Acc: 0.97\n",
      "Boubaker, Epoch 18, Train Loss: 0.0391, Test Loss: 0.1021, Test Acc: 0.97\n",
      "Boubaker, Epoch 19, Train Loss: 0.0404, Test Loss: 0.0973, Test Acc: 0.97\n",
      "Boubaker, Epoch 20, Train Loss: 0.0379, Test Loss: 0.0945, Test Acc: 0.97\n",
      "Boubaker, Epoch 21, Train Loss: 0.0352, Test Loss: 0.0948, Test Acc: 0.97\n",
      "Boubaker, Epoch 22, Train Loss: 0.0335, Test Loss: 0.0969, Test Acc: 0.97\n",
      "Boubaker, Epoch 23, Train Loss: 0.0319, Test Loss: 0.0905, Test Acc: 0.98\n",
      "Boubaker, Epoch 24, Train Loss: 0.0303, Test Loss: 0.0976, Test Acc: 0.97\n",
      "Boubaker, Epoch 25, Train Loss: 0.0293, Test Loss: 0.0967, Test Acc: 0.97\n",
      "Boubaker, Epoch 26, Train Loss: 0.0288, Test Loss: 0.0905, Test Acc: 0.98\n",
      "Boubaker, Epoch 27, Train Loss: 0.0281, Test Loss: 0.1009, Test Acc: 0.97\n",
      "Boubaker, Epoch 28, Train Loss: 0.0255, Test Loss: 0.1116, Test Acc: 0.97\n",
      "Boubaker, Epoch 29, Train Loss: 0.0243, Test Loss: 0.1219, Test Acc: 0.97\n",
      "Boubaker, Epoch 30, Train Loss: 0.0230, Test Loss: 0.0951, Test Acc: 0.97\n",
      "Boubaker, Epoch 31, Train Loss: 0.0226, Test Loss: 0.0958, Test Acc: 0.98\n",
      "Boubaker, Epoch 32, Train Loss: 0.0234, Test Loss: 0.1112, Test Acc: 0.97\n",
      "Boubaker, Epoch 33, Train Loss: 0.0204, Test Loss: 0.1162, Test Acc: 0.97\n",
      "Boubaker, Epoch 34, Train Loss: 0.0210, Test Loss: 0.0980, Test Acc: 0.97\n",
      "Boubaker, Epoch 35, Train Loss: 0.0191, Test Loss: 0.1149, Test Acc: 0.97\n",
      "Boubaker, Epoch 36, Train Loss: 0.0198, Test Loss: 0.1025, Test Acc: 0.97\n",
      "Boubaker, Epoch 37, Train Loss: 0.0174, Test Loss: 0.1111, Test Acc: 0.97\n",
      "Boubaker, Epoch 38, Train Loss: 0.0177, Test Loss: 0.1100, Test Acc: 0.97\n",
      "Boubaker, Epoch 39, Train Loss: 0.0180, Test Loss: 0.1032, Test Acc: 0.97\n",
      "Boubaker, Epoch 40, Train Loss: 0.0158, Test Loss: 0.1037, Test Acc: 0.97\n",
      "Boubaker, Epoch 41, Train Loss: 0.0146, Test Loss: 0.1065, Test Acc: 0.97\n",
      "Boubaker, Epoch 42, Train Loss: 0.0164, Test Loss: 0.1265, Test Acc: 0.97\n",
      "Boubaker, Epoch 43, Train Loss: 0.0157, Test Loss: 0.1114, Test Acc: 0.97\n",
      "Boubaker, Epoch 44, Train Loss: 0.0135, Test Loss: 0.1243, Test Acc: 0.97\n",
      "Boubaker, Epoch 45, Train Loss: 0.0156, Test Loss: 0.1092, Test Acc: 0.97\n",
      "Boubaker, Epoch 46, Train Loss: 0.0114, Test Loss: 0.1153, Test Acc: 0.97\n",
      "Boubaker, Epoch 47, Train Loss: 0.0148, Test Loss: 0.1058, Test Acc: 0.97\n",
      "Boubaker, Epoch 48, Train Loss: 0.0107, Test Loss: 0.1087, Test Acc: 0.97\n",
      "Boubaker, Epoch 49, Train Loss: 0.0131, Test Loss: 0.1220, Test Acc: 0.97\n",
      "Boubaker, Epoch 50, Train Loss: 0.0128, Test Loss: 0.1153, Test Acc: 0.97\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1cAAAHWCAYAAACbsXOkAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy80BEi2AAAACXBIWXMAAA9hAAAPYQGoP6dpAACV5UlEQVR4nOzdd3iT5foH8G+Spkn33hTaUlYLFGjZsqeAIkOKogwFF6iI44jnd0BcOBDxuBAUQRRFRDwKiOw9CpQiq2V0UujeTdukyfv7I02gtIWOpBn9fq6L69g3b973SfucNnee+74fkSAIAoiIiIiIiKhJxKYeABERERERkTVgcEVERERERGQADK6IiIiIiIgMgMEVERERERGRATC4IiIiIiIiMgAGV0RERERERAbA4IqIiIiIiMgAGFwREREREREZAIMrIiIiIiIiA2BwRUREREREZAAMroiILIBIJKrXv/379zf5XgqFAm+++Wa9r7V//36IRCL8+uuvTb53c7h27RqefvpphISEQC6Xw9nZGf3798enn36KsrIyUw+PiIgsmI2pB0BERPe2fv36al9///332LVrV43jnTp1avK9FAoFlixZAgAYPHhwk69nTrZt24aHH34YMpkM06dPR+fOnaFUKnH48GG8+uqruHDhAlatWmXqYRIRkYVicEVEZAEee+yxal8fP34cu3btqnGc6paUlISpU6eiTZs22Lt3L/z8/PSPzZ07F1evXsW2bdtMOMKmKy8vh62tLcRiJqYQEZkCf/sSEVkJjUaDFStWIDw8HHK5HD4+Pnj66aeRn59f7bxTp05h1KhR8PT0hJ2dHYKDg/HEE08AAJKTk+Hl5QUAWLJkiT7d8M0332zy+BITE/Hwww/D3d0d9vb26NOnT63BzGeffYbw8HDY29vDzc0NUVFR2LBhg/7x4uJizJ8/H0FBQZDJZPD29saIESMQGxt71/t/+OGHKCkpwbffflstsNIJDQ3Fiy++qP+6srISb7/9Ntq2bQuZTIagoCC88cYbqKioqPa8oKAgjBs3DocPH0avXr0gl8sREhKC77//Xn/OqVOnIBKJsG7duhr3/fvvvyESibB161b9sfT0dDzxxBPw8fGBTCZDeHg41qxZU+15unTMn3/+Gf/3f/+HgIAA2Nvbo6ioCACwadMmhIWFQS6Xo3PnztiyZQtmzpyJoKCgatep77ypz+vUKSgowEsvvaT/GbVq1QrTp09HTk6O/pyKigosXrwYoaGhkMlkCAwMxGuvvVbj+0tEZEm4ckVEZCWefvpprF27FrNmzcILL7yApKQkfP755zhz5gyOHDkCqVSKrKwsjBw5El5eXnj99dfh6uqK5ORk/PbbbwAALy8vfPXVV3j22WcxYcIETJw4EQDQtWvXJo0tMzMT/fr1g0KhwAsvvAAPDw+sW7cODz74IH799VdMmDABALB69Wq88MILmDx5Ml588UWUl5fjn3/+wYkTJ/Doo48CAJ555hn8+uuvmDdvHsLCwpCbm4vDhw/j0qVL6NGjR51j+PPPPxESEoJ+/frVa8yzZ8/GunXrMHnyZLz88ss4ceIEli5dikuXLmHLli3Vzr169SomT56MJ598EjNmzMCaNWswc+ZMREZGIjw8HFFRUQgJCcEvv/yCGTNmVHvuxo0b4ebmhlGjRum/V3369IFIJMK8efPg5eWFv/76C08++SSKioowf/78as9/++23YWtri1deeQUVFRWwtbXFtm3bEB0djS5dumDp0qXIz8/Hk08+iYCAgBqvsz7zpr6vEwBKSkowYMAAXLp0CU888QR69OiBnJwc/PHHH7h+/To8PT2h0Wjw4IMP4vDhw3jqqafQqVMnnDt3Dp988gkuX76M33//vV4/IyIisyMQEZHFmTt3rnD7r/BDhw4JAIQff/yx2nk7duyodnzLli0CAOHkyZN1Xjs7O1sAICxevLheY9m3b58AQNi0aVOd58yfP18AIBw6dEh/rLi4WAgODhaCgoIEtVotCIIgjB8/XggPD7/r/VxcXIS5c+fWa2w6hYWFAgBh/Pjx9To/Li5OACDMnj272vFXXnlFACDs3btXf6xNmzYCAOHgwYP6Y1lZWYJMJhNefvll/bGFCxcKUqlUyMvL0x+rqKgQXF1dhSeeeEJ/7MknnxT8/PyEnJycaveeOnWq4OLiIigUCkEQbn3fQ0JC9Md0unTpIrRq1UooLi7WH9u/f78AQGjTpo3+WH3nTUNe56JFiwQAwm+//SbcSaPRCIIgCOvXrxfEYnG1+SAIgrBy5UoBgHDkyJEazyUisgRMCyQisgKbNm2Ci4sLRowYgZycHP2/yMhIODo6Yt++fQAAV1dXAMDWrVuhUqmabXzbt29Hr169cN999+mPOTo64qmnnkJycjIuXryoH9/169dx8uTJOq/l6uqKEydO4MaNG/W+vy5VzsnJqd7jBYAFCxZUO/7yyy8DQI10xrCwMAwYMED/tZeXFzp06IDExET9sejoaKhUKv0qIQDs3LkTBQUFiI6OBgAIgoDNmzfjgQcegCAI1X6Wo0aNQmFhYY30xxkzZsDOzk7/9Y0bN3Du3DlMnz4djo6O+uODBg1Cly5dqj23vvOmIa9z8+bNiIiI0K9G3k4kEunv26lTJ3Ts2LHafYcOHQoANe5LRGQpGFwREVmBK1euoLCwEN7e3vDy8qr2r6SkBFlZWQC0b7AnTZqEJUuWwNPTE+PHj8d3331n9DqXlJQUdOjQocZxXXfDlJQUAMC//vUvODo6olevXmjXrh3mzp2LI0eOVHvOhx9+iPPnzyMwMBC9evXCm2++We3NfW2cnZ0BaOu16jtesViM0NDQasd9fX3h6uqqH69O69ata1zDzc2tWt1SREQEOnbsiI0bN+qPbdy4EZ6envqgIjs7GwUFBVi1alWNn+OsWbMAQP+z1AkODq4xdgA1xl7bsfrOm4a8zmvXrqFz5841zrvzvhcuXKhxz/bt29f6GomILAVrroiIrIBGo4G3tzd+/PHHWh/XNanQ7Ud1/Phx/Pnnn/j777/xxBNP4OOPP8bx48errXSYQqdOnZCQkICtW7dix44d2Lx5M7788kssWrRI3x5+ypQpGDBgALZs2YKdO3fio48+wgcffIDffvsN999/f63XdXZ2hr+/P86fP9+g8ehWWu5FIpHUelwQhGpfR0dH491330VOTg6cnJzwxx9/4JFHHoGNjfbPsUajAaDtDnlnbZbOnfVvt69aNVR9541OfV9nfe7bpUsXLF++vNbHAwMDG3Q9IiJzweCKiMgKtG3bFrt370b//v3r9Wa7T58+6NOnD959911s2LAB06ZNw88//4zZs2fXO6BoiDZt2iAhIaHG8fj4eP3jOg4ODoiOjkZ0dDSUSiUmTpyId999FwsXLoRcLgcA+Pn54bnnnsNzzz2HrKws9OjRA++++26dwRUAjBs3DqtWrcKxY8fQt2/fe45Xo9HgypUr1fYOy8zMREFBQbXxNkR0dDSWLFmCzZs3w8fHB0VFRZg6dar+cS8vLzg5OUGtVmP48OGNuodubFevXq3x2J3HGjpv6qNt27b3DGLbtm2Ls2fPYtiwYUaZb0REpsK0QCIiKzBlyhSo1Wq8/fbbNR6rrKxEQUEBACA/P7/GKkO3bt0AQJ8aaG9vDwD65xjCmDFjEBMTg2PHjumPlZaWYtWqVQgKCkJYWBgAIDc3t9rzbG1tERYWBkEQoFKpoFarUVhYWO0cb29v+Pv73zO18bXXXoODgwNmz56NzMzMGo9fu3YNn376qX68ALBixYpq5+hWWsaOHVuPV11Tp06d0KVLF2zcuBEbN26En58fBg4cqH9cIpFg0qRJ2Lx5c60BSnZ29j3v4e/vj86dO+P7779HSUmJ/viBAwdw7ty5aufWd940xKRJk3D27NkaHRWBWytcU6ZMQXp6OlavXl3jnLKyMpSWljb4vkRE5oArV0REVmDQoEF4+umnsXTpUsTFxWHkyJGQSqW4cuUKNm3ahE8//RSTJ0/GunXr8OWXX2LChAlo27YtiouLsXr1ajg7O+sDCjs7O4SFhWHjxo1o37493N3d0blz53vW0WzevFm/EnW7GTNm4PXXX8dPP/2E+++/Hy+88ALc3d2xbt06JCUlYfPmzfpNb0eOHAlfX1/0798fPj4+uHTpEj7//HOMHTsWTk5OKCgoQKtWrTB58mRERETA0dERu3fvxsmTJ/Hxxx/fdXxt27bFhg0bEB0djU6dOmH69Ono3LkzlEoljh49ik2bNmHmzJkAtPVRM2bMwKpVq1BQUIBBgwYhJiYG69atw0MPPYQhQ4Y04qekFR0djUWLFkEul+PJJ5+sseHv+++/j3379qF3796YM2cOwsLCkJeXh9jYWOzevRt5eXn3vMd7772H8ePHo3///pg1axby8/Px+eefo3PnztUCrvrOm4Z49dVX8euvv+Lhhx/GE088gcjISOTl5eGPP/7AypUrERERgccffxy//PILnnnmGezbtw/9+/eHWq1GfHw8fvnlF/z999+Iiopq0H2JiMyCCTsVEhFRI93Zil1n1apVQmRkpGBnZyc4OTkJXbp0EV577TXhxo0bgiAIQmxsrPDII48IrVu3FmQymeDt7S2MGzdOOHXqVLXrHD16VIiMjBRsbW3v2ZZd1xK8rn+6dtvXrl0TJk+eLLi6ugpyuVzo1auXsHXr1mrX+vrrr4WBAwcKHh4egkwmE9q2bSu8+uqrQmFhoSAI2tblr776qhARESE4OTkJDg4OQkREhPDll1/W+3t3+fJlYc6cOUJQUJBga2srODk5Cf379xc+++wzoby8XH+eSqUSlixZIgQHBwtSqVQIDAwUFi5cWO0cQdC2KB87dmyN+wwaNEgYNGhQjeNXrlzRf28OHz5c6xgzMzOFuXPnCoGBgYJUKhV8fX2FYcOGCatWrdKfc68W+D///LPQsWNHQSaTCZ07dxb++OMPYdKkSULHjh1rnHuvedPQ15mbmyvMmzdPCAgIEGxtbYVWrVoJM2bMqNZeXqlUCh988IEQHh4uyGQywc3NTYiMjBSWLFmi/3kTEVkakSA0sAqViIiILFK3bt3g5eWFXbt2mXooRERWiTVXREREVkalUqGysrLasf379+Ps2bMYPHiwaQZFRNQCcOWKiIjIyiQnJ2P48OF47LHH4O/vj/j4eKxcuRIuLi44f/48PDw8TD1EIiKrxIYWREREVsbNzQ2RkZH45ptvkJ2dDQcHB4wdOxbvv/8+AysiIiPiyhUREREREZEBsOaKiIiIiIjIABhcERERERERGQBrrmqh0Whw48YNODk5QSQSmXo4RERERERkIoIgoLi4GP7+/jU2fr8Tg6ta3LhxA4GBgaYeBhERERERmYm0tDS0atXqrucwuKqFk5MTAO030NnZ2SDXVKlU2LlzJ0aOHAmpVGqQa1LLwflDTcH5Q43FuUNNwflDTWFO86eoqAiBgYH6GOFuGFzVQpcK6OzsbNDgyt7eHs7OziafIGR5OH+oKTh/qLE4d6gpOH+oKcxx/tSnXIgNLYiIiIiIiAyAwRUREREREZEBMLgiIiIiIiIyANZcEREREVGDqdVqqFSqOh9XqVSwsbFBeXk51Gp1M46MrEFzzh+JRAIbGxuDbMHE4IqIiIiIGqSkpATXr1+HIAh1niMIAnx9fZGWlsZ9Q6nBmnv+2Nvbw8/PD7a2tk26DoMrIiIiIqo3tVqN69evw97eHl5eXnW+8dVoNCgpKYGjo+M9N14lulNzzR9BEKBUKpGdnY2kpCS0a9euSfdjcEVERERE9aZSqSAIAry8vGBnZ1fneRqNBkqlEnK5nMEVNVhzzh87OztIpVKkpKTo79lYnOlERERE1GBM9SNrYqgAjsEVERERERGRATC4IiIiIiIiMgAGV0REREREBhQUFIQVK1Y06Rpr166Fq6urQcZDzYfBFRERERFZvZkzZ0IkEun/eXh4YPTo0fjnn39MPTSTWbt2bbXvSW3/kpOTG33t+gSH1hZEMrgiIiIiohZh9OjRuHnzJm7evIk9e/bAxsYG48aNM/WwmoUgCKisrKx2LDo6Wv/9uHnzJvr27Ys5c+ZUOxYYGGiiEVsmBldm7p2tFzF8+QFs++emqYdCREREVIMgCFAoK2v9V6ZU1/mYIf7dbRPj2shkMvj6+sLX1xfdunXD66+/jrS0NGRnZ+vPOXfuHIYOHQo7Ozt4eHjgqaeeQklJif7xwYMHY/78+dWu+9BDD2HmzJnVjhUXF+ORRx6Bg4MDAgIC8MUXX1R7fPny5ejSpQscHBwQGBiI5557rtp97pSdnY2oqChMmDABFRUV0Gg0WLp0KYKDg2FnZ4eIiAj8+uuv+vP3798PkUiEv/76C5GRkZDJZDh8+HC1a9rZ2em/H76+vrC1tYW9vb3+a7lcjqeffhpeXl5wdnbG0KFDcfbsWf3zz549iyFDhsDJyQnOzs6IjIzEqVOnsH//fsyaNQuFhYX6FbA333zzXj+eWqWmpmL8+PFwdHSEs7MzpkyZgszMzHuOAQBSUlLwwAMPwM3NDQ4ODggPD8f27dsbNY764j5XZi6ruAJXs0pwo6DM1EMhIiIiqqFMpUbYor9Ncu+Lb42CvW3j3s6WlJTghx9+QGhoKDw8PAAApaWlGDVqFPr27YuTJ08iKysLs2fPxrx587B27doGXf+jjz7CG2+8gSVLluDvv//Giy++iPbt22PEiBEAtK2///vf/yI4OBiJiYl47rnn8Nprr+HLL7+sca20tDSMGDECffr0wbfffguJRIJ3330XP/zwA1auXIl27drh4MGDeOyxx+Dl5YVBgwbpn/v6669j2bJlCAkJgZubW4New8MPPww7Ozv89ddfcHFxwddff41hw4bh8uXLcHd3x7Rp09C9e3d89dVXkEgkiIuLg1QqRb9+/bBixQosWrQICQkJAABHR8cG3RvQ7nU1YcIEODo64sCBA6isrMTcuXMRHR2N/fv3A0CdYwCAuXPnQqlU4uDBg3BwcMDFixcbNY6GYHBl5jwcbQEAuaVKE4+EiIiIyLJt3bpV/+a6tLQUfn5+2Lp1q36Pow0bNqC8vBzff/89HBwcAACff/45HnjgAXzwwQfw8fGp97369++P119/HQDQvn17HDlyBJ988ok+uLp99SsoKAjvvPMOnnnmmRrBVUJCAkaMGIEJEyZgxYoVEIlEqKiowHvvvYfdu3ejb9++AICQkBAcPnwYX3/9dbXg6q233tLfsyEOHz6MmJgYZGVlQSaTAQCWLVuG33//Hb/++iueeuoppKam4tVXX0XHjh0BAO3atdM/38XFBSKRCL6+vg2+t86BAwdw7tw5JCUl6dMTv//+e4SHh+PkyZPo2bPnXceQmpqKSZMmoUuXLgC03yNjY3Bl5jwctMFVXmmFiUdCREREVJOdVIKLb42qcVyj0aC4qBhOzk4G26C1tns3xJAhQ/DVV18BAPLz8/Hll1/i/vvvR0xMDNq0aYNLly4hIiJCH1gB2iBJo9EgISGhQcGVLui5/evbOwju3r0bS5cuRXx8PIqKilBZWYny8nIoFArY29sDAMrKyjBgwAA8+uij1Z579epVKBSKGkGTUqlE9+7dqx2Lioqq95hvd/bsWZSUlOhX9XTKyspw7do1AMCCBQswe/ZsrF+/HsOHD8fDDz+Mtm3bNup+tbl8+TICAwOr1X2FhYXB1dUVly5dQs+ePe86hhdeeAHPPvssdu7cieHDh2PSpEno2rWrwcZXG9ZcmTl3B+0nBbklXLkiIiIi8yMSiWBva1PrPztbSZ2PGeKfSCRq0FgdHBwQGhqK0NBQ9OzZE9988w1KS0uxevXqel9DLBbXqPVSqVQNGkdycjLGjRuHrl27YvPmzTh9+rS+JkupvPWeTyaTYfjw4di6dSvS09P1x3W1Wdu2bUNcXJz+38WLF6vVXelec2OUlJTAz8+v2vXj4uKQkJCAV199FQDw5ptv4sKFCxg7diz27t2LsLAwbNmypVH3a6y7jWH27NlITEzE448/jnPnziEqKgqfffaZUcfD4MrMMS2QiIiIyDhEIhHEYjHKyrS17Z06dcLZs2dRWlqqP+fIkSMQi8Xo0KEDAMDLyws3b95qNKZWq3H+/Pka1z5+/HiNrzt16gQAOH36NDQaDT7++GP06dMH7du3x40bN2pcQywWY/369YiMjMSQIUP054SFhUEmkyE1NVUfLOr+Gaq7X48ePZCRkQEbG5sa9/D09NSf1759e7z00kvYuXMnJk6ciO+++w4AYGtrC7Va3aQxtG/fHmlpaUhLS9Mfu3jxIgoKChAWFnbPMQBAYGAgnnnmGfz22294+eWXGxRINwaDKzN3Ky2QwRURERFRU1RUVCAjIwMZGRm4dOkSnn/+eZSUlOCBBx4AoG2OIJfLMWPGDJw/fx779u3D888/j8cff1yfEjh06FBs27YN27ZtQ3x8PJ599lkUFBTUuNeRI0fw4Ycf4vLly/jiiy+wadMmvPjiiwCA0NBQqFQqfPbZZ0hMTMT69euxcuXKWscskUjw448/IiIiAkOHDkVGRgacnJzwyiuv4KWXXsK6detw7do1xMbG4rPPPsO6desM8r0aPnw4+vbti4ceegg7d+5EcnIyjh49in//+984deoUysrKMG/ePOzfvx8pKSk4cuQITp48qQ8gg4KCUFJSgj179iAnJwcKhaLOe6nV6horZJcuXcLgwYPRpUsXTJs2DbGxsYiJicH06dMxaNAgREVF3XMM8+fPx99//42kpCTExsZi3759+seMhTVXZs69KrjKLWHNFREREVFT7NixA35+fgAAJycndOzYEZs2bcLgwYMBAPb29vrOfj179oS9vT0mTZqE5cuX66/xxBNP4OzZs5g+fTpsbGzw0ksvYciQITXu9fLLL+PUqVNYsmQJnJ2dsXz5cowapa1Ni4iIwPLly/HBBx9g4cKFGDhwIJYuXYrp06fXOm4bGxv89NNPiI6OxtChQ7F//368/fbb8PLywtKlS5GYmAhXV1f06NEDb7zxhkG+VyKRCNu3b8e///1vzJo1C9nZ2fD19cXAgQPh4+MDiUSC3NxcTJ8+HZmZmfD09MTEiROxZMkSAEC/fv3wzDPPIDo6Grm5uVi8eHGd7dhLSkpq1Iq1bdsWp06dwpYtW/Diiy9i4MCBEIvFGD16tD61715jUKvVmDt3Lq5fvw5nZ2eMHj0an3zyiUG+P3URCQ3dIKAFKCoqgouLCwoLC+Hs7GyQa6pUKmzfvh1jxozRt4esj8IyFSKW7AQAxL89GvIGFm6SdWjs/CECOH+o8Th3qDbl5eVISkpCcHAw5HJ5nedpNBoUFRXB2dnZaA0tyHo19/y527xuSGzAmW7mnOU2kEq0xZpMDSQiIiIiMl8MrsycSCSCm70uNZDBFRERERGRuWJwZQE8HKvasXOvKyIiIiIis8XgygKwYyARERERkfljcGUB9HtdMS2QiIiIzAR7opE1MdR8ZnBlAfTt2LlyRURERCYmkWg7FyuVfF9C1kO3D1dTO6NynysL4MG9roiIiMhM2NjYwN7eHtnZ2ZBKpXW2ydZoNFAqlSgvL2crdmqw5po/giBAoVAgKysLrq6u+g8PGovBlQXQNbRgzRURERGZmkgkgp+fH5KSkpCSklLneYIgoKysDHZ2dhCJRM04QrIGzT1/XF1d4evr2+TrMLiyAEwLJCIiInNia2uLdu3a3TU1UKVS4eDBgxg4cCA3oaYGa875I5VKm7xipcPgygLo0wLZip2IiIjMhFgshlwur/NxiUSCyspKyOVyBlfUYJY6f5gAawH0aYHsFkhEREREZLYYXFkAXVpgqVKNcpXaxKMhIiIiIqLaMLiyAM5yG0gl2kI+1l0REREREZknBlcWQCQS6VevmBpIRERERGSeGFxZCHcHbd0Vm1oQEREREZknBlcW4tZGwly5IiIiIiIyRwyuLISHY1VaIGuuiIiIiIjMEoMrC8GNhImIiIiIzBuDKwtxKy2QNVdEREREROaIwZWF0G8kzJUrIiIiIiKzxODKQjAtkIiIiIjIvDG4shD6tEC2YiciIiIiMktmEVx98cUXCAoKglwuR+/evRETE1PnuatXr8aAAQPg5uYGNzc3DB8+vMb5giBg0aJF8PPzg52dHYYPH44rV64Y+2UYlT4tkK3YiYiIiIjMksmDq40bN2LBggVYvHgxYmNjERERgVGjRiErK6vW8/fv349HHnkE+/btw7FjxxAYGIiRI0ciPT1df86HH36I//73v1i5ciVOnDgBBwcHjBo1CuXl5c31sgxOlxZYqlSjXKU28WiIiIiIiOhONqYewPLlyzFnzhzMmjULALBy5Ups27YNa9asweuvv17j/B9//LHa19988w02b96MPXv2YPr06RAEAStWrMD//d//Yfz48QCA77//Hj4+Pvj9998xderUGtesqKhARcWtdLuioiIAgEqlgkqlMsjr1F2nsdezkwiQSkRQqQVkFpTC39XOIOMiy9DU+UMtG+cPNRbnDjUF5w81hTnNn4aMwaTBlVKpxOnTp7Fw4UL9MbFYjOHDh+PYsWP1uoZCoYBKpYK7uzsAICkpCRkZGRg+fLj+HBcXF/Tu3RvHjh2rNbhaunQplixZUuP4zp07YW9v39CXdVe7du1q9HPtxRIUqkX4c+c+BDoacFBkMZoyf4g4f6ixOHeoKTh/qCnMYf4oFIp6n2vS4ConJwdqtRo+Pj7Vjvv4+CA+Pr5e1/jXv/4Ff39/fTCVkZGhv8ad19Q9dqeFCxdiwYIF+q+Lior06YbOzs71fj13o1KpsGvXLowYMQJSqbRR1/gq6RgKM4rRqXsvDGznaZBxkWUwxPyhlovzhxqLc4eagvOHmsKc5o8uq60+TJ4W2BTvv/8+fv75Z+zfvx9yubzR15HJZJDJZDWOS6VSg/8wm3JNT0cZgGIUlKlNPsnINIwxJ6nl4PyhxuLcoabg/KGmMIf505D7m7ShhaenJyQSCTIzM6sdz8zMhK+v712fu2zZMrz//vvYuXMnunbtqj+ue15jrmnuPBy1TS24kTARERERkfkxaXBla2uLyMhI7NmzR39Mo9Fgz5496Nu3b53P+/DDD/H2229jx44diIqKqvZYcHAwfH19q12zqKgIJ06cuOs1LQE3EiYiIiIiMl8mTwtcsGABZsyYgaioKPTq1QsrVqxAaWmpvnvg9OnTERAQgKVLlwIAPvjgAyxatAgbNmxAUFCQvo7K0dERjo6OEIlEmD9/Pt555x20a9cOwcHB+M9//gN/f3889NBDpnqZBqHfSLiEGwkTEREREZkbkwdX0dHRyM7OxqJFi5CRkYFu3bphx44d+oYUqampEItvLbB99dVXUCqVmDx5crXrLF68GG+++SYA4LXXXkNpaSmeeuopFBQU4L777sOOHTuaVJdlDvQbCXPlioiIiIjI7Jg8uAKAefPmYd68ebU+tn///mpfJycn3/N6IpEIb731Ft566y0DjM58MC2QiIiIiMh8mbTmihpGnxZYyrRAIiIiIiJzw+DKgujTAku4ckVEREREZG4YXFkQXVpgqVKNcpXaxKMhIiIiIqLbMbiyIM5yG0glIgCsuyIiIiIiMjcMriyISCTSr14xNZCIiIiIyLwwuLIw7g7auis2tSAiIiIiMi8MrizMrY2EuXJFRERERGROGFxZGA/HqrRA1lwREREREZkVBlcWhhsJExERERGZJwZXFuZWWiBrroiIiIiIzAmDKwuj30iYK1dERERERGaFwZWFYVogEREREZF5YnBlYfRpgWzFTkRERERkVhhcWRh9WiBbsRMRERERmRUGVxZGlxZYqlSjXKU28WiIiIiIiEiHwZWFcZbbQCoRAWDdFRERERGROWFwZWFEIpF+9YqpgURERERE5oPBlQVyd9DWXbGpBRERERGR+WBwZYE8HXUbCXPlioiIiIjIXDC4skD6tEDWXBERERERmQ0GVxaIGwkTEREREZkfBlcWyLNqr6vcEtZcERERERGZCwZXFohpgURERERE5ofBlQViWiARERERkflhcGWB9N0C2YqdiIiIiMhsMLiyQLp9rriJMBERERGR+WBwZYF0aYGlSjXKVWoTj4aIiIiIiAAGVxbJWW4DqUQEgHVXRERERETmgsGVBRKJRLc6BjI1kIiIiIjILDC4slC6uis2tSAiIiIiMg8MriyUvmMgV66IiIiIiMwCgysLxY2EiYiIiIjMC4MrC6ULrnKYFkhEREREZBYYXFkoT0fudUVEREREZE4YXFkopgUSEREREZkXBlcW6lZaIIMrIiIiIiJzwODKQum6Beax5oqIiIiIyCwwuLJQun2uWHNFRERERGQeGFxZKF1aYKlSjXKV2sSjISIiIiIiBlcWylluA6lEBADIZd0VEREREZHJMbiyUCKR6FbHQKYGEhERERGZHIMrC6aru+JGwkREREREpsfgyoLpOwZy5YqIiIiIyOQYXFkwbiRMRERERGQ+GFxZsFsbCTMtkIiIiIjI1EweXH3xxRcICgqCXC5H7969ERMTU+e5Fy5cwKRJkxAUFASRSIQVK1bUOEetVuM///kPgoODYWdnh7Zt2+Ltt9+GIAhGfBWm4enIva6IiIiIiMyFSYOrjRs3YsGCBVi8eDFiY2MRERGBUaNGISsrq9bzFQoFQkJC8P7778PX17fWcz744AN89dVX+Pzzz3Hp0iV88MEH+PDDD/HZZ58Z86WYBNMCiYiIiIjMh0mDq+XLl2POnDmYNWsWwsLCsHLlStjb22PNmjW1nt+zZ0989NFHmDp1KmQyWa3nHD16FOPHj8fYsWMRFBSEyZMnY+TIkXddEbNUt9ICGVwREREREZmajalurFQqcfr0aSxcuFB/TCwWY/jw4Th27Fijr9uvXz+sWrUKly9fRvv27XH27FkcPnwYy5cvr/M5FRUVqKi4VbdUVFQEAFCpVFCpVI0ey+101zHU9QDAVS4BAOSWVBj0umR+jDF/qOXg/KHG4tyhpuD8oaYwp/nTkDGYLLjKycmBWq2Gj49PteM+Pj6Ij49v9HVff/11FBUVoWPHjpBIJFCr1Xj33Xcxbdq0Op+zdOlSLFmypMbxnTt3wt7evtFjqc2uXbsMdq2sMgCwQXahAtu3bzfYdcl8GXL+UMvD+UONxblDTcH5Q01hDvNHoVDU+1yTBVfG8ssvv+DHH3/Ehg0bEB4ejri4OMyfPx/+/v6YMWNGrc9ZuHAhFixYoP+6qKgIgYGBGDlyJJydnQ0yLpVKhV27dmHEiBGQSqUGuWZRmQrvxu1DhUaEoSNGQS6VGOS6ZH6MMX+o5eD8ocbi3KGm4PyhpjCn+aPLaqsPkwVXnp6ekEgkyMzMrHY8MzOzzmYV9fHqq6/i9ddfx9SpUwEAXbp0QUpKCpYuXVpncCWTyWqt4ZJKpQb/YRrymu42NpBKRFCpBRQpBTjZ8xeXtTPGnKSWg/OHGotzh5qC84eawhzmT0Pub7KGFra2toiMjMSePXv0xzQaDfbs2YO+ffs2+roKhQJicfWXJZFIoNFoGn1NcyUSiW51DGQ7diIiIiIikzJpWuCCBQswY8YMREVFoVevXlixYgVKS0sxa9YsAMD06dMREBCApUuXAtA2wbh48aL+v9PT0xEXFwdHR0eEhoYCAB544AG8++67aN26NcLDw3HmzBksX74cTzzxhGlepJG5O8iQWVTBjYSJiIiIiEzMpMFVdHQ0srOzsWjRImRkZKBbt27YsWOHvslFampqtVWoGzduoHv37vqvly1bhmXLlmHQoEHYv38/AOCzzz7Df/7zHzz33HPIysqCv78/nn76aSxatKhZX1tz8XTkyhURERERkTkweUOLefPmYd68ebU+pguYdIKCgiAIwl2v5+TkhBUrVmDFihUGGqF540bCRERERETmwaSbCFPT3dpImGmBRERERESmxODKwnk6arscMi2QiIiIiMi0GFxZOKYFEhERERGZBwZXFu5WWiCDKyIiIiIiU2JwZeH03QJZc0VEREREZFIMriycuwNrroiIiIiIzAGDKwvnUbVyVapUo1ylNvFoiIiIiIhaLgZXFs5JZgOpRAQAyGXdFRERERGRyTC4snAikehWx0CmBhIRERERmQyDKyvgUVV3xY2EiYiIiIhMh8GVFdDVXXHlioiIiIjIdBhcWQFuJExEREREZHoMrqwA0wKJiIiIiEyPwZUVYFogEREREZHpMbiyAkwLJCIiIiIyPQZXVsCjKrjKYXBFRERERGQyDK6sgD4tkDVXREREREQmw+DKCrhXNbRgzRURERERkekwuLICupWrUqUa5Sq1iUdDRERERNQyMbiyAk4yG0glIgBALuuuiIiIiIhMgsGVFRCJRLc6BjI1kIiIiIjIJBhcWQluJExEREREZFoMrqwENxImIiIiIjItBldWghsJExERERGZFoMrK8G0QCIiIiIi02JwZSWYFkhEREREZFoMrqwE0wKJiIiIiEyLwZWV8KgKrnIYXBERERERmQSDKyuhTwtkzRURERERkUkwuLIS7lUNLVhzRURERERkGgyurIRu5apUqUa5Sm3i0RARERERtTwMrqyEk8wGUokIAJDLuisiIiIiombH4MpKiESiWx0DmRpIRERERNTsGFxZEW4kTERERERkOgyurAg3EiYiIiIiMh0GV1ZElxaYy5UrIiIiIqJmx+DKiujSAtnQgoiIiIio+TG4siJMCyQiIiIiMh0GV1bkVloggysiIiIioubG4MqKeDC4IiIiIiIyGQZXVkSfFsiGFkREREREzY7BlRVx1zW0YM0VEREREVGzY3BlRXQrVwqlGuUqtYlHQ0RERETUsjC4siJOMhtIJSIArLsiIiIiImpuDK6siEgkutUxsIR1V0REREREzcnkwdUXX3yBoKAgyOVy9O7dGzExMXWee+HCBUyaNAlBQUEQiURYsWJFreelp6fjscceg4eHB+zs7NClSxecOnXKSK/AvHAjYSIiIiIi0zBpcLVx40YsWLAAixcvRmxsLCIiIjBq1ChkZWXVer5CoUBISAjef/99+Pr61npOfn4++vfvD6lUir/++gsXL17Exx9/DDc3N2O+FLPBjYSJiIiIiEzDxpQ3X758OebMmYNZs2YBAFauXIlt27ZhzZo1eP3112uc37NnT/Ts2RMAan0cAD744AMEBgbiu+++0x8LDg42wujN062NhJkWSERERETUnEwWXCmVSpw+fRoLFy7UHxOLxRg+fDiOHTvW6Ov+8ccfGDVqFB5++GEcOHAAAQEBeO655zBnzpw6n1NRUYGKilvBSFFREQBApVJBpVI1eiy3013HUNeri5ud9keaXVRu9HtR82mu+UPWifOHGotzh5qC84eawpzmT0PGYLLgKicnB2q1Gj4+PtWO+/j4ID4+vtHXTUxMxFdffYUFCxbgjTfewMmTJ/HCCy/A1tYWM2bMqPU5S5cuxZIlS2oc37lzJ+zt7Rs9ltrs2rXLoNe7U/Z1EQAJ/klIxHb1VaPei5qfsecPWTfOH2oszh1qCs4fagpzmD8KhaLe55o0LdAYNBoNoqKi8N577wEAunfvjvPnz2PlypV1BlcLFy7EggUL9F8XFRUhMDAQI0eOhLOzs0HGpVKpsGvXLowYMQJSqdQg16xN8anr2JZ2EXZu3hgzpofR7kPNq7nmD1knzh9qLM4dagrOH2oKc5o/uqy2+jBZcOXp6QmJRILMzMxqxzMzM+tsVlEffn5+CAsLq3asU6dO2Lx5c53PkclkkMlkNY5LpVKD/zCNcc3beTvbAQDyyypNPhHJ8Iw9f8i6cf5QY3HuUFNw/lBTmMP8acj9TdYt0NbWFpGRkdizZ4/+mEajwZ49e9C3b99GX7d///5ISEioduzy5cto06ZNo69pSfTdAtnQgoiIiIioWZk0LXDBggWYMWMGoqKi0KtXL6xYsQKlpaX67oHTp09HQEAAli5dCkDbBOPixYv6/05PT0dcXBwcHR0RGhoKAHjppZfQr18/vPfee5gyZQpiYmKwatUqrFq1yjQvspnp97liK3YiIiIiomZl0uAqOjoa2dnZWLRoETIyMtCtWzfs2LFD3+QiNTUVYvGtxbUbN26ge/fu+q+XLVuGZcuWYdCgQdi/fz8Abbv2LVu2YOHChXjrrbcQHByMFStWYNq0ac362kzFvWrlSqFUo1ylhlwqMfGIiIiIiIhaBpM3tJg3bx7mzZtX62O6gEknKCgIgiDc85rjxo3DuHHjDDE8i+Mks4FUIoJKLSC3VIkAVztTD4mIiIiIqEUwWc0VGYdIJLotNZB1V0REREREzYXBlRVyd9CmBuaWsu6KiIiIiKi5MLiyQvqOgWxqQURERETUbBhcWSEP/coV0wKJiIiIiJoLgysr5K6ruWJaIBERERFRs2FwZYWYFkhERERE1PwYXFkhDza0ICIiIiJqdgyurBC7BRIRERERNT8GV1ZInxbIhhZERERERM2GwZUVurWJMFeuiIiIiIiaC4MrK+RetXKlUKpRrlKbeDRERERERC1Do4KrtLQ0XL9+Xf91TEwM5s+fj1WrVhlsYNR4TjIbSCUiAKy7IiIiIiJqLo0Krh599FHs27cPAJCRkYERI0YgJiYG//73v/HWW28ZdIDUcCKR6LbUQNZdERERERE1h0YFV+fPn0evXr0AAL/88gs6d+6Mo0eP4scff8TatWsNOT5qJHYMJCIiIiJqXo0KrlQqFWQy7crI7t278eCDDwIAOnbsiJs3bxpudNRo3EiYiIiIiKh5NSq4Cg8Px8qVK3Ho0CHs2rULo0ePBgDcuHEDHh4eBh0gNc6tjYSZFkhERERE1BwaFVx98MEH+PrrrzF48GA88sgjiIiIAAD88ccf+nRBMi13Xc0V0wKJiIiIiJqFTWOeNHjwYOTk5KCoqAhubm7640899RTs7e0NNjhqPKYFEhERERE1r0atXJWVlaGiokIfWKWkpGDFihVISEiAt7e3QQdIjePBhhZERERERM2qUcHV+PHj8f333wMACgoK0Lt3b3z88cd46KGH8NVXXxl0gNQ4Ho7atMDMonITj4SIiIiIqGVoVHAVGxuLAQMGAAB+/fVX+Pj4ICUlBd9//z3++9//GnSA1Did/JwAAAkZxVAoK008GiIiIiIi69eo4EqhUMDJSfvmfefOnZg4cSLEYjH69OmDlJQUgw6QGqeVmz0CXO1QqREQm1Jg6uEQEREREVm9RgVXoaGh+P3335GWloa///4bI0eOBABkZWXB2dnZoAOkxusV7A4AiEnKNfFIiIiIiIisX6OCq0WLFuGVV15BUFAQevXqhb59+wLQrmJ1797doAOkxtMFVyeS8kw8EiIiIiIi69eoVuyTJ0/Gfffdh5s3b+r3uAKAYcOGYcKECQYbHDVN76rg6kxaAcpVasilEhOPiIiIiIjIejUquAIAX19f+Pr64vr16wCAVq1acQNhMxPs6QBPRxlySirwz/VC/UoWEREREREZXqPSAjUaDd566y24uLigTZs2aNOmDVxdXfH2229Do9EYeozUSCKRSL96xborIiIiIiLjatTK1b///W98++23eP/999G/f38AwOHDh/Hmm2+ivLwc7777rkEHSY3XK9gd287dxImkPMwz9WCIiIiIiKxYo4KrdevW4ZtvvsGDDz6oP9a1a1cEBATgueeeY3BlRnqHaFeuTqfkQ6XWQCpp1GIlERERERHdQ6Peaefl5aFjx441jnfs2BF5eexMZ07aezvBxU4KhVKNCzeKTD0cIiIiIiKr1ajgKiIiAp9//nmN459//jm6du3a5EGR4YjFIvQMYt0VEREREZGxNSot8MMPP8TYsWOxe/du/R5Xx44dQ1paGrZv327QAVLT9Q52x+5LmYhJysNTA9uaejhERERERFapUStXgwYNwuXLlzFhwgQUFBSgoKAAEydOxIULF7B+/XpDj5GaSFd3FZOUB7VGMPFoiIiIiIisU6P3ufL396/RuOLs2bP49ttvsWrVqiYPjAwnzM8ZDrYSFJVXIiGjGGH+zqYeEhERERGR1WHruBbARiJGJOuuiIiIiIiMisFVC6HfTDiZ3RyJiIiIiIyBwVULoQ+ukvIgCKy7IiIiIiIytAbVXE2cOPGujxcUFDRlLGREXVq5QGYjRk6JEteySxHq7WjqIRERERERWZUGBVcuLi73fHz69OlNGhAZh8xGgu6tXXE8MQ8xSXkMroiIiIiIDKxBwdV3331nrHFQM+gV7FEVXOXi0d6tTT0cIiIiIiKrwpqrFqRPVd3VCdZdEREREREZHIOrFqR7azfYiEW4WViO6/llph4OEREREZFVYXDVgtjZStC1lbZu7kQSW7ITERERERkSg6sWplewBwBuJkxEREREZGgMrlqY3iG36q6IiIiIiMhwzCK4+uKLLxAUFAS5XI7evXsjJiamznMvXLiASZMmISgoCCKRCCtWrLjrtd9//32IRCLMnz/fsIO2UJFt3CAWASm5CmQUlpt6OEREREREVsPkwdXGjRuxYMECLF68GLGxsYiIiMCoUaOQlZVV6/kKhQIhISF4//334evre9drnzx5El9//TW6du1qjKFbJGe5FGH+zgCAmGSuXhERERERGYrJg6vly5djzpw5mDVrFsLCwrBy5UrY29tjzZo1tZ7fs2dPfPTRR5g6dSpkMlmd1y0pKcG0adOwevVquLm5GWv4FqlXEOuuiIiIiIgMrUGbCBuaUqnE6dOnsXDhQv0xsViM4cOH49ixY0269ty5czF27FgMHz4c77zzzl3PraioQEVFhf7roqIiAIBKpYJKpWrSOHR01zHU9ZoiqrUL1hwBjl/LNYvx0L2Z0/why8P5Q43FuUNNwflDTWFO86chYzBpcJWTkwO1Wg0fH59qx318fBAfH9/o6/7888+IjY3FyZMn63X+0qVLsWTJkhrHd+7cCXt7+0aPoza7du0y6PUao0QFADa4ml2KX/63HY5SU4+I6ssc5g9ZLs4faizOHWoKzh9qCnOYPwqFot7nmjS4Moa0tDS8+OKL2LVrF+Ryeb2es3DhQixYsED/dVFREQIDAzFy5Eg4OzsbZFwqlQq7du3CiBEjIJWaPppZm3oEV7JK4douEiPDfO79BDIpc5s/ZFk4f6ixOHeoKTh/qCnMaf7ostrqw6TBlaenJyQSCTIzM6sdz8zMvGezirqcPn0aWVlZ6NGjh/6YWq3GwYMH8fnnn6OiogISiaTac2QyWa31W1Kp1OA/TGNcszF6h3jgSlYpTqcWYWxEK1MPh+rJXOYPWSbOH2oszh1qCs4fagpzmD8Nub9JG1rY2toiMjISe/bs0R/TaDTYs2cP+vbt26hrDhs2DOfOnUNcXJz+X1RUFKZNm4a4uLgagVVL1btqM+ETbGpBRERERGQQJk8LXLBgAWbMmIGoqCj06tULK1asQGlpKWbNmgUAmD59OgICArB06VIA2iYYFy9e1P93eno64uLi4OjoiNDQUDg5OaFz587V7uHg4AAPD48ax1uyXsHazYQv3ixCUbkKznJ+okRERERE1BQmD66io6ORnZ2NRYsWISMjA926dcOOHTv0TS5SU1MhFt9aYLtx4wa6d++u/3rZsmVYtmwZBg0ahP379zf38C2Wj7McQR72SM5V4HRyPoZ09Db1kIiIiIiILJrJgysAmDdvHubNm1frY3cGTEFBQRAEoUHXZ9BVu17B7kjOVeBEUh6DKyIiIiKiJjL5JsJkOrq6K24mTERERETUdAyuWjBd3dU/1wuhUFaaeDRERERERJaNwVUL1srNDv4uclRqBJxJLTD1cIiIiIiILBqDqxZMJBLpV69OJOWZeDRERERERJaNwVUL1zuEdVdERERERIbA4KqF061cnUktQEWl2sSjISIiIiKyXAyuWrgQTwd4OtqiolKDf64Xmno4REREREQWi8FVC3d73VUM666IiIiIiBqNwRXp97tiUwsiIiIiosZjcEX6lavTyXmoVGtMPBoiIiIiIsvE4IrQwccJznIblCrVuHCjyNTDISIiIiKySAyuCGIx666IiIiIiJqKwRUBYN0VEREREVFTMbgiALfqrk4m50GjEUw8GiIiIiIiy8PgigAA4f7OsLeVoLBMhYTMYlMPh4iIiIjI4jC4IgCAjUSMyDZuAFh3RURERETUGAyuSK9PiLbuisEVEREREVHDMbgiPV3d1YmkXAgC666IiIiIiBqCwRXpdW3lAlsbMXJKlEjMKTX1cIiIiIiILAqDK9KT2UjQPdAVAFMDiYiIiIgaisEVVdObdVdERERERI3C4Iqq6a2ru0pk3RURERERUUMwuKJqurd2hY1YhBuF5TiWmGvq4RARERERWQwGV1SNva0NxncLAAA892MsktnYgoiIiIioXhhcUQ3vPNQZEa1cUKBQ4Yl1J1GoUJl6SEREREREZo/BFdVgZyvB6ulR8HORIzG7FHM3xEKl1ph6WEREREREZo3BFdXK21mOb2ZEwd5WgsNXc7D4jwtscEFEREREdBcMrqhO4f4u+HRqd4hEwIYTqfjuSLKph0REREREZLYYXNFdjQjzwRv3dwIAvLPtIvbGZ5p4RERERERE5onBFd3T7AHBeKRXIDQC8PyGM4jPKDL1kIiIiIhqSMktRVZxuamHQS0Ygyu6J5FIhLfGd0a/th4oVarx5NpT/MVFREREZiWjsByjVxzCwyuPQa1hnTiZBoMrqhepRIyvpkUixNMB6QVleOr70yhXqU09LCIiIiIAwMEr2ShTqZGSq8CJxFxTD4daKAZXVG8u9lJ8O7MnXOykiEsrwCubzrKDIBEREZmF49duBVR//nPThCOhlozBFTVIsKcDVj4WCRuxCFv/uYkVu6+YekhERETUwgmCgKO3BVc7zt/kHp1kEgyuqMH6tvXAuxM6AwA+3XMF/4tLN/GIiIiIqCVLzlUgo6gcthIx3B1ska9Q4dg1pgZS82NwRY0S3bM1nh4YAgB49dd/cDol38QjIiIiopbq6LUcAEC31q4Y08UXALD1nxvNcu+Sikpo2ECDqjC4okb71+iOGBnmA2WlBk+vP4W0PIWph0REREQtkG6Vql9bD4zr6g8A2HE+A8pK46YGHr2Wg4glO/HB3/FGvQ9ZDgZX1GhisQgrpnZDuL8zckqUmL3uFIrLVaYeFhEREbUggiDog6u+IR7oGeQObycZisorcfhqtlHvvepgItQaARtOpBo9kCPLwOCKmsTe1gbfzIiCt5MMCZnFeP6nM6hkASkRERE1k8uZJcgtVUIuFaNba1dIxCKM6eIHANh61nhdA9MLynDgsjZ4Ky6vxKErxg3kyDIwuKIm83OxwzczoiCXirE/IRtfH0w09ZCIiIiohThWVW8V1cYdMhsJAGBcV21wtfNiptH25dx4Mg2370izle3fCQyuyEC6tnLFOw91AaBdImd6IBERETUHXQv2vm099Md6tHaDn4scJRWV+tUlQ6pUa/DLyTQAwMx+QQCAXUYM5MhyMLgig5nQPQAhXg4oLFNh/fEUUw+HiIiIrJxaI+BEUh6A6sGVWCzCWF1qoBFWlA5czkZGUTnc7KV4/f6O8DdiIEeWhcEVGYxELMK8IaEAgG8OJUGhrDTxiIiICADySpUoVDCjgKzPpZtFKCxTwVFmg64BLtUeGxeh7Rq451ImypSGXVH6KSYVADA5shXkUgnGdjVeIGfJ0vIUWLAxDufTC009lGbD4IoM6sEIf7R2t0deqRIbTqSaejhERC1ebkkFRiw/gDH/PcSGQ2R1dF0CewW7w0ZS/W1tRCsXBLrbQaFUY298lsHuebOwTH+96J6tAQBjuxovkLNky3ddxm9n0vHE2pPILq4w9XCaBYMrMigbiRjPDW4LQFt7xdxjIiLTWnUoEbmlSqQXlOFadqmph0NkULrNg/uGeNR4TCQSYWwXbdBjyA2Ffzl5HRpBG9CFejsCqB7I7UswXCBnyYrKVdh+TruSl1VcgZc2xkHdAjZbNovg6osvvkBQUBDkcjl69+6NmJiYOs+9cOECJk2ahKCgIIhEIqxYsaLGOUuXLkXPnj3h5OQEb29vPPTQQ0hISDDiK6DbTezRCv4ucmQVV2DTqTRTD4eIqMXKLanA+mO3amAv3mw5qTlk/VRqDWJqqbe6na5r4N74LJRWNL1cQa0R8EvVe5tHe7XWHzdWIGfJ/oi7gYpKDQJc7WAnleDw1Rx8tveKqYdldCYPrjZu3IgFCxZg8eLFiI2NRUREBEaNGoWsrNqjfoVCgZCQELz//vvw9fWt9ZwDBw5g7ty5OH78OHbt2gWVSoWRI0eitJSf2DUHWxsxnqlavVp5IJGb6hERmcjqQ0lQ3JaidCG9yISjITKsc+mFKFWq4WInRZifc63nhPs7I9jTARWVGuy+lNnkex68ko30gjK42EkxunP196GGDuQsne4D9ln9g/DuhM4AgE/3XMGRqzmmHJbRmTy4Wr58OebMmYNZs2YhLCwMK1euhL29PdasWVPr+T179sRHH32EqVOnQiaT1XrOjh07MHPmTISHhyMiIgJr165FamoqTp8+bcyXQreZEhUILycZ0gvKsOXMdVMPh4ioxckrVeL7Y8kAgFHhPgCACzcYXJH10NVb9Qlxh1gsqvUckUikD3oM0Wzip6p68ok9AiCXSqo9Fu7vjCAPe5SrDBPIWbL4jCKcvV4IG7EID3UPwMQerRAdFQhBAF78+QyyispNPUSjsTHlzZVKJU6fPo2FCxfqj4nFYgwfPhzHjh0z2H0KC7VpEO7u7rU+XlFRgYqKW0V2RUXaPz4qlQoqlWG6K+muY6jrmTsJgNn922Dpjsv4Yt9VPNjFp0ahKdVfS5s/ZFicPy3T1/uvQqFUI9zfCc8ODMbfFzJx4UYhlEolRKLa34jeiXOHmsLY8+fIVW3b815Bbne9x+gwL3y29yr2J2Qhr1gBJ7m0UffLKq7AnqpGFg9396/1nvd39sFXB5Kw9ewNjAn3btR9rMHGqm6KQzp4wUUmhkqlwv+NaY+zafmIzyzBvA2xWDcz8q7vDc3p909DxmDS4ConJwdqtRo+Pj7Vjvv4+CA+Pt4g99BoNJg/fz769++Pzp0713rO0qVLsWTJkhrHd+7cCXt7e4OMQ2fXrl0GvZ45c1MDDjYSpOaV4d0f/kZPL+svYjS2ljR/yPA4f1qOEhXwXawEgAh9nQpwLfYwJCIJisor8cOWv+Ahb9j1OHeoKYwxfyo1wMkk7RxXXT+P7dvP13muIAA+dhJklgEfb9yNXo18P7LzughqjQTBTgKunD6I2qqHnEsBwAb74jPx2x/bITfpO23TqNQAv5zW/myChZvYvv1WDdokP2BZtgQxyfl4YfVOjGt979IRc/j9o1Ao6n2u1f/I586di/Pnz+Pw4cN1nrNw4UIsWLBA/3VRURECAwMxcuRIODvXnsPbUCqVCrt27cKIESMglTbuExNLlOmciI93X8XRAmf85/F+dS7b09211PlDhsH50/J8vOsKlJokhPk54bVpfSASifBt6jFcvFkMn46RGBnmc++LgHOHmsaY8ycmOQ+qE6fg4WCLJyaNuOdqbKLdNfx33zVcF/vgzTE9Gnw/jUbAR58cAlCOZ0Z0xpjuAbWeJwgCNt04isScUohbd8OYbv4Nvpel+/tCJkpPnIW3kwwvTR1QY3XKq30G5v/yD3alixE9NBKD2nvVeh1z+v2jy2qrD5MGV56enpBIJMjMrJ6XmpmZWWezioaYN28etm7dioMHD6JVq1Z1nieTyWqt35JKpQb/YRrjmuZs5n0hWH04GYk5pdidkKvfZI8ap7nmz+mUPLz26z94ZWQH3N+FPzNr0dJ+/7RU+aVKrD+uTcmZP7w9bG1tAQCdA1xw8WYxEjJLMTaiYfOAc4eawhjz52SKtuSjb1sP/Ry/mwe7t8J/913Dkau5KFUJcLW/93Nud/ByNq4XlMNJboMHuwVCeke91e0eiPDHp3uuYMeFLDzcs02D7mMNfovT1rZNimwFO3nN99cP9QjE6dRCrD+eglc3n8e2FwbA39WuzuuZw++fhtzfpEUwtra2iIyMxJ49e/THNBoN9uzZg759+zb6uoIgYN68ediyZQv27t2L4OBgQwyXGsFJLsWs/trv/2d7r0DTAvY3sAaf7LqCa9mleGXTWaTl1X8pnIhM75vDiShVqhHm54wRt61Q6bqpsakFWYOjVc0s6mrBfqdQb0d09HVCpUbA3xcyGny/n09WNbLoHgA727oDKwB4IEL7oeTBK9koVJi+Xqg5ZRSWY3/VPl8PR9a9sPF/4zqhc4Az8hUqPP/TGaisaINzk3cYWLBgAVavXo1169bh0qVLePbZZ1FaWopZs2YBAKZPn16t4YVSqURcXBzi4uKgVCqRnp6OuLg4XL16VX/O3Llz8cMPP2DDhg1wcnJCRkYGMjIyUFZW1uyvj7QtOB1lNojPKNYXgpL5Ss4pxeGqNqmlSjUW/NIyNv0jsgb5pUqsPZIMAJg/vF21VKnwABcADK7I8pUp1YhLLQAA9GvrWe/nPRCh24eqYV0Ds4srsPOCNstq6m17W9Ul1NsJHX2doFIL+PtiwwM5S7Y5VrvBcs8gN4R4OdZ5nsxGgi8fjYST3AanU/Lx0d/Wsx+tyYOr6OhoLFu2DIsWLUK3bt0QFxeHHTt26JtcpKam4ubNW/8nuHHjBrp3747u3bvj5s2bWLZsGbp3747Zs2frz/nqq69QWFiIwYMHw8/PT/9v48aNzf76CHC1t8XjfbXL4p/tvQJB4Bt1c/ZT1adzEa1c4CizwcnkfKw6mGjiURFRfdS1agUAnfycIRIBGUXlyC2pqOMKRObvdEo+lGoNfJ3lCPKof+MxXUv2o9dyG/T/gV9PX0elRkC3QFd0qmM/rTuN7WK49u+WQhAE/d5WD0cF3vP81h72+GhyBABg1cFE7LpoHe3rTR5cAdraqJSUFFRUVODEiRPo3bu3/rH9+/dj7dq1+q+DgoIgCEKNf/v379efU9vjgiBg5syZzfeiqJrZ9wVDLhXjn+uFOHjFujePs2TKSg1+PaXdl+y5IaFY/EAYAGD5rgRcuFFoyqER0T3klyqx7mgKAODFO1atAMBRZoMgDwcAwMWbXL0iy3X0mvZ9RL+2HvXeVgAA2ng4oEuAC9QaAX+dr9+KkkYj6FMCH63HqpXOuKpVsiNXc5Bfqqz38yzZyeR8JOcq4GAr0QeX9zK6sy+eqCofefmXOKsoRTCL4Iqsn4ejDNN6V61e7eHqlbnaeTEDuaVK+DjLMKyjNyZHtsKocB+o1AJe2hiHcpXa1EMkojp8ezgJJRWV6OTnXGc3wDB/1l2R5TuW2LB6q9vd2lD4xj3O1DqemIuUXAUcZTYYF1H/Bk/Bng4I93eGWiNgRyNqvCzRL1WrVuO6+sNBVv+eea/f3xERga4oKq/EvA2xUFZadv0VgytqNk8NDIGtjRinUvJxPDHP1MOhWmyo2nk+OioQNhIxRCIR3pvQBZ6OMlzOLLGqnGgia1KgUGLt0WQAwIvDaq5a6YQzuCILV1JRiX+u3+oU2FC6rsUnkvKQVVx+z/M3VG2GO76bP+xtG9Zke2wDAzlLVlJRiW1VKZBTetbdyKI2tjZifPFod7jYSXH2eiHe237JGENsNgyuqNn4OMsRXZWD+9ne2rbeI1NKyinF0Wu5EImAKT1v5Up7OMrw4eQuALSfjB+5yrROInNTn1UrAAj31zW1sL4039KKSuw4fxNlSq6wW7OTSXlQawS0drdHK7f611vptHKzR/fWrhAE4K9zd19Ryi2p0HcWfKQBKYE647poUwOPXctFjpXXOW775wbKVGqEeDmgR2u3Bj+/lZs9lk/R1l+tPZqM7ecst1aNwRU1q2cGt4VUIsLRa7k4ncLVK3Pyc9Wnc4Pbe9X4gzW0ow+m9db+YXll01kUlrWs1rJE5qxAcatD4IvDQu+6WbuuHXtSTilKKyqbY3jNokChxNRVx/HMD7H4cv/Vez+BLJau3qpvSMNXrXTGddV1Dbz7itJvselQqQV0beWCzlXdNhuitYc9Ilq5QCOg3jVelmrjSW1K4JSowAbVwd1uWCcfPD0oBADwr1//QUquZdZfMbiiZhXgaodJPbTLxZ/t5R9Ac1FRqcam09pGFnV9OvfvsZ0Q5GGPm4XlWPy/8805PCK6izWHk1BcUYmOvk4YGeZ713O9nGTwdpJBEID4DOtIDcwtqcAjq0/gXLp2Nc5aOo5R7XT1Vv1CGx9cjemi/f/JyeR83CysfZseQRDwU9WHjlN7NnzVSkcfyJ213tTAq1nFiE0tgEQswsTuAU261isjO6BnkBuKKyrx/M9nobLA8isGV9Tsnh3cFhKxCPsTsnHuuvWlpliinRcykVeqhK+zHEM7etd6jr2tDT6J7gaJWITf427gTyv+Q0FkKQoVKnynX7Vqd9dVKx1rqrvKKi7HI6uP49LNIng62kIkAuIzipFZdO9aGrI8BQqlft42ZeXKz8UOPYO0qWvb6miVfiIpD4k5pbC3leDBbv6NvteYqrqrmOQ8g83Lo9dy8OwPp82ms96mqi7DQzp4wdtZ3qRrSSVifPZID7g72OJSRjG2JFteqGJ5IyaL18bDAeOrWpSy9so86BpZTOmpbWRRl+6t3TB3SCgA4P9+P4+MQr6BITKlb4/cWrUaFX73VSsdfd1VumUHVxmF5Zi66jguZ5bAx1mGjU/3Rdeq1K2Dl7ObZQxnUvPx1f5rqFRb4MfrFuhEUh4EAWjr5dDkN/G3UgNrD65+uq2RhWMDOt/dKcDVDj30NV5NryO6dLMIc9adwl/nM7D4jwtNvl5TqdQabI5NB1C/va3qw9dFjhXR3SASAccyRUjMLjXIdZsLgysyieeGhEIkAnZezMQl7rdiUonZJTiWmAuxCIjuee9fjM8PDUXXVi4oLFPh1V/PQqNhW30iUyhUqPDd4SQA9V+1Am5bubppuZkD6QVliF51DInZpQhwtcMvT/dFWy9HDGzvBQA40AzBlSAImL8xDh/siMfGqhbUZFzHrjW+Bfud7u/iC7EIiEsrqLEClF+q1NdINaaRxZ3uFcjVV1ZxOWavO4XSqqYte+OzcLwqTdJU9idkI6ekAp6OtnVmvjTGwPZeWDi6A+aFa5tkWBIGV2QSod6OGFO1wdwX+1h7ZUo/VxWhDu7gjQBXu3ueL5WI8Ul0N8ilYhy6koPvjyUbeYREVJs1VatWHXzqv2oF3Fq5upxRApUFrrik5SkQ/fUxpOQqEOhuh41P90Gbqs2RB1UFV4eu5EBt5A9+rmSV6Avu1x9L4f6NzUAXXPVr69nka3k7ydE7WBukbbtjRem3M+lQVmoQ5ueMLo1oZHGnMV38IBIBp1LycaOg9hqveylXqfHU96eRXlCGEE8HTKiqbXr/r3iTzj3d3lYTe7SC9C6ZL40xq18btHU26CWbBYMrMpl5Vell287dxNWsEhOPpmWqqFTj16pGFg3Zeb6tlyPeGNMJALD0r3hczSo2yviIqHaFZSqsOVK1ajW8/qtWABDobgcnuQ2Uao3F/e5NyinFlK+P4Xp+GYI9HfDL032rdTftFugKJ7kNCstUOHu9wKhj2X3pVuOM+IxixKbmG/V+LV1OSQUSMrV/a/o0od7qdrpNgW/vGnh7I4tHerdudOe72/m6yNGzjTsANKrFuCAIeGXTWcSlFcDFTopvZ/bEwjEdYW8rQVxagb5dfHPLKi7H3vgsAMDDkQ3b28qaMbgik+nk54wRYT4QBLB1ronsOJ+hb2QxuINXg577eJ82GNjeCxWVGszfGGfxO6oTWZI1h5NQXK5dtRrdgFUrABCJRPqW7JbU1OJqVjGivz6Gm4XlCPV2xMan+sDPpfpqu41EjAHttKsaxq672l3VldDNXgpAu3pFxqNLf+vo6wR3B1uDXPP+zn6QiEU4n16E5BxtXc/plHxczSqBnVSC8U1oZHGnW4Fcw4OrFbuvYOs/N2EjFmHlY5EI9nSAt5Mcs+8LBgB8uCPBJHV/W2LTodYI6N7aFe18nJr9/uaKwRWZ1PNDtatX/4u7gVQL3c/Akuk+nYu+RyOL2ohEInw0uStc7aU4n17E5iREzeT2VasXGlBrdTtL20w4IaMYU1cdR1ZxBTr6OuHnp/rU2dBgYDvj113llFTgTFoBAOD9SV0BANvPZSDXyjeKNaWjBqy30nF3sEW/quvpVq82VP1dHNfVD85yqcHuNbpz3TVed/O/uHR8ukf79/XdCZ2rvf45A0Pg7mCLxJxS/FLVsa+5CIKgTwmcYqBGFtaCwRWZVNdWrhjU3gtqjYD3d1xCiRVtamnurmWX4HhiXr0bWdTGx1mO9yZ0AaCtnePG0ETG990R7apVex9H3N+5YatWOpbUjv18eiGmrjqGnBIlwv2d8dOcPvB0lNV5vq6pxdm0AuSXKo0ypr3xWRAEoHOAM0aF+6JrKxco1Zpmf4Pbkhw3YL3V7R64rdlEoUKlb83+SO+mN7K43e01XvVNDTydko9Xf/0HAPD0wBBE37HflpNcqv+QesXuy1Aom+89VGxqAa5ll0IuFWNcVbt50mJwRSb3wjDtL4bt5zLQb+kefLwzgZ/+NYOfqtqvD+ngDf96NLKoy5gufpjYPQAaAXhp41mUMkAmMprCMhXWHG7aqhUAhFUFV5duFJl1x8+zaQV4dPVx5CtUiAh0xYbZfeB2j5Qwf1c7tPdxhEYADl/NMcq49lTVWw3r6AMAeKxPGwDAhpgUozfSaIkyCsuRmFMKsQjoFexu0GuPCveFVCJCfEYxlu1MQEWlBh19ndA90NWg9wEalhqYlqfA0+tPQVmpwYgwH7w2umOt5z3auzUC3e2QVVyh3/OuOWyqWrUa08UPTgZc4bMGDK7I5CLbuOPzR7sjxNMBReWV+GzvVfT/YC8W/e+82WyQZ23KVWpsjq1qZGGAT+feHB+OAFc7pOYp8M62i02+HhHVbu2RZBSVV6KdtyPGdG78p8Wh3o6wtRGjuKISafnm+Xv2dEoeHvvmBIrKKxHZxg0/PNkLLvb1exM3yIgt2ctVahy8rA3aRoRpg6sHuvrDWW6DtLyyZttjqyU5lqj9fncOcIGLnWHfyLvYS3FfqHY1bP1xbd3c1J6BBmlkcSddjde59EJ9jVdtistVmL3uFHJKlAjzc8aK6G6Q1PFBisxGgldGdgAArNx/DXlGWq29nUJZiT/PatMoo5kSWAODKzIL47r6Y9eCQVj5WA9EtHJBuUqD74+lYPCy/Zj/8xnuhWVgf1/IQL5CBT8Xuf5NSFM4y6VY9nAERCLgp5g0faE3ERlOUbkK3x5OBNC0VStAu6VCh6oCdHNMDTyRmIvHv41BcUUlege74/snejXo0/FB7bX77Ry8nG3wNtXHEnNRplLD11muT6+0s5XoN1D94TgbWxja0auGr7e6nW4fKgCQ2YgxobtxOt/dXuN1Z/t3nUq1Bs//dAYJmcXwdpLh25lRcLjHJsYPdPVHmJ8ziisqm2V7m+3nMlCqVCPIw97gK4nWgMEVmQ2JWITRnf3w+9z+2DCnNwa084RaI+D3uBu4/9NDmPVdDGKS8riXiAFsONH4RhZ16dvWQ9+56I0t51BRqTbIdYlIa93tq1Zdml7jcKvuyryaWhy9moMZ38VAoVTjvlBPrJ3V655vLu8UFeQGuVSMrOIKxGcYdqsI3YdHwzp5V1vdmFaVBbA3IcuoWRdqjdDiurPqm1kYqAX7nUaE+8C26m/h2K5+9V4hbQxdfVJdqYHvbr+E/QnZkEvF+GZGVI2OmLURi0V4/X5t2uD6YylGz/rRNbJ4OMo4K3yWjsEVmR2RSIR+bT2x/sne2Pr8fRjX1Q9iEbAvIRtTvj6GSV8dxa6LmWZdJ2DOrmaV4ERS0xpZ1OXlkR3g5yJHVnEFtp5t2k70RHRLaUWlvkPgvKGhdaYINYQ5NrVQqTV44eczKFdpMLiDF76ZEQU7W0mDryOXSvRvxA2ZGigIAvZc0u7rM7yTT7XHQrwccV+oJwThVsc5QyutqMSITw4gfPEOjP/iCN784wL+F5eOtDyF1X7wmJanQHpBGWzEIvQMMs4qibNcislRrSCXivFE/2Cj3ENnVLgvbMQiXLpZhGvZ1feZW388RV83tXxKN3Rt5Vrv6w5o54n+oR5QqjX4ZNdlA464uqScUsRUvYeY2CPAaPexZAyuyKx1DnDB54/2wN6XB+PR3q1hayNGbGoB5nx/CqNWHMSvp69DZYK9HSyZrv360I7e9fpErCHkUgke76st7F5zJMlq/9gTNbefYlKRr1AhyMO+WgpTU4Tp27GbT3B17FouckqU8HCwxdePR0IubXhgpaOvu0owXHB14UYRMorKYSeV1Jqipmts8cvJNKOs3n++7yoSs0uhUgs4m1aAtUeT8eLPcRjw4T70fHcP5nx/Cl/uv4pj13KbtXOcMR29pq23igh0bfAKZkO8M74z4haNROcAF6PdAwBc7W31e7Hd/iHkoSvZePOPCwCAV0d1aPDqtEgkwr+qml5siUs3WjmFrpHFwPZeBn8PYS0YXJFFCPJ0wHsTuuDwv4bg2cFt4SSzwZWsEryy6SyGfryfe2TVk6EbWdTmkZ6tIZeKceFGEU4m5xvlHkQtSblKjVUHtbVWzwxqa5BVKwDo5OcEkQjILq5AVnG5Qa7ZVLoW1aM6+0Jm0/jACgAGddDWXZ1KyTNYF9PdVV0CB7TzrDXwG97JG77OcuSWKrHjfIZB7qmTmF2Cbw5p58H7E7vg06ndMLNfECJaucBGLEJOSQV2XczEhzsS8Mjq4+i8+G+M+fQQ/r3lHH49fR3Xskss8gOvY/oW7MZJCdQRi0VNCuYbYmzVByTbzmmbQlzNKsZzP8ZCrREwsXsAnhvctlHX7drKFeO6+kEQgA93xBtsvDqVao3+PQT3tqobgyuyKN5OcvxrdEccWTgUr9/fEV5OMqTlleGVX88yTbAedpzPQIFCBX8Xub7g29DcHGz1xcC6ltFE1Hi/nr6OrOIK+LnIMbGH4Qrt7W1tEOLpAAC4aAarV5VqDf6+oA1IxhqgpizIwx6B7nZQqQX9G/Sm0gVXw8N8an3cRiLGI720H1wZsrGFIAh4a+tFqNQCBrX3QnTPQIzvFoA3HwzH/+bdh/NLRmHzs33xf2M7YWwXP/i7yKERgIs3i/DjiVS8sukshn18AH2W7rGo/QgFQTB6vZUpjKyq8bqcWYLjibl4Yu0pFJdXIqqNG5ZO6tKkOqZXRnaAjViEfQnZBpv3Ooeu5CCzqAJu9lIM62Sc9xDWgMEVWSRnuRTPDGqL357tB3tbCWKS8vDDCXZoupdbjSxaG+zT79o80T8IALDzYgbb6RM1gUqtwcoD1wAATw0Mga2NYf9sh5tRauDxxDzkK1Rwd7BFbwN0IBOJRAZtyX6zsAzn04sgEmnTqusytVcgbMQinEzON1hq1p5LWdifkA2pRITFD4TVePMtl0oQ2cYdsweE4ItpPXB04TAcXzgMX03rgacGhqBnkBtkNmJkFlXguR9jkWMhe0km5pQiq7gCtjZi9GjjZurhGIyzXKrf7HrmdzFIzVMg0N0OXz8e2eQV2yBPB31myvs74g26WqlrZDGhe6smj9OaMbgiixbobq/PMX7/r3i+kb+Lq1nFiEk2TiOLO7XzccKAdp7QCMD3x5KNei8ia/ZH3A1czy+Dh4MtpvY0fCqvbjNhc1i50rWmHhXuY7AuproVekMEV7pGFt0DXeHpKKvzPB9nOUaGa1e2DLF6Va5S462t2v0Dn7wvBCFejvV6nq+LHPd38cMbYzph0zP9EPufEQj1dkRmUQVe2hhnEdkeulWrHq1dmy1lr7k8ULWhcLlKAyeZDdbM6AmPu8yrhnh+aDvY20pwNq3AYOmpuSUV+pXbKT2N06reWjC4Iov3eJ826BXsDoVSjX9t/scic8qbw4YT2k+chnb0ga+L3Oj303Vc+vlkGkoMVO9A1JJoNAK+3K/ds+bJAcGN6pp3L+bSjr1SrcHOqpRAQ7SZ1+nb1gNSiQipeYq7btpaH3su6Vqw154SeDtdY4vfz6SjuFzVpPuuPpiI1DwFfJxleH5oaKOv4yCzwZfTekAuFePQlRz93DJnx/X1Vp4mHonhDevkAyeZDSRiET6f1gPtqvadMwQvJxnmDAgBAHz0d4JBGn/9HncDKrWArq1c0NHXucnXs2YMrsjiicUifDipK+RSMY5eyzVaC1xLdnsji2lGamRxp0HtvRDi6YDi8kpsPn29We5JZE3+vpCBa9mlcJbb4PGqN+uGpksLTM5VNDkIaIqYpDzklirhZi9FHwPW1jjKbBDVRpti2JTVK4WyEkeq3uiPqKPe6nZ9QzzQ1ssBpUo1fj+T3uj7pheU4YuqIOiNMZ2a3C2vvY8T3h7fGQCwfNdlHE80bE2OIWk0Ao4lNk8zC1NwlNngt+f64X9z++vTVw1pzsAQeDjYIjGnVJ/O1xgqtQbrjibj093a9u4Ps5HFPTG4IqsQ5OmAV0dp0wOXbo9HekGZiUdkXv46fxOFZSoEuNrp87yNTSwWYWZV7dXao8kWkYJCZC4EQcDn+7Rvqmf2C4KT3Dibmro72MKvaiX70k3DbrbbELqUwJFhvpAaKCVQZ6AB6q4OXcmBslKDQHc7tPO+d1qeSCTSr179cDy10RkV7227hHKVBr2C3PFghGFa8D8cFYhJPVpBIwAv/HQG2cXmWX91OasYeaVK2EklDdrvyZK083EyWut3R5kNXhjWDgCwYveVBrfmFwQB++KzMHrFQSz+4wKKyisR7u+MCd25t9W9MLgiqzGzXxAi27ihpKISrzM9sJpbjSwCjdrI4k6TerSCk9wGSTml2H85q9nuS2Tp9l/OxoUbRbC3lWCWkTc1NXVqoFoj6LsEjulquJRAHd2qwLFruShXNW7vqd0Xq7oEdvKpdye3iT1awU4qQUJmcaO2pTh6NQfbzt2EWAS8+WB4kzrI3enth8LRztsRWcXa+iu1GX74dfSqdtWqZ7C7wRu5tBSP9GqN1u72yC6uaFD33oSMYkxfE4NZa0/iWnYp3B1s8fZDnfG/uf3haMS9xqwFZytZDYlYhA8nd4XMRptPvukUU9EA4HLVH3aJWNTs+1I4yGz0bYnXHE5u1nsTWSpBEPDFXu2q1bTereHmYGvU+5l6M+GYpDzklCjhYic1SvpXJz8neDnJUKZS41Qjghy1RsDeeO2HQ8PrUW+l42Inxfhu2tWm9Q1sbKFSa7C4akPZx/q00TceMRR7W239lZ1UgsNXc/DFPvOrv9KlBFpTC/bmZmsjxiujOgAAVh5IRF6p8q7n55ZU4N9bzuH+Tw/i0JUcSCUiPDUwBPteGYzH+7QxWKMZa8fvElmVtl6OWDCiPQDg7W0XkVFoHhtjmtJPVTVoQzt6N0sjiztN79sGYhFw+GoOLmeaLu2IyFKcSMrDqZR82NqI9UXpxnRr5co0wdV2fUqgj8FTAoHqLdkPXml4amBcWgFyS5VwktugVwNbxOtSA3ecv9mg9Lvvj6XgSlYJ3Oyl+r9phtbOxwnvPKStv1qx+zKOXssxyn0aQ60R9PVg1lhv1ZzGdfFD5wBnlFRU4vO9tQfRFZVqfH3gGgZ/tB8/nkiFRgBGh/ti94JBeGNMJ7jYGSct2VoxuCKrM3tACCICXVFcXok3tpxr0emB5Sq1vpnEo83UyOJOrdzsMSrcFwDw3RFuKkx0L7pVhClRreDtbPwPRHTB1ZXMYlRUNi5trrHUGgE7jJgSqKOvu0poeHCl6xI4uIN3g4O/zgEu6BboCpVaqHdTgeziCqzYpW0e8NrojnC1N97K5aTIVng4Ult/9eLPcWZTf3XxRhGKyyvhJLPRz09qHLFYpN+yZv3x5Gpb1giCgL/O3cSI5Qex9K94FFdUonOAM35+qg9WPh6JNh4Ophq2RWNwRVZHIhZh2eSusJWIsTc+C7/FNr5Tk6Xb9s9NFJVXahtZtGueRha1eeI+bc3Ib7Hp90xLIGrJzqYV4NCVHEjEIjw9sG2z3DPA1Q4udlJUagRcySxplnvqnErOQ3ZxBZzlNuhvxHbbA0I9IRIBCZnFuFnYsIZHur19hneqe+Pgu9F1etxwIrVetU0f7tC+ye0S4NIsqdxvje+M9j6OyC6uwPyNZ0xef3U+vRBL/tSmRPYOcWcqmgEMaOeF+0I9oVILWF4VuJ+7XojoVcfx7I+xSM1TwNtJho8md8Ufc+8zaMfOlogzlqxSOx8nvDhc2yVnyZ8XkFXU8tIDlZUarKlaKZrazI0s7hTVxg2dA5xRUanRpykSUU26DoEPdQtAoLt9s9xTJBIhzM80mwnrUgJHhPkatWmBm4MtIqo6zh1sQNfA1FwFLmeWQCIWYXD7xgVXY7v6wdVeivSCMuyLv3tjn9jUfGyqyjZYMj68WX5v29lK9PVXR67m4rO9V4x+z9pcz1fgpY1xGPfZYW1arESMGf2CTDIWa6Rbvfo9Lh3P/nAaD35xGDFJeZBLxXhhaCj2vTIYD0cFQmzC9wrWgsEVWa2nB4agS4ALisor8e/fz7e49MC3t17EhRtFcJLZILqXafelEIlE+k2Fvz+WbJANDYmsTXxGEXZdzIRIBDw3pHlWrXRM0TFQoxHw13ltSuDYrr5Gv5++7upy/WuLdKtWPYPc4GLfuLoTuVSiX4G6W2MLjUbAm1VNLCb1aIUerd0adb/GCPV2wrsTtPVXn+65gqNXm6/+qlChwnvbL2HosgPYUrUn2Phu/tjz8iAMMGHGhbXp0soFD0T4QxCAv85nQBCAh7r5Y+/Lg7FgZIcm76FGtzC4IqtlIxHjo4e7QioRYdfFTPxx9oaph9RsfjmZpv8jvmJqN3g7NX8jizuN7eoHT0cZMosq9J9WE9EtX+67BgAY09kPbb3uvZeSIYUHNH9Ti9Op+cgqroCT3Ab9Q42XEqgzqIP2jfqhK9morOcHPHvib7Vgbwrd5u0Hr2QjJbe01nN+OZWGf64Xwklmg3/d36FJ92uMiT1aIToqEIIAvPBzHLKKjZvxUa5SY/XBRAz8aB9WHUyEUq1B3xAP/DnvPnw6tXuzrdy2JK+N6oDW7vboFeyO3+f2x4qp3eHvamfqYVkdBldk1Tr6OmPeEG164Jt/XDCbYl1jiksrwP/9fh4A8NLw9hjWxDcFhiKzkehrD747kmzawRCZmeScUmz9R/sBUHOvWgFAeFU79ks3i5ptw+9t/1SlBHbygcxGYvT7RbRyhYudFEXllTh7veCe5xeVq3AiMQ9A04OrNh4OGNjeC4Jwa9/B2xUqVPjw7wQAwIvD25nsA7E3HwxHBx8n5JRU4MWfjLP/lUYj4Pcz6Rj28QG8u/0SCstU6ODjhO9m9cSGOb3RpZVxNtUlINDdHgdfG4Jfnu6LboGuph6O1WJwRVbvuSFt0cnPGfkKFRb977yph2NU2cUVeGb9aSjVGowI88HzQ0NNPaRqpvVpDVuJGHFpBYhNbfh+M0TW6qv916ARtFsm6AKd5hTi6QCZjRilSjWS61hZMSSNRsCOqpTAMV2M1yXwdhKxCPe1066Q1adr4IGEbFRqBIR6OyLIs+ld03QfLv1yKq3GZsbLdyUgr1SJdt6OJq0zsrOV4ItpPWBvK8GxxFz8d49h66+OXM3BA58fxvyNcUgvKIOvsxwfTe6K7S8OwJAO3gbdKJnIVBhckdWTSsT4aHJX2IhF+Ot8hv7TUmujUmsw98dYZBSVI8TLAcunRJhdYaqnowwPVm2q2ZDd4oms2Y2CMvx2RtvEYO4Q03wgYiMRo6Nf86UGnknLR0ZRORxlNhjQ3vgpgTq6uqsD9Whqoau3GtbILoF3GtrRGwGudshXqKqlRl+6WaRP437zwXCj7PXVEKHejnhvQhcAwH/3XsHhK02vv4rPKMaMNTGY9s0JfS3wa6M76JsomLLhEpGhMbiiFqFzgAueHaxNtVn0v/PILbG+9MB3tl5ETHIeHGU2WPV4FJzk5rnp36z+QQC0BbUNbYlMZI1WHUyESi2gb4gHIts0XxODOzXnZsLb/tGuWg3v5N0sKYE6uuDqn/TCu24LoVJr9J39RhgotVoiFuGRquZCP1QFU4IgYPEfF6ARgDFdfJul9qw+HuoegKk9tfVX8zeeaXDHXWWlBukFZYhNLcCPV8V48MtjOHA5G1KJCLP6B+HAa0Pw3OBQ2Nk238+eqLmwNQi1GPOGhuLvCxm4nFmCN/+8iM8e6W7qIRnMplNpWHdM+8f6k+huCPVu3mL4hgj3d0HvYHecSMrD98dS9O1hiVqi7OIK/fYE80ycxttcHQO1XQK1KzfNlRKo4+MsR0dfJ8RnFOPQlWyM7xZQ63mnkvNRVF4JdwdbdDdg174pPQPx6Z4riE0twPn0QiTmlOrbYf97bJjB7mMIbz4Yjri0AsRnFOOFn8/ghyd7QywSIbdUicyi8qp/FcgsKkdWcTkyCrVfZxWXI6fk9sBV+zn+uK5+eHVUB25MS1aPwRW1GDIbCZY9HIEJXx7Fn2dvYFxXP4wKN377X2M7m1aAf1c1sHhxWDuMCDOPBhZ388R9wTiRlIcNJ1LxwtB2/PSSWqxvDyeholKDboGu6NfWtBt33r7XlSAIRqt/ibtegJuF5XCwlWBg++ZvtT2ovRfiM4px4HLdwdWeqpTAIR28DZqy5u0kx6hwX2z95yZWHUxETJK2YcZzg0MRYGZd2+RSbf3Vg58dxvHEPPR+bw8Ky1SorGeTC6lEBG8nGVxFCiyZ0hdRweaxKkdkbAyuqEXp2soVTw0MwVf7r+HfW86jd7A7XO1tTT2sRsspqcAzP5yGslKD4Z288eKwdqYeUr0M7+SDQHc7pOWVYcuZdDxa1aaYqCUpVKj06WHzhoSavJi/o68zxCJUrUxUwNfFOB3rtlfVvQ7r5AO5tPk/WBnU3gtfH0zEwcs50GiEGrWpgiDo662GG6je6naP92mDrf/c1G8P0trdHk8NDDH4fQyhrZcj3pvYBS/+HIfcqjRKkUhbP+vjLIOPkxw+LnLt/zrLqv23m70t1OpKbN++HRHsAEgtCIMranFeHNYOuy5m4mpWCaZ8fQz/Gt0RQztaXpciXQOLm4XlCPF0wPLobmbXwKIuErEIM/oG4Z1tl7DmSBIe6RVocd9/oqZadywZJRWV6OjrZLCmCU1hZytBWy9HXMkqwcWbhUYJrgTh1sbBzZ0SqBMZ5AZ7WwlySipwKaOoRnfGa9mlSM5VwFYixgAjrKz1CnZHex9HXM4sAQD8Z1yYSYLM+hrfLQCh3o5QqQX4OMvg5SiDTT2bbqjV9z6HyNqwoQW1OHKpBMunRMDFTorLmSV4ct0pTPn6GE4l55l6aA3y7rZLOJFU1cBieiSczbSBRV2m9AyEg60EV7NKcMgA3agsQWGZCs+sP43lOxOabS8hMk+lFZVYc0TbMfM5M1i10tHXXaUbp6nF2euFSC8og72tBIM7NH9KIKBNEdelYNbWNVC3atWnrQccZYb/DFokEmFW/2AA2g6CxlgdM7Rwfxd0C3SFn4tdvQMropaK/w+hFqlrK1cceHUwnh4UApmNGCeT8zF55THMXncS8RnG75TVVJtPX8fao8kAgI+nRCDU28m0A2oEZ7kUD0dpO2d9d6RltGX/fO8V7LiQgf/uvYoFv8ShUq0x9ZDIRDacSEWBQoVgTweMNdEKTm10qzjG6hioa0E+tKO3SVdrdLVete13pau3GmHEoGdqz0D8/FQffPFoD7MJrInIMBhcUYvlam+Lhfd3woFXh+CRXtp9NnZfysL9nx7Cgl/ikJanMPUQa3XueiHe2HIOAPDC0FCLbsoxs18QRCJgX0I2rmWXmHo4RpWWp8C6o9r6GrEI+D3uBp77MRYVlcybaWnKVWqsOpQIAHh2UFuz2uNHv3J10/AdAwVB0AdXpg4odS3ZT6fko7hcpT+eV6rE6RTtBudDDdSCvTYikQh9QjzYzIfICplFcPXFF18gKCgIcrkcvXv3RkxMTJ3nXrhwAZMmTUJQUBBEIhFWrFjR5GtSy+brIsfSiV2x86WBGNPFF4IA/BabjmEfH8CSPy+Y1Z5YuaVKPPPDaVRUajCsozfmD29v6iE1SZCnA4Z11H46vPZIsmkHY2TLdiZAqdagf6gHVj0eBVsbMXZezMTsdadQpmSA1ZJsOn0d2cUV8HeR46HutXerM5WwquAqLa8MhWWqe5zdMOfSC3E9vwx2UgkGdzBtKlwbDwcEedijUiPg6LVc/fF98VnQCNrOiebWvY+ILIPJg6uNGzdiwYIFWLx4MWJjYxEREYFRo0YhKyur1vMVCgVCQkLw/vvvw9e39k/sG3pNIkDbFenLaZH439z+6NfWA0q1Bt8dScbAD/dhxe7LKKmoNOn41ALw4sazSC8oQ7CFNbC4G13twa+nr6NQYdg3c+bin+sF+F+ctjPYwvs7YXiYD76b2RP2thIcupKD6WtOoKjcOl87VadSa7By/zUAwNOD2sLWxuR/hqtxtbfVBxUXDZwauO22lEBzWLHRrV4dvK3uyphdAomoZTD5b/Xly5djzpw5mDVrFsLCwrBy5UrY29tjzZo1tZ7fs2dPfPTRR5g6dSpkMplBrkl0u4hAV2yY0wfrn+yFzgHOKFWqsWL3FQz6cB++O5JksjSu/6WIcSIpHw62Eqx6PBIudpbVwKIu/dp6oIOPE8pUamw8lWrq4RicIAh4b/slAMCE7gHoHKCtaekf6on1T/aGk9wGJ5PzMW31CeSXKu92KbICW2LTkV5QBi8nGaJ7Bpp6OLUyxmbCgiDgr3Om7RJ4J33d1eVsCIKAikq1PtAabgH7BRKReTJpK3alUonTp09j4cKF+mNisRjDhw/HsWPHmu2aFRUVqKi4lfpVVKT9tE6lUkGlMsynybrrGOp6ZHx9glyx+ane2HEhE5/suYrkXAWW/HkR3x5KxCsj22Nsl+arddp8Og0Hbmo/C/lgYmcEucutai7N6BuIN36/iLVHkjEm3BveTjKrKfLel5CN44l5sLUR48WhIdV+bl39HbF+VhRmrTuNc+mFmPL1UaydGQVvp9o/OGos/v4xD5VqDT7bewUAMLt/G0iggUplfk1NOvo4YufFTJy/XgCVSvt7rqlz58KNIqTmKSCXinFfW1ezmIuRgc6QSkS4nl+Gyze1XQxLlWp4O8nQwcveLMZo6fi7h5rCnOZPQ8Zg0uAqJycHarUaPj7VPyHy8fFBfHx8s11z6dKlWLJkSY3jO3fuhL29faPGUZddu3YZ9HrUPF4IBY67iLAjTYzrBeWY/8s/+PNwHMa11sDYmXnn80RYe1kMQISRARqoU05je4px79ncpGrAwUaCG4XluO+jg7CXCPC1B3ztBfjZVf23nQAnqXYDS0uhFoAPz0oAiHCfdyXOHt2Hs7Wc90w74MuLElzJKsX4T/djbrga7oaNrwDw94+pxWSLkJYvgaONALe8i9i+/aKph1QrRZ4IgAQnLt/Arl1pAJo+d/5MEQMQo4NTJfbv3tn0QRpIsKMYlwvFWPnHQeSUiQCIEWpXhh07/jL10KwKf/dQU5jD/FEo6t/kjJsIA1i4cCEWLFig/7qoqAiBgYEYOXIknJ2dDXIPlUqFXbt2YcSIEZBKrSOdq6V5AMD/KdX48kAiVh5Mwp4bYkjd/PDRpM5GaSksCAJWH07GN8evQBCAzm4afDxrKOQyW4PfyxxI2tzEZ3uvISVPAYVahMRiILG4eiTlZi9FqLcj2ns7op23A0K9HdHO2xHuDub5Pfnl1HVklF2Eq50UH828D853SeUcnqfAjLWncT2/DKuuOWLdzEgEezoYZBz8/WN6ao2AFf89AkCBZ4a0x4SBwaYeUp26F5bjm4SDyKoQY8DgQTi0f2+T5o4gCFi+QvvaZw3vhjHNuOp/Lzeck/HB35eRK/XG1fxSAOWYMTISQ020B5e14e8eagpzmj+6rLb6MGlw5enpCYlEgszMzGrHMzMz62xWYYxrymSyWuu3pFKpwX+YxrgmNR+pVIrXx4Shg58zXvv1H+y4kInM4gp8Mz0KHo6GW2qoqFTj37+fx6+nrwMAHu3VClGiZMhltlY7fyZGtsbEyNYoV6mRmF2KK1nFSMgoxuXMElzJKkZqngL5ChVOJufjZHJ+ted6Osrw2ugOmBJlPjUsCmUlPt2rbVzw/LB28HC++yp4Wx8X/PpMP0z75jiuZZfi0W9PYf2TvdDJzzAf8AD8/WNK2+PSkZSrgKu9FDPvC4FUar6fbQZ62MDNXop8hQrJ+dqU+abMnQs3CpGSp4DMRowR4X5m9dqHdvLFB39fxqGruRAEQC4VY1AHH0hNuAeXNeLvHmoKc5g/Dbm/SRta2NraIjIyEnv27NEf02g02LNnD/r27Ws21yS604TurbD+yd5wsZPiTGoBJnx51GD7NOWWVOCxb07g19PXIRYBSx4Mx5IHwiAxefuZ5iGXShDm74zx3QLw2uiO+GZGFA68OgQXl4zG1ufvw/IpEXhmUFsM6+iNVm7armY5JRVYuv0SylXm09L8m0NJyCquQKC7HR7r07pez/F1kWPj030R5ueMnJIKTF11HHFpBcYdKBmdRiPg871XAQBP9g+Go8x8govaiEQi/WbCF28WN/l6ur2tBnfwgoOZvfb2Po7wdZZDELRf3xfqZdLNjYnI8pn87dqCBQuwevVqrFu3DpcuXcKzzz6L0tJSzJo1CwAwffr0as0plEol4uLiEBcXB6VSifT0dMTFxeHq1av1viaRIfQJ8cBvz/VDa3d7pOYpMPHLoziRmHvvJ95FQkYxxn9xBCeT8+Ekt8HaWb0wo1+QYQZs4exsJegc4IKJPVrh9fs74tuZPXH4X0NxfskoBLjaIV+hwv/i0k09TABAdnEFvj6gXbV6bVRHyGzq/2bN01GGn57qg+6tXVFYpsK01cdxvInzikxrx4UMXMkqgZPcBjP6B5l6OPWi6xh48WbT2rFrNw42ry6BtxOJRBjY3lP/9YgwtmAnoqYxeXAVHR2NZcuWYdGiRejWrRvi4uKwY8cOfUOK1NRU3Lx5U3/+jRs30L17d3Tv3h03b97EsmXL0L17d8yePbve1yQylLZejtjyXD/9G+HHv43B72ca9wZ/z6VMTPzyCK7nlyHIwx5bnuuvbxVMdXOU2WBmVQD63ZFkCLqPoE3o0z2XUapUI6KVC8Z1bfgbShc7KX54sjf6tfVAqVKNGWtisD+B+/TdTVG5Ckk5pahUm1f3PY1GwH/3aDsEzuofDGe5ZaRGhemDq6atXMVnFCMppxS2NmIM62Sef4MHtb8VUA3pyOCKiJrGLNbn582bh3nz5tX62P79+6t9HRQUVK83T3e7JpEheTjK8NOcPnhpYxz+Op+B+RvjkJqnwPNDQ+vVTlwQBKw+lIilf8VDEIC+IR746rEecLU3zyYN5mhKz0B8svsy4jOKcexaLvqFet77SUZyNasEP8VoO6y9MaZTo1vKO8hssGZmT8z9MRZ74rMw5/tT+O/U7rjfDD/9N7X8UiXGfXYY6QVlkNmI0c7HER19ndHR1wkdfZ3RwdcJXgZub19fuy9lIj6jGI4yGzxhIatWAPRpgQkZxdAENP46upTAQe29zDYdcnAHL/QKckdHPyd4O8lNPRwisnDm+ZuOyMLIpRJ88WgPfLAjHl8fTMTyXZeRmqfAexO6wNam7gViZaUG/95yDpuqGlc80qs13hofDmlLKbAyEBc7KSb1aIX1x1Ow5kiySYOrD3bEQ60RMLyTD3qHeDTpWnKpBCsfj8T8jXHY9s9NPPtjLB6M8MfLI9ujjYdhOgkaQ0WlGnmlSvi52Bn9XoIg4F+b/0F6QVnVvTU4n16E8+nV09k8HW3RoSrY0gVd7XwcjVpfIwgCPquqtZrRr41FfWAS7OkAO6kEZSo1ssoadw1BELCtKrgaa8YfCjjIbPDLM6zJJiLDYHBFZCBisQgLx3RCoLs9Fv1P2+nvRkEZvnosEi61tODOLanAsz/EIiY5D2IR8J9xYZjZL8hqNs9tbjP7B2H98RTsic9ESm6pSYKPmKQ87LqYCYlYhNfv72CQa0olYvx3and4ONji+2Mp+OPsDWw/dxOP9m6N54e2M9mKTG2Sc0rxU0wqNp2+jrxSJT6c1BVTehq3g+OGmFTsvJgJqUSE357tDye5DeIzihCfUYz4m8VIyCxGcm4pckqUyLmaiyNXb9WviUVAkKcDuga44OWRHRDobth9DfcnZONceiHsbSV48r4Qg17b2CRiETr6OeFMagGulzbud9LlzBIkZpfCViLGsE5MtyOiloHBFZGBPdanDQLc7DDvx1gcvZaLyV8dxZqZPau9cUvIKMaT607ien4ZnGQ2+OzR7hjcgW8+mqKtlyMGd/DC/oRsrD2ajMUPhDfr/QVBwHvbLwEAonsGItTbyWDXlohFeGt8Z0yJCsSHfyfg4OVsfH8sBb+evo7Z9wVjzsAQOJmolkel1mDPpUz8eCIVh67kVHvs/34/j1AfR/Ro7WaUe1/JLMbbW7Ub8f5rdEd0aaVNZQvydMDozrdWShTKSlzJLKkWdMVnFCFfoUJidikSs0vxT3ohtjzXv9YPQhpDEAR8WlVr9XifNma7F9vdhPs740xqAdIbGVzpVq0Gtvc02fwkImpuzD0iMoIhHbyx6Zl+8HWW40pWCSZ8eRRnq1pq743PxKSvjuJ6fhlau9tjy9x+DKwM5In+2o1ZN526juJyVbPee/u5DMSlFcDeVoL5w9sZ5R6dA1zw/RO9sGFOb0S0coFCqcZ/917FoI/249vDSaiobL5W9OkFZVi+MwH939+LZ36IxaErORCJgCEdvLB6ehRGhftAqdbg2R9OI6u43OD3L1ep8cLPcShXaTCgnaf+Z18be1sbRAS6Irpnayx+IBw/PdUHsf8ZgZg3hmHtrJ7wc5EjMbsUL/x0BmqNYRqiHL6ag7i0AsilYsweYFmrVjq6uqvrisY9/6+q4MocuwQSERkLV66IjCTM3xlb5vbDE2tP4dLNIkSvOoZJPVrhp5hUaASgd7A7Vj4WCTcL/ETbXA1o54lQb0dczSrBplPX8cR9db/hNiRlpQYf/h0PAHhqYIjRi+L7tfXE73P7Y8f5DHz0dwISc0rx9taLWHM4CQtGtMdD3QMgERs+vVStEXDwcjZ+PJGCvfFZ0MUhno62iO4ZiKk9W+tXaPu29cC1L47galYJnvshFhvm9Llr/WFDfbgjAZduFsHDwRYfT4mAuIGvVyQSwdtZDm9nOVZPj8LklUdx4HI2PtgRjzfGdGrS2AThVofAR3u1MavUzYbQtWO/Xiqqs5GUIAjIKVEiNa8UKbkKpOQqkJqnQHJuKa5klUAqEWF4mHl2CSQiMgYGV0RG5Odih03P9MXcH2Nx4HI2fjyRCgCY2jMQb43vbNA3m6R9wzyzXxD+7/fzWHcsGTP6BRklyLjTjydSkJKrgJeTDHOaaZVCJBLh/i5+GBHmg02nr2PF7stILyjDy5vOYvWhRLw2ugOGdPA2SA1fVnE5Np26jg0nUvWNIwCgX1sPTOvdBiPCfGrMZUeZDVY9HonxXxzBqZR8vLX1At55qEuTxwIA+xKysOZIEgDgo4e7NjmY7RzggmUPR2DehjNYdTARHXycMCmyVaOvdzwxDyeT82FrI8bTgyxz1QoA2vs4QSIWQVEJnEkrhEojQkpeKVKrgqiUPAVSc0tRqqx7xXRsFz+LaT9PRGQIDK6IjMxRZoNvZ0Th7a0X8VtsOl4c3g5P3hfMxhVGMrFHAD76OwEpuQrsjc/CCCN/al5YptKvUrw0vD0cmrndtI1EjEd6tcZD3QKw9mgyvtp/FfEZxXhi7Sn0CnLHv+7vgK7+d6//EgQB5SoNistVKCqvRFG5CsXllShQKLHzQib+vpCByqplKhc7KR6ObIVHerdGWy/Hu143xMsRn07thifXncIPx1PRJcAF0T1bN+n1ZhdX4NVNZwEAM/sFYWhHw/x8x3X1R0JGMT7bexULfzuHYC+HRteK6ebD1J6B8HG23NbecqkEoV4OSMgsQfTqmDrPE4kAfxc7tHa3RxsPe7TxcEAbD3u0drdHJz/nZhwxEZHpMbgiagY2EjGWjO+MxQ+ENzh9iRrG3tYGU3sF4usDifjuSJLRg6uv9l9DvkKFUG9HTIlq/GpHU9nZSvDs4LZ4pFcgvtp/DWuPJiMmOQ+TvjqGIR08oS4UY//mcyhRaoOo4vLKqn/a/668R61RZBs3TOvdGmO6+DWoffnQjj54aXh7LN91Gf/5/QLa+ziheyODFo1GwCubziKnRIkOPk54/f6OjbpOXV4a3h7xGcXYdTETT68/jT/n3Qdfl4YFRyeT83AsMRdSiQjPDGpr0PGZwohO3kjI1Kb3aYMnh9uCKO3XrdzsILMxXkt7IiJLwuCKqBkxsGoe0/sG4ZtDSTh6LRfxGUXo6GucT8/TC8r06Wmvj+4IGzPYn8zV3hYLx3TCzP5BWLHrCjadTsO+hBwAYiDj5l2fKxZpV1qd5FI4yW3gLJeik58TpvZq3aQViHlDQnE+vRA7L2bimR9O48/n72tUKt/ao8k4cDkbMhsx/vtId4PvUSUWi/BJdDdM+vIoEjKL8dT6U/jl6b4Nuo9u1WpyZCD8XY2/z5exvTgsFP4llzHxgfshl7E+lIjoXhhcEZHVCXC1w+hwX2w7dxPfHU7GB5O7GuU+H+9MgLJSg97B7ma3j4+fix0+mNwVcwYGY9PJNFxLvIZuYR3g6iDTBk52UjjLbwVSTnIpHGwlRklXFYtFWB7dDQ9VNbiY+2MsfpzdsAYXF28U4f2/tE1D/m9sJ3TwNVyr+9s5ymzwzYwoPPj5YfxzvRD/2vwPVkR3q9f35UxqPg5dyYFELMJzgy1/1UrHQYpmqV0kIrIGpv+YlYjICGb1DwIAbIlLR25JhcGvf+FGIbacSQcAvDGmk9nW0IV6O+GVke3wQGsNnh4YjMf6tMH4bgEY0sEbkW3c0d7HCX4udnCU2Rj1NTjKbPD145FwktngZHK+fn+q+ihTqvHCz2egVGswvJM3HuvTxmjjBIBAd3t8OS0SNmIR/hd3AysPJNbreZ/tvQoAmNg9wOAbEhMRkWVgcEVEVimyjRu6tnKBslKDn2JSDX799/+KhyAAD0T4IyLQ1eDXt0ZtvRyxYmo3iETA+uMp+OVkWr2e9862i7iaVQJvJxk+nBzRLIFs37YeWPygdiPqD/+Ox55LmXc9/9z1QuyNz4JYBMwdEmr08RERkXlicEVEVkkkEulXr9YfT4FKrTHYtfclZOHQlRxIJSK8NqqDwa7bEgzrpG1wAQD/9/t5nEnNv+v5f1/I0G9hsHxKN7g3475wj/dpg2m9W0MQgBd/jsOVzOI6z/1sr7bWany3AAR5OjTXEImIyMwwuCIiqzW2iz+8nGTILKrA9nN3b+ZQX1cyizH/5zgA2sYZTP9quHlDQjEyzAdKtQbP/hCLrOLyWs/LKCzHvzb/AwB4emAI7mvn2ZzDBAAsfiAcvYPdUVJRidnfn0KBQlnjnEs3i7DzYiZEXLUiImrxGFwRkdWytRHj8ar6nDVHkpt8vRsFZZi+JgaFZSp0C3TFyyPbN/maLZFYLMLHUyLQ1ssBGUXlmPtjLJSV1VcW1RoBL22MQ4FChS4BLnh5pGlWCG1txPhyWg+0crNDSq4CczfEovKOVdDPq2qtxnbxQ6j33ff+IiIi68bgiois2qO9W8NWIsbZtALE3iMF7W4KFEpMXxODm4XlaOvlgO9m9oS9LRuuNpaTXIpV06PqbHCx6mAijiXmwt5Wgk+ndmtQZ0FD83CUYfX0KNjbSnDkai7e2XZJ/9iVzGJsP69dFX1+aDtTDZGIiMwEgysismqejjKM7+YPAFhzOKlR1yhTqvHE2pO4mlUCX2c5vn+yN9yasfbHWrX1csQn0d0AVG9wcTatAB/vTAAAvPlgOEK8TL8a1MnPGcundAOg3W/r56omKZ/vuwpBAEaH+xqtPTwREVkOBldEZPVm9Q8GAPx1PgM3C8sa9FyVWoO5G2IRm1oAZ7kNvn+yFwKsYHNYczE8rHqDi6NXc/Diz2dQqREwtqsfHo5sZeIR3jK6sy8WjNCO9T//O49fTqbhz7M3AADPD2OtFRERMbgiohYgzN8ZfULcodYIWH8spd7PEwQBC387h73xWZDZiLFmZk+09+HqhKE9PzQUI6oaXEz79gSScxUIcLXDew91Mbv9w54fGoqxXfygUgt4bfM/0AjA8E7eCPd3MfXQiIjIDDC4IqIWQbd6tSEmFWVKdb2e88GOBPx6+jokYhG+eLQHooLcjTnEFkssFmF5VYMLQQDEIuCT6G5wsZeaemg1iEQifPRwV4T5OeuPsdaKiIh0GFwRUYswvJMPAt3tUKBQ4fe49Hue/82hRKw8cA0AsHRiFwwP8zH2EFs0J7kUq6dHYUA7T7w3oQt6BZtvIGtva4PVM6IQ7u+Mx/q05ibSRESkx1ZXRNQiSMQizOgbhHe2XcJ3R5IwtWdgnSlnv59J13eEe210B0yJCmzOobZYIV6OWP9kb1MPo14CXO2w7YUBph4GERGZGa5cEVGLMaVnIBxsJbicWYIjV3NrPefA5Wy8suksAGBW/yA8O6htcw6RiIiILBiDKyJqMZzlUjxctQr13ZGabdnj0grw7A+nUakR8GCEP/4zNszsGioQERGR+WJwRUQtyox+QRCJgD3xWUjKKdUfv5ZdglnfxUChVGNAO08sezgCYjEDKyIiIqo/BldE1KIEezpgSAdvAMC6o8kAgIzCckz/Ngb5ChW6tnLBV49FwtaGvx6JiIioYfjugYhanCeq2rJvOpWG6/kKzFgTg/SCMgR7OuC7mT3hKGOvHyIiImo4BldE1OL0D/VAex9HlCrV/9/e3cdUWf9/HH9dCOd4uJUbhYOKN5PwpsF3guDJXEuYSs3CaFljDbXNmUe/mnNruhRdbrpqZTajW+37mymFG2b+NCMzWs67cBgWMm2WbqjkSkEK5ee5fn84z35nGvGTS69z4PnYznbO9bm8eONeO9tr1w16dP13arzQqgExTv3X7FwlRjvtHg8AAIQoyhWAXscwDP8fFb78V4dinOH6z+xcDU6ItHkyAAAQyihXAHqlon8NVHKsU87wML1fmqNR7li7RwIAACGOGwsA9EouRx/9978n6tr/+JTaz2X3OAAAoAegXAHotZK4vwoAAFiIywIBAAAAwAKUKwAAAACwAOUKAAAAACxAuQIAAAAAC1CuAAAAAMAClCsAAAAAsADlCgAAAAAsQLkCAAAAAAtQrgAAAADAApQrAAAAALAA5QoAAAAALEC5AgAAAAALUK4AAAAAwAKUKwAAAACwQLjdAwQj0zQlSS0tLZYds6OjQ3/++adaWloUERFh2XHRO5AfdAf5wZ0iO+gO8oPuCKb83OwENztCZyhXt9Ha2ipJGjx4sM2TAAAAAAgGra2tiouL63Qfw+xKBetlfD6fmpqaFBMTI8MwLDlmS0uLBg8erLNnzyo2NtaSY6L3ID/oDvKDO0V20B3kB90RTPkxTVOtra1KTU1VWFjnd1Vx5uo2wsLCNGjQoLty7NjYWNsDgtBFftAd5Ad3iuygO8gPuiNY8vNPZ6xu4oEWAAAAAGAByhUAAAAAWIBydY84nU6VlZXJ6XTaPQpCEPlBd5Af3Cmyg+4gP+iOUM0PD7QAAAAAAAtw5goAAAAALEC5AgAAAAALUK4AAAAAwAKUKwAAAACwAOXqHtmwYYOGDh2qvn37Ki8vT4cPH7Z7JAShb7/9VtOmTVNqaqoMw9D27dsD1k3T1IoVK+R2u+VyuVRQUKCTJ0/aMyyCypo1azRu3DjFxMRowIABKioqUmNjY8A+7e3t8nq9SkxMVHR0tIqLi3XhwgWbJkYwKS8vV2Zmpv+PdXo8Hu3evdu/TnbQVWvXrpVhGFq0aJF/G/nB31m5cqUMwwh4jRw50r8eitmhXN0Dn3zyiRYvXqyysjIdPXpUWVlZmjJlipqbm+0eDUGmra1NWVlZ2rBhw23XX3nlFa1fv17vvPOODh06pKioKE2ZMkXt7e33eFIEm5qaGnm9Xh08eFDV1dXq6OjQ5MmT1dbW5t/nhRde0Oeff67KykrV1NSoqalJTzzxhI1TI1gMGjRIa9euVW1trb7//ntNmjRJjz/+uH788UdJZAddc+TIEb377rvKzMwM2E5+0JkxY8bo3Llz/td3333nXwvJ7Ji463Jzc02v1+v/fP36dTM1NdVcs2aNjVMh2Ekyq6qq/J99Pp+ZkpJivvrqq/5tly5dMp1Op7l161YbJkQwa25uNiWZNTU1pmneyEpERIRZWVnp36ehocGUZB44cMCuMRHE4uPjzQ8++IDsoEtaW1vN9PR0s7q62nzooYfMhQsXmqbJdw86V1ZWZmZlZd12LVSzw5mru+zatWuqra1VQUGBf1tYWJgKCgp04MABGydDqDl9+rTOnz8fkKW4uDjl5eWRJdzi8uXLkqSEhARJUm1trTo6OgLyM3LkSKWlpZEfBLh+/boqKirU1tYmj8dDdtAlXq9Xjz76aEBOJL578M9Onjyp1NRUDR8+XCUlJTpz5oyk0M1OuN0D9HQXL17U9evXlZycHLA9OTlZJ06csGkqhKLz589L0m2zdHMNkCSfz6dFixZpwoQJuv/++yXdyI/D4VC/fv0C9iU/uKm+vl4ej0ft7e2Kjo5WVVWVRo8erbq6OrKDTlVUVOjo0aM6cuTILWt896AzeXl5+uijj5SRkaFz585p1apVmjhxoo4fPx6y2aFcAUAP4/V6dfz48YDr1oF/kpGRobq6Ol2+fFnbtm1TaWmpampq7B4LQe7s2bNauHChqqur1bdvX7vHQYgpLCz0v8/MzFReXp6GDBmiTz/9VC6Xy8bJ7hyXBd5lSUlJ6tOnzy1PNrlw4YJSUlJsmgqh6GZeyBI6M3/+fO3cuVP79u3ToEGD/NtTUlJ07do1Xbp0KWB/8oObHA6HRowYoezsbK1Zs0ZZWVl68803yQ46VVtbq+bmZo0dO1bh4eEKDw9XTU2N1q9fr/DwcCUnJ5MfdFm/fv1033336dSpUyH73UO5usscDoeys7O1d+9e/zafz6e9e/fK4/HYOBlCzbBhw5SSkhKQpZaWFh06dIgsQaZpav78+aqqqtLXX3+tYcOGBaxnZ2crIiIiID+NjY06c+YM+cFt+Xw+Xb16leygU/n5+aqvr1ddXZ3/lZOTo5KSEv978oOuunLlin7++We53e6Q/e7hssB7YPHixSotLVVOTo5yc3O1bt06tbW1adasWXaPhiBz5coVnTp1yv/59OnTqqurU0JCgtLS0rRo0SKtXr1a6enpGjZsmJYvX67U1FQVFRXZNzSCgtfr1ZYtW/TZZ58pJibGfz16XFycXC6X4uLi9Nxzz2nx4sVKSEhQbGysFixYII/Ho/Hjx9s8Pey2dOlSFRYWKi0tTa2trdqyZYu++eYb7dmzh+ygUzExMf57O2+KiopSYmKifzv5wd9ZsmSJpk2bpiFDhqipqUllZWXq06ePnnnmmdD97rH7cYW9xVtvvWWmpaWZDofDzM3NNQ8ePGj3SAhC+/btMyXd8iotLTVN88bj2JcvX24mJyebTqfTzM/PNxsbG+0dGkHhdrmRZG7atMm/z19//WXOmzfPjI+PNyMjI83p06eb586ds29oBI3Zs2ebQ4YMMR0Oh9m/f38zPz/f/PLLL/3rZAf/H//3UeymSX7w92bMmGG63W7T4XCYAwcONGfMmGGeOnXKvx6K2TFM0zRt6nUAAAAA0GNwzxUAAAAAWIByBQAAAAAWoFwBAAAAgAUoVwAAAABgAcoVAAAAAFiAcgUAAAAAFqBcAQAAAIAFKFcAAAAAYAHKFQAA3WQYhrZv3273GAAAm1GuAAAhbebMmTIM45bX1KlT7R4NANDLhNs9AAAA3TV16lRt2rQpYJvT6bRpGgBAb8WZKwBAyHM6nUpJSQl4xcfHS7pxyV55ebkKCwvlcrk0fPhwbdu2LeDf19fXa9KkSXK5XEpMTNScOXN05cqVgH02btyoMWPGyOl0yu12a/78+QHrFy9e1PTp0xUZGan09HTt2LHDv/bHH3+opKRE/fv3l8vlUnp6+i1lEAAQ+ihXAIAeb/ny5SouLtaxY8dUUlKip59+Wg0NDZKktrY2TZkyRfHx8Tpy5IgqKyv11VdfBZSn8vJyeb1ezZkzR/X19dqxY4dGjBgR8DNWrVqlp556Sj/88IMeeeQRlZSU6Pfff/f//J9++km7d+9WQ0ODysvLlZSUdO/+AwAA94RhmqZp9xAAANypmTNnavPmzerbt2/A9mXLlmnZsmUyDENz585VeXm5f238+PEaO3as3n77bb3//vt68cUXdfbsWUVFRUmSdu3apWnTpqmpqUnJyckaOHCgZs2apdWrV992BsMw9NJLL+nll1+WdKOwRUdHa/fu3Zo6daoee+wxJSUlaePGjXfpfwEAEAy45woAEPIefvjhgPIkSQkJCf73Ho8nYM3j8aiurk6S1NDQoKysLH+xkqQJEybI5/OpsbFRhmGoqalJ+fn5nc6QmZnpfx8VFaXY2Fg1NzdLkp5//nkVFxfr6NGjmjx5soqKivTAAw/c0e8KAAhelCsAQMiLioq65TI9q7hcri7tFxEREfDZMAz5fD5JUmFhoX799Vft2rVL1dXVys/Pl9fr1WuvvWb5vAAA+3DPFQCgxzt48OAtn0eNGiVJGjVqlI4dO6a2tjb/+v79+xUWFqaMjAzFxMRo6NCh2rt3b7dm6N+/v0pLS7V582atW7dO7733XreOBwAIPpy5AgCEvKtXr+r8+fMB28LDw/0PjaisrFROTo4efPBBffzxxzp8+LA+/PBDSVJJSYnKyspUWlqqlStX6rffftOCBQv07LPPKjk5WZK0cuVKzZ07VwMGDFBhYaFaW1u1f/9+LViwoEvzrVixQtnZ2RozZoyuXr2qnTt3+ssdAKDnoFwBAELeF198IbfbHbAtIyNDJ06ckHTjSX4VFRWaN2+e3G63tm7dqtGjR0uSIiMjtWfPHi1cuFDjxo1TZGSkiouL9frrr/uPVVpaqvb2dr3xxhtasmSJkpKS9OSTT3Z5PofDoaVLl+qXX36Ry+XSxIkTVVFRYcFvDgAIJjwtEADQoxmGoaqqKhUVFdk9CgCgh+OeKwAAAACwAOUKAAAAACzAPVcAgB6Nq98BAPcKZ64AAAAAwAKUKwAAAACwAOUKAAAAACxAuQIAAAAAC1CuAAAAAMAClCsAAAAAsADlCgAAAAAsQLkCAAAAAAv8LwplI4rIoEjGAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 1000x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Boubaker processing time: 807.36 seconds\n"
     ]
    }
   ],
   "source": [
    "import time\n",
    "\n",
    "epochs = 50\n",
    "\n",
    "def train_and_validate(models, model_names, train_loader, test_loader, criterion, optimizers, device, epochs):\n",
    "    train_losses = [[] for _ in range(len(models))]\n",
    "    test_losses = [[] for _ in range(len(models))]\n",
    "    best_weights = [None] * len(models)\n",
    "    best_test_losses = [float('inf')] * len(models)\n",
    "    model_times = [0.0] * len(models)  # Initialize a list to store processing time for each model\n",
    "\n",
    "    for epoch in range(epochs):\n",
    "        for i, (model, optimizer) in enumerate(zip(models, optimizers)):\n",
    "            start_time = time.time()  # Record the start time\n",
    "\n",
    "            train_loss = train(model, train_loader, criterion, optimizer, device)\n",
    "            test_loss, test_accuracy = validate(model, test_loader, criterion, device)\n",
    "\n",
    "            end_time = time.time()  # Record the end time\n",
    "            model_times[i] += end_time - start_time  # Accumulate the processing time for the current model\n",
    "\n",
    "            train_losses[i].append(train_loss)\n",
    "            test_losses[i].append(test_loss)\n",
    "\n",
    "            if test_loss < best_test_losses[i]:\n",
    "                best_test_losses[i] = test_loss\n",
    "                best_weights[i] = model.state_dict()\n",
    "\n",
    "            print(f'{model_names[i]}, Epoch {epoch+1}, Train Loss: {train_loss:.4f}, '\n",
    "                  f'Test Loss: {test_loss:.4f}, Test Acc: {test_accuracy:.2f}')\n",
    "\n",
    "    return train_losses, test_losses, best_weights, model_times\n",
    "\n",
    "\n",
    "optimizers = [optim.Adam(model.parameters(), lr=0.0002) for model in models]\n",
    "\n",
    "train_losses, test_losses, best_weights, model_times = train_and_validate(models, Model_Names, train_loader, test_loader, criterion, optimizers, device, epochs)\n",
    "\n",
    "plt.figure(figsize=(10, 5))\n",
    "for i in range(len(models)):\n",
    "    plt.plot(range(1, epochs + 1), test_losses[i], label=f'{Model_Names[i]} Test Loss')\n",
    "plt.title('Test Loss Convergence')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend()\n",
    "plt.grid(True)\n",
    "plt.show()\n",
    "\n",
    "# Save the best weights for each model\n",
    "for i, model in enumerate(models):\n",
    "    model.load_state_dict(best_weights[i])\n",
    "    torch.save(model.state_dict(), f'{Model_Names[i]}_best_weights.pth')\n",
    "\n",
    "# Print the processing time for each model\n",
    "for i, model_name in enumerate(Model_Names):\n",
    "    print(f\"{model_name} processing time: {model_times[i]:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# **Model Test**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def predict(model, test_loader, device):\n",
    "    model.eval()\n",
    "    predictions = []\n",
    "    test_labels=[]\n",
    "\n",
    "    with torch.no_grad():\n",
    "        for data, target in test_loader:\n",
    "            data = data.to(device)\n",
    "            output = model(data)\n",
    "            pred = output.argmax(dim=1)\n",
    "            test_labels.extend(target.cpu().numpy())\n",
    "            predictions.extend(pred.cpu().numpy())\n",
    "    return predictions,test_labels\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model Boubaker - Overall Accuracy: 0.9731, Kappa: 0.9701, F1 Score: 0.9731\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import f1_score, cohen_kappa_score, accuracy_score\n",
    "metrics = {\n",
    "    'Accuracy': [],\n",
    "    'Kappa': [],\n",
    "    'F1 Score': []\n",
    "}\n",
    "\n",
    "for i, model in enumerate(models):\n",
    "    model.load_state_dict(best_weights[i])\n",
    "    model.eval()\n",
    "\n",
    "    test_preds,test_labels = predict(model, test_loader, device)\n",
    "    f1 = f1_score(test_labels, test_preds, average='micro')\n",
    "    kappa = cohen_kappa_score(test_labels, test_preds)\n",
    "    overall_acc = accuracy_score(test_labels, test_preds)\n",
    "    metrics['Accuracy'].append(overall_acc)\n",
    "    metrics['Kappa'].append(kappa)\n",
    "    metrics['F1 Score'].append(f1)\n",
    "    \n",
    "    print(f\"Model {Model_Names[i]} - \"\n",
    "          f\"Overall Accuracy: {overall_acc:.4f}, \"\n",
    "          f\"Kappa: {kappa:.4f}, \"\n",
    "          f\"F1 Score: {f1:.4f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(10, 6))\n",
    "bar_width = 0.2\n",
    "opacity = 0.8\n",
    "index = np.arange(len(metrics.keys()))\n",
    "\n",
    "colors = ['#1f77b4', '#ff7f0e', '#2ca02c', '#d62728']  # Custom colors for each model\n",
    "\n",
    "for i, model_name in enumerate(Model_Names):\n",
    "    model_metrics = [metrics['Accuracy'][i], metrics['Kappa'][i], metrics['F1 Score'][i]]\n",
    "    ax.bar(index + i * bar_width, model_metrics, bar_width,\n",
    "           alpha=opacity, color=colors[i], label=model_name)\n",
    "\n",
    "ax.set_xlabel('Metrics')\n",
    "ax.set_ylabel('Scores')\n",
    "ax.set_title('Model Metrics Comparison')\n",
    "ax.set_xticks(index + bar_width * (len(Model_Names) - 1) / 2)\n",
    "ax.set_xticklabels(metrics.keys())\n",
    "ax.legend()\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "TPU",
  "colab": {
   "collapsed_sections": [],
   "machine_shape": "hm",
   "name": "ISPRS-2020-HyperspectralChange DetectionUSA.ipynb",
   "provenance": []
  },
  "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": "3.11.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}