Mid-Term.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "4e77284c",
   "metadata": {},
   "outputs": [],
   "source": [
    "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"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "dab0f999",
   "metadata": {},
   "outputs": [],
   "source": [
    "# MNIST dataset\n",
    "transform = transforms.Compose([transforms.ToTensor(),\n",
    "                                transforms.Normalize((0.5,), (0.5,))])\n",
    "\n",
    "train_all = datasets.MNIST('../data', train=True, download=True, transform=transform)\n",
    "train_data, val_data = torch.utils.data.random_split(train_all, [50000, 10000])\n",
    "test_data = datasets.MNIST('../data', train=False, transform=transform)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "85d838aa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10, Loss: 2.262241159253718, Validation Accuracy: 0.4674\n",
      "Epoch 2/10, Loss: 0.6941344764112206, Validation Accuracy: 0.9042\n",
      "Epoch 3/10, Loss: 0.22859272902921948, Validation Accuracy: 0.9466\n",
      "Epoch 4/10, Loss: 0.15283074486486214, Validation Accuracy: 0.9589\n",
      "Epoch 5/10, Loss: 0.11993379757055998, Validation Accuracy: 0.9643\n",
      "Epoch 6/10, Loss: 0.1011524717728405, Validation Accuracy: 0.9664\n",
      "Epoch 7/10, Loss: 0.08823109541178854, Validation Accuracy: 0.9712\n",
      "Epoch 8/10, Loss: 0.07897574133823251, Validation Accuracy: 0.9732\n",
      "Epoch 9/10, Loss: 0.07165178829444396, Validation Accuracy: 0.9731\n",
      "Epoch 10/10, Loss: 0.06582594430967308, Validation Accuracy: 0.977\n",
      "Test Accuracy: 0.9824\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "'''\n",
    "Qeustion 1:\n",
    "    Implement LeNet architecture, and train on all the 50K training samples with vanilla SGD.\n",
    "    Attach learning curves and report test accuracy\n",
    "'''\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Set seed for reproducibility\n",
    "torch.manual_seed(0)\n",
    "\n",
    "# Define the LeNet architecture\n",
    "class LeNet(nn.Module):\n",
    "    def __init__(self):\n",
    "        super(LeNet, self).__init__()\n",
    "        self.conv1 = nn.Conv2d(1, 6, kernel_size=5)\n",
    "        self.pool = nn.MaxPool2d(kernel_size=2, stride=2)\n",
    "        self.conv2 = nn.Conv2d(6, 16, kernel_size=5)\n",
    "        self.fc1 = nn.Linear(16*4*4, 120)\n",
    "        self.fc2 = nn.Linear(120, 84)\n",
    "        self.fc3 = nn.Linear(84, 10)\n",
    "\n",
    "    def forward(self, x):\n",
    "        x = self.pool(torch.relu(self.conv1(x)))\n",
    "        x = self.pool(torch.relu(self.conv2(x)))\n",
    "        x = x.view(-1, 16*4*4)\n",
    "        x = torch.relu(self.fc1(x))\n",
    "        x = torch.relu(self.fc2(x))\n",
    "        x = self.fc3(x)\n",
    "        return x\n",
    "\n",
    "# Initialize the network\n",
    "lenet = LeNet()\n",
    "\n",
    "# Loss function\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "\n",
    "# Optimizer (SGD)\n",
    "optimizer = optim.SGD(lenet.parameters(), lr=0.01)\n",
    "\n",
    "# MNIST dataset\n",
    "transform = transforms.Compose([transforms.ToTensor(),\n",
    "                                transforms.Normalize((0.5,), (0.5,))])\n",
    "# Prepare data loaders\n",
    "train_loader = DataLoader(train_data, batch_size=64, shuffle=True)\n",
    "val_loader = DataLoader(val_data, batch_size=64, shuffle=False)\n",
    "test_loader = DataLoader(test_data, batch_size=64, shuffle=False)\n",
    "\n",
    "# Function to train and validate the model\n",
    "def train_and_validate(model, train_loader, val_loader, optimizer, criterion, epochs):\n",
    "    model.train()\n",
    "    training_history = {'train_loss': [], 'val_accuracy': []}\n",
    "    \n",
    "    for epoch in range(epochs):\n",
    "        total_loss = 0\n",
    "        for data, target in train_loader:\n",
    "            optimizer.zero_grad()\n",
    "            output = model(data)\n",
    "            loss = criterion(output, target)\n",
    "            loss.backward()\n",
    "            optimizer.step()\n",
    "            total_loss += loss.item()\n",
    "        \n",
    "        # Validation\n",
    "        correct = 0\n",
    "        total = 0\n",
    "        with torch.no_grad():\n",
    "            for data, target in val_loader:\n",
    "                output = model(data)\n",
    "                _, predicted = torch.max(output.data, 1)\n",
    "                total += target.size(0)\n",
    "                correct += (predicted == target).sum().item()\n",
    "        \n",
    "        val_accuracy = correct / total\n",
    "        training_history['train_loss'].append(total_loss / len(train_loader))\n",
    "        training_history['val_accuracy'].append(val_accuracy)\n",
    "        print(f'Epoch {epoch+1}/{epochs}, Loss: {total_loss / len(train_loader)}, Validation Accuracy: {val_accuracy}')\n",
    "    \n",
    "    return training_history\n",
    "\n",
    "# Train the model\n",
    "history = train_and_validate(lenet, train_loader, val_loader, optimizer, criterion, epochs=10)\n",
    "\n",
    "# Test the model\n",
    "lenet.eval()\n",
    "correct = 0\n",
    "total = 0\n",
    "with torch.no_grad():\n",
    "    for data, target in test_loader:\n",
    "        output = lenet(data)\n",
    "        _, predicted = torch.max(output.data, 1)\n",
    "        total += target.size(0)\n",
    "        correct += (predicted == target).sum().item()\n",
    "\n",
    "test_accuracy = correct / total\n",
    "print(f'Test Accuracy: {test_accuracy}')\n",
    "\n",
    "# Plot the learning curves\n",
    "plt.plot(history['train_loss'], label='Training Loss')\n",
    "plt.plot(history['val_accuracy'], label='Validation Accuracy')\n",
    "plt.title('Training Progress')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "2df3ffbe",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10, Loss: 0.08352566969490437, Validation Accuracy: 0.9791\n",
      "Epoch 2/10, Loss: 0.05423970473960013, Validation Accuracy: 0.9793\n",
      "Epoch 3/10, Loss: 0.04073130600819843, Validation Accuracy: 0.9794\n",
      "Epoch 4/10, Loss: 0.03209313697388867, Validation Accuracy: 0.9864\n",
      "Epoch 5/10, Loss: 0.026233508646035034, Validation Accuracy: 0.983\n",
      "Epoch 6/10, Loss: 0.019385182253240495, Validation Accuracy: 0.9865\n",
      "Epoch 7/10, Loss: 0.01718150594306257, Validation Accuracy: 0.9867\n",
      "Epoch 8/10, Loss: 0.012682007738155388, Validation Accuracy: 0.9882\n",
      "Epoch 9/10, Loss: 0.008565306662815362, Validation Accuracy: 0.9875\n",
      "Epoch 10/10, Loss: 0.007298587362392029, Validation Accuracy: 0.99\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "'''\n",
    "Question 2:\n",
    "    How can we make the training loss converge faster? Implement your idea and compare the new\n",
    "    training curve against that in question 1.\n",
    "\n",
    "Answer:\n",
    "    - Use a more advanced optimization algorithm: Replace SGD with optimizers like Adam or RMSprop,\n",
    "    which use adaptive learning rates and momentum.\n",
    "'''\n",
    "# Adjusted optimizer - using Adam\n",
    "optimizer = optim.Adam(lenet.parameters(), lr=0.001)\n",
    "\n",
    "# Learning rate scheduler\n",
    "scheduler = torch.optim.lr_scheduler.StepLR(optimizer, step_size=1, gamma=0.9)\n",
    "\n",
    "# Adjusted training function to include scheduler step\n",
    "def train_and_validate_with_scheduler(model, train_loader, val_loader, optimizer, criterion, scheduler, epochs):\n",
    "    model.train()\n",
    "    training_history = {'train_loss': [], 'val_accuracy': []}\n",
    "    \n",
    "    for epoch in range(epochs):\n",
    "        model.train()\n",
    "        total_loss = 0\n",
    "        for data, target in train_loader:\n",
    "            optimizer.zero_grad()\n",
    "            output = model(data)\n",
    "            loss = criterion(output, target)\n",
    "            loss.backward()\n",
    "            optimizer.step()\n",
    "            total_loss += loss.item()\n",
    "        \n",
    "        # Step through the scheduler\n",
    "        scheduler.step()\n",
    "        \n",
    "        # Validation\n",
    "        model.eval()\n",
    "        correct = 0\n",
    "        total = 0\n",
    "        with torch.no_grad():\n",
    "            for data, target in val_loader:\n",
    "                output = model(data)\n",
    "                _, predicted = torch.max(output.data, 1)\n",
    "                total += target.size(0)\n",
    "                correct += (predicted == target).sum().item()\n",
    "        \n",
    "        val_accuracy = correct / total\n",
    "        training_history['train_loss'].append(total_loss / len(train_loader))\n",
    "        training_history['val_accuracy'].append(val_accuracy)\n",
    "        print(f'Epoch {epoch+1}/{epochs}, Loss: {total_loss / len(train_loader)}, Validation Accuracy: {val_accuracy}')\n",
    "    \n",
    "    return training_history\n",
    "\n",
    "# Training the model with the new optimizer and scheduler\n",
    "history_adam = train_and_validate_with_scheduler(lenet, train_loader, val_loader, optimizer, criterion, scheduler, epochs=10)\n",
    "\n",
    "# Plot the learning curves for comparison\n",
    "plt.plot(history['train_loss'], label='SGD Training Loss')\n",
    "plt.plot(history_adam['train_loss'], label='Adam Training Loss')\n",
    "plt.title('Training Loss Comparison')\n",
    "plt.xlabel('Epoch')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "5211f222",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10, Loss: 0.8615707757936162, Validation Accuracy: 0.918\n",
      "Epoch 2/10, Loss: 0.24470016011481832, Validation Accuracy: 0.949\n",
      "Epoch 3/10, Loss: 0.1686678464364284, Validation Accuracy: 0.9552\n",
      "Epoch 4/10, Loss: 0.1273876990482306, Validation Accuracy: 0.9682\n",
      "Epoch 5/10, Loss: 0.10685116737179316, Validation Accuracy: 0.9735\n",
      "Epoch 6/10, Loss: 0.0851632842651693, Validation Accuracy: 0.9755\n",
      "Epoch 7/10, Loss: 0.07641411028731211, Validation Accuracy: 0.9769\n",
      "Epoch 8/10, Loss: 0.06377924093086819, Validation Accuracy: 0.9754\n",
      "Epoch 9/10, Loss: 0.059759097559389414, Validation Accuracy: 0.9798\n",
      "Epoch 10/10, Loss: 0.04872731560586127, Validation Accuracy: 0.9787\n",
      "Epoch 1/10, Loss: 0.08336549068389036, Validation Accuracy: 0.9793\n",
      "Epoch 2/10, Loss: 0.0583576578676672, Validation Accuracy: 0.9852\n",
      "Epoch 3/10, Loss: 0.04428397804674933, Validation Accuracy: 0.9832\n",
      "Epoch 4/10, Loss: 0.03498240782710476, Validation Accuracy: 0.9859\n",
      "Epoch 5/10, Loss: 0.028208372759563692, Validation Accuracy: 0.9881\n",
      "Epoch 6/10, Loss: 0.022280022573774642, Validation Accuracy: 0.9853\n",
      "Epoch 7/10, Loss: 0.017999274736937836, Validation Accuracy: 0.9872\n",
      "Epoch 8/10, Loss: 0.013712635104101725, Validation Accuracy: 0.9887\n",
      "Epoch 9/10, Loss: 0.010576028496954742, Validation Accuracy: 0.9888\n",
      "Epoch 10/10, Loss: 0.007723657456232974, Validation Accuracy: 0.9886\n",
      "Epoch 1/10, Loss: 0.04266078871947247, Validation Accuracy: 0.9883\n",
      "Epoch 2/10, Loss: 0.02940689832502976, Validation Accuracy: 0.9889\n",
      "Epoch 3/10, Loss: 0.022654672555904837, Validation Accuracy: 0.988\n",
      "Epoch 4/10, Loss: 0.017245422463899013, Validation Accuracy: 0.9834\n",
      "Epoch 5/10, Loss: 0.013299009753047722, Validation Accuracy: 0.9894\n",
      "Epoch 6/10, Loss: 0.009941101939289364, Validation Accuracy: 0.9891\n",
      "Epoch 7/10, Loss: 0.008198404861646122, Validation Accuracy: 0.9896\n",
      "Epoch 8/10, Loss: 0.0058389387442730365, Validation Accuracy: 0.9893\n",
      "Epoch 9/10, Loss: 0.0048696443426728365, Validation Accuracy: 0.9905\n",
      "Epoch 10/10, Loss: 0.001934151139710957, Validation Accuracy: 0.9912\n",
      "Epoch 1/10, Loss: 0.025030199998404303, Validation Accuracy: 0.9867\n",
      "Epoch 2/10, Loss: 0.01590904278470852, Validation Accuracy: 0.9905\n",
      "Epoch 3/10, Loss: 0.011885325078648859, Validation Accuracy: 0.9882\n",
      "Epoch 4/10, Loss: 0.009042178333854683, Validation Accuracy: 0.9881\n",
      "Epoch 5/10, Loss: 0.007020050342362664, Validation Accuracy: 0.9892\n",
      "Epoch 6/10, Loss: 0.004889401430792041, Validation Accuracy: 0.9904\n",
      "Epoch 7/10, Loss: 0.004582198943650153, Validation Accuracy: 0.9898\n",
      "Epoch 8/10, Loss: 0.003263692550352776, Validation Accuracy: 0.991\n",
      "Epoch 9/10, Loss: 0.0026572405818434766, Validation Accuracy: 0.9891\n",
      "Epoch 10/10, Loss: 0.0011343942657642287, Validation Accuracy: 0.9911\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1000x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test accuracies for different training sizes: [0.9787, 0.9886, 0.9912, 0.9911]\n"
     ]
    }
   ],
   "source": [
    "'''\n",
    "Question 3:\n",
    "    Train on 20%, 50%, 80% of the full 50K training samples, and report test accuracies. On a 2D\n",
    "    coordinate (x-axis: training size; y-axis: test accuracy), plot the accuracies, along with the case of\n",
    "    100% training samples (question 1). Discuss the result.\n",
    "\n",
    "'''\n",
    "# Define training sizes and store accuracies\n",
    "training_sizes = [0.2, 0.5, 0.8, 1.0]  # Representing 20%, 50%, 80%, and 100% of the training data\n",
    "accuracies = []\n",
    "\n",
    "# Initialize the network\n",
    "lenet = LeNet()\n",
    "\n",
    "# Criterion\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "\n",
    "# Train and evaluate for each specified training size\n",
    "for size in training_sizes:\n",
    "    # Calculate the number of samples for the subset\n",
    "    subset_size = int(size * 50000)\n",
    "    \n",
    "    # Split the data into training and validation subsets\n",
    "    train_subset, _ = torch.utils.data.random_split(train_all, [subset_size, len(train_all) - subset_size])\n",
    "    \n",
    "    # Create data loaders\n",
    "    train_loader = DataLoader(train_subset, batch_size=64, shuffle=True)\n",
    "    \n",
    "    # Optimizer (Adam) and Scheduler\n",
    "    optimizer = optim.Adam(lenet.parameters(), lr=0.001)\n",
    "    scheduler = torch.optim.lr_scheduler.StepLR(optimizer, step_size=1, gamma=0.9)\n",
    "    \n",
    "    # Train the model using the training subset and evaluate on the test set\n",
    "    accuracy = train_and_validate_with_scheduler(lenet, train_loader, test_loader, optimizer, criterion, scheduler, epochs=10)\n",
    "    accuracies.append(accuracy['val_accuracy'][-1])\n",
    "\n",
    "# Plot accuracies against the training size\n",
    "plt.figure(figsize=(10, 5))\n",
    "plt.plot([size * 50000 for size in training_sizes], accuracies, marker='o')\n",
    "plt.title('Test Accuracy vs Training Data Size')\n",
    "plt.xlabel('Training Data Size')\n",
    "plt.ylabel('Test Accuracy (%)')\n",
    "plt.grid(True)\n",
    "plt.xticks([size * 50000 for size in training_sizes])\n",
    "plt.show()\n",
    "\n",
    "# Output the accuracies for reference\n",
    "print(\"Test accuracies for different training sizes:\", accuracies)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "444cae70",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/20, Loss: 0.8471982892911145, Validation Accuracy: 0.9242\n",
      "Epoch 2/20, Loss: 0.19877341795404246, Validation Accuracy: 0.9522\n",
      "Epoch 3/20, Loss: 0.13621242300482692, Validation Accuracy: 0.9614\n",
      "Epoch 4/20, Loss: 0.10766857177920781, Validation Accuracy: 0.97\n",
      "Epoch 5/20, Loss: 0.0905456470479821, Validation Accuracy: 0.9738\n",
      "Epoch 6/20, Loss: 0.07196890083467884, Validation Accuracy: 0.9724\n",
      "Epoch 7/20, Loss: 0.06287798736019025, Validation Accuracy: 0.9773\n",
      "Epoch 8/20, Loss: 0.05528331271396463, Validation Accuracy: 0.9743\n",
      "Epoch 9/20, Loss: 0.04735447820216702, Validation Accuracy: 0.9781\n",
      "Epoch 10/20, Loss: 0.04332656442976681, Validation Accuracy: 0.978\n",
      "Epoch 11/20, Loss: 0.03722084014807015, Validation Accuracy: 0.9799\n",
      "Epoch 12/20, Loss: 0.03359064374157601, Validation Accuracy: 0.9799\n",
      "Epoch 13/20, Loss: 0.028893327752757034, Validation Accuracy: 0.9789\n",
      "Epoch 14/20, Loss: 0.027306704660343708, Validation Accuracy: 0.9798\n",
      "Epoch 15/20, Loss: 0.02430276854992933, Validation Accuracy: 0.9809\n",
      "Epoch 16/20, Loss: 0.022969238955385177, Validation Accuracy: 0.9811\n",
      "Epoch 17/20, Loss: 0.02136552024181861, Validation Accuracy: 0.9806\n",
      "Epoch 18/20, Loss: 0.019445575928689236, Validation Accuracy: 0.9807\n",
      "Epoch 19/20, Loss: 0.01808945670728364, Validation Accuracy: 0.9808\n"
     ]
    }
   ],
   "source": [
    "'''\n",
    "Question 4:\n",
    "    How can we improve the accuracies for cases with fewer training samples? Implement your idea\n",
    "    and compare against what you get for question 3.\n",
    "Answer:\n",
    "    - Increasing training iteration\n",
    "'''\n",
    "\n",
    "\n",
    "# Define training sizes and store accuracies\n",
    "training_size = 0.2  # Representing 20%, 50%, 80%, and 100% of the training data\n",
    "\n",
    "# Initialize the network\n",
    "lenet = LeNet()\n",
    "\n",
    "\n",
    "# Criterion\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "\n",
    "# Train and evaluate for each specified training size\n",
    "# Calculate the number of samples for the subset\n",
    "subset_size = int(training_size * 50000)\n",
    "    \n",
    "# Split the data into training and validation subsets\n",
    "train_subset, _ = torch.utils.data.random_split(train_all, [subset_size, len(train_all) - subset_size])\n",
    "    \n",
    "# Create data loaders\n",
    "train_loader = DataLoader(train_subset, batch_size=64, shuffle=True)\n",
    "    \n",
    "# Optimizer (Adam) and Scheduler\n",
    "optimizer = optim.Adam(lenet.parameters(), lr=0.001)\n",
    "scheduler = torch.optim.lr_scheduler.StepLR(optimizer, step_size=1, gamma=0.9)\n",
    "    \n",
    "        \n",
    "# Train the model using the training subset and evaluate on the test set\n",
    "accuracy = train_and_validate_with_scheduler(lenet, \n",
    "                                             train_loader, \n",
    "                                             test_loader, \n",
    "                                             optimizer,\n",
    "                                             criterion, \n",
    "                                             scheduler, \n",
    "                                             epochs=20)\n",
    "inc_accuracies = accuracy['val_accuracy'][-1]\n",
    "\n",
    "# Data preparation\n",
    "categories = ['Original', 'Increased Iterations']\n",
    "accuracies = [accuracies[0], inc_accuracies]\n",
    "\n",
    "# Bar positions\n",
    "x = range(len(categories))\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "# Space between bars\n",
    "bar_width = 0.4\n",
    "offset = 0.1\n",
    "\n",
    "rects1 = ax.bar(x[0], accuracies[0], bar_width, label=categories[0], color='skyblue')\n",
    "rects2 = ax.bar(x[0] + bar_width + offset, accuracies[1], bar_width, label=categories[1], color='orange')\n",
    "\n",
    "# Adding text for labels, title, and custom x-axis tick labels, etc.\n",
    "ax.set_ylabel('Test Accuracy (%)')\n",
    "ax.set_title('Test Accuracy by Training Size and Method')\n",
    "ax.set_xticks([i + bar_width / 2 for i in x])\n",
    "ax.legend()\n",
    "\n",
    "# Narrow the y-axis range to magnify the difference\n",
    "ax.set_ylim([0.9785, 0.9810])\n",
    "\n",
    "# Function to add a label above each bar\n",
    "def autolabel(rects):\n",
    "    \"\"\"Attach a text label above each bar displaying its height.\"\"\"\n",
    "    for rect in rects:\n",
    "        height = rect.get_height()\n",
    "        ax.annotate(f'{height:.4f}',\n",
    "                    xy=(rect.get_x() + rect.get_width() / 2, height),\n",
    "                    xytext=(0, 3),  # 3 points vertical offset\n",
    "                    textcoords=\"offset points\",\n",
    "                    ha='center', va='bottom')\n",
    "\n",
    "autolabel(rects1)\n",
    "autolabel(rects2)\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "77e2748f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAnYAAAHWCAYAAAD6oMSKAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjYuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8o6BhiAAAACXBIWXMAAA9hAAAPYQGoP6dpAABnlklEQVR4nO3deVyN6f8/8Nc5bSelRasalESEMpYYxhopjG3so2SdIYbGGD6orNnHksEYkxmyjjTDzETC2BJD+ZhhCC00LUhFpvXcvz/8ur9znKKTEvfn9Xw8zuPhXOe6r/t931pe3ct1ywRBEEBEREREbz15TRdARERERFWDwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6I3hgnT56ETCbDDz/8UNOlvLKuXbuia9eulVp2zJgxsLOzq9J6qlrp/9XJkydrupRqYWdnhzFjxtR0GS8kk8ng5+dX7evZvn07ZDIZkpKSqn1d9OoY7OitIpPJKvSqil82T58+RVBQUKXG+uWXXyCTyWBjYwOlUvnKtVDVSEpKqvDX0P/qLzGlUonvv/8ebm5uqFOnDmrXro3GjRvD29sb58+fr+ny3jilXy/jx48v8/O5c+eKfR48eKDx+OfOnUNQUBCys7NfsVL6X6Fd0wUQaWLHjh0q77///ntERUWptTdt2vSV1/X06VMsWLAAADQ+8hIWFgY7OzskJSXh+PHjcHd3f+V66NVZWFiofa2sXr0a9+7dw5dffqnW91UcPXq00stu3bq1xv4gmDZtGjZu3Ij+/ftj1KhR0NbWxo0bN/Drr7+iYcOGaN++PQCgc+fO+Oeff6Crq1sjdb5JFAoFDhw4gK+++kptf+zevRsKhQL5+fmVGvvcuXNYsGABxowZAxMTkyqolqSOwY7eKh999JHK+/PnzyMqKkqtvSbl5eXhxx9/RHBwMEJDQxEWFvbGBru8vDwYGBjUdBmvjYGBgdrXyp49e/Do0aMXfg0JgoD8/Hzo6+tXeF2vEnh0dHQqveyryMjIwFdffYUJEybg66+/Vvls7dq1uH//vvheLpdDoVC87hLfSL1798ZPP/2EX3/9Ff379xfbz507h8TERAwePBgHDhyowQrpfwlPxZLkKJVKrF27Fs7OzlAoFLCyssKkSZPw6NEjlX6///47PDw8YG5uDn19fdjb22Ps2LEAnp2yKz1is2DBAvFUSlBQ0EvXf/DgQfzzzz8YMmQIhg8fjvDw8DL/Ws/Pz0dQUBAaN24MhUKBunXrYtCgQbh9+7bKtqxbtw4tWrSAQqGAhYUFevfujd9//12sUyaTYfv27WrjP19vUFAQZDIZrl27hpEjR8LU1BSdOnUCAPz3v//FmDFj0LBhQygUClhbW2Ps2LF4+PCh2ripqakYN24cbGxsoKenB3t7e3zyyScoLCzEnTt3IJPJ1I5+Ac9+yclkMuzevful+7CkpAT/+c9/YG1tDQMDA3zwwQe4e/eu+HlgYCB0dHRUgkapiRMnwsTEpNJHSIBn11f17dsXR44cQZs2baCvr48tW7YAAEJDQ9G9e3dYWlpCT08PzZo1w6ZNm9TGeP4au9Jr0vbt24clS5bgnXfegUKhQI8ePXDr1i2VZZ+/xq70/3nVqlX4+uuv4eDgAD09PbRt2xYXL15UW/f+/fvRrFkzKBQKNG/eHAcPHqzQdXuJiYkQBAEdO3ZU+0wmk8HS0lJte0ovVSi9Dqus1/NHvHfu3InWrVtDX18fderUwfDhw1X+f8uTnJyMyZMno0mTJtDX14eZmRmGDBmidtq8tJazZ8/C398fFhYWMDAwwMCBA9W+ZgRBwOLFi/HOO++gVq1a6NatG/7888+X1vJvtra26Ny5M3bt2qXSHhYWhhYtWqB58+ZlLhcbG4vevXvD2NgYtWrVQpcuXXD27Fnx86CgIHz++ecAAHt7+3IvE4iIiEDz5s2hp6cHZ2dnREZGqq0rLi4Onp6eMDIygqGhIXr06FHmqfU///wT3bt3h76+Pt555x0sXryYl5O8ZXjEjiRn0qRJ2L59O3x9fTFt2jQkJiYiJCQEcXFxOHv2LHR0dJCZmYlevXrBwsICs2fPhomJCZKSkhAeHg7g2Wm4TZs24ZNPPsHAgQMxaNAgAEDLli1fuv6wsDB069YN1tbWGD58OGbPno1Dhw5hyJAhYp+SkhL07dsX0dHRGD58OD799FM8fvwYUVFR+OOPP+Dg4AAAGDduHLZv3w5PT0+MHz8excXFOH36NM6fP482bdpUav8MGTIEjo6OWLp0KQRBAABERUXhzp078PX1hbW1Nf788098/fXX+PPPP3H+/HnIZDIAwN9//4127dohOzsbEydOhJOTE1JTU/HDDz/g6dOnaNiwITp27IiwsDDMmDFDbb/Url1b5YhGeZYsWQKZTIYvvvgCmZmZWLt2Ldzd3REfHw99fX2MHj0aCxcuxN69e1UuHi8sLMQPP/yAwYMHv/LRpBs3bmDEiBGYNGkSJkyYgCZNmgAANm3aBGdnZ3zwwQfQ1tbGoUOHMHnyZCiVSkyZMuWl4y5btgxyuRwzZ85ETk4OVqxYgVGjRiE2Nvaly+7atQuPHz/GpEmTIJPJsGLFCgwaNAh37twRj/L9/PPPGDZsGFq0aIHg4GA8evQI48aNg62t7UvHb9CgAYBnwXDIkCGoVavWS5cp1blzZ7XT3MnJyZg3b55KIFyyZAnmz5+PoUOHYvz48bh//z42bNiAzp07Iy4u7oWnGy9evIhz585h+PDheOedd5CUlIRNmzaha9euuHbtmlq9U6dOhampKQIDA5GUlIS1a9fCz88Pe/fuFfsEBARg8eLF8PLygpeXFy5fvoxevXqhsLCwwtsOACNHjsSnn36KJ0+ewNDQEMXFxdi/fz/8/f3L/CPj+PHj8PT0ROvWrREYGAi5XC7+0XD69Gm0a9cOgwYNws2bN7F79258+eWXMDc3B6B6mcCZM2cQHh6OyZMno3bt2li/fj0GDx6MlJQUmJmZAXgW1t5//30YGRlh1qxZ0NHRwZYtW9C1a1f89ttvcHNzAwCkp6ejW7duKC4uxuzZs2FgYICvv/5aoyPV9AYQiN5iU6ZMEf79ZXz69GkBgBAWFqbSLzIyUqX94MGDAgDh4sWL5Y59//59AYAQGBhY4XoyMjIEbW1tYevWrWLbe++9J/Tv31+l37fffisAENasWaM2hlKpFARBEI4fPy4AEKZNm1Zun8TERAGAEBoaqtbn+doDAwMFAMKIESPU+j59+lStbffu3QIA4dSpU2Kbt7e3IJfLy9xvpTVt2bJFACBcv35d/KywsFAwNzcXfHx81Jb7txMnTggABFtbWyE3N1ds37dvnwBAWLdundjWoUMHwc3NTWX58PBwAYBw4sSJF67n3/r06SM0aNBApa1BgwYCACEyMlKtf1n7ysPDQ2jYsKFKW5cuXYQuXbqobVvTpk2FgoICsX3dunUCAOHq1atim4+Pj0pNpf/PZmZmQlZWltj+448/CgCEQ4cOiW0tWrQQ3nnnHeHx48di28mTJwUAattZFm9vbwGAYGpqKgwcOFBYtWqVyv/l89tT3r7+559/hNatWws2NjZCWlqaIAiCkJSUJGhpaQlLlixR6Xv16lVBW1tbrf15Ze37mJgYAYDw/fffi22hoaECAMHd3V38uhQEQZgxY4agpaUlZGdnC4IgCJmZmYKurq7Qp08flX7/+c9/BAAv/XoVhGffZ1OmTBGysrIEXV1dYceOHYIgCMLPP/8syGQyISkpSfzeu3//viAIz75XHB0dBQ8PD5X1Pn36VLC3txd69uwptq1cuVIAICQmJpa5bl1dXeHWrVti25UrVwQAwoYNG8S2AQMGCLq6usLt27fFtr///luoXbu20LlzZ7Ft+vTpAgAhNjZWbMvMzBSMjY3LrYHePDwVS5Kyf/9+GBsbo2fPnnjw4IH4at26NQwNDXHixAkAEI8KHD58GEVFRVW2/j179kAul2Pw4MFi24gRI/Drr7+qnAo+cOAAzM3NMXXqVLUxSo+OHThwADKZDIGBgeX2qYyPP/5Yre3ff5Hn5+fjwYMH4kXyly9fBvDstHBERAT69etX5tHC0pqGDh0KhUKBsLAw8bMjR47gwYMHFb4W0tvbG7Vr1xbff/jhh6hbty5++eUXlT6xsbEqp67DwsJQr149dOnSpULreRF7e3t4eHiotf97X+Xk5ODBgwfo0qUL7ty5g5ycnJeO6+vrq3L93fvvvw8AuHPnzkuXHTZsGExNTctd9u+//8bVq1fh7e0NQ0NDsV+XLl3QokWLl44PPDvVHBISAnt7exw8eBAzZ85E06ZN0aNHD6SmplZoDACYPHkyrl69igMHDsDa2hoAEB4eDqVSiaFDh6p8f1pbW8PR0VH8/izPv/d9UVERHj58iEaNGsHExET8Ov23iRMnqnyvvP/++ygpKUFycjIA4NixYygsLMTUqVNV+k2fPr3C21nK1NQUvXv3Fi812LVrF9577z3xKOi/xcfHIyEhASNHjsTDhw/F/ZCXl4cePXrg1KlTFT796e7uLh7hB56dVTAyMhK/JkpKSnD06FEMGDAADRs2FPvVrVsXI0eOxJkzZ5Cbmwvg2d387du3R7t27cR+FhYWGDVqlMb7g2oOgx1JSkJCAnJycmBpaQkLCwuV15MnT5CZmQng2S+6wYMHY8GCBTA3N0f//v0RGhqKgoKCV1r/zp070a5dOzx8+BC3bt3CrVu30KpVKxQWFmL//v1iv9u3b6NJkybQ1i7/aojbt2/DxsYGderUeaWanmdvb6/WlpWVhU8//RRWVlbQ19eHhYWF2K80rNy/fx+5ubnlXi9UysTEBP369VO53igsLAy2trbo3r17hWp0dHRUeS+TydCoUSOVa4uGDRsGPT09MUDm5OTg8OHDGDVq1CsF31Jl7ScAOHv2LNzd3WFgYAATExNYWFjgP//5j1jDy9SvX1/lfWlQe/4a0MosWxpYGjVqpLZsWW1lkcvlmDJlCi5duoQHDx7gxx9/hKenJ44fP47hw4dXaIwtW7YgNDQUGzZsEP9AAJ59fwqCAEdHR7Xvz+vXr4vfn+X5559/EBAQgHr16kFPTw/m5uawsLBAdnZ2mfu+ovvr+a83CwsLlQBdUSNHjkRUVBRSUlIQERGBkSNHltkvISEBAODj46O2H7755hsUFBRU6GsJUN9G4Nl2lm7j/fv38fTpU/FSgn9r2rQplEqleH1jcnKy2r4AUOay9ObiNXYkKUqlEpaWlipHi/6t9NqU0klwz58/j0OHDuHIkSMYO3YsVq9ejfPnz6sc7aiohIQE8UL2sn44hoWFYeLEiRqP+yLlBZiSkpJylynrepmhQ4fi3Llz+Pzzz+Hq6gpDQ0MolUr07t27UhdOe3t7Y//+/Th37hxatGiBn376CZMnT4ZcXnV/S5qamqJv374ICwtDQEAAfvjhBxQUFFTZHdJl7afbt2+jR48ecHJywpo1a1CvXj3o6uril19+wZdfflmhfaWlpVVmu/D/r3esrmUrw8zMDB988AE++OAD8Xqs5OTkMo9Clbpw4QI+/fRTjB8/Xu3rXalUQiaT4ddffy1zW172fTd16lSEhoZi+vTp6NChA4yNjSGTyTB8+PAy9/3r3l8ffPAB9PT04OPjg4KCAgwdOrTMfqW1rly5Eq6urmX2qejPoNe9jfTmY7AjSXFwcMCxY8fQsWPHCl3w2759e7Rv3x5LlizBrl27MGrUKOzZswfjx4/X+KhPWFgYdHR0sGPHDrUftmfOnMH69euRkpKC+vXrw8HBAbGxsSgqKip3agsHBwccOXIEWVlZ5R61Kz2q8PzkpaVHIiri0aNHiI6OxoIFCxAQECC2lx5VKGVhYQEjIyP88ccfLx2zd+/esLCwQFhYGNzc3PD06VOMHj26wjU9v25BEHDr1i21m1e8vb3Rv39/XLx4EWFhYWjVqhWcnZ0rvB5NHTp0CAUFBfjpp59UjpS87BTi61IauJ6/y7a8Nk20adMGv/32G9LS0soNdvfv38eHH34IV1dXbNy4Ue1zBwcHCIIAe3t7NG7cWOMafvjhB/j4+GD16tViW35+fqUn7y3djoSEBJXTlPfv36/QEdTn6evrY8CAAdi5cyc8PT3Fmx2eV3rq1MjI6KVTIb3q0WcLCwvUqlULN27cUPvsr7/+glwuR7169QA82x/Pf+8BKHNZenPxVCxJytChQ1FSUoJFixapfVZcXCz+Anj06JHaX7SlfzmXno4tvcOuor80wsLC8P7772PYsGH48MMPVV6lUxaUXn8zePBgPHjwACEhIWrjlNY1ePBgCIIgTpJcVh8jIyOYm5vj1KlTKp9/9dVXFaoZ+L+/+J/fH2vXrlV5L5fLMWDAABw6dEicbqWsmgBAW1sbI0aMwL59+7B9+3a0aNGiQncUl/r+++/x+PFj8f0PP/yAtLQ0eHp6qvQr/eW5fPly/Pbbb9U+n2FZ+yonJwehoaHVut6KsrGxQfPmzfH999/jyZMnYvtvv/2Gq1evvnT59PR0XLt2Ta29sLAQ0dHRkMvl5Z7SLSkpwfDhw1FYWIgDBw6UOY/foEGDoKWlhQULFqh9vQmCUOb0Ov+mpaWlttyGDRteeIT6Rdzd3aGjo4MNGzaojPv8174mZs6cicDAQMyfP7/cPq1bt4aDgwNWrVql8v9U6t9TspTOM1nZ8KqlpYVevXrhxx9/VLmUISMjA7t27UKnTp1gZGQEAPDy8sL58+dx4cIFlVrKOwNCbyYesSNJ6dKlCyZNmoTg4GDEx8ejV69e0NHRQUJCAvbv349169bhww8/xHfffYevvvoKAwcOhIODAx4/foytW7fCyMgIXl5eAJ799d2sWTPs3bsXjRs3Rp06ddC8efMyrzGLjY3FrVu3yn1uo62tLd59912EhYXhiy++gLe3N77//nv4+/vjwoULeP/995GXl4djx45h8uTJ6N+/P7p164bRo0dj/fr1SEhIEE+Lnj59Gt26dRPXNX78eCxbtgzjx49HmzZtcOrUKdy8ebPC+8zIyAidO3fGihUrUFRUBFtbWxw9ehSJiYlqfZcuXYqjR4+iS5cumDhxIpo2bYq0tDTs378fZ86cUZmqwtvbG+vXr8eJEyewfPnyCtcDAHXq1EGnTp3g6+uLjIwMrF27Fo0aNcKECRNU+uno6GD48OEICQmBlpYWRowYodF6NNWrVy/o6uqiX79+mDRpEp48eYKtW7fC0tISaWlp1bruilq6dCn69++Pjh07wtfXF48ePUJISAiaN29eZoj4t3v37qFdu3bo3r07evToAWtra2RmZmL37t24cuUKpk+fXu5RqM2bN+P48eP4+OOP1Y5gWllZoWfPnnBwcMDixYsxZ84cJCUlYcCAAahduzYSExNx8OBBTJw4ETNnziy3vr59+2LHjh0wNjZGs2bNEBMTg2PHjonTemjKwsICM2fORHBwMPr27QsvLy/ExcXh119/LXc7X8bFxQUuLi4v7COXy/HNN9/A09MTzs7O8PX1ha2tLVJTU3HixAkYGRnh0KFDAJ6FQODZo8mGDx8OHR0d9OvXT6OJxRcvXoyoqCh06tQJkydPhra2NrZs2YKCggKsWLFC7Ddr1izs2LEDvXv3xqeffipOd9KgQQP897//rcTeoBrxum/DJapKz093Uurrr78WWrduLejr6wu1a9cWWrRoIcyaNUv4+++/BUEQhMuXLwsjRowQ6tevL+jp6QmWlpZC3759hd9//11lnHPnzgmtW7cWdHV1Xzj1ydSpUwUAKtMJPC8oKEgAIFy5ckUQhGdTG8ydO1ewt7cXdHR0BGtra+HDDz9UGaO4uFhYuXKl4OTkJOjq6goWFhaCp6encOnSJbHP06dPhXHjxgnGxsZC7dq1haFDhwqZmZnlTndSOuXCv927d08YOHCgYGJiIhgbGwtDhgwR/v777zK3OTk5WfD29hYsLCwEPT09oWHDhsKUKVNUpvAo5ezsLMjlcuHevXvl7pd/K51CY/fu3cKcOXMES0tLQV9fX+jTp4+QnJxc5jIXLlwQAAi9evWq0DqeV950J3369Cmz/08//SS0bNlSUCgUgp2dnbB8+XJx+pp/TwdR3nQn+/fvVxmvrClrypvuZOXKlWr1lPV/tGfPHsHJyUnQ09MTmjdvLvz000/C4MGDBScnpxfui9zcXGHdunWCh4eH8M477wg6OjpC7dq1hQ4dOghbt25VmZrj+elOSr++ynr9ez8IgiAcOHBA6NSpk2BgYCAYGBgITk5OwpQpU4QbN268sL5Hjx4Jvr6+grm5uWBoaCh4eHgIf/31l9CgQQOVqUlKpzt5flqesqZoKSkpERYsWCDUrVtX0NfXF7p27Sr88ccfamOWB/9/upMXKe97Ly4uThg0aJBgZmYm6OnpCQ0aNBCGDh0qREdHq/RbtGiRYGtrK8jlcpWvs/LWXVbtly9fFjw8PARDQ0OhVq1aQrdu3YRz586pLfvf//5X6NKli6BQKARbW1th0aJFwrZt2zjdyVtEJgi8wpKIqkerVq1Qp04dREdHV9s6rly5AldXV3z//fcaXcf3v8bV1RUWFhaIioqq6VKIqBrxGjsiqha///474uPj4e3tXa3r2bp1KwwNDcWng/yvKyoqQnFxsUrbyZMnceXKFbVHexGR9PAaOyKqUn/88QcuXbqE1atXo27duhg2bFi1rOfQoUO4du0avv76a/j5+Wl0zZGUpaamwt3dHR999BFsbGzw119/YfPmzbC2ti5zcmoikhYGOyKqUj/88AMWLlyIJk2aYPfu3a/8zNbyTJ06FRkZGfDy8irzzuH/VaampmjdujW++eYb3L9/HwYGBujTpw+WLVtW6ZsMiOjtUeOnYjdu3Ag7OzsoFAq4ubmp3Gb9vKKiIixcuBAODg5QKBRwcXFBZGSkSp+SkhLMnz8f9vb20NfXh4ODAxYtWqRyK3t4eDh69eoFMzMzyGQyxMfHq60rPz8fU6ZMgZmZGQwNDTF48GBkZGRU2XYTSVVQUBCUSiWuX79eJY/2Kk9SUhL++ecfREREqDx+7H+dsbEx9u7di3v37qGgoABZWVnYv3+/ymOniEi6ajTY7d27F/7+/ggMDMTly5fh4uICDw+Pch8rM2/ePGzZsgUbNmzAtWvX8PHHH2PgwIGIi4sT+yxfvhybNm1CSEgIrl+/juXLl2PFihXYsGGD2CcvLw+dOnV64RQMM2bMwKFDh7B//3789ttv+Pvvv3kNDxEREb3RavSuWDc3N7Rt21acpFWpVKJevXqYOnUqZs+erdbfxsYGc+fOxZQpU8S2wYMHQ19fHzt37gTwbJ4jKysrbNu2rdw+pZKSkmBvb4+4uDiVx7rk5OTAwsICu3btwocffgjg2QzdTZs2RUxMjMqzD4mIiIjeFDV2jV1hYSEuXbqEOXPmiG1yuRzu7u6IiYkpc5mCggK163X09fVx5swZ8f17772Hr7/+Gjdv3kTjxo1x5coVnDlzBmvWrKlwbZcuXUJRUZHKo16cnJxQv379Fwa7goIClYfIK5VKZGVliad8iUg6tm7divXr1yMjIwPNmzfHypUrxclkn1dUVIQ1a9Zg165dSEtLg6OjIxYsWKDyM6akpATBwcHYu3cvMjMzYW1tjVGjRuHzzz8Xf34IgoClS5fiu+++Q05ODtzc3PDll1+qnGaNj49HYGAg4uLiIJfL8cEHH2Dp0qWVev4xEVUfQRDw+PFj2NjYVOlztGtsguLU1FQBgNoEiZ9//rnQrl27MpcZMWKE0KxZM+HmzZtCSUmJcPToUUFfX1/Q1dUV+5SUlAhffPGFIJPJBG1tbUEmkwlLly4tc7zSST/j4uJU2sPCwlTGLNW2bVth1qxZ5W7Tiybo5Isvvvjiiy+++Hr+dffu3XJzRWW8VXfFrlu3DhMmTICTkxNkMhkcHBzg6+uLb7/9Vuyzb98+hIWFYdeuXXB2dkZ8fDymT58OGxsb+Pj4VGt9c+bMgb+/v/g+JycH9evXx927d8Vn8RHR26979+549913sWrVKgDPjs43a9YMEydOVPkZUKpJkyaYOXOmyiPRPvroI+jr62Pr1q0Anj3n2MLCAhs3biyzjyAIaNKkCfz8/DBt2jQAz37GODo64quvvsKHH36I0NBQLFmyBDdv3hSPAPz555947733cPnyZd5AQfQGyc3NRb169ar85q8aC3bm5ubQ0tJSu9M0IyMD1tbWZS5jYWGBiIgI5Ofn4+HDh7CxscHs2bPRsGFDsc/nn3+O2bNnY/jw4QCAFi1aIDk5GcHBwRUOdtbW1igsLER2drbKsy9fVBsA6OnpQU9PT63dyMiIwY5IIgoLCxEfH4958+apfF/37NkTcXFxZX6vFxYWwsTEROUzIyMjnDlzRmzr3Lkzvv76a6Snp4uXkcTGxmLNmjUwMjLCnTt3kJGRgb59+4rLGBkZwc3NDVeuXMHYsWMhl8uhp6en8nPLwsICwLMndLRq1ao6dgkRvYKqvlSrxu6K1dXVRevWrVUeNaRUKhEdHY0OHTq8cFmFQgFbW1sUFxfjwIED6N+/v/jZ06dP1c5Va2lpQalUVri21q1bQ0dHR6W2GzduICUl5aW1EZG0PXjwACUlJbCyslJpt7KyQnp6epnLeHh4YM2aNUhISIBSqURUVBTCw8ORlpYm9in9g9TJyQk6Ojpo1aoVpk+fjlGjRgGAOPaL1tu9e3ekp6dj5cqVKCwsxKNHj8Qb0f69LiKSrhqd7sTf3x9bt27Fd999h+vXr+OTTz5BXl4efH19AQDe3t4qN1fExsYiPDwcd+7cwenTp9G7d28olUrMmjVL7NOvXz8sWbIEP//8M5KSknDw4EGsWbMGAwcOFPtkZWUhPj4e165dA/AstMXHx4s/HI2NjTFu3Dj4+/vjxIkTuHTpEnx9fdGhQwfeEUtEGlu3bh0cHR3h5OQEXV1d+Pn5wdfXV+WP0H9fRnL58mV89913WLVqFb777rsKr8fZ2RnfffcdVq9ejVq1asHa2hr29vawsrKq2ouziejNVaVX7FXChg0bhPr16wu6urpCu3bthPPnz4ufdenSRfDx8RHfnzx5UmjatKmgp6cnmJmZCaNHjxZSU1NVxsvNzRU+/fRToX79+oJCoRAaNmwozJ07VygoKBD7hIaGlnkBY2BgoNjnn3/+ESZPniyYmpoKtWrVEgYOHCikpaVptG05OTkCACEnJ0eznUJEb6yCggJBS0tLOHjwoEq7t7e38MEHH7xw2X/++Ue4d++eoFQqhVmzZgnNmjUTP3vnnXeEkJAQlf6LFi0SmjRpIgiCINy+fVsA1G/26ty5szBt2jS1daWnpwuPHz8Wnjx5IsjlcmHfvn0abCURVbfqygg1Oo+d1OXm5sLY2Bg5OTkvvMaupKQERUVFr7EyotdHR0cHWlpaNV1GlXJzc0O7du3Eic+VSiXq168PPz+/MufgfF5RURGaNm2KoUOHYunSpQAAMzMzLF68GJ988onYLzg4GKGhobh58yYEQYCNjQ1mzpyJzz77DMCznzGWlpbYvn27eF3x87799ltMnToVqampKtfeEVHNqmhG0NRbdVes1AiCgPT0dGRnZ9d0KUTVysTEBNbW1pKZz9Hf3x8+Pj5o06YN2rVrh7Vr16pdRmJra4vg4GAAzy4jSU1NhaurK1JTU8XHrpV1GUn9+vXh7OyMuLg4rFmzBmPHjgXw7ALr6dOnY/HixXB0dIS9vT3mz58PGxsbDBgwQBwnJCQE7733HgwNDREVFYXPP/8cy5YtY6gj+h/BYFeDSkOdpaUlatWqJZlfekSlBEHA06dPxccE1q1bt4YrqhrDhg3D/fv3ERAQgPT0dLi6uiIyMlK8sSElJUXlmrb8/HzMmzcPd+7cgaGhIby8vLBjxw6VsLVhwwbMnz8fkydPRmZmJmxsbDBp0iQEBASIfWbNmoW8vDxMnDgR2dnZ6NSpEyIjI1Umbr9w4QICAwPx5MkTODk5YcuWLRg9enT17xQieiPwVGw1etFh1pKSEty8eROWlpYwMzOroQqJXo+HDx8iMzMTjRs3ltxpWSKiyqiuU7G8TaqGlF5TV6tWrRquhKj6lX6d81pSIqLqxWBXw3j6lf4X8OuciOj1YLAjIiIikggGO3qtkpKSIJPJEB8fX+Fltm/fXuV39FWmDiIiojcd74p9wyyLe/Ba1ze7lXmllrt79y4CAwMRGRmJBw8eoG7duhgwYAACAgJeeDNIvXr1kJaWBnPziq932LBh8PLyqlSdRERE/0t4xI40dufOHbRp0wYJCQnYvXs3bt26hc2bN4vP+c3KyipzucLCQmhpacHa2hra2hX/m0JfXx+WlpZVVT4REZFk8YgdaWzKlCnQ1dXF0aNHoa+vDwCoX78+WrVqBQcHB8ydOxebNm2CnZ0dxo0bh4SEBERERGDQoEEICgqCvb094uLi4OrqCgD46aef8Nlnn+Hu3bvo0KEDxowZgzFjxuDRo0cwMTHB9u3bMX36dHEi56CgIEREROCzzz7D/Pnz8ejRI3h6emLr1q2oXbs2ACAyMhKLFy/GH3/8AS0tLXTo0AHr1q2Dg4NDTewyKnXYCShMrukq6G02kjN0Eb0Ij9iRRrKysnDkyBFMnjxZDHWlrK2tMWrUKOzduxel0yOuWrUKLi4uiIuLw/z589XGS0xMxIcffogBAwbgypUrmDRpEubOnfvSOm7fvo2IiAgcPnwYhw8fxm+//YZly5aJn+fl5cHf3x+///47oqOjIZfLMXDgQCiVylfcA0RERG8uHrEjjSQkJEAQBDRt2rTMz5s2bYpHjx7h/v37AIDu3buLz7UEnt208G9btmxBkyZNsHLlSgBAkyZN8Mcff2DJkiUvrEOpVGL79u3iEbrRo0cjOjpaXG7w4MEq/b/99ltYWFjg2rVraN68ecU3mIiI6C3CI3ZUKRV9YEmbNm1e+PmNGzfQtm1blbZ27dq9dFw7Ozsx1AHPHlVV+tgq4FkAHTFiBBo2bAgjIyPY2dkBePaoJyIiIqlisCONNGrUCDKZDNevXy/z8+vXr8PU1BQWFhYAAAMDg2qpQ0dHR+W9TCZTOc3ar18/ZGVlYevWrYiNjUVsbCyAZzdwEBERSRWDHWnEzMwMPXv2xFdffYV//vlH5bP09HSEhYVh2LBhFX7SQJMmTfD777+rtF28ePGVanz48CFu3LiBefPmoUePHuLpYSIiIqljsCONhYSEoKCgAB4eHjh16hTu3r2LyMhI9OzZE7a2ti+9Pu7fJk2ahL/++gtffPEFbt68iX379mH79u0AKv8YKlNTU5iZmeHrr7/GrVu3cPz4cfj7+1dqLCIiorcJgx1pzNHREb///jsaNmyIoUOHwsHBARMnTkS3bt0QExODOnXqVHgse3t7/PDDDwgPD0fLli2xadMm8a5YPT29StUnl8uxZ88eXLp0Cc2bN8eMGTPEmzOIiIikTCZU9Cp40lhubi6MjY2Rk5MDIyMjlc/y8/ORmJgIe3t7KBSKGqrwzbRkyRJs3rwZd+/erelSqIqIX+9/ekLBeezoVXAeO5KIF2WEV8HpTqjGffXVV2jbti3MzMxw9uxZrFy5En5+fjVdFhER0VuHwY5qXEJCAhYvXoysrCzUr18fn332GebMmVPTZREREb11GOyoxn355Zf48ssva7oMIiKitx5vniAiIiKSCAY7IiIiIolgsCMiIiKSCAY7IiIiIolgsCMiIiKSCAY7IiIiIolgsCOqoKSkJMhkMsTHx9d0KRqTyWSIiIio6TKIiKiacR67N82uyj34vtI0fDzPmDFjkJ2dzZBQjq5du8LV1RVr164FANjZ2WH69OmYPn36a1l/UFAQIiIi1MJnWloaTE1NX0sNRERUc3jEjt4oJSUlUCqVNV3GG6ewsPCVlre2toaenl4VVUNERG8qBjt6JV27dsW0adMwa9Ys1KlTB9bW1ggKClLpk52djUmTJsHKygoKhQLNmzfH4cOHAQDbt2+HiYkJfvrpJzRr1gx6enpISUlBQUEBZs6cCVtbWxgYGMDNzQ0nT54Ux3z48CFGjBgBW1tb1KpVCy1atMDu3btV1vvDDz+gRYsW0NfXh5mZGdzd3ZGXlyd+/s0336Bp06ZQKBRwcnLCV199pbL8hQsX0KpVKygUCrRp0wZxcXEa75vk5GTMmDEDMpkMMtn/HY09c+YM3n//fejr66NevXqYNm2aSm12dnZYtGgRvL29YWRkhIkTJwIAvvjiCzRu3Bi1atVCw4YNMX/+fBQVFYn7csGCBbhy5Yq4vu3btwNQPxV79epVdO/eXdw3EydOxJMnT8TPx4wZgwEDBmDVqlWoW7cuzMzMMGXKFHFdwLNn/Do6OkKhUMDKygoffvihRvuHiIiqHk/F0iv77rvv4O/vj9jYWMTExGDMmDHo2LEjevbsCaVSCU9PTzx+/Bg7d+6Eg4MDrl27Bi0tLXH5p0+fYvny5fjmm29gZmYGS0tL+Pn54dq1a9izZw9sbGxw8OBB9O7dG1evXoWjoyPy8/PRunVrfPHFFzAyMsLPP/+M0aNHw8HBAe3atUNaWhpGjBiBFStWYODAgXj8+DFOnz4NQXh26jksLAwBAQEICQlBq1atEBcXhwkTJsDAwAA+Pj548uQJ+vbti549e2Lnzp1ITEzEp59+qtF+CQ8Ph4uLCyZOnIgJEyaI7bdv30bv3r2xePFifPvtt7h//z78/Pzg5+eH0NBQsd+qVasQEBCAwMBAsa127drYvn07bGxscPXqVUyYMAG1a9fGrFmzMGzYMPzxxx+IjIzEsWPHAADGxsZqdeXl5cHDwwMdOnTAxYsXkZmZifHjx8PPz08MggBw4sQJ1K1bFydOnMCtW7cwbNgwuLq6YsKECfj9998xbdo07NixA++99x6ysrJw+vRpjfYPERFVPQY7emUtW7YUw4ejoyNCQkIQHR2Nnj174tixY7hw4QKuX7+Oxo0bAwAaNmyosnxRURG++uoruLi4AABSUlIQGhqKlJQU2NjYAABmzpyJyMhIhIaGYunSpbC1tcXMmTPFMaZOnYojR45g3759YrArLi7GoEGD0KBBAwBAixYtxP6BgYFYvXo1Bg0aBACwt7fHtWvXsGXLFvj4+GDXrl1QKpXYtm0bFAoFnJ2dce/ePXzyyScV3i916tSBlpYWateuDWtra7E9ODgYo0aNEq+7c3R0xPr169GlSxds2rQJCoUCANC9e3d89tlnKmPOmzdP/LednR1mzpyJPXv2YNasWdDX14ehoSG0tbVV1ve8Xbt2IT8/H99//z0MDAwAACEhIejXrx+WL18OKysrAICpqSlCQkKgpaUFJycn9OnTB9HR0ZgwYQJSUlJgYGCAvn37onbt2mjQoAFatWpV4X1DRETVg8GOXlnLli1V3tetWxeZmZkAgPj4eLzzzjtiqCuLrq6uyhhXr15FSUmJ2jIFBQUwMzMD8OxavKVLl2Lfvn1ITU1FYWEhCgoKUKtWLQCAi4sLevTogRYtWsDDwwO9evXChx9+CFNTU+Tl5eH27dsYN26cypG04uJi8QjX9evX0bJlSzFkAUCHDh0qs3vUXLlyBf/9738RFhYmtgmCAKVSicTERDRt2hQA0KZNG7Vl9+7di/Xr1+P27dt48uQJiouLYWRkpNH6r1+/DhcXFzHUAUDHjh2hVCpx48YNMdg5OzurHFmtW7curl69CgDo2bMnGjRogIYNG6J3797o3bs3Bg4cKO5/IiKqGQx29Mp0dHRU3stkMvEGCH19/Zcur6+vr3L92ZMnT6ClpYVLly6pBAsAMDQ0BACsXLkS69atw9q1a9GiRQsYGBhg+vTp4k0GWlpaiIqKwrlz53D06FFs2LABc+fORWxsrBg+tm7dCjc3N5Xxn19fdXjy5AkmTZqEadOmqX1Wv3598d//Dl4AEBMTg1GjRmHBggXw8PCAsbEx9uzZg9WrV1dLnS/6f61duzYuX76MkydP4ujRowgICEBQUBAuXrwIExOTaqmHiIhejsGOqlXLli1x79493Lx584VH7f6tVatWKCkpQWZmJt5///0y+5w9exb9+/fHRx99BABQKpW4efMmmjVrJvaRyWTo2LEjOnbsiICAADRo0AAHDx6Ev78/bGxscOfOHYwaNarM8Zs2bYodO3YgPz9fPGp3/vx5TTYdwLOjkSUlJSpt7777Lq5du4ZGjRppNNa5c+fQoEEDzJ07V2xLTk5+6fqe17RpU2zfvh15eXlieDx79izkcjmaNGlS4Xq0tbXh7u4Od3d3BAYGwsTEBMePHxdPbxMR0evHu2KpWnXp0gWdO3fG4MGDERUVhcTERPz666+IjIwsd5nGjRtj1KhR8Pb2Rnh4OBITE3HhwgUEBwfj559/BvDsurTSI3LXr1/HpEmTkJGRIY4RGxuLpUuX4vfff0dKSgrCw8Nx//598TTnggULEBwcjPXr1+PmzZu4evUqQkNDsWbNGgDAyJEjIZPJMGHCBFy7dg2//PILVq1apfH229nZ4dSpU0hNTcWDBw8APLuz9dy5c/Dz80N8fDwSEhLw448/ws/P74VjOTo6IiUlBXv27MHt27exfv16HDx4UG19iYmJiI+Px4MHD1BQUKA2zqhRo6BQKODj44M//vgDJ06cwNSpUzF69GjxNOzLHD58GOvXr0d8fDySk5Px/fffQ6lUahQMiYio6jHYUbU7cOAA2rZtixEjRqBZs2aYNWvWS48qhYaGwtvbG5999hmaNGmCAQMG4OLFi+Kpynnz5uHdd9+Fh4cHunbtCmtrawwYMEBc3sjICKdOnYKXlxcaN26MefPmYfXq1fD09AQAjB8/Ht988w1CQ0PRokULdOnSBdu3b4e9vT2AZ6d8Dx06hKtXr6JVq1aYO3culi9frvG2L1y4EElJSXBwcICFhQWAZ0cxf/vtN9y8eRPvv/8+WrVqhYCAAPFGkfJ88MEHmDFjBvz8/ODq6opz585h/vz5Kn0GDx6M3r17o1u3brCwsFCbAgYAatWqhSNHjiArKwtt27bFhx9+iB49eiAkJKTC22ViYoLw8HB0794dTZs2xebNm7F79244OztXeAwiIqp6MqF0/geqcrm5uTA2NkZOTo7aBe75+flITEyEvb29ygX6RFIkfr3/6QlFYfLLFyAqj4ZPyyF6U70oI7wKHrEjIiIikggGOyIiIiKJYLAjIiIikggGOyIiIiKJYLAjIiIikggGuxpWOpM/kZSJX+fCi6e5ISKiV8MnT9QQXV1dyOVy/P3337CwsICurq7KY7WIpEAQBBQWFuL+/fuQy+XQLUqv6ZKIiCSNwa6GyOVy2NvbIy0tDX///XdNl0NUrWrVqoX69etDfqm4pkshIpI0BrsapKuri/r166O4uPilT2IgeltpaWlBW1ubR6SJiF4DBrsaJpPJoKOjAx0dnZouhYiIiN5yvHmCiIiISCIY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCLeiGC3ceNG2NnZQaFQwM3NDRcuXCi3b1FRERYuXAgHBwcoFAq4uLggMjJSpU9JSQnmz58Pe3t76Ovrw8HBAYsWLYIgCGIfQRAQEBCAunXrQl9fH+7u7khISFAZx87ODjKZTOW1bNmyqt14IiIioipS48Fu79698Pf3R2BgIC5fvgwXFxd4eHggMzOzzP7z5s3Dli1bsGHDBly7dg0ff/wxBg4ciLi4OLHP8uXLsWnTJoSEhOD69etYvnw5VqxYgQ0bNoh9VqxYgfXr12Pz5s2IjY2FgYEBPDw8kJ+fr7K+hQsXIi0tTXxNnTq1enYEERER0SuSCf8+jFUD3Nzc0LZtW4SEhAAAlEol6tWrh6lTp2L27Nlq/W1sbDB37lxMmTJFbBs8eDD09fWxc+dOAEDfvn1hZWWFbdu2ldlHEATY2Njgs88+w8yZMwEAOTk5sLKywvbt2zF8+HAAz47YTZ8+HdOnT6/UtuXm5sLY2Bg5OTkwMjKq1BhEkrJLVtMV0NtuZI3+yiKqMtWVEWr0iF1hYSEuXboEd3d3sU0ul8Pd3R0xMTFlLlNQUACFQqHSpq+vjzNnzojv33vvPURHR+PmzZsAgCtXruDMmTPw9PQEACQmJiI9PV1lvcbGxnBzc1Nb77Jly2BmZoZWrVph5cqVKC4uLnd7CgoKkJubq/IiIiIiel20a3LlDx48QElJCaysrFTarays8Ndff5W5jIeHB9asWYPOnTvDwcEB0dHRCA8PR0lJidhn9uzZyM3NhZOTE7S0tFBSUoIlS5Zg1KhRAID09HRxPc+vt/QzAJg2bRreffdd1KlTB+fOncOcOXOQlpaGNWvWlFlbcHAwFixYoPmOICIiIqoCNRrsKmPdunWYMGECnJycIJPJ4ODgAF9fX3z77bdin3379iEsLAy7du2Cs7Mz4uPjMX36dNjY2MDHx6fC6/L39xf/3bJlS+jq6mLSpEkIDg6Gnp6eWv85c+aoLJObm4t69epVckuJiIiINFOjp2LNzc2hpaWFjIwMlfaMjAxYW1uXuYyFhQUiIiKQl5eH5ORk/PXXXzA0NETDhg3FPp9//jlmz56N4cOHo0WLFhg9ejRmzJiB4OBgABDH1mS9wLPrAYuLi5GUlFTm53p6ejAyMlJ5EREREb0uNRrsdHV10bp1a0RHR4ttSqUS0dHR6NChwwuXVSgUsLW1RXFxMQ4cOID+/fuLnz19+hRyueqmaWlpQalUAgDs7e1hbW2tst7c3FzExsa+cL3x8fGQy+WwtLTUaDuJiIiIXocaPxXr7+8PHx8ftGnTBu3atcPatWuRl5cHX19fAIC3tzdsbW3Fo22xsbFITU2Fq6srUlNTERQUBKVSiVmzZolj9uvXD0uWLEH9+vXh7OyMuLg4rFmzBmPHjgUAyGQyTJ8+HYsXL4ajoyPs7e0xf/582NjYYMCAAQCAmJgYxMbGolu3bqhduzZiYmIwY8YMfPTRRzA1NX29O4mIiIioAmo82A0bNgz3799HQEAA0tPT4erqisjISPHGhpSUFJWjb/n5+Zg3bx7u3LkDQ0NDeHl5YceOHTAxMRH7bNiwAfPnz8fkyZORmZkJGxsbTJo0CQEBAWKfWbNmIS8vDxMnTkR2djY6deqEyMhI8Y5bPT097NmzB0FBQSgoKIC9vT1mzJihcg0dERER0ZukxuexkzLOY0f0HM5jR6+K89iRREhyHjsiIiIiqjoMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSwWBHREREJBEMdkREREQSoV2ZhVJSUpCcnIynT5/CwsICzs7O0NPTq+raiIiIiEgDFQ52SUlJ2LRpE/bs2YN79+5BEATxM11dXbz//vuYOHEiBg8eDLmcBwKJiIiIXrcKJbBp06bBxcUFiYmJWLx4Ma5du4acnBwUFhYiPT0dv/zyCzp16oSAgAC0bNkSFy9erO66iYiIiOg5FTpiZ2BggDt37sDMzEztM0tLS3Tv3h3du3dHYGAgIiMjcffuXbRt27bKiyUiIiKi8smEf59TpSqVm5sLY2Nj5OTkwMjIqKbLIap5u2Q1XQG97UbyVxZJQ3VlhErdPFHqwYMHiI2NRUlJCdq2bYu6detWVV1EREREpKFKB7sDBw5g3LhxaNy4MYqKinDjxg1s3LgRvr6+VVkfEREREVVQhW9fffLkicr7BQsW4MKFC7hw4QLi4uKwf/9+zJ07t8oLJCIiIqKKqXCwa926NX788Ufxvba2NjIzM8X3GRkZ0NXVrdrqiIiIiKjCKnwq9siRI5gyZQq2b9+OjRs3Yt26dRg2bBhKSkpQXFwMuVyO7du3V2OpRERERPQiFQ52dnZ2+Pnnn7F792506dIF06ZNw61bt3Dr1i2UlJTAyckJCoWiOmslIiIiohfQ+BERI0aMwMWLF3HlyhV07doVSqUSrq6uDHVERERENUyju2J/+eUXXL9+HS4uLvjmm2/w22+/YdSoUfD09MTChQuhr69fXXUSERER0UtU+IjdZ599Bl9fX1y8eBGTJk3CokWL0KVLF1y+fBkKhQKtWrXCr7/+Wp21EhEREdELVPjJE2ZmZjh69Chat26NrKwstG/fHjdv3hQ/v3btGiZNmoTTp09XW7FvGz55gug5fPIEvSo+eYIkoroyQoWP2BkYGCAxMREAcPfuXbVr6po1a1bpULdx40bY2dlBoVDAzc0NFy5cKLdvUVERFi5cCAcHBygUCri4uCAyMlKlT0lJCebPnw97e3vo6+vDwcEBixYtwr8zrCAICAgIQN26daGvrw93d3ckJCSojJOVlYVRo0bByMgIJiYmGDdunNp8fkRERERvigoHu+DgYHh7e8PGxgZdunTBokWLqqSAvXv3wt/fH4GBgbh8+TJcXFzg4eGhMkfev82bNw9btmzBhg0bcO3aNXz88ccYOHAg4uLixD7Lly/Hpk2bEBISguvXr2P58uVYsWIFNmzYIPZZsWIF1q9fj82bNyM2NhYGBgbw8PBAfn6+2GfUqFH4888/ERUVhcOHD+PUqVOYOHFilWw3ERERUVWr8KlYAHj48CHu3LkDR0dHmJiYVEkBbm5uaNu2LUJCQgAASqUS9erVw9SpUzF79my1/jY2Npg7dy6mTJkitg0ePBj6+vrYuXMnAKBv376wsrLCtm3byuwjCAJsbGzw2WefYebMmQCAnJwcWFlZYfv27Rg+fDiuX7+OZs2a4eLFi2jTpg0AIDIyEl5eXrh37x5sbGxeum08FUv0HJ6KpVfFU7EkETV+KhZ4dp1d27ZtqyzUFRYW4tKlS3B3d/+/guRyuLu7IyYmpsxlCgoK1E4D6+vr48yZM+L79957D9HR0eI1gFeuXMGZM2fg6ekJAEhMTER6errKeo2NjeHm5iauNyYmBiYmJmKoAwB3d3fI5XLExsa+4pYTERERVb0KBbuPP/4Y9+7dq9CAe/fuRVhYWIX6PnjwACUlJbCyslJpt7KyQnp6epnLeHh4YM2aNUhISIBSqURUVBTCw8ORlpYm9pk9ezaGDx8OJycn6OjooFWrVpg+fTpGjRoFAOLYL1pveno6LC0tVT7X1tZGnTp1yq2toKAAubm5Ki8iIiKi16VC89hZWFjA2dkZHTt2RL9+/dCmTRvY2NhAoVDg0aNHuHbtGs6cOYM9e/bAxsYGX3/9dbUVvG7dOkyYMAFOTk6QyWRwcHCAr68vvv32W7HPvn37EBYWhl27dsHZ2Rnx8fGYPn06bGxs4OPjU221BQcHY8GCBdU2PhEREdGLVOiI3aJFi3Dz5k107NgRX331Fdq3b4/69evD0tISTZo0gbe3N+7cuYOvv/4a58+fR8uWLSu0cnNzc2hpaSEjI0OlPSMjA9bW1mUuY2FhgYiICOTl5SE5ORl//fUXDA0N0bBhQ7HP559/Lh61a9GiBUaPHo0ZM2YgODgYAMSxX7Rea2trtRs4iouLkZWVVW5tc+bMQU5Ojvi6e/duhfYDERERUVWo8DV2VlZWmDt3Lq5evYoHDx7g8uXLOHv2LG7cuIFHjx7hhx9+QO/evTVaua6uLlq3bo3o6GixTalUIjo6Gh06dHjhsgqFAra2tiguLsaBAwfQv39/8bOnT59CLlfdNC0tLSiVSgCAvb09rK2tVdabm5uL2NhYcb0dOnRAdnY2Ll26JPY5fvw4lEol3NzcyqxJT08PRkZGKi8iIiKi10WjR4qVMjU1hampaZUU4O/vDx8fH7Rp0wbt2rXD2rVrkZeXB19fXwCAt7c3bG1txaNtsbGxSE1NhaurK1JTUxEUFASlUolZs2aJY/br1w9LlixB/fr14ezsjLi4OKxZswZjx44FAMhkMkyfPh2LFy+Go6Mj7O3tMX/+fNjY2GDAgAEAgKZNm6J3796YMGECNm/ejKKiIvj5+WH48OEVuiOWiIiI6HWrVLCrSsOGDcP9+/cREBCA9PR0uLq6IjIyUryxISUlReXoW35+PubNm4c7d+7A0NAQXl5e2LFjh8qduhs2bMD8+fMxefJkZGZmwsbGBpMmTUJAQIDYZ9asWcjLy8PEiRORnZ2NTp06ITIyUuWO27CwMPj5+aFHjx6Qy+UYPHgw1q9fX/07hYiIiKgSNJrHjjTDeeyInsN57OhVcR47kog3Yh47IiIiInpzMdgRERERSYTGwS4wMBDJycnVUQsRERERvQKNg92PP/4IBwcH9OjRA7t27UJBQUF11EVEREREGtI42MXHx+PixYtwdnbGp59+Cmtra3zyySe4ePFiddRHRERERBVUqWvsWrVqhfXr1+Pvv//Gtm3bcO/ePXTs2BEtW7bEunXrkJOTU9V1EhEREdFLvNLNE4IgoKioCIWFhRAEAaampggJCUG9evWwd+/eqqqRiIiIiCqgUsHu0qVL8PPzQ926dTFjxgy0atUK169fx2+//YaEhAQsWbIE06ZNq+paiYiIiOgFNA52LVq0QPv27ZGYmIht27bh7t27WLZsGRo1aiT2GTFiBO7fv1+lhRIRERHRi2n8SLGhQ4di7NixsLW1LbePubk5lErlKxVGRERERJrRONjNnz+/OuogIiIiolek8anYwYMHY/ny5WrtK1aswJAhQ6qkKCIiIiLSnMbB7tSpU/Dy8lJr9/T0xKlTp6qkKCIiIiLSnMbB7smTJ9DV1VVr19HRQW5ubpUURURERESaq9RdsWXNUbdnzx40a9asSooiIiIiIs1V6uaJQYMG4fbt2+jevTsAIDo6Grt378b+/furvEAiIiIiqhiNg12/fv0QERGBpUuX4ocffoC+vj5atmyJY8eOoUuXLtVRIxERERFVgMbBDgD69OmDPn36VHUtRERERPQKXulZsURERET05tD4iF1JSQm+/PJL7Nu3DykpKSgsLFT5PCsrq8qKIyIiIqKK0/iI3YIFC7BmzRoMGzYMOTk58Pf3x6BBgyCXyxEUFFQNJRIRERFRRWgc7MLCwrB161Z89tln0NbWxogRI/DNN98gICAA58+fr44aiYiIiKgCNA526enpaNGiBQDA0NAQOTk5AIC+ffvi559/rtrqiIiIiKjCNA5277zzDtLS0gAADg4OOHr0KADg4sWL0NPTq9rqiIiIiKjCNA52AwcORHR0NABg6tSpmD9/PhwdHeHt7Y2xY8dWeYFEREREVDEyQRCEVxng/PnzOHfuHBwdHdGvX7+qqksScnNzYWxsjJycHBgZGdV0OUQ1b5espiugt93IV/qVRfTGqK6MoNF0J0VFRZg0aRLmz58Pe3t7AED79u3Rvn37KiuIiIiIiCpHo1OxOjo6OHDgQHXVQkRERESvQONr7AYMGICIiIhqKIWIiIiIXoXGT55wdHTEwoULcfbsWbRu3RoGBgYqn0+bNq3KiiMiIiKiitP45onSa+vKHEwmw507d165KKngzRNEz+HNE/SqePMEScQbcfMEACQmJlbZyomIiIio6mh8jR0RERERvZk0PmL3skmIv/3220oXQ0RERESVp3Gwe/Tokcr7oqIi/PHHH8jOzkb37t2rrDAiIiIi0ozGwe7gwYNqbUqlEp988gkcHByqpCgiIiIi0lyVXGMnl8vh7++PL7/8siqGIyIiIqJKqLKbJ27fvo3i4uKqGo6IiIiINKTxqVh/f3+V94IgIC0tDT///DN8fHyqrDAiIiIi0ozGwS4uLk7lvVwuh4WFBVavXv3SO2aJiIiIqPpoHOxOnDhRHXUQERER0SvS+Bq7xMREJCQkqLUnJCQgKSmpKmoiIiIiokrQONiNGTMG586dU2uPjY3FmDFjqqImIiIiIqoEjYNdXFwcOnbsqNbevn17xMfHV0VNRERERFQJGgc7mUyGx48fq7Xn5OSgpKSkSooiIiIiIs1pHOw6d+6M4OBglRBXUlKC4OBgdOrUqUqLIyIiIqKK0/iu2OXLl6Nz585o0qQJ3n//fQDA6dOnkZubi+PHj1d5gURERERUMRofsWvWrBn++9//YujQocjMzMTjx4/h7e2Nv/76C82bN6+OGomIiIioAjQ+YgcANjY2WLp0aVXXQkRERESvQOMjdqGhodi/f79a+/79+/Hdd99VSVFEREREpDmNg11wcDDMzc3V2i0tLXkUj4iIiKgGaRzsUlJSYG9vr9beoEEDpKSkVElRRERERKQ5jYOdpaUl/vvf/6q1X7lyBWZmZlVSFBERERFpTuNgN2LECEybNg0nTpxASUkJSkpKcPz4cXz66acYPnx4ddRIRERERBWg8V2xixYtQlJSEnr06AFt7WeLK5VKeHt7Y8mSJVVeIBERERFVjEwQBKEyCyYkJCA+Ph76+vpo0aIFGjRoUNW1vfVyc3NhbGyMnJwcGBkZ1XQ5RDVvl6ymK6C33chK/coieuNUV0ao1Dx2AODo6AhHR0exuE2bNmHbtm34/fffq6w4IiIiIqq4Sgc7ADhx4gS+/fZbhIeHw9jYGAMHDqyquoiIiIhIQxoHu9TUVGzfvh2hoaHIzs7Go0ePsGvXLgwdOhQyGU+zEBEREdWUCt8Ve+DAAXh5eaFJkyaIj4/H6tWr8ffff0Mul6NFixYMdUREREQ1rMJH7IYNG4YvvvgCe/fuRe3atauzJiIiIiKqhAofsRs3bhw2btyI3r17Y/PmzXj06FF11kVEREREGqpwsNuyZQvS0tIwceJE7N69G3Xr1kX//v0hCAKUSmV11khEREREFaDRkyf09fXh4+OD3377DVevXoWzszOsrKzQsWNHjBw5EuHh4dVVJxERERG9hMaPFCvl6OiIpUuX4u7du9i5cyeePn2KESNGVGVtRERERKSBV5rHDgDkcjn69euHfv36ITMzsypqIiIiIqJKqPQRu7JYWlpW5XBEREREpIEqDXZEREREVHMY7IiIiIgkgsGOiIiISCI0DnYNGzbEw4cP1dqzs7PRsGHDShWxceNG2NnZQaFQwM3NDRcuXCi3b1FRERYuXAgHBwcoFAq4uLggMjJSpY+dnR1kMpnaa8qUKWKf27dvY+DAgbCwsICRkRGGDh2KjIyMl46zbNmySm0jERERUXXTONglJSWhpKRErb2goACpqakaF7B37174+/sjMDAQly9fhouLCzw8PMq9w3bevHnYsmULNmzYgGvXruHjjz/GwIEDERcXJ/a5ePEi0tLSxFdUVBQAYMiQIQCAvLw89OrVCzKZDMePH8fZs2dRWFiIfv36qU22vHDhQpWxpk6dqvE2EhEREb0OFZ7u5KeffhL/feTIERgbG4vvS0pKEB0dDTs7O40LWLNmDSZMmABfX18AwObNm/Hzzz/j22+/xezZs9X679ixA3PnzoWXlxcA4JNPPsGxY8ewevVq7Ny5EwBgYWGhssyyZcvg4OCALl26AADOnj2LpKQkxMXFwcjICADw3XffwdTUFMePH4e7u7u4bO3atWFtba3xdhERERG9bhUOdgMGDAAAyGQy+Pj4qHymo6MDOzs7rF69WqOVFxYW4tKlS5gzZ47YJpfL4e7ujpiYmDKXKSgogEKhUGnT19fHmTNnyl3Hzp074e/vD5lMJo4hk8mgp6cn9lMoFJDL5Thz5oxKsFu2bBkWLVqE+vXrY+TIkZgxYwa0tV95+j8iIiKiKlfhhFJ6itLe3h4XL16Eubn5K6/8wYMHKCkpgZWVlUq7lZUV/vrrrzKX8fDwwJo1a9C5c2c4ODggOjoa4eHhZZ4eBoCIiAhkZ2djzJgxYlv79u1hYGCAL774AkuXLoUgCJg9ezZKSkqQlpYm9ps2bRreffdd1KlTB+fOncOcOXOQlpaGNWvWlLmugoICFBQUiO9zc3MruiuIiIiIXpnG19glJiaqhbrs7Oyqquel1q1bB0dHRzg5OUFXVxd+fn7w9fWFXF72pmzbtg2enp6wsbER2ywsLLB//34cOnQIhoaGMDY2RnZ2Nt59912Vcfz9/dG1a1e0bNkSH3/8MVavXo0NGzaohLd/Cw4OhrGxsfiqV69e1W48ERER0QtoHOyWL1+OvXv3iu+HDBmCOnXqwNbWFleuXNFoLHNzc2hpaandjZqRkVHudW0WFhaIiIhAXl4ekpOT8ddff8HQ0LDMO3KTk5Nx7NgxjB8/Xu2zXr164fbt28jMzMSDBw+wY8cOpKamvvDOXjc3NxQXFyMpKanMz+fMmYOcnBzxdffu3RdsPREREVHV0jjYbd68WTwSFRUVhWPHjiEyMhKenp74/PPPNRpLV1cXrVu3RnR0tNimVCoRHR2NDh06vHBZhUIBW1tbFBcX48CBA+jfv79an9DQUFhaWqJPnz7ljmNubg4TExMcP34cmZmZ+OCDD8rtGx8fD7lcXu6j0/T09GBkZKTyIiIiInpdNL4LID09XQx2hw8fxtChQ9GrVy/Y2dnBzc1N4wL8/f3h4+ODNm3aoF27dli7di3y8vLEu2S9vb1ha2uL4OBgAEBsbCxSU1Ph6uqK1NRUBAUFQalUYtasWSrjKpVKhIaGwsfHp8ybHUJDQ9G0aVNYWFggJiYGn376KWbMmIEmTZoAAGJiYhAbG4tu3bqhdu3aiImJwYwZM/DRRx/B1NRU4+0kIiIiqm4aBztTU1PcvXsX9erVQ2RkJBYvXgwAEASh3BsYXmTYsGG4f/8+AgICkJ6eDldXV0RGRoo3VKSkpKhc95afn4958+bhzp07MDQ0hJeXF3bs2AETExOVcY8dO4aUlBSMHTu2zPXeuHEDc+bMQVZWFuzs7DB37lzMmDFD/FxPTw979uxBUFAQCgoKYG9vjxkzZsDf31/jbSQiIiJ6HWSCIAiaLODn54fDhw/D0dERcXFxSEpKgqGhIfbs2YMVK1bg8uXL1VXrWyc3NxfGxsbIycnhaVkiANglq+kK6G03UqNfWURvrOrKCBofsfvyyy9hZ2eHu3fvYsWKFTA0NAQApKWlYfLkyVVWGBERERFpRuMjdlRxPGJH9BwesaNXxSN2JBHVlRE0visWePZYr06dOsHGxgbJyckAgLVr1+LHH3+sssKIiIiISDMaB7tNmzbB398fnp6eyM7OFm+YMDExwdq1a6u6PiIiIiKqII2D3YYNG7B161bMnTsXWlpaYnubNm1w9erVKi2OiIiIiCquUo8Ua9WqlVq7np4e8vLyqqQoIiIiItKcxsHO3t4e8fHxau2RkZFo2rRpVdRERERERJVQ4elOFi5ciJkzZ8Lf3x9TpkxBfn4+BEHAhQsXsHv3bgQHB+Obb76pzlqJiIiI6AUqPN2JlpYW0tLSYGlpibCwMAQFBeH27dsAABsbGyxYsADjxo2r1mLfNpzuhOg5nO6EXhWnOyGJqPEJiv+d/0aNGoVRo0bh6dOnePLkCSwtLausICIiIiKqHI2ePCGTqf61XatWLdSqVatKCyIiIiKiytEo2DVu3Fgt3D0vKyvrlQoiIiIiosrRKNgtWLAAxsbG1VULEREREb0CjYLd8OHDeT0dERER0RuqwvPYvewULBERERHVrAoHuwrOikJERERENaTCp2KVSmV11kFEREREr0jjR4oRERER0ZuJwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCSCwY6IiIhIIhjsiIiIiCTijQh2GzduhJ2dHRQKBdzc3HDhwoVy+xYVFWHhwoVwcHCAQqGAi4sLIiMjVfrY2dlBJpOpvaZMmSL2uX37NgYOHAgLCwsYGRlh6NChyMjIUBknKysLo0aNgpGREUxMTDBu3Dg8efKkajeeiIiIqIrUeLDbu3cv/P39ERgYiMuXL8PFxQUeHh7IzMwss/+8efOwZcsWbNiwAdeuXcPHH3+MgQMHIi4uTuxz8eJFpKWlia+oqCgAwJAhQwAAeXl56NWrF2QyGY4fP46zZ8+isLAQ/fr1g1KpFMcZNWoU/vzzT0RFReHw4cM4deoUJk6cWI17g4iIiKjyZIIgCDVZgJubG9q2bYuQkBAAgFKpRL169TB16lTMnj1brb+NjQ3mzp2rcvRt8ODB0NfXx86dO8tcx/Tp03H48GEkJCRAJpPh6NGj8PT0xKNHj2BkZAQAyMnJgampKY4ePQp3d3dcv34dzZo1w8WLF9GmTRsAQGRkJLy8vHDv3j3Y2Ni8dNtyc3NhbGyMnJwccT1E/9N2yWq6AnrbjazRX1lEVaa6MkKNHrErLCzEpUuX4O7uLrbJ5XK4u7sjJiamzGUKCgqgUChU2vT19XHmzJly17Fz506MHTsWMplMHEMmk0FPT0/sp1AoIJfLxXFiYmJgYmIihjoAcHd3h1wuR2xsbLm15ebmqryIiIiIXpcaDXYPHjxASUkJrKysVNqtrKyQnp5e5jIeHh5Ys2YNEhISoFQqERUVhfDwcKSlpZXZPyIiAtnZ2RgzZozY1r59exgYGOCLL77A06dPkZeXh5kzZ6KkpEQcJz09HZaWlipjaWtro06dOuXWFhwcDGNjY/FVr169iu4KIiIioldW49fYaWrdunVwdHSEk5MTdHV14efnB19fX8jlZW/Ktm3b4OnpqXLq1MLCAvv378ehQ4dgaGgIY2NjZGdn49133y13nIqYM2cOcnJyxNfdu3crPRYRERGRprRrcuXm5ubQ0tJSuxs1IyMD1tbWZS5jYWGBiIgI5Ofn4+HDh7CxscHs2bPRsGFDtb7Jyck4duwYwsPD1T7r1asXbt++jQcPHkBbWxsmJiawtrYWx7G2tla7gaO4uBhZWVnl1qanp6dyepeIiIjodarRI3a6urpo3bo1oqOjxTalUono6Gh06NDhhcsqFArY2tqiuLgYBw4cQP/+/dX6hIaGwtLSEn369Cl3HHNzc5iYmOD48ePIzMzEBx98AADo0KEDsrOzcenSJbHv8ePHoVQq4ebmpummEhEREVW7Gj1iBwD+/v7w8fFBmzZt0K5dO6xduxZ5eXnw9fUFAHh7e8PW1hbBwcEAgNjYWKSmpsLV1RWpqakICgqCUqnErFmzVMZVKpUIDQ2Fj48PtLXVNzM0NBRNmzaFhYUFYmJi8Omnn2LGjBlo0qQJAKBp06bo3bs3JkyYgM2bN6OoqAh+fn4YPnx4he6IJSIiInrdajzYDRs2DPfv30dAQADS09Ph6uqKyMhI8YaKlJQUleve8vPzMW/ePNy5cweGhobw8vLCjh07YGJiojLusWPHkJKSgrFjx5a53hs3bmDOnDnIysqCnZ0d5s6dixkzZqj0CQsLg5+fH3r06AG5XI7Bgwdj/fr1VbsDiIiIiKpIjc9jJ2Wcx47oOZzHjl4V57EjiZDkPHZEREREVHUY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCIY7Oi12bhxI+zs7KBQKODm5oYLFy6U27eoqAgLFy6Eg4MDFAoFXFxcEBkZqdLHzs4OMplM7TVlyhSxT3p6OkaPHg1ra2sYGBjg3XffxYEDB8TPT548WeYYMpkMFy9erPqdQEREVI0Y7Oi12Lt3L/z9/REYGIjLly/DxcUFHh4eyMzMLLP/vHnzsGXLFmzYsAHXrl3Dxx9/jIEDByIuLk7sc/HiRaSlpYmvqKgoAMCQIUPEPt7e3rhx4wZ++uknXL16FYMGDcLQoUPFcd577z2VMdLS0jB+/HjY29ujTZs21bhHiIiIqp5MEAShpouQqtzcXBgbGyMnJwdGRkY1XU6NcnNzQ9u2bRESEgIAUCqVqFevHqZOnYrZs2er9bexscHcuXNVjr4NHjwY+vr62LlzZ5nrmD59Og4fPoyEhATIZDIAgKGhITZt2oTRo0eL/czMzLB8+XKMHz9ebYyioiLY2tpi6tSpmD9//ittM5Vhl6ymK6C33Uj+yiJpqK6MwCN2VO0KCwtx6dIluLu7i21yuRzu7u6IiYkpc5mCggIoFAqVNn19fZw5c6bcdezcuRNjx44VQx3w7Ijc3r17kZWVBaVSiT179iA/Px9du3Ytc5yffvoJDx8+hK+vr4ZbSUREVPMY7KjaPXjwACUlJbCyslJpt7KyQnp6epnLeHh4YM2aNUhISIBSqURUVBTCw8ORlpZWZv+IiAhkZ2djzJgxKu379u1DUVERzMzMoKenh0mTJuHgwYNo1KhRmeNs27YNHh4eeOeddzTfUCIiohrGYEdvpHXr1sHR0RFOTk7Q1dWFn58ffH19IZeX/SW7bds2eHp6wsbGRqV9/vz5yM7OxrFjx/D777/D398fQ4cOxdWrV9XGuHfvHo4cOYJx48ZVyzYRERFVN+2aLoCkz9zcHFpaWsjIyFBpz8jIgLW1dZnLWFhYICIiAvn5+Xj48CFsbGwwe/ZsNGzYUK1vcnIyjh07hvDwcJX227dvIyQkBH/88QecnZ0BAC4uLjh9+jQ2btyIzZs3q/QPDQ2FmZkZPvjgg1fZXCIiohrDI3ZU7XR1ddG6dWtER0eLbUqlEtHR0ejQocMLl1UoFLC1tUVxcTEOHDiA/v37q/UJDQ2FpaUl+vTpo9L+9OlTAFA7yqelpQWlUqnSJggCQkND4e3tDR0dHY22j4iI6E3BI3b0Wvj7+8PHxwdt2rRBu3btsHbtWuTl5Yk3KXh7e8PW1hbBwcEAgNjYWKSmpsLV1RWpqakICgqCUqnErFmzVMZVKpUIDQ2Fj48PtLVVv5ydnJzQqFEjTJo0CatWrYKZmRkiIiIQFRWFw4cPq/Q9fvw4EhMTy7xTloiI6G3BYEevxbBhw3D//n0EBAQgPT0drq6uiIyMFG+oSElJUTmylp+fj3nz5uHOnTswNDSEl5cXduzYARMTE5Vxjx07hpSUFIwdO1ZtnTo6Ovjll18we/Zs9OvXD0+ePEGjRo3w3XffwcvLS6Xvtm3b8N5778HJyanqN56IiOg14Tx21Yjz2BE9h/PY0aviPHYkEZzHjoiIiIheiMGOiIiISCIY7IiIiIgkgsGOiIiISCIY7IiIiIgkgsGOiIiISCI4j90bblncg5ougd5is1uZ13QJRET0GvGIHREREZFEMNgRERERSQSDHREREZFEMNgRERERSQSDHREREZFEMNgRERERSQSDHREREZFEMNgRERERSQSDHREREZFEMNgRERERSQSDHREREZFEMNgRERERSQSDHREREZFEMNgRERERSQSDHREREZFEMNgRERERSQSDHREREZFEMNgRERERSQSDHREREZFEMNgRERERSQSDHREREZFEMNgRERERSQSDHREREZFEMNgRERERSQSDHREREZFEMNgRERERSQSDHREREZFEvBHBbuPGjbCzs4NCoYCbmxsuXLhQbt+ioiIsXLgQDg4OUCgUcHFxQWRkpEofOzs7yGQytdeUKVPEPunp6Rg9ejSsra1hYGCAd999FwcOHHjpOMuWLavajSciIiKqIjUe7Pbu3Qt/f38EBgbi8uXLcHFxgYeHBzIzM8vsP2/ePGzZsgUbNmzAtWvX8PHHH2PgwIGIi4sT+1y8eBFpaWniKyoqCgAwZMgQsY+3tzdu3LiBn376CVevXsWgQYMwdOhQlXEAYOHChSpjTZ06tRr2AhEREdGrq/Fgt2bNGkyYMAG+vr5o1qwZNm/ejFq1auHbb78ts/+OHTvwn//8B15eXmjYsCE++eQTeHl5YfXq1WIfCwsLWFtbi6/Dhw/DwcEBXbp0EfucO3cOU6dORbt27dCwYUPMmzcPJiYmuHTpksr6ateurTKWgYFB9ewIIiIiolekXZMrLywsxKVLlzBnzhyxTS6Xw93dHTExMWUuU1BQAIVCodKmr6+PM2fOlLuOnTt3wt/fHzKZTGx/7733sHfvXvTp0wcmJibYt28f8vPz0bVrV5Xlly1bhkWLFqF+/foYOXIkZsyYAW3tsndbQUEBCgoKxPc5OTkAgNzc3PJ3wkvkP3lc6WWJcnN1a7oEVU9rugB6673Cz1OiN0lpNhAEoWoHFmpQamqqAEA4d+6cSvvnn38utGvXrsxlRowYITRr1ky4efOmUFJSIhw9elTQ19cXdHV1y+y/d+9eQUtLS0hNTVVpf/TokdCrVy8BgKCtrS0YGRkJR44cUemzevVq4cSJE8KVK1eETZs2CSYmJsKMGTPK3Z7AwEABAF988cUXX3zxxVeFXrdv365IZKqwGj1iVxnr1q3DhAkT4OTkBJlMBgcHB/j6+pZ76nbbtm3w9PSEjY2NSvv8+fORnZ2NY8eOwdzcHBERERg6dChOnz6NFi1aAAD8/f3F/i1btoSuri4mTZqE4OBg6Onpqa1rzpw5KssolUpkZWXBzMxM5Wgh0euQm5uLevXq4e7duzAyMqrpcoiI6F9ycnJQv3591KlTp0rHrdFgZ25uDi0tLWRkZKi0Z2RkwNrausxlLCwsEBERgfz8fDx8+BA2NjaYPXs2GjZsqNY3OTkZx44dQ3h4uEr77du3ERISgj/++APOzs4AABcXF5w+fRobN27E5s2by1y3m5sbiouLkZSUhCZNmqh9rqenpxb4TExMyt1+otfByMiIwY6I6A0ll1ft7Q41evOErq4uWrdujejoaLFNqVQiOjoaHTp0eOGyCoUCtra2KC4uxoEDB9C/f3+1PqGhobC0tESfPn1U2p8+fXahz/M7U0tLC0qlstx1xsfHQy6Xw9LS8qXbRkRERPS61fipWH9/f/j4+KBNmzZo164d1q5di7y8PPj6+gJ4Ni2Jra0tgoODAQCxsbFITU2Fq6srUlNTERQUBKVSiVmzZqmMq1QqERoaCh8fH7WbHZycnNCoUSNMmjQJq1atgpmZGSIiIhAVFYXDhw8DAGJiYhAbG4tu3bqhdu3aiImJwYwZM/DRRx/B1NT0NewZIiIiIs3UeLAbNmwY7t+/j4CAAKSnp8PV1RWRkZGwsrICAKSkpKgcWcvPz8e8efNw584dGBoawsvLCzt27FA75Xns2DGkpKRg7NixauvU0dHBL7/8gtmzZ6Nfv3548uQJGjVqhO+++w5eXl4Anp1W3bNnD4KCglBQUAB7e3vMmDFD5Ro6ojeZnp4eAgMDy7welIiIalZ1/YyWCUJV32dLRERERDWhxicoJiIiIqKqwWBHREREJBEMdkREREQSwWBHREREJBEMdkRvsY0bN8LOzg4KhQJubm64cOFCuX23bt2K999/H6ampjA1NYW7u/sL+xMRUeWdOnUK/fr1g42NDWQyGSIiIl7YPzw8HD179oSFhQWMjIzQoUMHHDlyROP1MtgRvaX27t0Lf39/BAYG4vLly3BxcYGHhwcyMzPL7H/y5EmMGDECJ06cQExMDOrVq4devXohNTX1NVdORCR9eXl5cHFxwcaNGyvU/9SpU+jZsyd++eUXXLp0Cd26dUO/fv0QFxen0Xo53QnRW8rNzQ1t27ZFSEgIgGeTcterVw9Tp07F7NmzX7p8SUkJTE1NERISAm9v7+oul4jof5ZMJsPBgwcxYMAAjZZzdnbGsGHDEBAQUOFleMSO6C1UWFiIS5cuwd3dXWyTy+Vwd3dHTExMhcZ4+vQpioqKqvwB1ERE9OqUSiUeP36s8c9oBjuit9CDBw9QUlIiPqGllJWVFdLT0ys0xhdffAEbGxuVcEhERG+GVatW4cmTJxg6dKhGy9X4I8WI6PVbtmwZ9uzZg5MnT0KhUNR0OURE9C+7du3CggUL8OOPP8LS0lKjZRnsiN5C5ubm0NLSQkZGhkp7RkYGrK2tX7jsqlWrsGzZMhw7dgwtW7aszjKJiEhDe/bswfjx47F///5KnVHhqViit5Curi5at26N6OhosU2pVCI6OhodOnQod7kVK1Zg0aJFiIyMRJs2bV5HqUREVEG7d++Gr68vdu/ejT59+lRqDB6xI3pL+fv7w8fHB23atEG7du2wdu1a5OXlwdfXFwDg7e0NW1tbBAcHAwCWL1+OgIAA7Nq1C3Z2duK1eIaGhjA0NKyx7SAikqInT57g1q1b4vvExETEx8ejTp06qF+/PubMmYPU1FR8//33AJ6dfvXx8cG6devg5uYm/ozW19eHsbFxhdfL6U6I3mIhISFYuXIl0tPT4erqivXr18PNzQ0A0LVrV9jZ2WH79u0AADs7OyQnJ6uNERgYiKCgoNdYNRGR9J08eRLdunVTa/fx8cH27dsxZswYJCUl4eTJkwCe/cz+7bffyu1fUQx2RERERBLBa+yIiIiIJILBjoiIiEgiGOyIiIiIJILBjoiIiEgiGOyIiIiIJILBjoiIiEgiGOyIiIiIJILBjoiIiEgiGOyIiIiIJILBjoiIiEgiGOyIiIiIJILBjoiIiEgi/h+5B2DXfoN8TgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Example accuracies\n",
    "original_accuracy = 0.9787  # Original test accuracy for 20% training data size\n",
    "increased_accuracy = 0.9809  # Test accuracy with increased training iterations for 20% data size\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1accd9f7",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.8.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}