PyTorch_28_NeuralNetwork_ActivationFunctionsMNIST.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "PyTorch_28_NeuralNetwork_ActivationFunctionsMNIST.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyNNm0UsLKxIbLkjZIQjsCtE",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/codeplinth/pytorchbootcamp/blob/master/PyTorch_28_NeuralNetwork_ActivationFunctionsMNIST.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-ZIdPvTMPs_C",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import torch\n",
        "from torch import nn,optim\n",
        "from torch.utils.data import Dataset,DataLoader\n",
        "import torchvision\n",
        "from torchvision import datasets,transforms\n",
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline"
      ],
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fLX_AZmFRGem",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class Net_Sigmoid(nn.Module):\n",
        "  def __init__(self,D_in,H,D_out):\n",
        "    super(Net_Sigmoid,self).__init__()\n",
        "    self.l1 = nn.Linear(D_in,H)\n",
        "    self.l2 = nn.Linear(H,D_out)\n",
        "    self.sigmoid = nn.Sigmoid()\n",
        "\n",
        "  def forward(self,x):\n",
        "    out1 = self.sigmoid(self.l1(x))\n",
        "    z_pred = self.l2(out1)\n",
        "    return z_pred\n"
      ],
      "execution_count": 25,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FoqLx2HiSJFC",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class Net_Tanh(nn.Module):\n",
        "  def __init__(self,D_in,H,D_out):\n",
        "    super(Net_Tanh,self).__init__()\n",
        "    self.l1 = nn.Linear(D_in,H)\n",
        "    self.l2 = nn.Linear(H,D_out)\n",
        "    self.tanh = nn.Tanh()\n",
        "\n",
        "  def forward(self,x):\n",
        "    out1 = self.tanh(self.l1(x))\n",
        "    z_pred = self.l2(out1)\n",
        "    return z_pred"
      ],
      "execution_count": 26,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "idIbJ6I2SUEG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class Net_ReLU(nn.Module):\n",
        "  def __init__(self,D_in,H,D_out):\n",
        "    super(Net_ReLU,self).__init__()\n",
        "    self.l1 = nn.Linear(D_in,H)\n",
        "    self.l2 = nn.Linear(H,D_out)\n",
        "    self.relu = nn.ReLU()\n",
        "\n",
        "  def forward(self,x):\n",
        "    out1 = self.relu(self.l1(x))\n",
        "    z_pred = self.l2(out1)\n",
        "    return z_pred"
      ],
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GgI-L6DFSbsW",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def train_model(model,optimizer,criterion,train_loader,validation_loader,EPOCHS=100):\n",
        "  metrics = {'training_loss':[],'validation_accuracy':[]}\n",
        "  for epoch in range(EPOCHS):\n",
        "    for x,y in train_loader:\n",
        "      z = model(x.view(-1,1*28*28))\n",
        "      loss = criterion(z,y)\n",
        "      optimizer.zero_grad()\n",
        "      loss.backward()\n",
        "      optimizer.step()\n",
        "      metrics['training_loss'].append(loss.item())\n",
        "\n",
        "    correct = 0\n",
        "    for x_val,y_val in validation_loader:\n",
        "      z = model(x_val.view(-1,1*28*28))\n",
        "      _,y_pred = torch.max(z,1)\n",
        "      correct += (y_pred == y_val).sum().item()\n",
        "    accuracy = ( 100 * correct ) / len(validation_dataset)\n",
        "    metrics['validation_accuracy'].append(accuracy)\n",
        "  return metrics"
      ],
      "execution_count": 32,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "znbzIPW_YvTQ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "input_dim = 1*28*28\n",
        "hidden_dim = 100\n",
        "output_dim = 10\n",
        "learning_rate = 0.01 \n",
        "criterion = nn.CrossEntropyLoss()"
      ],
      "execution_count": 33,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c3eRGCZTP1Ha",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "train_dataset = datasets.MNIST(root = '../mnist_data',train = True,download=True,transform = transforms.ToTensor())\n",
        "validation_dataset = datasets.MNIST(root = '../mnist_data',train = False,download=True,transform = transforms.ToTensor())\n",
        "train_loader = DataLoader(dataset = train_dataset,shuffle=True,batch_size =2000)\n",
        "validation_loader = DataLoader(dataset = validation_dataset,shuffle = False,batch_size = 5000 )"
      ],
      "execution_count": 34,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zR8n4Qb3ZDS0",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model_sigmoid = Net_Sigmoid(input_dim,hidden_dim,output_dim)\n",
        "optimizer = optim.SGD(model_sigmoid.parameters(),learning_rate)\n",
        "training_result_sigmoid = train_model(model_sigmoid,optimizer,criterion,train_loader,validation_loader,EPOCHS=10)"
      ],
      "execution_count": 36,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9r61Krg2Zo8K",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model_tanh = Net_Tanh(input_dim,hidden_dim,output_dim)\n",
        "optimizer = optim.SGD(model_tanh.parameters(),learning_rate)\n",
        "training_result_tanh = train_model(model_tanh,optimizer,criterion,train_loader,validation_loader,EPOCHS=10)"
      ],
      "execution_count": 38,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "l5jb4pqoZpny",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model_relu = Net_ReLU(input_dim,hidden_dim,output_dim)\n",
        "optimizer = optim.SGD(model_relu.parameters(),learning_rate)\n",
        "training_result_relu = train_model(model_relu,optimizer,criterion,train_loader,validation_loader,EPOCHS=10)"
      ],
      "execution_count": 40,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fLLQThq2dGEF",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "outputId": "ffe5a469-578f-4dd2-b1fb-e77add978a08"
      },
      "source": [
        "plt.plot(training_result_sigmoid['training_loss'],label = 'Sigmoid')\n",
        "plt.plot(training_result_tanh['training_loss'],label = 'Tanh')\n",
        "plt.plot(training_result_relu['training_loss'],label = 'ReLU')\n",
        "plt.xlabel('Iterations')\n",
        "plt.ylabel('LOSS')\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HAQuyiRPdlx9",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "outputId": "4acc580d-0ea2-43e3-8ba8-5c6393315781"
      },
      "source": [
        "plt.plot(training_result_sigmoid['validation_accuracy'],label = 'Sigmoid')\n",
        "plt.plot(training_result_tanh['validation_accuracy'],label = 'Tanh')\n",
        "plt.plot(training_result_relu['validation_accuracy'],label = 'ReLU')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('ACCURACY')\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX4AAAEGCAYAAABiq/5QAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeXhU1f3H8ffJHrKH7CSQyL4mgRBUVEAUEZFFBbdWqP7EorVVqwXbWq2ioqV1aRUEVMAV6wJKFVEUN5CQkAXCvgQIZCV7yD7n98cdQoIhBMjkJpnv63nmmZk7M3e+GeUzZ8499xyltUYIIYT9cDC7ACGEEG1Lgl8IIeyMBL8QQtgZCX4hhLAzEvxCCGFnnMwuoCUCAgJ0ZGSk2WUIIUSHkpSUlK+1Djx9e4cI/sjISBITE80uQwghOhSl1KGmtktXjxBC2BkJfiGEsDM2DX6l1INKqXSl1Hal1HtKKTelVJRSarNSap9SaqVSysWWNQghhGjMZsGvlOoG/B6I01oPAhyBW4DngBe01r2AQuAuW9UghBDil2zd1eMEuCulnIAuQBZwJfCh9fHlwBQb1yCEEKIBmwW/1voosAA4jBH4xUASUKS1rrU+LRPo1tTrlVKzlFKJSqnEvLw8W5UphBB2x5ZdPX7AZCAKCAM8gPEtfb3WerHWOk5rHRcY+IthqEIIIc6TLcfxXwUc1FrnASilPgZGAr5KKSdrqz8cOGrDGoQQov2rq4ETx6E8D8rzjcuJfOP+pfeDu1+rvp0tg/8wcLFSqgtQAYwFEoFvgZuA94EZwGob1iCEEG3PUgcnCozgPhng5SeD/eS2k5c8qCxqej/KEQZP6zjBr7XerJT6ENgK1ALJwGLgf8D7Sql51m2v26oGIYRoFRaLEc4ng7qp8G7YSj9RADS1yJWCLl3BIwA8AiF4oHHtEXBqW5eAU9vcfMGh9XvkVUdYgSsuLk7LlA1CiPNSVwNVpVBdbr2UGZeqMut962MN71eVWbteTgb5cdB1Te/f3a9BWHe1Xp8M8NMCvYs/ODi22Z+ulErSWsedvr1DzNUjhLATWkPNCWsIn7yUt+B+aYNQP+3xuuoWvrkCF09w8QBXT6Nl7hcJ4XENWuWB1hb7yXD3B0dnW34iNiHBL4RoHZY6I4CrSqCy5LTr4mYeO+26yS6SJji6ngpplwYXr5BTAX5ym2sL7ju526RbpT2S4BdCQF1tg4A+GcINg7r4zEFdVWrcri49+/s4OIObN7h6n7r2j2pw38u4uHiAi1cTwd7gfgdsabcXEvxCdBZaG90cFYXGpbLIervozNsqi4zbVSVn37+TW+PAdvUCr2Bw9fllmJ9+ffK2kxsoZfvPQjRLgl+I9qau5rRgPkt4N9xmqT3zfh2cjQOR7n7g7gveYRA04NR9N58zhLd1u5PMp9hZSPALYUtaG10h9Sfm5J66XZZ7asRIRdGp8K4ua36frj5GULv7GqHtHdY40N39jGGAp29z7iKtbQFI8Atx7hqdZZkHZXmnbje6WIcC1lY2vR8331OjRHzCIWTw2cPb1Rsc5Z+tuDDyf5AQWhut7LLcBifj5DY+WadhuFcUNL0fR5fGw/4C+5+67REInoGNx3hL14lohtaa45XH8XP1w7GVx/5L8IvOr7Yaio9AYUbjS/GRlrfKPQIhqB94XN4g3INOPeYRYPSFS1eKOEc1dTUcLj1MRnEGB0sOcrD4oHG7+CClNaV8PvVzIrwjWvU9JfhFx6e1EeCFGVB0CAoPWsP9kHFdchS05dTzHV3Brwf4RDRulXsGNW6hS6tctKLCykIj1EuMUD95O7M0k7oGZwUHuQcR5RPFhIsmEOUThYeLR6vXIsEvOoaaCig63KDFfqhx672mvPHzPUOMsy57jDSu6y89jMfs5EQd0bZqLDUcLT1qBHvJqZb7wZKDFFcV1z/PxcGFHj496OPXh2siryHKJ4oo7yh6ePfA08XT5nVK8Iv2wWKBspzGYV7UINxLsxo/39nDCHG/SLhoVONw9+0Ozu5tWr6wL8VVxY1a7ye7aY6UHKFWnxpS29WtK1E+UYzrMY5I70iifKKI9IkkzCOs1fvtz4UEv2hbFUVweBMUHGjcei86dFo/uzJGuvhFQs+xp7XaI40uGelPFzZUZ6njWNmx+n73ht0zBZWnDvA7OTjRw6sHF/lcxNjuY41w944k0icSbxdvE/+CM5PgF7alNeTugL3rYO9XcPjnU7McunobIR7YF/pcc6oF7xdlhL6Tq5mVCztSZ6njQPEB0vLSSMtPY1v+NjKKM6ix1NQ/x8/VjyifKMZEjGnUeu/m2Q0nh44VpR2rWtExVJXBwe9h75dG2JdYF1kLGQKXPQi9xkJgP2NcurTahQkKKgvYlreN1LxU0vLT2J6/nXLrcSJfV18GBQzism6XEeUdVd+C93XzNbnq1iPBLy6c1nB8v7VVvw4O/WRMheviBT3HwOhHoddV4B1qdqXCDtVYathTuMdozeelkZqXypHSIwA4Kkf6+PVh4kUTiQ6MJjowmgivCFQnb5BI8IvzU1MJGT+eCvvCg8b2wH4w4h7oPQ4iLpbhkKLN5ZTnkJafVh/06cfTqaqrAiDQPZDowGim9ZnGkMAhDOg6AHcn+xsIIMEvWq7o8Km++gPfQW2FMYd51BVwyX1G2Pv1MLtKYUeq6qrYeXyn0WVjbc3nnMgBwNnBmQFdBzC97/T61nxwl+BO35pvCQl+cWZ1NcYInJNhn7fL2O4XCUPvMII+cqQMnRRtQmtNZllmfUs+LS+NXYW7qLXOSNrNsxtDg4cSHRjNkIAh9PXvi4uj/OJsigS/aKw02wj5vetg/7fG4hoOzkbAnwz7rr3koKywufKactLz00nLT6tv0Z8cRunu5M6ggEHMGDCD6MBoBgcOJsA9wOSKOw4JfntnqYPMxFN99dlpxnbvbjD4RiPoo64wFt0Qwka01hwuPUxybnJ9yO8r2ofFOtVGlE8Ul3e7nCGBQ4gOjKanb88ON4SyPbHZJ6eU6gusbLDpIuBvwArr9kggA5iutS60VR2iCeXHYf96I+j3fW3MAa8cIWIEjH3cCPvggdKqFzZTY6lhd8FutuZsJTk3ma25W+tb814uXgwJGMLY7mOJDoxmUMAgfFx9TK64c7FZ8GutdwMxAEopR+Ao8AkwF1ivtZ6vlJprvT/HVnUIq9oq+Hkh7FpjtPDRxiRkfcYbQd9zjDGuXggbKK8pJzUvleTcZJJzkknLT6OitgKAcM9wRoaNJDY4ltjAWC7yvQgHJXMp2VJb/VYaC+zXWh9SSk0GRlu3Lwc2IMFvW7k74aP/g5ztEDYURs+F3ldDaKxMViZsIvdELltzt5Kck0xybjK7C3dj0RYclAN9/fpyQ+8biA2KJTYolqAuQWaXa3faKvhvAd6z3g7WWp+ccSsbCG7qBUqpWcAsgO7du9u8wE5Ja9iyFNb9FVw84daV0He82VWJTsaiLRwsPlgf9Ftzt3K0zDhb293JnSEBQ5g1ZBaxQbFEB0bj4dz60wyLc6O01rZ9A6VcgGPAQK11jlKqSGvt2+DxQq11s30McXFxOjEx0aZ1djplubD6PqMfv9fVMOVVY755IS5QdV016cfT6/vnk3OTKakuAYzZKIcGDyU2KJahQUPp498HZwdnkyu2X0qpJK113Onb26LFfy2wVWudY72fo5QK1VpnKaVCgdw2qMG+7PkSVt1rLPJ97T8g/m45UCvOW3FVMal5qfVBvz1/O9WWagAivSO5qsdV9UFvD9MddAZtEfy3cqqbB+BTYAYw33q9ug1qsA81FbDuMdiyBIIHwcw1ENTf7KpEB6K15lj5sUat+X1F+wBwUk4MCBjAbf1vIyYohtigWPzd/E2uWJwPmwa/UsoDuBq4p8Hm+cAHSqm7gEPAdFvWYDey0owDuPm74eL7YOzfwNnN7KpEO6e15mDxQTZnb2Zrzla25m4l94TxI9zT2ZPooGiujbqW2KBYBgUMsst5bTojmwa/1roc6HratuMYo3xEa7BY4OdXYf3fwd0ffvWxMe2xEE3QWpNZmklCdgKbszezJXsL+RX5AAR3CWZY0DBig41um16+vUxdJUrYjpz61pGVHINVs+HABuh7HUz6N3h0PevLhH3JLs8mITuBhKwEErITyCo3BtUFuAcQHxJvXELjCfcMl/55OyHB31Ht/Aw+vd84Mev6l2DoDDmAKwA4XnGcLdlb6lv0h0oOAeDj6kN8SDx3DrqT+JB4onyiJOjtlAR/R1NVBl8+CltXQGgM3LgUAnqbXZUwUXFVMYk5ifUt+pMHYz2dPRkWPIzpfaYzInQEvf16yxmxApDg71iOJsFHdxsLlV/2IIz+syx0YofKa8pJykkyWvVZm9lVsAuNxs3RjdigWK676DpGhIygf9f+MpGZaJL8X9ERWOrgpxfh22fAMxhmfAZRl5tdlWgjlbWVpOSl1Lfot+dvp07X4ezgTHRgNLOjZxMfGs/ggMEy/7xoEQn+9q7oCHxyj7GO7cCpMPEFmUytk6upq2Fb/jY2Z28mISuB1LxUaiw1OCpHBgYM5M5BdzI8ZDgxQTEyvFKcFwn+9mzbh7DmIdB1MGURRN8iB3A7oVpLLTuP7zRG3mQnkJybTEVtBQpFP/9+3NbvNuJD4xkaNBRPF0+zyxWdgAR/e1RZAp8/AmnvQ/hwuGEx+F9kdlWiFVm0hc1Zm1m1bxXfZ35PWU0ZAL18ezGl1xRGhIwgLiRO5qEXNiHB394c3gwf3w3FR2DUXLjiEXCU/0ydRWZpJqv3r2b1vtVklWfh5eLFuMhxXBx6McNDhsvygaJNSKK0F3W18P0/4PvnwSccfrMWuo8wuyrRCk7UnODrw1+zat8qtmRvQaG4JOwSHhz2IFd2vxJXR1ezSxR2RoK/PSg4aLTyM7fAkFtgwj/AzdvsqsQF0FqTmpfKqn2rWJuxlvKaciK8Irg/9n4m9ZxEiEeI2SUKOybBbyatIfV9+PxhY83bG1+HwTeZXZW4ALkncvls/2es2reKjJIM3J3cGddjHFN6TWFY8DA5U1a0CxL8ZqkohDUPQvon0GMkTF0EvrLSWEdUU1fDhswNfLL3E3469hMWbWFo0FDuHHQn4yLHyYpTot2R4DfDwR+MsfllOcb0ySMfAJkFscPZXbCbT/Z9wv8O/I+iqiKC3IO4c9CdTO45mUifSLPLE+KMJPjbUm01bHgGfnzRGJ551zroNszsqsQ5KKos4n8H/8fqfavZWbATZwdnxkSMYWrvqVwSeolMYyw6BAn+tpK/11goJSsFht4B1zwLrnIyTkdQZ6lj47GNrNq3im+PfEuNpYb+/v15NP5RJkRNwNfN9+w7EaIdkeBvC6krYc0D4OQK09+CAZPMrki0QEZxBqv3r+bTfZ+SW5GLr6svN/e9mSm9ptDXv6/Z5Qlx3iT4ba08Hz79ndGlc9Mb4B1mdkWiGeU15azLWMcn+z4hOTcZB+XAZd0uY26vuYwOH42zo7PZJQpxwST4bW3rcqirNhZLkdBvl7TWJOYksmrfKr469BUVtRVEekfywNAHuL7n9QR1CTK7RCFalQS/LdXVwpY34KLREChdA+1Ndnk2q/etZvX+1RwpPYKHswcToiYwpdcUogOjZcy96LRsGvxKKV9gKTAI0MCdwG5gJRAJZADTtdaFtqzDNLs/h5JMmPC82ZWIBvYU7uG11Nf46tBXaDTDQ4YzO3o2Y7uPpYtzF7PLE8LmbN3ifwlYq7W+SSnlAnQB/gys11rPV0rNBeYCc2xchzkSFoNPd+gz3uxKBMa4+9fSjMD3cPbgzkF3cmOfG4nwijC7NCHalM2CXynlA1wBzATQWlcD1UqpycBo69OWAxvojMGfswMyfoCr/i4nZ5lsV8EuFqUuYv3h9Xg6ezJryCzuGHCHTHks7JYtW/xRQB7wplIqGkgC/gAEa62zrM/JBoKberFSahYwC6B79w44lcGWJeDkZozZF6bYeXwni1IX8c2Rb/By9uK30b/lV/1/JYEv7J4tg98JGArcr7XerJR6CaNbp57WWiuldFMv1lovBhYDxMXFNfmcdquiyJh8bfBN0MXf7GrsTvrxdBalLmLDkQ14OXtxb/S93D7gdrxdZMZTIcC2wZ8JZGqtN1vvf4gR/DlKqVCtdZZSKhTItWEN5kh5F2pOwPC7za7ErqTnp7MwdSHfZX6Hl4sX98bcy+39JfCFOJ3Ngl9rna2UOqKU6qu13g2MBXZYLzOA+dbr1baqwRQWi9HNEzECwmLMrsYubMvbxsLUhfxw9Ae8Xbz5XczvuK3/bXi5eJldmhDtkq1H9dwPvGMd0XMA+A3gAHyglLoLOARMt3ENbWv/eig4AGP+YnYlnV5aXhoLUxfy49Ef8XH14fexv+fWfrfKguRCnIVNg19rnQLENfHQWFu+r6kSFoNnMPSX+XhsJSU3hUWpi/jp2E/4uvryh6F/4NZ+t8q890K0kJy525qO74e9X8GoOeDkYnY1nU5ybjILUxayKWsTfq5+PDjsQW7pe4ucdCXEOZLgb01bXjfG7Mf9xuxKOpWknCQWpi5kc9Zm/N38eWjYQ9zc92YJfCHOkwR/a6kuh+S3YcBk8JKFtFtDYnYii1IXsTnbCPyH4x5mWp9pEvhCXCAJ/taSthKqiiH+HrMr6fC2ZG9hYepCtmRvIcA9gEfiHmFa32m4O7mbXZoQnYIEf2vQGhKWQMgQiIg3u5oOSWvNluwtvJr6Kkk5SQS6BzJn+Bxu6nMTbk5uZpcnRKciwd8aMn6E3B0w6T8gU/meE601m7M3szBlIVtztxLkHsTc+Lnc2PtGCXwhbESCvzUkLAZ3P2OKBtEiWms2ZW1iUeoiknOTCeoSxKPxj3JjnxtxdXQ1uzwhOjUJ/gtVnAm7/geX/g6cpQ/6bLTWbDq2iVdTXyU1L5XgLsH8ZcRfuKH3Dbg4yhBYIdqCBP+FSnwD0BB3l9mVtHvlNeXM+3keaw6sIcQjhMcufowpvaZI4AvRxiT4L0RNJSQtgz7Xgl8Ps6tp13YV7OLh7x7mSOkRZkfP5u7Bd8vC5UKYRIL/QqR/AieOQ7zMwnkmWmve2/UeCxIX4Ofqx9JxSxkeMtzssoSwaxL8FyJhMQT0MRZTF79QXFXM4xsfZ/3h9Vze7XLmXTYPfzdZn0AIs0nwn6/MJDi2FSYskCGcTUjJTeFP3/+JvIo8Ho57mF8P+DUOysHssoQQGFMkN0kptVApJStYnEnCa+DiBdG3mF1Ju2LRFpZuW8rMtTNxUA6sGL+CGQNnSOgL0Y4096/xAJCklLqtrYrpMMpyjf79mNvAVRb7OCm/Ip/ZX8/mpa0vMbb7WP57/X8ZHDjY7LKEEKc5Y1eP1vofSql3gX9ZF01ZCFgaPP5xG9TXPiUth7pqOajbwM9ZP/PoD49SWl3K3y75Gzf1vgklXWBCtEvN9vFrrY8qpf4HPA1cz6ng14B9Bn9djTF2v+eVENDb7GpMV2up5dWUV1m6bSlRPlG8dvVr9PHrY3ZZQohmnDH4lVIDMVr5x4B4rXVWm1XVnu36H5Qeg4n/MrsS02WXZzPn+zlszd3K1F5TmRs/V6ZMFqIDaK7F/yHwB631urYqpkNIWAy+3aH3OLMrMdWGIxv4609/paauhmcvf5aJF000uyQhRAs1F/z3AL9YxFQpNQHI0Von2ayq9ip7Oxz6Ca5+ylhpyw5V11XzQtILvL3zbfr79+cfo/5BD285a1mIjqS54H8CaGoNwXTgTeDKs+1cKZUBlAJ1QK3WOk4p5Q+sBCKBDGC61rrwXIo2zZYl4OQGsb8yuxJTHC45zCPfP8KO4zu4rd9t/DHujzLPjhAdUHPDOb201odO32jdFnAO7zFGax2jtY6z3p8LrNda9wbWW++3fxWFkPYBDJ4GXezv7NMvDn7B9DXTySzN5MUxL/LoiEcl9IXooJpr8fs189iFHMGbDIy23l4ObADmXMD+2kbyO1BzAuJnmV1Jm6qoreC5hOf4aO9HxATG8PwVzxPqGWp2WUKIC9Bci/9rpdTTqsFgbGV4EvimhfvXwDqlVJJS6mRiBjcYIZQNBJ9z1W3NUmd083S/BEKHmF1Nm9lXuI/b/ncbH+/9mLsH380b49+Q0BeiE2iuxf9HYCmwTymVYt0WA2wBWnrm0mXWcwGCgK+UUrsaPqi11kop3dQLrV8UswC6d+/ewrezkX1fQ2EGjH3c3DraiNaaj/d+zPyE+XRx7sKiqxdxadilZpclhGglzZ25Ww7cqpS6CBho3ZyutT6glGrRROpa66PW61yl1CdAPJCjlArVWmcppUKB3DO8djGwGCAuLq7JL4c2k7AYPEOg//WmltEWyqrLeHLTk3yR8QUXh17Ms5c/S4D7uRzSEUK0d2edOUtrfUBr/RmwBohSSr0OZJ7tdUopD6WU18nbwDhgO/ApMMP6tBnA6vOsvW3k7zNa/HF3QidfOCQ9P53pa6az7tA6fh/7e167+jUJfSE6obNOy6yUuhi4DZgC+AP3AQ+3YN/BwCfWQwROwLta67VKqS3AB9b5fw4B08+z9raxZSk4OMOwmWZXYjNaa97e+Tb/SvoXAe4BvDn+TWKDYs0uSwhhI81N2fAMMA04DLwH/B1I1Fovb8mOtdYHgOgmth8Hxp5XtW2tqgxS3oGBU8Cr/R+DPh9FlUU89tNjbMjcwOiI0cwbOQ8fVx+zyxJC2FBzLf7/A/ZgzNfzmda66kwHYjuttPehqgTi7zG7EptIyklizvdzKKgsYG78XG7rd5vMqCmEHWgu+EOBq4FbgReVUt8C7kopJ611bZtUZyatIWEJhMZAeNzZn9+B1FnqWLptKa+mvkq4ZzhvT3ibAV0HmF2WEKKNNDeqpw5YC6xVSrkCEwF34KhSar3WunMv0HLwe8jbBZNf7VRLK+adyOPRHx9lc9ZmJkRN4G+X/A0P519MySSE6MRatOau1roK+Aj4yDpS5wGbVtUeJCwGd38YdKPZlbSajUc38uiPj3Ki5gRPXvokU3pNka4dIexQc2vuOiqlblVKPayUGmTdNhH4ErihrQo0RdFh2P05DJsBzm5mV9MqPtv/GbPXz8bfzZ/3J77P1N5TJfSFsFPNtfhfByKABOBlpdQxIA6Yq7Ve1RbFmSbxDeM67k5z62glq/et5rGfHiM+NJ5/X/lv3J3czS5JCGGi5oI/DhiitbYopdww5tXpaR2O2XnVVBpr6vadYCy40sF9svcTHt/4OBeHXszLV76Mm1Pn+AUjhDh/zZ25W621tgBorSuBA50+9AG2fwQVBZ1iFs6P937M4xsf59KwSyX0hRD1mmvx91NKpVlvK6Cn9b7CmF+t801TqTUkvAaB/SDqCrOruSAf7vmQv2/6OyO7jeSlMS/h6uhqdklCiHaiueDv32ZVtBeZiZCVCtf9s0MP4fxg9wc89fNTXN7tcl4Y84KEvhCikebG8f9i9a1OL+E1cPWGIbeYXcl5W7lrJfM2z2NU+Cj+NfpfskqWEOIXmhvOWaqUKmlwKVZK7VdKLVVKdW3LIttEaQ6kr4KY28HV0+xqzst7u95j3uZ5jI4YLaEvhDijMwa/1tpLa+3d4OKDMdInHVjUZhW2laRlYKmB4f9ndiXn5Z2d7/DM5mcYEzGGf42S0BdCnNlZ5+NvSGtdqLV+Aehpo3rMUVdjjN3vORYCepldzTl7a8dbzE+Yz9juY/nnqH/i3MnXDRBCXJhzCn4A6+pbLZrqocPY+RmUZcOIjjcL54r0FTy/5Xmu7nE1/xj1Dwl9IcRZNTcff1PTMvgBNwMf2qwiMyQsBr9I6HWV2ZWck+Xpy1mQuIBxPcYx/4r5ODtI6Ashzq65lvvpC8xq4Djwktb6f7YrqY1lpcHhTTDuaXBwNLuaFntj+xu8kPQC4yPH8+zlz+Lk0Ll+hAkhbKe54Zy/actCTJOwGJzcIfZ2sytpsaXblvLS1pe4NvJanrn8GQl9IcQ5aW445z+UUr/o9FZK3aOUmm/bstrIiQLY9l8YMh3c/cyupkWWpC3hpa0vMSFqgoS+EOK8NHdw90pgcRPbl2AsytLxJb8NtZUdZl6e11Jf4+Xkl5l40USeuUxCXwhxfpoLflet9S/W2LVO3NZx5zM4yVIHW5ZAj5EQMsjsas5qYcpC/pPyHyb1nMS8kfNw7EDHI4QQ7UtzwV+hlOp9+kbrtoqWvoF1QZdkpdQa6/0opdRmpdQ+pdRKpZQ5ZxrtXWcsuBJ/tylv31Jaa15JeYVXU19lcs/JPHnpkxL6QogL0lzw/w34Qik1Uyk12Hr5DfA/62Mt9QdgZ4P7zwEvaK17AYXAXedadKvY/Bp4hUG/9ttrdTL0F6UuYmqvqTw5UkJfCHHhmpuy4QtgCjAGWGa9jAZu1Fp/3pKdK6XCgeuApdb7CuPYwcnzAJZb36Nt5e2BA98aK2y10xOetNb8O/nfvJb2Gjf2vpEnLn0CB3XO59sJIcQvNHcClxuQo7Wecdr2QKWUm3VxlrN5EfgT4GW93xUo0lrXWu9nAt3O8P6zgFkA3bu38kpYW5aCowsMm9m6+20lWmteTn6ZpduWclOfm3js4sck9IUQraa5NHkZuLyJ7ZcBL5xtx9aF2XO11knnU5jWerHWOk5rHRcYGHg+u2haVSmkvAsDp4JnK+63lWiteWHrCyzdtpTpfaZL6AshWl1z4wGHaa1/Mc5Ra/2JUmpeC/Y9EpiklJoAuAHewEuAr1LKydrqDweOnkfd5y/1fagubZdDOLXW/CvpXyxLX8bNfW/mLyP+gurAC8IIIdqn5pqSXc7zdQBorR/VWodrrSOBW4BvtNa3A98CN1mfNgNY3cJaL5zWxpm6YUMhPK7N3rYltNYsSFzAsvRl3NrvVgl9IYTNNBfguUqp+NM3WrflXcB7zgEeUkrtw+jzf/0C9nVuDmyA/D3trrWvteb5Lc+zYscKbu9/O4/GPyqhL4Swmea6eh4BPlBKLQNO9tPHAXdgtOBbTGu9AdhgvX0A+MUXSptIWAJdAoz+/XZCa81zW57jnZ3v8Kv+v+JPw/8koS+EsKnmhnMmACMwztKdidEtg/X6DptX1toKD8GeL2DYDHB2M7sawIgu4soAAB/qSURBVAj9ZzY/wzs73+GOAXdI6Ash2kSzk71orXOAx5VSQ4FbMUL/CuCjNqitdSW+Dihj7H47YNEWntn8DCt3r2TmwJk8NOwhCX0hRCO5pZUEebV+Q7W5cfx9MML+ViAfWAkorfWYVq/C1moqYOsK6Hcd+ISbXQ0WbeHpn5/mgz0f8JtBv+HBoQ9K6AshANifV8ba7dmsS88mNbOYH/40hgj/5sbanLvmWvy7gB+AiVrrfQBKqQdb9d3byrYPoaKwXRzUtWgLT256ko/2fsRdg+7iD0P/IKEvhB3TWrP9aAlfpmezNj2bfbllAERH+PKn8X3p4tL607Q0F/w3YBzE/VYptRZ4n444K6fWkPAaBA2AyMtMLcWiLfx909/5eO/H3D34bu6PvV9CXwg7VGfRJGYUsDY9m3XpORwtqsDRQREf6c+vL+7BuIHBhPq42+z9m1uBaxWwSinlAUwGHgCClFILgU+01utsVlVrOrIZsrfBxBfAxJC1aAuPb3ycVftWcc+Qe7gv5j4JfSHsSFVtHRv3HefL9Gy+2pHD8fJqXJwcuKJ3AH+4qjdX9Q/G36NtJis+60oeWuty4F3gXaWUHzANYyx+xwj+hMXg6gODp5tWQp2ljsc3Ps7q/auZHT2be2PuNa0WIUTbKa+qZcPuPNamZ/PtrlzKqmrxdHViTL8gxg8MYVTfQDxd235BpXN6R611IcaqXE2tzNX+lGbDjtUQfw+4eppSgtaaJzY9wer9q7k3+l5mx8w2pQ4hRNsoKK/m6505rEvP5vu9+VTXWujq4cLEIaFcMzCES3t1xdXJ3OnVO/fafYlvgqUWhpsz5T/A2zvfru/ekdAXonPKKq5gXXoOa7dnk5BRQJ1F083XnV+N6ME1A4OJi/TH0aH9dO127uAv2A+9x0HXnqa8fVJOEv9M/CdXRlzJfTH3mVKDEMI29ueV8WV6Nl+m55B6pAiA3kGezB7Vk/GDQhgY5t1uj+N17uC/cSnUVpny1rkncvnjhj8S4RXBvMvmtdv/AYQQLaO1Jv1YCWu3Z/NlejZ7Tw67DPfhT+P7cs3AEHoGmtOlfK46d/ADOLm2+VvW1NXwxw1/5ETtCZaOW4qXi9fZXySEaHdODrv8Mj2HL9OzOVpUgYOCEVFduX1Ed8YNDCHM13bDLm2l8we/CRYkLiAlL4V/XPEPevn1MrscIcQ5KKuqJeHgcdal5zQadnl5r7YfdmkrEvytbM2BNby7611+PeDXjI8ab3Y5QoizyC+rYsvBAhIyCtiSUcCOYyVYNPXDLq8ZGMzovkGmDLu0lc7zl7QDuwt28/eNf2dY8DAeHNYxZ7cQojPTWnOkoMII+YNG0B/ILwfAzdmBmAhffjemF8Oj/ImP8jd92KWtSPC3kpLqEh7c8CDeLt4sGLUAZwdns0sSwu5ZLJrdOaVsySggwRr0OSXGgA8fd2fievgxfXgE8VH+DArzwcXJPta3luBvBRZt4c8//JmssizeHP8mAe4BZpckhF2qrrWw7WgRCQcL2ZJRQGJGASWVtQCEeLsRH9WV+Eg/hkf50yfIC4d2NLa+LUnwt4IlaUv4LvM7Ho1/lJigGLPLEcJulFXVsvVQYX2LPuVIEVW1FgB6BnowYXAowyONbptwP3cZVm0lwX+Bfjz6I6+kvMLEiyZya79bzS5HiE4tv6yKxIyC+hb9jqwS6iwaBwUDw3y4fUQP4qP8iIv0J8Cz7YdydxQS/BcgszSTOd/Pobdfb/52yd+kNSFEK9Jak1lYUd83n5BRwIE840Csq5NxIPbe0T0ZHunP0B5+nWrUja3Z7JNSSrkB3wOu1vf5UGv9uFIqCmNu/64Yi7j/Wmtdbas6bKWytpKHNjyE1poXR7+Iu1PHO4lDiPZEa83+vDI27T9OQkYhWw4WkF1SCYC3mxNxkf5MGxZBfJQfg7r5dNoRN23Bll+RVcCVWusypZQz8KNS6gvgIeAFrfX7SqlFwF3AQhvW0eq01sz7eR47C3bynyv/Q4R3hNklCdEhaa1JyyxmbXo2X27Prh9aGeztWt83PzzSn77B9nsg1hZsFvxaaw2UWe86Wy8auBK4zbp9OfAEHSz4/7vnv6zev5rfRv+WURGjzC5HiA6l4epTX27P5lhxJY4Oikt7duU3l0UxqncgEf5yINaWbNopppRyxOjO6QW8AuwHirTWtdanZALdzvDaWcAsgO7du9uyzHOSlpfGswnPMrLbSH475LdmlyNEh1Bda2Hj/ny+tC41eGr1qUAeGteXq/oH4dulY0+D0JHYNPi11nVAjFLKF/gE6HcOr61f8CUuLk7bpsJzc7ziOA9teIjgLsE8d/lzODpIH6MQZ1JRXcd3e/JYuz2L9btyKa2sxcPFkSv7BzN+YAij+wbiIQdkTdEmn7rWukgp9S1wCeCrlHKytvrDgaNtUcOFqrXU8qfv/0RRVRFvXfsWPq4+ZpckRLtTXFHDt7tyWbs9mw17cqmsseDbxZnxA0MYPyiEkb0CcHOWBpPZbDmqJxCosYa+O3A18BzwLXATxsieGcBqW9XQml5OfpmE7ATmjZxH/679zS5HiHYjv6yKr3YYq09t3J9PTZ0m2NuV6XERjB8YQnyUP06O9jEVQkdhyxZ/KLDc2s/vAHygtV6jlNoBvK+UmgckA6/bsIZW8dWhr3hz+5tM7zOdyb0mm12OEKY7VlTBl+nZrN2ezZaMAiwauvt34c6RUVwzKISYcF8ZhdOO2XJUTxoQ28T2A0C8rd63tR0oPsBff/wrQwKGMCd+jtnlCGGaA3ll9SNxUjOLAegb7MXvruzN+IEh9A/1kpE4HYQcWWlGeU05D377IG5Obvxz9D9xcZRRB8J+aK3ZmVXK2vRs1m7PYk/OqaUG54zvxzUDg7mogyw1KBqT4D8DrTWP/fQYGSUZLLl6CSEeIWaXJITNWSya5CNF9d04hwtO4KBgeKQ/j18/gHEDQ+jWAZcaFI1J8J/Bih0r+OrQVzw07CHiQztMz5QQ50xrTdKhQj5NPcaX6dnklFTh7KgY2SuAe0f35KoBwTLhWScjwd+ELdlbeCHpBa7ucTUzB840uxwhbKKypo7VKUdZtvEQO7NKcHN2YHSfIMYPCmFMvyB83GUxoc5Kgv802eXZPPzdw3T37s5TI5+Sg1Wi0zlScIK3Nx9i5ZYjFJ2ooV+IF8/eMJjJMWF0cZFIsAfyX7mB6rpq/vjdH6msreTNa97Ew9nD7JKEaBVaazbuP86yjRms35mDUoprBgYz45JI4qP8210Dp6amhszMTCorK80upUNwc3MjPDwcZ+eW/UqT4G/g+S3Pk5aXxj9H/ZOLfC8yuxwhLlh5VS0fJx9lxcYM9uaW4e/hwuzRPbl9RA/C2vFB2szMTLy8vIiMjGx3X0rtjdaa48ePk5mZSVRUVIteI8Fv9en+T1m5eyW/GfgbxkWOM7scIS5IRn45KzYd4r9JRyitrGVwNx8WTItm4pDQDjFlQmVlpYR+Cyml6Nq1K3l5eS1+jQQ/sKtgF09uepL4kHh+P/T3ZpcjxHmxWDTf7c1j+cYMNuzOw9lRMWFwKDMujSQ2wrfDhWhHq9dM5/pZ2X3wF1cV88C3D+Dj6sPzVzyPk4PdfySigymprOHDxEze+vkQB/PLCfRy5YGrenNbfHeCvN3MLk+0Q3adchZtYe4Pc8k5kcOy8cvo6t7V7JKEaLG9OaUs35TBx1uPcqK6jqHdfXnglhiuHRSKi5NMitYann76ad59910cHR1xcHDgtddeY8mSJTz00EMMGDDAZu87YcIE3n33XXx9fRttf+KJJ/D09OThhx++oP3bdfAvSl3Ej0d/5LGLHyM6MNrscoQ4qzqLZv3OHJZvyuCnfcdxcXJgUnQYMy6JZHC4TBXemjZt2sSaNWvYunUrrq6u5OfnU11dzdKlS23+3p9//rlN92+3wf995vcsTF3IpJ6TmNZnmtnlCNGsohPVrNxyhLd+PkRmYQWhPm48ck1fbhkeQddOflbt3z9LZ8exklbd54Awbx6/fmCzz8nKyiIgIABXV+PzDQgIAGD06NEsWLCAuLg4Xn/9dZ577jl8fX2Jjo7G1dWV//znP8ycORN3d3eSk5PJzc3ljTfeYMWKFWzatIkRI0awbNkyAN577z2eeeYZtNZcd911PPfccwBERkaSmJhIQEAATz/9NMuXLycoKIiIiAiGDRt2wX+/XQb/kZIjzP1hLv38+/HYxY/JQSTRbu04VsLyjRmsSjlKVa2FEVH+/GVCf64eECxz3NvYuHHjePLJJ+nTpw9XXXUVN998M6NGnVpj+9ixYzz11FNs3boVLy8vrrzySqKjT/UcFBYWsmnTJj799FMmTZrETz/9xNKlSxk+fDgpKSkEBQUxZ84ckpKS8PPzY9y4caxatYopU6bU7yMpKYn333+flJQUamtrGTp0qAT/+aioreDBDQ+iULww+gXcnOTgl2hfauosrEvPYfnGDBIyCnBzduCGod2445JI+od6m11emztby9xWPD09SUpK4ocffuDbb7/l5ptvZv78+fWPJyQkMGrUKPz9/QGYNm0ae/bsqX/8+uuvRynF4MGDCQ4OZvDgwQAMHDiQjIwMDh06xOjRowkMDATg9ttv5/vvv28U/D/88ANTp06lS5cuAEyaNKlV/ja7Cn6tNU9teoo9hXt4ZewrhHuFm12SEPXyy6p4P+Ewb/98mOySSsL93PnzhH5Mj4uQhchN4ujoyOjRoxk9ejSDBw9m+fLlLX7tyS4iBweH+tsn79fW1rb4LFtbsKvfiit3r+SzA58xO2Y2l4dfbnY5QgCQeqSIh1amcOmz37Bg3R56B3uy9I44vntkDLOu6Cmhb5Ldu3ezd+/e+vspKSn06NGj/v7w4cP57rvvKCwspLa2lo8++uic9h8fH893331Hfn4+dXV1vPfee426kgCuuOIKVq1aRUVFBaWlpXz22WcX9kdZ2U2LPyU3hee2PMcV4Vdwz5B7zC5H2Lmq2jo+35bF8o2HSDlShIeLI7fER3DHJZH0CpLFTdqDsrIy7r//foqKinBycqJXr14sXryYm266CYBu3brx5z//mfj4ePz9/enXrx8+Pi0fWRUaGsr8+fMZM2ZM/cHdyZMbL+06dOhQbr75ZqKjowkKCmL48OGt8rcprXWr7MiW4uLidGJi4nm/Pr8in5s/uxkXRxfen/g+Pq4y7E2YI7u4knc2H+K9hMPkl1UTFeDBHZf04KZh4Xi5yTTIJ+3cuZP+/fubXcZZlZWV4enpSW1tLVOnTuXOO+9k6tSpptTS1GemlErSWsed/txO3+KvtdTyyHePUFJdwtsT3pbQF21Oa82WjEKWb8xgbXo2Fq25sm8QMy6N5LJeAbIoeQf2xBNP8PXXX1NZWcm4ceMaHZhtzzp98L+Y9CKJOYk8c9kz9PXva3Y5wo5UVBsLnSzfZCx04u3mxJ0jI/n1xZF079rF7PJEK1iwYIHZJZwXmwW/UioCWAEEAxpYrLV+SSnlD6wEIoEMYLrWutAWNazNWMvyHcu5td+tXN/zelu8hRC/cPj4qYVOiiuMhU7m3zCYyTHdcHdp/zNjis7Pli3+WuCPWuutSikvIEkp9RUwE1ivtZ6vlJoLzAXmtPaba635bP9nxATG8EjcI629eyEasVg0P+7LZ/nGDL7ZnYuDUowfGMKMSyMZHuknJwmKdsVmwa+1zgKyrLdLlVI7gW7AZGC09WnLgQ3YIPiVUrw45kXKq8txdpSDZsI2Sitr+CgpkxWbDnEgv5wATxfuH9OL20b0IMRHTg4U7VOb9PErpSKBWGAzEGz9UgDIxugKauo1s4BZAN27dz+v93V2cMbXzffsTxTiHO3LLWXFpkN8lJRJeXUdMRG+vHhzDNcODsHVSbpzRPtm8+BXSnkCHwEPaK1LGv7k1VprpVST40m11ouBxWAM57R1nUKczcmZMVdsOsSP+/JxcXRgYnQoMy6JJDpCGhidyfHjxxk7diwA2dnZODo61k+tkJCQgIvL2U+q27BhAwsWLGDNmjU2rfV82DT4lVLOGKH/jtb6Y+vmHKVUqNY6SykVCuTasgYhLlRheTUrE4/w1qZDHC2yr5kx7VXXrl1JSUkBWm8O/PbElqN6FPA6sFNr/a8GD30KzADmW69X26oGIS7E9qPFrNiUweqUY1TVWrj4In8em9ifq/rLzJht6ou5kL2tdfcZMhiunX/25zWwZMkSFi9eTHV1Nb169eKtt96iS5cuzJw5E29vbxITE8nOzub555+vP7u3rKyMm266ie3btzNs2DDefvvtdnGg35Yt/pHAr4FtSqkU67Y/YwT+B0qpu4BDwHQb1iDEOamutbA2PZsVGzNIPFSIu7MjNw4LZ8YlkfQN8TK7PGGiG264gbvvvhuAv/71r7z++uvcf//9gDF3/48//siuXbuYNGlSffAnJyeTnp5OWFgYI0eO5KeffuKyyy4z7W84yZajen4EzvTVNtZW7yvE+cgtqeTdhMO8u/kwuaVV9OjahccmDuCmYeH4uMuoMFOdY8vcVrZv385f//pXioqKKCsr45prrql/bMqUKTg4ODBgwABycnLqt8fHxxMebswCHBMTQ0ZGRucOfiHaO601Ww8XsnzjIb7YnkVNnWZ030CeuzSSUb0DZSoF0cjMmTNZtWoV0dHRLFu2jA0bNtQ/1nDa5YbznzXc7ujoSG1tbZvUejYS/MLu5JdV8fm2LD5IPML2oyV4uTnx64sjueOSHkQGeJhdnminSktLCQ0NpaamhnfeeYdu3bqZXdJ5k+AXdqG0soZ16Tl8mnqMH/flU2fR9AvxYt6UQUyN7YaHq/xTEM176qmnGDFiBIGBgYwYMYLS0lKzSzpvdjEts7BPlTV1bNidx2epx/h6Zw5VtRbC/dyZFB3GpJgw+oXY3zKGHUVHmZa5PZFpmYXdqrNofj5wnNUpR/liezallbUEeLpwy/AIJsV0Y2h333YxnE4IM0nwiw5Pa01qZjGrU46yJi2LvNIqPF2duGZgCJNjwri0Z1cZdy9EAxL8osPal1vK6pRjfJp6jEPHT+Di5MCVfYOYHBPGmH5BuDnLnDlCNEWCX3QoR4sq+Cz1GJ+mHGNHVgkOCkb2CuB3Y3pxzaAQvGX5QiHOSoJftHsF5dV8vi2LT1OOkZBRAEBsd18ev34A1w0JJchLpj8W4lx06uDffrQYFycHegV6ysk4HUx5VS1f7chhdcpRftibT61F0zvIk4fH9WFSdDdZulCIC9Cpg/+5tbv4YW8+Xq5ODInwITbCj5gIX2K6+xIgsyq2O9W1Fr7bk8fqlKN8vTOHyhoL3Xzd+b/LL2JyTBj9QrxkRI5oM46OjgwePJja2lqioqJ466238PU98/TbTc3iOXPmTCZOnFg/dw+Ap6cnZWVlNq39bDp18D85eRBJhwpJOVJIypEiFn63nzqLcd5ChL97oy+CgWHesoCGCeosms0Hj/NpyjG+2J5NcUUN/h4uTBsWweSYMIZ295Nfa8IU7u7u9VMzz5gxg1deeYW//OUvJlfVOjp18EcFeBAV4MFNw4xJkiqq69h+rJjkw8YXwZaMAj5NPQaAi6MD/cO8iY3wJba7L7ERfkT4u0sL0wa01mw7WszqlGOsSTtGTkkVHi6OXDMwhOtjwrisVwDOMvxSWD2X8By7Cna16j77+fdjTnzLV3y95JJLSEtLA2D//v3cd9995OXl0aVLF5YsWUK/fv1atT5b69TBfzp3F0eGR/ozPNK/flt2cSUpRwpJPlJE8uEiVm45wrKNGQD4e7gQE+FLrPVXQXSEr4waaaE6iyavtIqjRRUcK6qovz5WVMGenDIOF5zAxdGB0X0DmRQTxth+wbi7yC8u0f7U1dWxfv167rrrLgBmzZrFokWL6N27N5s3b+bee+/lm2++MbnKc2NXwd+UEB83xvuEMn5QKAC1dRb25JSRfKSQlMNFJB8p4ptdxiJhSkHPQM/6L4KYCF/6BnvZ5clBZVW1vwj0Y0WV9feziyuptTSeDsTbzYkwX3f6hnhx35iejB8Yik8X+SIVzTuXlnlrqqioICYmhqNHj9K/f3+uvvpqysrK2LhxI9OmTat/XlVV1Rn30VSPQXvoRbD74D+dk6MDA8K8GRDmze0jegBQXFFDWmZR/RfB+l25/DcpEwB3Z0cGh/tYu4d8ie3uR7B3xx5eWGfR5JZWWoO9sj7YjxaeCvqSysbTyzo5KEJ83AjzdWd4pD9hvsbtMF93uvm6E+rjhpf8WhIdyMk+/hMnTnDNNdfwyiuvMHPmTHx9fev7/s+ma9euFBYW1t8vKCggICDAViW3mAR/C/i4O3N570Au720stqy15nDBCVKs3UPJR4p448eD1NQZLdxQHzdirb8IYiL8GNzNp111Y5RW1nCsqPKMLfbsksr6g+An+bg7E+brTrifO/FR/g1C3Qj4IC83HOUgrOiEunTpwssvv8yUKVO49957iYqK4r///S/Tpk1Da01aWhrR0dFNvnb06NG8+OKLzJgxAxcXF5YtW8aYMWPa+C/4JQn+86CUokdXD3p09WByjDEnd2VNHTuySup/FaQcKeTzbdkAODoovN2c6l+r6vcDJxcpU4pG21WT21WD1zX/vPoIbrBdA3mlVZQ20VoP9XUjzMedEQ1CPczXzWit+7rjKdMWCzsWGxvLkCFDeO+993jnnXeYPXs28+bNo6amhltuuaU++OfNm8eLL75Y/7rMzEySkpIYNmwYjo6O9OzZk0WLFpn1Z9STaZltKL+sipTDRaQcKaKksgatQWN83sZt6m/TcLv1AY1ucJtGr6fR63WjfTW1/eSNQC/XX3TDBHi6SmtdtCsyLfO5k2mZ24kAT1euGhDMVQOCzS5FCCHq2Ww4ilLqDaVUrlJqe4Nt/kqpr5RSe63XfrZ6fyGEEE2z5TjEZcD407bNBdZrrXsD6633hRDiFzpCN3R7ca6flc2CX2v9PVBw2ubJwHLr7eXAFFu9vxCi43Jzc+P48eMS/i2gteb48eO4ubV8GHlb9/EHa62zrLezgTN2fiulZgGzALp3794GpQkh2ovw8HAyMzPJy8szu5QOwc3NjfDw8BY/37SDu1prrZQ649e51noxsBiMUT1tVpgQwnTOzs5ERUWZXUan1dZzDeQopUIBrNe5bfz+Qghh99o6+D8FZlhvzwBWt/H7CyGE3bPlcM73gE1AX6VUplLqLmA+cLVSai9wlfW+EEKINtQhztxVSuUBh87z5QFAfiuW09HJ53GKfBaNyefRWGf4PHporQNP39ghgv9CKKUSmzpl2V7J53GKfBaNyefRWGf+POxvInkhhLBzEvxCCGFn7CH4F5tdQDsjn8cp8lk0Jp9HY5328+j0ffxCCCEas4cWvxBCiAYk+IUQws506uBXSo1XSu1WSu1TStntFNBKqQil1LdKqR1KqXSl1B/Mrqk9UEo5KqWSlVJrzK7FbEopX6XUh0qpXUqpnUqpS8yuySxKqQet/062K6XeU0q1fNrLDqLTBr9SyhF4BbgWGADcqpQaYG5VpqkF/qi1HgBcDNxnx59FQ38AdppdRDvxErBWa90PiMZOPxelVDfg90Cc1noQ4AjcYm5Vra/TBj8QD+zTWh/QWlcD72OsB2B3tNZZWuut1tulGP+ou5lblbmUUuHAdcBSs2sxm1LKB7gCeB1Aa12ttS4ytypTOQHuSiknoAtwzOR6Wl1nDv5uwJEG9zOx87ADUEpFArHAZnMrMd2LwJ8Ai9mFtANRQB7wprXra6lSysPsosygtT4KLAAOA1lAsdZ6nblVtb7OHPziNEopT+Aj4AGtdYnZ9ZhFKTURyNVaJ5ldSzvhBAwFFmqtY4Fy7HRZVOs64JMxvgzDAA+l1K/Mrar1debgPwpENLgfbt1ml5RSzhih/47W+mOz6zHZSGCSUioDowvwSqXU2+aWZKpMIFNrffJX4IcYXwT26CrgoNY6T2tdA3wMXGpyTa2uMwf/FqC3UipKKeWCcYDmU5NrMoVSSmH03+7UWv/L7HrMprV+VGsdrrWOxPj/4hutdadr1bWU1jobOKKU6mvdNBbYYWJJZjoMXKyU6mL9dzOWTnig27SlF21Na12rlPod8CXGkfk3tNbpJpdllpHAr4FtSqkU67Y/a60/N7Em0b7cD7xjbSQdAH5jcj2m0FpvVkp9CGzFGA2XTCecukGmbBBCCDvTmbt6hBBCNEGCXwgh7IwEvxBC2BkJfiGEsDMS/EIIYWck+IXdUkrVKaVSGlxa7WxVpVSkUmp7a+1PiNbUacfxC9ECFVrrGLOLEKKtSYtfiNMopTKUUs8rpbYppRKUUr2s2yOVUt8opdKUUuuVUt2t24OVUp8opVKtl5On+DsqpZZY53Zfp5Rytz7/99a1EdKUUu+b9GcKOybBL+yZ+2ldPTc3eKxYaz0Y+A/GTJ4A/waWa62HAO8AL1u3vwx8p7WOxpjj5uQZ4r2BV7TWA4Ei4Ebr9rlArHU/v7XVHyfEmciZu8JuKaXKtNaeTWzPAK7UWh+wTm6XrbXuqpTKB0K11jXW7Vla6wClVB4QrrWuarCPSOArrXVv6/05gLPWep5Sai1QBqwCVmmty2z8pwrRiLT4hWiaPsPtc1HV4HYdp46pXYexOtxQYIt1wQ8h2owEvxBNu7nB9Sbr7Y2cWobvduAH6+31wGyoX8fX50w7VUo5ABFa62+BOYAP8ItfHULYkrQ0hD1zbzBbKRhrzp4c0umnlErDaLXfat12P8YqVY9grFh1cgbLPwCLlVJ3YbTsZ2Os3tQUR+Bt65eDAl6282UOhQmkj1+I01j7+OO01vlm1yKELUhXjxBC2Blp8QshhJ2RFr8QQtgZCX4hhLAzEvxCiP9vpw4EAAAAAAT5Ww9yQcSM+AFmxA8wEyjPJgLdwOpJAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}