mnist_dataset_KerasAPI_model.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "authorship_tag": "ABX9TyMv7fUMXpfnAGzWMt1tsITb",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/aaolcay/Convolutional-Neural-Network-Examples/blob/main/mnist_dataset_KerasAPI_model.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## **CNN for `mnist` Dataset**\n",
        "In this notebook, we applied one simple CNN architecture by using Keras API. "
      ],
      "metadata": {
        "id": "fs4ij0aKHtcQ"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 48,
      "metadata": {
        "id": "FEKNIwwCuKEt"
      },
      "outputs": [],
      "source": [
        "import tensorflow.keras as keras\n",
        "from keras.datasets import mnist\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### **Load Dataset**\n",
        "Data chosen in this implementation is quite famous, and most of the tutorials on CNNs include this dataset as it is very easy to load and to use with CNN model considering model's high accuracy score."
      ],
      "metadata": {
        "id": "tpG06b5EumyA"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "(x_train, y_train), (x_test,y_test) = mnist.load_data()"
      ],
      "metadata": {
        "id": "fZCme5B0ulvU"
      },
      "execution_count": 49,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### **Data Shapes**\n",
        "We need to know the data shape and which type of data we deal with. This should also be noticed that 2 dimensional convolutional layer (`Conv2D`) needs a shape of (**#samples**, **height**, **width**, **#channels**). For example, imagine that we have 200 training data consist of images whose heights (i.e., #rows) are 28 and widths are 28 (i.e., #columns) as well. For colorful image the number of channels is 3 (RGB: Red, Green, Blue) but gray images have only one channel that is what we consider in this implementation for making prediction.  "
      ],
      "metadata": {
        "id": "opHme_NRvbfr"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print(f'The shape of training data is : {x_train.shape}.\\nThe shape of test data is {x_test.shape}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "w0NWf-fUvbAe",
        "outputId": "6cdec1e8-fde9-436d-c10a-166e78e7b955"
      },
      "execution_count": 50,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The shape of training data is : (60000, 28, 28).\n",
            "The shape of test data is (10000, 28, 28)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "The data shapes are not appropriate to be used within the 2D convolutional layers. Therefore, we need to add channel number to the datasets.\n",
        "\n",
        "                                  (60000, 28, 28) -> (60000, 28, 28, 1)\n",
        "\n",
        "                                  (10000, 28, 28) -> (10000, 28, 28, 1)\n"
      ],
      "metadata": {
        "id": "j9KRFQ55xO8D"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "x_train = x_train[..., np.newaxis]\n",
        "x_test = x_test[..., np.newaxis]\n",
        "print(f'The shape of training data is : {x_train.shape}.\\nThe shape of test data is {x_test.shape}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7fJliZANxNOl",
        "outputId": "94747c0e-8931-4751-fac3-9237db768944"
      },
      "execution_count": 51,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The shape of training data is : (60000, 28, 28, 1).\n",
            "The shape of test data is (10000, 28, 28, 1)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Let's now look at the labels and convert them to categorical form. Watch out CNN is a deep learning model and needs labeled data that makes it a supervised method. "
      ],
      "metadata": {
        "id": "YFCul16mycQn"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print(f'Labels of train dataset: {y_train[:20]}\\nLabels of test dataset: {y_test[:20]}')\n",
        "print(f'The shape of training data is : {y_train.shape}.\\nThe shape of test data is {y_test.shape}')\n",
        "ytrain = y_train.copy()\n",
        "ytest = y_test.copy()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xfTyR3dayObo",
        "outputId": "ac0280ef-a64f-4d79-e702-36afbbfc0050"
      },
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Labels of train dataset: [5 0 4 1 9 2 1 3 1 4 3 5 3 6 1 7 2 8 6 9]\n",
            "Labels of test dataset: [7 2 1 0 4 1 4 9 5 9 0 6 9 0 1 5 9 7 3 4]\n",
            "The shape of training data is : (60000,).\n",
            "The shape of test data is (10000,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Visualize each label with the data (images of numbers)."
      ],
      "metadata": {
        "id": "TY-MymH_zVMz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "fig = plt.figure(figsize=(5,5))\n",
        "for i in range(20):\n",
        "  x_test[i,:,:,:]\n",
        "  fig.add_subplot(4,5,i+1)\n",
        "  plt.imshow(x_train[i,:,:,:], cmap='gray')\n",
        "  plt.title(f'{y_train[i]}')\n",
        "  fig.tight_layout(pad=0.2)\n",
        "\n",
        "fig = plt.figure(figsize=(5,5))\n",
        "for k in range(20):\n",
        "  fig.add_subplot(4,5,k+1)\n",
        "  plt.imshow(x_test[k,:,:,:], cmap='gray')\n",
        "  plt.title(f'{y_test[k]}')\n",
        "  fig.tight_layout(pad=0.2)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 729
        },
        "id": "g1-Xo9wxy-v_",
        "outputId": "ce7c6f22-60aa-466a-a503-105785c53981"
      },
      "execution_count": 53,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 360x360 with 20 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 360x360 with 20 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXIAAAFkCAYAAAA9q7CPAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/NK7nSAAAACXBIWXMAAAsTAAALEwEAmpwYAACTdklEQVR4nOz9eXRb13mvjz8bI0EABECC80yJ1EBS82R5kqcocpzYcdLEcZumN21ykzZt06ZdTYfvXb1tf71t721zu3rTIandxElax07sxHFiO55kW5I1SxzEUZxHcARBzNP+/UHhhNRAURJJANR51sISeQjgvOejfd6z97vf/W4hpURFRUVFJX3RJNsAFRUVFZVbQ3XkKioqKmmO6shVVFRU0hzVkauoqKikOaojV1FRUUlzVEeuoqKikuaojlxFRUUlzUl7Ry6E8F72igkh/inZdqUCQgijEOIpIUSfEGJWCHFeCHEo2XalCkKILwkhTgshQkKIbyXbnlRECJEthHhRCOG71I6eTLZNqYYQoloIERRCfDdZNuiSdeLlQkppSfwshLAAo8DzybMopdABA8C9QD/wMPCcEKJeStmbTMNShGHgr4CDgCnJtqQqXwfCQD6wDfipEKJBSnkhqValFl8HTiXTgLTvkV/Gx4Ax4L1kG5IKSCl9Uso/l1L2SinjUsqXgR5gZ7JtSwWklC9IKX8ETCbbllRECGFm7p76/6SUXinlEeAl4NPJtSx1EEI8AbiBN5Npx1pz5J8BnpFq3YGrIoTIB2oAtTelshRqgKiUsmPesQagNkn2pBRCiCzgL4DfT7Yta8aRCyHKmQshfDvZtqQiQgg98D3g21LKtmTbo5IWWADPZcdmAGsSbElF/hJ4Sko5mGxD0j5GPo9PA0eklD3JNiTVEEJogO8wF+v8UpLNUUkfvEDWZceygNkk2JJSCCG2AQ8C25NsCrC2HPmvAn+TbCNSDSGEAJ5ibrLqYSllJMkmqaQPHYBOCFEtpey8dGwramgO4ABQAfTP3WJYAK0QYrOUcsdqG7MmHLkQYj9QjJqtcjX+BdgEPCilDCTbmFRCCKFj7h7QMncTZjAXE44m17LUQErpE0K8APyFEOI3mMtaeRTYn1TDUoNvAM/O+/0PmHPsX0yGMWslRv4Z4AUp5W0/5JvPpXmD/87cDTg6L9f+l5NrWcrwZ0AA+CrwK5d+/rOkWpR6/CZzqZljwH8BX1RTD0FK6ZdSjiZezIWhglLK8WTYI9QEDxUVFZX0Zq30yFVUVFRuW1RHrqKiopLm3JIjF0J8UAjRLoS4KIT46nIZtVZQ9bk2qjaLo+qzOKo+C7npGLkQQstcetJDwCBztQY+JaVsWT7z0hdVn2ujarM4qj6Lo+pzJbfSI98DXJRSdkspw8yl4jy6PGatCVR9ro2qzeKo+iyOqs9l3EoeeTFzlfUSDAJ7F/uAECLdU2QmpJS5S3zvDemjaqO2nXncdvpIKcUNvF29ty5jxRcECSE+D3x+pc+zSvQt55ep2iyOqs/irDF9lpU1ps11286tOPIhoHTe7yWXji1ASvkN5lZBrYUn441wXX1UbRTUtrMQVZ/FUe+ty7iVGPkpoFoIUSmEMABPMFerWGUOVZ9ro2qzOKo+i6Pqcxk33SOXUkaFEF8CXmOuVsXT6tLdX5AMfbRaLUIILhXxQQihHLtkE7FYjHg8rvycDNS2sziqPouj6nMlq7pEfw0Mcc5IKXetxBffqjZarZYPfvCDOJ1ODAYDdrudvLw8tm7ditPpBGBoaIhjx44xPDyMy+Xi2LFjeDyXl5u+aVZMG1DbzvVId31ucLLzhkh3bVhC21kT1Q9vd3JycsjJyWHbtm3k5eUpjjw3N3eBI8/PzycYDOJ0OhkcHKSpqQm/3080qhb7m49Op0Ov11NVVYUQgvHxcXw+H16vN9mmrSoajYacnBzMZjN2ux2AYDDI5OQkMzMzRCIR1FpNqYHqyNcADz/8MA8//DD33XcfDocDQAmxaDQa5WbLz8/n0UcfZXx8HJfLxblz54hGo4yNjak35DwcDgf5+fl8+9vfxmKx8PWvf50zZ85w9OjRZJu2auh0OoxGIx/72MfYuXMnjz32GOFwmO7ubr797W/z6quv4nK5iETU8vapgOrI05j8/Hx27NjBHXfcwaZNmzCbzeh0c/+loVAIv9/P2NgYwWAQi8WCxWIhPz8fq9VKPB6npqaGYDDI+Pi46sjnYbVayc/Px2w2o9frCYVCt53D0uv1WCwWampq2LJlCxaLhWg0SllZGZWVlVRUVOB2u28bXRwOBzk5Odx55520tbVx4sSJJX/WarXicDjwer3Kfbnc95taNCuNKS0t5dd+7dd48MEHqa+vJzMzE5ib1PT5fIyMjHDmzBnee+89Ghsb6e/vB8BsNpObm0t9fT2bNm1Co1GbwXzsdjvFxcXo9Xqi0Shutxu/359ss1YVo9GI3W6nvr6eHTt2oNPpMJlMlJSUsH79empra8nIyEi2mauCRqMhPz+f7du386d/+qccPHjwhj6fnZ3Nxo0bKSgowGazKckHy4naI09DtFotpaWl1NXVsW/fPhwOB1JKXC4XfX19fPvb38btduPxeBgfH1d65Lt378bpdJKTk0NmZqbyUllITk4O5eXlAExNTXHixAkmJyeTbNXqoNFosFgs3HHHHTzyyCNUVlYm26Sko9Fo2LhxI5s2bcJoNN5wxyc/P589e/awd+9evF4v3/zmN5d9viUlHbnBYECn0ylP/Hg8Dsz1NEOh0IIUusTfbic0Gg12ux2n00lhYSEAkUiEoaEh2traeP311/F4PPj9fkKhEPF4HJvNRl5eHn6/H5vNBszpejvqdy0S6ZoOh4PCwkLC4TAzMzOMjo4SDAaTbd6KYzKZyMjIIDc3l5qaGvbs2aPMucxHp9NhMBgwm83KZHksFktaOutKI4SgoKCAgoICNBrNDTvyjIwMcnJycDqdBAIBDAbDstuYko58y5YtVFZWcujQITQajRKLC4fDvPPOO4yNjSk9zqmpqWSbu+oksgmysrKUrIqxsTH+4i/+gtbWVvr7+5UHnZQSs9nMxz/+cXbv3k1lZSVGo5FoNMr58+c5d+6c6swvYTabKSkp4b777uOxxx7j2LFjNDY2rlkHNR+dTsfBgweprKxk//79VFVVsWnTJmXOZT6lpaXs3LkTjUbD2NgYPT09jIyM0Ne37FUIko5Go8FgMLBt2zbq6+vRarU3/B1TU1O0trbywQ9+kLy8PMxmM16vl3A4vGx2pqQjr66uZuvWrUr81uv1EovFlIkVt9uN2+1menqasbGxG/7+cDhMMBhkcHCQQCCAz+db7ktYUeLxOBMTE3R2dvL6668zOTnJ5OQk3d3djI2NXZFOqNPpWLduHSUlJQuGhuFwWE09nEcibdNms2GxWBgeHqa/v/+2mAjWaDTKnEl1dTU5OTno9fqrvjc7O5vKyko0Gg3T09MUFBTQ3NzMzMwMXq93TbUpi8VCdnY2hYWFZGVlMTAwcNOdR6vVSlZWFnq9/qYeCIuRco5cCMF9993HwYMHKSoquuKCDx06RCwWw+/3Mzk5yejo6A19v5SSyclJxsfHeeaZZxgcHKSrq2s5L2HFiUQiNDQ00NnZybvvvksoFFLCAFfrPRqNRnbv3s26deuWvQGtJTIyMiguLsZut6PT6WhoaKChoeG2GLHodDoeffRRtm/fft3rraiooLy8nP379xMOhxkdHeXll19mcnKSrq4uZmZmVsnqlaewsJDq6mo2btxIZmYmr776Kp2dnTf0HQaDQcmEcjgcmEwm9Ho9gUBg2exMOUcupeS9995jenqayspKJWvAarVisViUoUlBQQGZmZlUVFTg8/mIxWJYrdYF8atEaCEWixEIBNDr9WRkZFBYWIjb7aa5uRmj0Zh2jjxBOBxmampKiU9e7QbcsmULNTU1FBcXK7HxqakpxsfH6e3tZXh4+LbocS6F3NxcHnzwQbKzs5mYmGBwcJDR0dE1r89dd93Fjh07yMvLU9YeXI1oNEo4HCYSiRCPx7Farej1enJycrj33nvJzc3l//yf/8PZs2dX+QqWn4yMDBwOB/fffz/33nsvGo2GgYEBXn/9dbq7u2/ou3JycqirqyMrK2uFrE1BRw5w7tw5RkdHWb9+vRICyc/PJycnh/Xr15OXl4fJZCIrKwuHw6E0sKysrCtietFolEgkQjQaxWq1kpOTA4DX66W6ujqtY+zRaJTZ2dmr/i1xQ27cuJEdO3bgdDoxm80ATE9PMzAwwMjICJOTk2veUS0FIQTZ2dns3r0bs9nM7Ows4+PjTE1NrWl9EiGVD3/4w9jtdoQQV73eWCxGOBzG7XYrITmDwYDRaMRqtVJfX09dXR3f//73aWhoSPt5BaPRSH5+Prt27eLBBx9UHuynT5++4YwTh8PB+vXrMZvNKzZpnpKOvKuri76+Ps6fP088HiccDqPX65XVZkajEafTSW5uLoWFhfT09OD3+6mpqbliRjgRDx8aGuLDH/4wn/vc58jMzCQWizE6OprWjnwxioqKWLduHU888QS7du0iKyuLeDzO9PQ0b775Jm+99RbDw8PLOuGSrmg0GnJzcyktLWXDhg14PB4mJiYIBoOEQqFkm7di2O12qqqq2LBhAyUlJeh0OmWSfD6BQIA333yToaEhWlpaCAaDRKNRtmzZQnl5OY888gg6nQ4pJfv27SMWi/HGG2+kdaZPYo1GYoLzm9/8Jg0NDczOzi55DiCRyllaWsq2bdswGo0rdr+lpCNPxI6u1dvUaDS4XC4cDgcjIyMMDAwQDAbx+XxX7ZFHo1H8fr8Su5u/6nF6enplLyZJmM1mCgsLKSoqIj8/H51Oh9frZWxsjL6+Pjo7OwkEAmu6t7lUNBoNDocDu92u9MbD4TCxWGzN6pMIiWzZsoWSkhKysrIWhFRisRjRaBSv14vb7aaxsZGBgQFaW1sJh8PE43E0Go2STZYYAVZVVeHxeGhubsbtdjMzM5NWGgohyMrKorCwkA0bNpCVlUUoFKKrq4uLFy8SjUaXfD0Gg4GioiJyc3OxWCz4/X5lMni5511S0pFfj3g8zuzsLF6vl8HBQSUW3tjYeNX32+127rrrLioqKrBarQwMDNDT08Pbb7/NyMjIKlu/OjgcDjZu3IjNZlMebuPj47z55pscP36cxsbG22ISbyno9Xpqa2uVxS+BQICZmZk1lX0xH61WS15eHnfccQf/83/+T2w2G2azeYEj9/l8TE9Pc/z4cTo7O/nmN7+J2+1WOllCCEZHRxkdHeWJJ55QHoKHDh1i3759ADQ1NfHaa69dc/4mFTEajezfv5/9+/ezb98+ZmZmGBwcpKOjg76+vht6KOXk5PDLv/zL7Nq1i3g8TldXFx0dHQt0XC7S0pEnSDjwBFdrLEajkezsbO6//342bNiAEIK2tjaampqYmZlJ6+Hf1bDZbOzfv59t27axa9cu7HY70WiUnp4empubee+99xgcHEybG2s10Gq1OJ1OpcJff38/Z86cWfabLZXQarUYjUaysrIwGo3KsvFwOMzY2Bitra2cOnWKzs5ORkZGlHtlfrvxeDyMjIxw6tQpNmzYoKx8tNls3HXXXeh0Oo4dO4bP50ubEJVOp2P9+vWUlZVhNBoZHR2lpaUFn893Q07cbDbjdDqpq6sjPz8fKSXt7e2cPXv2Ch2Xxe5l/bYUxGw2U1payuOPP66sUjt37hzvvvuuMnGzlnA6nfy3//bflJQpnU5HMBjk/PnzHD9+nJ/97Gdpc1OtFlqtloKCAqV9tLe3c/jw4WuG9tYCGo1GKYw1n0QY4bXXXuOZZ55RwkxXw+PxMDw8zFtvvQWgLCCyWq0cPHgQjUbDf/7nfxKNRtOmzRkMBurq6qiqqsJgMNDb28v7779/Q2tNhBDYbDaKiorYsWMHNpuNWCzG2bNneeedd1YkpLmmHblWq+Wuu+5i+/btOBwOtFotPp+PsbExXC5X2s+sX05WVha5ubls3LiRvLw8xYmPj4/z4x//mI6ODgKBgNobn0dxcTGlpaUcOHAAp9NJd3c3LS0tNDY2ruke+eXE43GGh4fp6urin//5n+ns7GR2dva61Q19Ph+nTp3C6XRSX19PUVGRkh2VbiQycNavX09xcTEALS0tvPbaaze0AYtWq+WOO+5g27Zt2O32BeUzWltbV6Ri5Jote6fVasnIyKCmpoaamhqMRiOhUIjR0VGlMP5acmhCCCWLJ5FqqNFolIVT7e3t9Pf3r+kJvJshUVelvLwcu92uLBYbHx9fszHyy0mkHPb399PW1saZM2fo6+sjHA5ft61EIhHGxsaYmJi4Yl5Br9djtVpXpLbISpDIisvOzsZsNhMIBHC5XPT399/QyF0IQUlJiRKeSayFmZycZHp6ekXuvzXbIy8uLqasrIxPfOITbNiwAZ1Ox3vvvcdTTz3F6dOnGR0dXVM9coPBwO/+7u+yfft2cnJylBWcp0+fprGxkZ6enjW14m65qK6uZseOHWRnZzM7O0tvby/T09NrLuR2NRKbjwghCIfD/I//8T84f/48s7OzN9XJmb+ZSTwep7Kyki984Qs8++yzHD58ePkvYJnRarUYDAacTiehUIgjR44opZ9vBCEE+fn5ygKrmZkZurq6VrQU8pp15A6Hg9LSUhwOBwaDQUno7+rqwuPxrCknnpWVRXZ2NhUVFRQXF6PVagmFQni9Xtra2mhsbFyRCZa1QKKKpFarJRgM0t/fv6Zj4zDX8ywrKyMvL2/B8UQNoxtBr9dTUFBAXl6eUtoggd/vZ3BwMG22yIvH40SjUXw+n+LUDQaDUpd+KT3pRCSgsLBQqZbo9/tXvILmmnXkpaWlbN++HYvFQjgcprm5mQsXLtDe3r7mhsxlZWVs3LiR6upqiouLEUIwMTHBxYsXefnllzl58uSay85ZLvLy8igpKUGr1TI5OcmxY8fWbEpqAqPRyIEDB9i2bZuS+XWzD3mz2czOnTupq6tTCmklFhUNDg7y7LPPMj4+vsxXsDKEw2F8Ph99fX3k5uZSVlaG0+nEYrEsufNnMplwOBzs2LGDuro6YK4kRltb24o+0NacIxdCoNPpKCoqYuPGjej1eqampnj11Ve5cOHCmooRJ1LI6uvruffee7Hb7UpIJZFjPzs7SygUWjPXvFwYDAYyMjKUWKbb7aa/v5/GxkYmJiaSbd6Kotfr2bhxI2VlZTf1+cTS/OzsbKqqqnjooYeorq5W/p7YoSoQCKzI4peVIpFdMzk5ic1mo6qqioMHD+J0Ojl+/PhVQ5OJ1NVEyRC73Y7D4SA7O3vB+1ZagzXnyBNDm/z8fMrLy5V65sePH6evry9tGtVSSKR6VVdXs3v3biwWC0IIYrEYs7OzDA8Pr7myosuFwWDAZrMpYQG3283o6GjaFlC7EXQ6HaWlpVeEVpZKRkYGVquV8vJyampq2LVrl1KQLVGkLlHSNp06TvF4nEgkwuTkJIWFhdhsNnbt2kVxcTHhcPiqIwudTkdlZSU2m02pbpgohZy47sQq2ZXUYc058pKSEg4ePMhdd91FVVUVHR0dNDQ00NTUlHZ1x69HSUkJjzzyCHfffTc1NTWYTCZmZ2c5d+4cP//5z3n++efXfJjgZqmoqFA2UDCbzZw6deqmJrZuR/bu3cvevXs5dOiQkiWVGAkm9gj453/+Z1paWnC5XGnVkfD5fDz77LPU1dURiUSU+ju/8zu/c9XQipRSKVGQ6FhlZmYu2M+0t7eXn/3sZyu6XeCaceRCCCwWC0VFRWzbto2CggL0ej39/f309PTg9XrX3I7fVquVLVu2KCV9YW55eXNzM52dnQwNDaXNQozVJlGLxmw2I4Rgdnb2tsobvxkSmtXU1LBhwwaqq6uv2ArO7/czNTVFS0sLPT09aZf9E4vF6O/vx2AwcObMGfx+P4FAgMzMzKuW900U30s8rBK1jeY78kAgwMTExIr6n+s6ciFEKfAMkA9I4BtSyn8UQmQD3wcqgF7gE1LKpFWgMhgM7Nq1i3vuuYfPfOYzxGIxPB4PP/7xjzl//vyKZakkU5+SkhI+8YlPYDQalWMul4t/+qd/Ynx8POmOKZXbTmZmJgUFBYp2Pp9vRdPDrkYq63M16urq+OpXv7pgrcLljI2N0dvbS2Nj4y1VFk2WNrFYjMHBQYaGhjh69CglJSXk5eVRXFx81Xz4YDDI22+/rSQTfOpTn+LgwYM8+OCDSsmHWCy24h2qpfTIo8BXpJRnhRBW4IwQ4nXg14A3pZR/I4T4KvBV4I9WztRro9frsdls3H333Wzbtg2DwcDIyAjDw8P09fXhcrlWMj6VNH0Sy6wTdTLgF3XIE5toXI3LN6LQaDRotVpleCyEULY9q6qquuLzUkqGh4fp7u6+Xgw0ZdtO4toyMzOJRqP09fXd8G5Ty0BS9dFoNErb0Wg0yoYrMJcJVVNTQ0ZGhtIuKioq2LBhAxaLhczMTHQ63YK2B3D8+HGOHTuG3++/1XsuadokMnnC4TCTk5OEw2FmZ2evun9pJBJZMA81OzvL1NTUgnBSYvPlxcod3CrXdeRSyhFg5NLPs0KIVqAYeBQ4cOlt3wYOkyRHnpGRQXZ2No888ghlZWVoNBql2E13d/eKxolTTR+NRoPZbF50VV4gEFiQyZLIfjGZTMp3JCZRP/zhDy+4WROTWUeOHGFwcJBQKHTN0U6qaTOf+bu2eL1e2tvbGRgYWE0TUkKfxP+tVqtl3bp1yvGHHnqIT37ykzgcDmXUkljwc3mxunnXw2uvvcbLL798y3algjYAMzMzSgXEpeDz+ZicnFwQRkmEpMLhcGrUIxdCVADbgRNA/iWxAUaZGwKtOhqNhieffJK9e/dSWVmJXq9ndHSUd999l5///OerWm88FfQpKyvj7//+75Vdka5GZ2cng4ODSjpVVlYWpaWlC1LIdDodFouF/PwrzZZSKrvJvPvuuwwPD1/XrlTQBuYe+ok9Jy9fwJJMVlOfYDDI4cOHcbvdHDp0CJj7//7c5z6nVPnLzc1V9LncaV/Nkff29tLS0nJTm6Ffj1RpO0tFo9EsGO14PB76+vpWNNS55FYshLAAPwS+LKX0XNZLk0KIq3b/hBCfBz5/q4ZejUQIoa6ujn379mGz2fD5fAwPD3Px4kVaW1tXbSHMzehzq9rEYjGCwSAZGRnKRIzFYuHuu+9e9HPFxcV0d3crDzm73c66deuora296vsjkQiRSGRBGCURN0z04hcjldqOwWCguLiY7OxsjEajsgNVYvuyZLDa+kSjUTo7O8nJySEWiymOJ7GA5XJHfbnTjsfjxONxJa1udnaW7u5uGhoabqi41FJIpbZzs0QikRVfLbwkRy6E0DMn5veklC9cOuwSQhRKKUeEEIXAVR/FUspvAN+49D3LGqi2Wq04nU6qq6tZv349Op2Oixcv8n//7//lzJkzDA0NrcpS/JvV51a1mZyc5OjRo2zevPmGFnds2LCBdevWLYiRJ3qmiVV58/Pt29ra6O3tVXYVAhgYGFCK5C9GqrUdq9XK3XffTXV1tbKhbnd3N+fOncPlci3HKW6IZOgTCoU4evQoBoOBxx9/HLPZvKQHcgK/368sv7948SJf//rXmZqaYnJy8oaX+C9GqrWdpZJ40K0mS8laEcBTQKuU8h/m/ekl4DPA31z698crYuHVbVKGyDt27FCG/+3t7bS0tNDe3q7sLr8KtiRNn4mJCd59910CgcBVl/9qNBpMJhNGo5G8vDyl167T6dBqtbhcLoLB4II6LFNTU/h8vgWF9Pv6+hgeHmZ4eFiZfZ+amlI+fy1Sre1oNBoyMjIoLy9XVt6Nj48zMDCA3+9f9fTUZOkTj8eVlawnTpxQNjYvLi5ekAF1LVwuFwMDAzQ0NNDd3U1PT4+S9bNco5pUazs3glarRaPRKB2i1XDqS+mR3wl8GmgSQpy/dOxPmBPyOSHErwN9wCdWxMKroNfrcTqdPPjgg/z2b/82ubm5+P1+nnvuOc6ePcuZM2dWczVZ0vS5ePEif/u3f8uDDz7Izp07r4j3CiEoLS0lPz+fhx56aEEWSzwe58SJE4yMjDA6Oqo0tmPHjjEwMEBvb+9y3JQp1XYSmTjbt2+noKAAmBttnD59mmAwmIwViEnRJx6PMzk5yZkzZ/j7v/97tmzZwoYNG/ilX/olcnNzr/v58+fP8+abb/LCCy+sZB2VlGo7S0Wn02EwGJRJ4URIcsXPe703SCmPAOIaf35gec1ZGhaLhTvuuINNmzaRnZ2NwWAgEAjQ09PDyMjIqt6QqaBPW1sbU1NTV12wkFhp9uqrry5w9LFYjN7eXvx+/4IVr6Ojo8zOzi5L40sFbeaTiOn6/X5GRkZwuVwcPXqU06dPJ2XhSrL18fl8XLx4EbfbrWyWXFZWRl1dHXl5eRQUFNDX18fExASNjY2EQiE0Gg3nzp2jra1tRSfvkq3NzVJRUcHu3bvRarWMjo5y/PhxLly4sOLnTY0p+xtACIHZbGbr1q1UVlYqW1UlVlit5DLYVGVwcHDJ6VG3M4laGh6Ph+npaWZnZ2lqauLChQtrbtXvUggGg0rITKPREAwGlQqGkUiErKws+vr66Onp4eWXX8bv9yOEwOVyMTk5qa4avgp5eXlUV1cTi8UYGxvjyJEjdHd3r/h508qRCyGoqamhvr6exx9/fEnDQBWVBLFYjO7ubr70pS8Bc9kb4+PjqkNi7iHX1NREe3s777//PhkZGWRkZCilLdxutxJ+i0ajyktlIS6Xi+bmZn7+85/T09PD0aNHV6Uee9o5cqfTSUFBwYL6Ionyk+lUaU0lOQSDQS5evJhsM1KSxCT3aq69WGsMDQ3R0NBAY2Mjg4ODK72qXCGtHLlGo6G6upqamhoyMzPR6/XAXBre6OgoPp9P7V2pqKgkjZdeeomf/vSnSgmM1epYppUjh4X7DCbS7l5//XUuXLig1N9WUVFRSQaJOkarTdo58gTxeFzZBef73/8+R48eZWZmRg2tqKio3HaklSOPxWK88cYbnDp1ildffZVIJEIwGFT2w1OduIqKyu1IWjlyKSUDAwMMDAzQ3NycbHNUVFRUUoLVduQTgO/Sv6mOkyvtLF/B86WTNnClPiupDaj6XA8v0L7C51gu1LazODesj1jtcIQQ4rSUcteqnvQmSIad6aINqPpcj9W2VdUm9c55s9yMrVeu6VZRUVFRSStUR66ioqKS5iTDkX8jCee8GZJhZ7poA6o+12O1bVW1Sb1z3iw3bOuqx8hVVFRUVJYXNbSioqKikuasmiMXQnxQCNEuhLgohPjqap13KQghSoUQbwshWoQQF4QQv3vp+J8LIYaEEOcvvR5eQRtSUh9Vm8VR9VkcVZ9rs6zaJDZaXckXoAW6gCrAADQAm1fj3Eu0rxDYcelnK9ABbAb+HPiD21kfVRtVH1Wf1NdmtXrke4CLUspuKWUYeBZ4dJXOfV2klCNSyrOXfp4FWoHiVTQhZfVRtVkcVZ/FUfW5NsupzWo58mJgYN7vg6zuf+aSEUJUANuBE5cOfUkI0SiEeFoI4Vih06aFPqo2i6PqsziqPtfmVrVRJzvnIYSwAD8Eviyl9AD/AqwDtgEjwN8nz7rkomqzOKo+i6Pqc22WQ5vVcuRDQOm830suHUsZhBB65sT8npTyBQAppUtKGZNSxoFvMjdMWwlSWh9Vm8VR9VkcVZ9rs1zarJYjPwVUCyEqhRAG4AngpVU693URQgjgKaBVSvkP844XznvbR4GVKrmYsvqo2iyOqs/iqPpcm+XUZlWqH0opo0KILwGvMTeL/LSU8sJqnHuJ3Al8GmgSQpy/dOxPgE8JIbYBEugF/vtKnDzF9VG1WRxVn8VR9bk2y6aNurJTRUVFJc1RJztVVFRU0hzVkauoqKikOaojV1FRUUlzVEeuoqKikuaojlxFRUUlzVEduYqKikqaozpyFRUVlTRHdeQqKioqaU7aO3IhxJeEEKeFECEhxLeSbU+qIoSoFkIEhRDfTbYtqYQQYpMQ4i0hxMyljQc+mmybUgkhxOFL7cZ76dWebJtShVRqO2nvyIFh4K+Ap5NtSIrzdebqTqhcQgihA34MvAxkA58HviuEqEmqYanHl6SUlkuvDck2JhVItbaT9o5cSvmClPJHwGSybUlVhBBPAG7gzSSbkmpsBIqAr12qNvcWcJS5+hcqKouRUm0n7R25yuIIIbKAvwB+P9m2pAkCqEu2ESnG/xJCTAghjgohDiTbmBQmaW1HdeRrn78EnpJSDibbkBSkHRgD/lAIoRdCfAC4F8hMrlkpxR8xt99lMfAN4CdCiHXJNSklSKm2ozryNcylUpgPAl9LsikpiZQyAjwGfAgYBb4CPMfcdmAqgJTyhJRyVkoZklJ+m7nwwYrteJ8upFrbWZV65CpJ4wBQAfTP1bDHAmiFEJullDuSaFfKIKVsZK4nBYAQ4hjw7eRZlPJI5kIItz2p1HbSvkcuhNAJITKYKxqvFUJkXJpRVpkbCif2/tsG/CvwU+Bg8kxKLYQQWy61mUwhxB8AhcC3kmxWSiCEsAshDibuKSHELwP3AK8m27ZUIJXaTto7cuDPgADwVeBXLv38Z0m1KEWQUvqllKOJF+AFglLK8WTblkJ8mrkNbseAB4CHpJSh5JqUMuiZS+0dByaA3wYek1J2JNWq1CFl2o66Q5CKiopKmrMWeuQqKioqtzWqI1dRUVFJc27JkQshPiiEaL9UZ+Cry2XUWkHV59qo2iyOqs/iqPos5KZj5EIILdABPMRc7uQp4FNSypblMy99UfW5Nqo2i6PqsziqPldyKz3yPcBFKWW3lDIMPAs8ujxmrQlUfa6Nqs3iqPosjqrPZdxKvnUxMDDv90Fg72IfEEKke4rMhJQyd4nvvSF9VG3UtjOP204fKeWNLDJS763LWPGFM0KIzzNX4nEt0LecX6ZqsziqPouzxvRZVtaYNtdtO7fiyIeA0nm/l1w6tgAp5TeYW2G4Fp6MN8J19VG1UVDbzkJUfRZHvbcu41Zi5KeAaiFEpRDCADwBvLQ8Zq0JVH2ujarN4qj6LI6qz2XcdI9cShkVQnwJeI25OidPSykvLJtlaY6qz7VRtVkcVZ/FUfW5klVdor8GhjhnpJS7VuKLVW0WR9VncdJdnxuc7Lwh0l0bltB21CqBKioqN4VGo0Gj0aDX6xccl1ISiUQwGo3Y7XYikYjyikajhEJqTbLlRnXkKioqN4XZbMZkMpGXl7fgeDweZ3R0lLKyMh555BFGR0eZmJhgcHAQt9tNV1cXarG+5WXNOvKcnBwcDgd33303Go2GM2fOMD4+zvDw8JprRHq9HqPRyP79+zGZTExMTDAxMcHIyAh+v59oNHpT35mZmYnD4WB6eprZ2Vni8fgKWK+SLmRmZpKXl4fT6cRut1NaWkpWVhaFhYVc2rgEgFgsxtDQELm5uezcuROfz4ff78ftdtPb28s///M/31SbVLk2a9aRFxYWsmnTJv7sz/4MvV7P1772Nc6cOcPo6CixWCzZ5i0rGRkZOBwOnnzySXJzc2lqaqKhoYFjx44pw9kbQQiByWTC6XSyefNmWltb8fl8qiO/jRFCkJWVRW1tLfX19axfv57t27fjdDopLi5e4Mij0Sj9/f0YDAZyc3PR6/VoNHMJcmfOnOGb3/ym6siXmTXnyHU6HSaTifvvv58PfehDOJ1O9Ho9n/3sZ8nIyODEiRNIKdeEU9JqtWRkZPDkk0/ygQ98gG3bthGLxZiYmODChQuMjY3dcDzSaDSSk5PDH/7hH1JUVEReXh7PPPMMfr+f8fFxIpHICl3N6iCEICMjg82bN7N7925yc3OJRqP8y7/8Cx6PZ020i+UgMzOTrKwsdu3aRWlpKfv378dut+N0OsnKysJsNmOz2TAYDAucOMy1y4KCAiV+nnDiKivHmnTkFouFyspKtm7dSmZmJlqtlnXr1uF0Oq9odOlKotecn5/Ptm3bOHDgABqNhunpaUKhEMFgkEAgcMPfa7FYyMvL45577lH0ys7ORq/Xp712RqORjIwMnE4nNTU17Nu3j5KSEgKBAN/61rfwer2qI7+EwWBQeuB1dXUcOnQIi8VyxcTm5cTjceLxuOK8w+Hwgr+HQqE1E9pMTPYKIdBqtRgMBnQ6HVqtFpgLMQWDQWKxGNFolHg8vmLXvuYcuc1mY8eOHVRVVZGbm4tGo8HtdvPee+/R0tKyJhqSEAKbzcb27dv5rd/6LbZs2YLZbOaVV16hsbGR//f//h9er/emvvuhhx5i165dVFZW0tfXxzPPPMM777zD4OBgWoektFot+/fvZ+PGjfzqr/4qeXl55OXlodVqcblc2Gw2Zmdnb1q3tUYiHv7ggw+yb98+TCbTkh7ko6OjeDwe+vr6rho+uXjx4pp5WDqdTiwWizJPkPA7BQUFxONxpqamePXVVxkaGqK/v5+xsTH8fv+K2LKmHLkQArvdTl1dneLEAYLBIJ2dnYyNjaW9E4e5noDdbqegoIDq6mocDgcAbW1tXLhwgampqRt2unq9HpPJRE1NDZs2bWJmZobBwUHa2tqYnJxMayeeuLatW7dSV1dHeXk5FouFzMxMYC77oq6uDoPBQGtrK7FYbM04m8UQQqDX65VwicfjIRgMMjU1BaCERvR6vdKz1mq1zM7O4vf78fl8hMNhPB4P4XAYv9/PxMQEs7OzDA8PX7XNjI6Opq22Go0Gs9mM3W4nJyeHdevW4XA4lJBTTU0NhYWF5OTkIKVkZmYGj8fDxMQELpeL4eFhPB4Pbrcbt9tNf3//smmxphy5Xq+ntLSUj3/845SVlSnHvV4v7733Hp2dnUm0bvnQ6XSUl5dTU1NDXV0dAH6/n5///OecP3/+ppyu1WqlqKiIAwcOsG/fPt5++22OHz/OkSNHCAaDy30Jq4rZbMbpdPKxj32Muro6bDYbgPJQN5lMPPnkk5w4cYLBwUF8Pt9tkeus1+uxWq3s37+f+vp6GhsbGRoa4uTJk0gpiUajBAIBfD4fwWAQjUaD0Wiko6ODvr4+uru7mZiYoKWlBZfLxeDgIOFwWPncWug0zcdoNFJWVsa2bdu4++67ueOOOyguLsZut19zHmDXrl3E43Elk2dqaorz589z5swZnnrqKSXkcqusGUeemGApKSmhrKwMq9UKzN2sPp+P1tZWxsbGkmzlrZORkYHNZqOuro7KykpgLhOgqamJoaGhGx666fV6srOzueOOOzh48CCVlZVEIhFaWlro6uoiFAqldW8coKamhq1bt5KXl4fJZLri70ajkS1btuB0OiksLKSpqYmLFy9y/vz5NZt2qdfrqamp4fHHH6euro7i4mIGBgZwuVwAzMzM0N3dzXe/+13eeustpqam0Ov12Gw2hoaGmJiYwO12Kz14v9+P1+slFoshpVwzTtxms2G1Wtm2bRtFRUXs2rWL4uJiysrKKCgowGQy4fV6mZ2dxeVy4fP5iEQi7Nq1C6PRiMvlwmKx4HA4MJlM2Gw2NmzYQDQaZd++fVy8eJHh4eFbtnNNOfLc3Fzy8/MXLFCIRqP4/X76+vrSvmcJc71Lh8NBdXU1xcXFxONxWlpaeOONNxgfH7/hnqROpyM3N5etW7fy0Y9+FKvVitfrpaurS+lhpTNCCMrLy9m5cyd2ux29Xq84GiklGo0GrVZLZWUlpaWl1NfX88Ybb3D8+HG6urrw+/1rzpEnQgSVlZV89KMfpaCgALPZTEZGhvIev9+vjPIMBgOjo6MYDAZycnKYmZnB6/WuGWd9NRJhJ4fDQVFREXfffTfV1dUcOHCAjIwMDAaDMmqZmppibGxMCUMGg0FqamqwWCwMDAxQUFCA3W7HYDBgNpspLy8nHA6zbds2ZmZmVEc+H4PBwL59+6itrVWOxeNxmpubuXDhwpq5GR955BF27tzJY489hlarpbOzk1OnTvHOO+/c1ESdyWSivr6eqqoqsrOzOXPmDO3t7bz66qtMTEyswBWsHkajkezsbHbv3s0HP/hBbDYboVCIwcFBAoEAfr+fmpoa7HY78IuMp3379lFWVsbRo0fxer243e6kXsdyotVqsdvtfPnLX1bywc+ePcuFCxc4cuTIFXHb6elphBBKBobL5VJ63WuZ6upqfumXfona2lqqqqooLi7GbDZjtVoZGxtjYGCAkZERXC4X3/3ud3G73Xi9XqLRKDqdjvXr12M0GvnmN7/Jvn37+MhHPqI8MGdmZli3bh1f+cpXiEQi9PX1MTs7e0sj3zXjyLVaLUVFReTk5CjHYrEYPT099Pb2rpmGl52dTXFxMU6nk0AgQH9/P1NTU0xPTy95kYUQAiEEOTk5FBUVUVNTQ0FBATqdju7ubhoaGpiYmEj7DI5EKCAnJ4ecnBzC4TBut5tz587h9/sJBAIIISgsLCQ7OxuDwYBer1dinjU1NQSDQS5cuJD2E6CX/59v2bKFyspKDAYDIyMjNDY2Mj4+fsX/+fw2laihslZJpBHm5ORQVVXFjh07qKmpoaSkBJPJRDweZ3x8nO7ubpqbmxkeHsblctHc3Kw4cZjrQHi9XsLhMAMDA+Tl5XHx4kXsdjtWqxW9Xo/BYMDhcGCz2dDpdLec2rtmHLler6e2tlaJGyd6EC+88ALnzp1bUyvJEmGBUCjExMTEDS/D12q1mEwmPvaxj1FbW8vjjz+O1WolHo/z/PPP89Of/nRN3LAWi4Xq6molNt7U1MTZs2f5vd/7PaXndM8997Bx40Y+/elPk5eXR35+PhaLBZPJxJe//GUaGhr46le/isfjuam8/FRBp9NhNBo5ePAg27Zt45577sFoNDIzM8P777/Pc889x8zMTNrPh9wKOp2OrKwsPvnJT7Jr1y4eeeQRtFotQghmZmYYHR3lpZde4sSJExw+fJhAIEA0Gr3i3hNCYLFYMBqNaDQaZX6uuLiY0tJSjEbjsq/JWBOOPBHHKi4uVnrksViMcDjM2NgY4+Pja6ZHPp/MzEwqKiq4++67MRqNBINBfD4fXV1dV+09mkwmpXeanZ3Nvffeq0wM+/1+hoaGmJmZWRNOHOZuKJ1Oh0ajUeYSmpublRswEZoKh8OMj4+TmZlJfn6+8tnEYo+VXMixWthsNkpKSti5cyfbtm0jGo0yMDDAm2++SWNj45qcC7hRnE4nRUVF7N69mw0bNihpl4FAgLfeeovu7m7ee+89ent7lUnN+e1CCIHRaMRisdDf309ubi6PPPIIJpMJq9VKbm6uMjKanZ1lcHBQmde6Ve3XhCPPz8+nsrKSyspKxZEnUqBGRkYYHx9PsoXLR6I3Ho/HsVqtbN68GaPRyM6dO/F6vYyMjPDjH//4qj30RO2U6upqysrKqK2tVXKph4eHaWhowOPxrPYlrRiJPGghBNFolNOnT3Pu3DklxhuNRmlvb1cKjDkcDuXGTIQRIpFI2odVYO7/vr6+nrvvvpv6+npGR0c5f/48f//3f8/09PSaSAS4VQoLC6mpqeHAgQM4nU4AJSvnBz/4AU1NTbS2tl61LSTCMhaLBbvdTnt7OwCf/exnMZvNZGZmLphMdrvdnD59WpmvUR05c/8BlZWV6HS/uJyWlhaampqYnZ1NomXLz6lTp/D5fOzYsUNZWVZYWIjdbicWixEKhaivr79qDzKx7NpkMpGZmYnRaFRm3U+ePMn3v/99BgYGrnLW9CQnJ4e7776b4uJitFottbW1BINBjhw5suDG0el0VFVVUVhYqByLx+NKrvTNVpBMJbKzs9m4cSMWi4VYLEZfXx99fX1MTk6mfWbScpFwxokyA4AyGisrK0Ov17N7927FWcMvHviJInN5eXnY7XZMJhMWi0UpGqbVapWRodfrpaOjg+eee4729vZlCWetCUeeWGmVSMqXUjIyMkJra+ua62kMDg6i1Wrp6+sjHA4rdVAyMjLIzMxEo9FQXFxMNBq94gZNhJsSPQSY63G43W4GBgZobm5eUw8+g8FAdna2EpPMz8+noKBAKeSUqI/hcDjIzs7GbDYveAAmKkeme40ZmEtbLS4uJiMjAykl4XCYeDyupGMmMlPSPYR0KyQ2vQgGg0QiEaUHrdFoKCoqwmKxKOWxEz32hJZms5nCwkJyc3Ox2Wx4PB6EEJjNZuX7w+Ew4XCYoaEhenp6aGlpYWpqalk0XxOOvLS0lJqaGnQ6nZI3furUKV588cU1lToGc47c5XLxG7/xG2RnZysrPCsqKvjoRz+K3W5nZmaGjo4Ozp49u+Czw8PDHD9+nC996Us89thjAAQCAZqamujo6GBwcDDtQwjzcblcvPbaa+Tm5rJu3Tqqqqrw+/1UVlZisVjIz89n586dVFRUUFhYiF6vZ3Z2FpPJhF6v57777iM3N5dXXnkFl8uV1iG6srIy7r//fnJycjAYDOzatQuz2YzH46GlpUUZfay1js+N0NnZyczMDKdPn6a6uloJPVosFn79139dKQaWeCV67okJ0USvO1FCYz7xeJzGxkY6Ozv5+te/zvDwMAMDA+oSfZhb5Wg2myktLaWsrEwZFiXqP0xNTaX9kPhyEvHa0dFRvF4vwWAQv9+vLNiwWCz4fD4GBgauKEkwPT2tbDYB4PP5GBsb49y5c/T396+5jIVEeqbb7SYWi5GVlUVZWRkPPfSQsmnGhg0byM3NZXZ2Fo/HQ3t7Oxs2bKC0tJTMzEycTifV1dVIKdPakRsMBmw2mzJnkJGRQX5+Prt27SI3N5fh4WHOnDmjbCKSGK0keqc+n0/pOWZkZKDX6/F4PEQikTVTziDhN44fP87AwAD9/f1KSnMiJTUxSguHw8rEZVZWFkajEaPRqHzX/FHczMwMk5OTnD17ltbWVvr6+pQ2uVyktSPPzs6mqqqKffv2sWfPHjQajVIbwu12K8V/1hpSSmXl3djYGM3NzQD8x3/8x6KfM5vN5OfnKw1oeHiYpqYmvvOd76y5kQvM3UCNjY0MDw8TCATIz8+nqKiIvXv3LhjOhkIhTp48yalTp/i3f/s3vvCFL/CRj3yEkpISCgsLefjhh9HpdFy4kL4btRuNRrKyspTJcp1OR1lZGWVlZUqc90c/+hG9vb20tLQoxaFGRkaU5fqJTlF+fj52u50LFy7gdrvXTFZYYr7on/7pn9Dr9ZjNZh566CH27NlDfX09GRkZSiEst9utpHQmUlyrq6uv+r39/f0cPXqUZ555hpaWFmZnZ5ddr7R25Hq9HovFoqSYAczOztLe3s709HSSrUs9cnNzOXjwIOXl5USjUQ4fPsy5c+eYmZlZM72q+cRiMXw+H0eOHEEIwUc/+lHy8vKUmLmUkubmZvr6+nj++efp7e1lfHyczs5OmpqacDqdmEwmNm7cSEdHB06nU6n0l24MDAzw1ltvUV9fv2DRHPwiTXPHjh2sX7+eHTt2KNsHJgqIzd90w2w2o9frGRoaYnp6mo6ODi5cuMDJkyeTcWkrQiwWw+/3c+7cOYaHh3nvvffQarVKnDsUCqHRaLBarTz55JMLeuMw17ufmpri3XffpaGhgePHj9Pb20swGFyRh17aO3Kz2bwgW8Xn89Hd3b2m0uiWA61WS15eHvv376egoIBYLMb58+c5d+7cmt3GLR6PEwqFOH/+PF6vl3379mE2m5XC//F4nM7OThoaGvjZz36m5ND39/fT3t7Onj17yMnJoby8nKKiIpxOJ8FgMC0d+cjICCdOnCAvL0/ZICIx/E/kzK9fv/6GvjNRjrWhoQFAWXi3FnrniRDtxYsXuXjx4lXfo9Vqyc7O5tChQ1fMLQQCAVwuF2+88QaNjY2cPn16Re+xtHbkJSUlSs5nosphT08PL730Ej09Pck2L2UwGo3s3buXO++8k0OHDqHX6/H7/bdF8SOY641OTEzwv//3/6aiooL9+/crK3+///3v09zczMzMjBI6OHnyJN3d3RQVFbF582bq6urYt28foVCIb33rW7S1tSX5im6cxsZGBgYGePfdd6msrOQ3f/M3sdlsyiYlmZmZN5ydY7VayczMVOLHGo2G119/nf7+/hW6itTinnvuob6+ng984ANKFgvMZTs988wzNDY28pOf/GRVdp5KS0eeGAomirtnZmYqS9YTBdvXUhrdraLVaikuLqaoqIjs7Gxl/sDj8eDz+ZJt3ooTCoWIRCK0trYyMzOD0WhUcu4vXrzI6OjogtWsMzMzhMNh+vr6cDgcbNu2TZn0tNlsaLXatJsYTuxkL4TA7XZz9uxZsrKy0Gg05ObmKs7YYDAo+9waDIZFv1Or1SqTooWFhaxbt46jR4+u0hUlj4yMDCwWCxs3bqS+vp7c3FwlzTASieD3+2lra6O1tZWJiYlVGe1e15ELIUqBZ4B8QALfkFL+oxAiG/g+UAH0Ap+QUq5KYDpRenXz5s08+OCD6HQ6ZdPhoaGha66+WglSUZ/LMRgM7Ny5k40bNwLQ09OjVIXs7+9fsR55KmmTWKLf1tbG4cOHleNXCwUkVgUfP36cSCTCwYMHycvLY8eOHRQWFmKxWPB4PLes22rqk5jkHBgYYHBwkHPnzil/q66uprS0FJ1OR1FREZ/5zGcoLi6mpKRkyd/vcDiU0q3LQSq1ncupqqpi7969PPnkk9TX12OxWJTRzNTUFMPDw7z//vur6oeWsr11FPiKlHIzsA/4LSHEZuCrwJtSymrgzUu/rwp6vZ7CwkIcDodSOSwajdLb28vIyMhqx3tTTp/L0el0VFRUkJeXRzQapaenhzNnzuD3+1c6rJJS2kgpicViytL7y2tlzCcej9PT00Nrayvt7e243W6ysrKoqamhtrb2upsQL5FV1ydR3iEUCikvl8tFZ2cnnZ2ddHR00NnZqaRaJha8tLW1cezYMZ599lmOHDlCZ2fnggnyRJx9GUmptgNz91FeXh6bN2/mvvvuUxZYJSbOExuyvPXWWzdUjXRZbLveG6SUI8DIpZ9nhRCtQDHwKHDg0tu+DRwG/mhFrLwMo9FIRUUFOTk5ypMwMXTu7e1dDRMUUlGfy9Hr9cpGFKFQiPb2dt55550VDz+lgzbXIh6P097eTiQS4dy5c9TV1VFSUsLWrVuJRCI0Njbe8qRnqugzPj6uOG63201jYyMOh4Pt27cjpSQYDHLmzBlaW1t5/fXXOXDgALt27cLpdF6RrbFcpIo280ls9bZr1y4ee+wxTCaTMnGeCNWdPn2aF198kYmJiVUNv91QjFwIUQFsB04A+ZfEBhhlbgi0Kmi1WqxWq9KIAoEAExMTvPbaa9ecYV4NUkWf+ezYsYNNmzaRm5uL0WgkFAoxOjpKT0/PqqYcpqI210NKydTUlFKErK6ujrvuuouysjJef/11hoeHl22tQqro4/f7aWhoUDJYNBoNmZmZHDhwgOrqaoqKiqivr2fdunULwijBYJDJyckVqZyZKtpkZmaya9cuqqqqMJlMyggkHA7T2trKf/3Xf3Hs2DHa29tXfYXskh25EMIC/BD4spTSM3+GW0ophRBXHaMKIT4PfP5WDZ1PIn8z4chDoRCzs7P09PQwOjq6nKdaMjejz0poczllZWVs2LABs9mMRqPB5/Ph8XiUWhCJIfFKlmpNpbZzowQCAVpbW6mvr8ftdit56CUlJQSDQaanp5cjVp4y+kQiEUZHR5V9ODMyMtDpdBQXF2MymYhGo1RUVFBUVJSwT0kyGBsbW/bOQapoI4TAZDJRVVVFXl7eghRWn8/H4OAgR44cobu7OzlrWObvX3itF6AHXgN+f96xdqDw0s+FQPsSvkcux6uqqkp+5zvfkefOnZNSStnd3S1ff/11WVRUJHU63bKc4xqv0yulz0rYK4SQf/d3fydPnz4tg8Gg7O/vl9/73vfkBz7wAQlIk8kks7KyZFFRkbTZbCmrzUrps9SXTqeT+/fvl3/+538u29ra5OzsrHz22Wfl7/7u78qMjAyp1WrXjD4ajUZaLBb58Y9/XP7oRz+Sw8PDMhaLyVgsJiORiAwEAjISiSi/T05Oytdee03+3u/9niwoKJBGo/Gmzpvq2phMJrlv3z45PDwsfT6fjMfjMh6PS5/PJ5977jn55S9/WRoMBqnRaFaiDV617cx/XXd2Qsw9Ap8CWqWU/zDvTy8Bn7n082eAH1/vu5aDxKao2dnZSgW/RFW/ZNSNTjV9LsdoNGIymRbsu2i1WikrK2Pv3r3s37+f+vr6BSVcl4tU12apRKNRxsbGaGlpYXx8nEgkQlVVFeXl5crWXTdDKuqT6GEPDg5y4sQJZbVrMBgkHo9jMBiUFbOJ7Kd3332XCxcuMD09vWyhlVTRJpGSuW/fPu68805sNptSFdHn8yk1VNrb25WKkslgKaGVO4FPA01CiPOXjv0J8DfAc0KIXwf6gE+siIWXodVqMRqNFBQUkJWVBczdaIliNkkQMqX0WQx5KWsjEd/85Cc/idVqpaenh1deeWUlFrqkjTbXY2hoCI/Hw6FDhyguLmb79u3Mzs5SWFhIPB6/2ZhoyukjL2VfNDY20tvbi8Viwev1KrXMbTYbs7OzTE9P8+qrr3LhwgWefvrplYiNp4Q2ie3f/viP/5itW7diMpmUv42Pj9PV1cV3v/tdXC7XSppxXZaStXIEuNaSrweW15zrk+gx9Pf3YzKZKCgoYGZmhqmpqaQ8DVNNn8XIzs7mjjvuYNOmTUoNjY6ODl5//fUVWa2YTtpcj3A4zOzsLD/60Y/o7u7mK1/5Crm5uTz00EO8+eabTE5O3vB3prI+iUqAP/nJTzh//jx33XUXTqeTkpISzp07R3t7O62trYyPj69Iml2qaJOXl0dJSQm5ublYrdYFf3v//fc5ffo0s7OzSV8glnYrO+cP/ZxOJ9FolNnZWdxu95qsF3KrJEpuxmIxMjMzqampUUJRr7zyCj09PRw/flwtMnYdYrEYgUCAo0ePMjQ0xBe/+EXsdjt79uyhpaWFhoaGNVXqILGp8PHjx2loaCASiShbof385z/n1KlTuN3uNVcm+nJycnKorKzEZrMpvfFYLEYsFqOpqYn3339/WbZqu1XSzpHH43HGxsZ46qmnuOOOO5icnKS1tZX+/v4136huFCklP/7xj2lra+PJJ58kGo3S3d3N2bNn6ejoYGRkBK/Xy9jYWNJ7FOmC2+1WaneXlZVx1113cfr0aVpbWxkeHk7LglrXIxgMcvjwYaUiosfjWRPb3y2F/fv38/jjjy+oGNne3s7Jkyc5fPgwTU1NKbFZedo5cpgb9g0ODtLR0UFubi4DAwOMj4+rzugqDA4OEovFOHfuHOFwmO7ubs6fP09nZ6dyM6q6LZ3EDlSJ2uTl5eVkZWVhs9mSHiddKaSUa7Je/VKw2+2UlJRgMBiQUioT342NjYyNjaVMraK0dOTRaJSJiQkOHz7Mu+++q6TgqA7pSnp6eujt7eXUqVMAyvZUyR4KpjMej4e/+7u/4/7776euro6MjAyKioro6ekhEAgk2zyVZSSxiXKiUNrMzAwNDQ08//zzTExMJNs8hWUtjrDaJOpmqL3KayMv1dZIFMRPUmbPmiKxCKStrY2nn36a9957T9kMW2VtMTAwwPnz5/H7/fh8Pk6ePEl7ezsejyelQktp2SNXUUk2oVCI5uZmZZs9lbVJZ2cn7733Htu2bSMWi/HGG2/Q1NSUchvXqI5cRUVF5Ro0NTUpZX+llHR3d6fkfIHqyFVUVFSuweTkJJOTk0ktxrcUVtuRTwC+S/+mOk6utLN8Bc+XTtrAlfqspDag6nM9vMzVIUkH1LazODesj1jtRQxCiNNSyl2retKbIBl2pos2oOpzPVbbVlWb1DvnzXIztqZ11oqKioqKiurIVVRUVNKeZDjybyThnDdDMuxMF21A1ed6rLatqjapd86b5YZtXfUYuYqKiorK8qKGVlRUVFTSnFVz5EKIDwoh2oUQF4UQX12t8y4FIUSpEOJtIUSLEOKCEOJ3Lx3/cyHEkBDi/KXXwytoQ0rqo2qzOKo+i6Pqc22WVZvr7QW3HC9AC3QBVYABaAA2r8a5l2hfIbDj0s9WoAPYDPw58Ae3sz6qNqo+qj6pr81q9cj3ABellN1SyjDwLPDoKp37ukgpR6SUZy/9PAu0AsWraELK6qNqsziqPouj6nNtllOb1XLkxcDAvN8HWd3/zCUjhKgAtgMnLh36khCiUQjxtBDCsUKnTQt9VG0WR9VncVR9rs2taqNOds5DCGEBfgh8WUrpAf4FWAdsA0aAv0+edclF1WZxVH0WR9Xn2iyHNqvlyIeA0nm/l1w6ljIIIfTMifk9KeULAFJKl5QyJqWMA99kbpi2EqS0Pqo2i6PqsziqPtdmubRZLUd+CqgWQlQKIQzAE8BLq3Tu6yKEEMBTQKuU8h/mHS+c97aPAitVfDpl9VG1WRxVn8VR9bk2y6nNqlQ/lFJGhRBfAl5jbhb5aSnlhdU49xK5E/g00CSEOH/p2J8AnxJCbAMk0Av895U4eYrro2qzOKo+i6Pqc22WTRt1ZaeKiopKmqNOdqqoqKikOaojV1FRUUlzVEeuoqKikuaojlxFRUUlzVEduYqKikqaozpyFRUVlTRHdeQqKioqaY7qyFVUVFTSnDXhyIUQ2UKIF4UQPiFEnxDiyWTblEoIIZ4QQrRe0qdLCHF3sm1KFYQQm4QQbwkhZi5tPPDRZNuUSqj31rURQnxJCHFaCBESQnwrmbasyhL9VeDrQBjIZ65i2E+FEA0ptBQ3aQghHgL+FvgkcJK5YvYqgBBCB/wY+FfgIeBe4CdCiO1Syo6kGpc6qPfWtRkG/go4CJiSaUjaL9EXQpiBaaAucfMJIb4DDEkpU2Zbp2QhhDgGPCWlfCrZtqQaQog64DhglZduBCHEz4ETUsr/L6nGpQDqvbU0hBB/BZRIKX8tWTashdBKDRC9rAfVANQmyZ6UQQihBXYBuZfCBoNCiP8nhEhq7yHFEUBdso1IEdR7K01YC47cAnguOzbD3B54tzv5gB74OHA3c0Pj7cCfJdGmVKIdGAP+UAihF0J8gLnwSmZyzUoZ1HsrTVgLjtwLZF12LAuYTYItqUbg0r//dGl/wAngH4AV27E8nZBSRoDHgA8Bo8BXgOeY2w5MRb230oa14Mg7AJ0Qonresa3AbT8ZI6WcZs4pzZ8ISe9JkWVGStkopbxXSpkjpTzI3G7rJ5NtV4qg3ltpQto7cimlD3gB+AshhFkIcSdzu2R/J7mWpQz/Afy2ECLv0iauvwe8nGSbUgYhxBYhRIYQIlMI8QfMZfV8K8lmpQTqvbU4QgidECKDuQ0rtJfaUVIyAdPekV/iN5lL/xkD/gv4opoepfCXzG131QG0AueA/19SLUotPs3cBrdjwAPAQ1LKUHJNSinUe+va/Blz4cuvAr9y6eekzD+lffqhioqKyu3OWumRq6ioqNy2qI5cRUVFJc25JUcuhPigEKL90mITdaXXZaj6XBtVm8VR9VkcVZ+F3HSM/NKqwQ7malQMMjeh9ikpZcvymZe+qPpcG1WbxVH1WRxVnyu5lR75HuCilLJbShkGnmUuNUllDlWfa6NqsziqPouj6nMZt5LzWAwMzPt9ENi72AeEEOmeIjMhpcxd4ntvSB9VG7XtzOO200dKKW7g7eq9dRkrnrwuhPg88PmVPs8q0becX6ZqsziqPouzxvRZVtaYNtdtO7fiyIeA0nm/l1w6tgAp5TeAb8CaeDLeCNfVR9VGQW07C1H1WRz13rqMW4mRnwKqhRCVQggD8ATw0vKYtSZQ9bk2qjaLo+qzOKo+l3HTPXIpZVQI8SXgNeZqDTytLt39BcnSRwiBwWBAo9EghCAejxOPxwmHwyt96iWjtp3FUfVZHFWfK1nVJfprYIhzRkq5ayW+eDm00el0VFRU8Kd/+qfk5+eTl5fHhQsX6O7u5mtf+xoez+WlpZeVFdMG1LZzPdJdnxuc7Lwh0l0bltB20nLPTo1Gg8lkIhqNEgqFMBgM6HQ6LBYLBoMBs9lMIBAgFJqrfRSPxwkEAoTD4ZTqmS4nGo2GwsJC1q9fz/bt2xVH7vV68Xg8aDTqIl4VleUkOzubzMxMnE4nUkoCgQBjY2O43e5VtyUtHbnJZKK+vp6xsTH6+vrIz88nJyeHO+64g4qKCrZv305rayu9vb0A+P1+Lly4wODgoHJsLaHVajGZTHzqU59i27ZtbNiwAb1ej5SSoaEhuru7iUajyTZTRWVN8eCDD7Jjxw4++9nPEgqFaGlp4etf/zovvbT64fq0ceRarRaDwcDevXspLi5m+/bteDwexsfHsdvtZGVlUVNTQ3Z2NqWlpZhMJiorKwEIhUJs3LiRqakpxsbGaGhoYHR0lIGBAWKxWJKv7NYpKSmhtLSUHTt2UFlZycjICKOjo/T29vLWW2/R2dm5ZkciKiuDVqtFp9Nhs9nIzMykoKAAp9PJ+vXrkVISj8cZHBzE7/fj8/kYGxtjcHCQUCi0Ju6ppZCXl0dZWRkWiwWTycS6desoLy+nuLiYsbExIpHIqtmSNo5cr9djsVj42Mc+xpYtW9ixYwdSSiKRCDqdDo1GQ2ZmJkLMhdqqqqqUzyYm+yKRCOFwmH/5l3/h5MmTjI6OrolGt379evbt28f+/fuxWq2cOXOGo0eP8rOf/Yz29vakDPVU0hu9Xk9mZiZlZWXk5uayf/9+6urqeOyxx5BSEo1GOXz4MC6XC5fLxenTp5mZmSEWi62Je+p6CCEoLCykoqJCGRFbLBaqq6uprKxkZmZGdeRXY+/evezYsYO7776bkpISMjIyADAajUxOThIIBPD7/cRiMSU2DnOCm81mqqur0el0GI1GHnvsMbZu3crg4CDDw8OMjY0l67KWBZPJhN1uR6fT4ff7aWpqorm5mfb2dnw+X7LNU0kTNBoNDoeDe+65h5qaGurq6sjPz8disZCdnY3VakVKiZQSjUbDli1bCIVChEIhqqqqKCkp4cUXX6SrqyvZl7IqlJWVsXHjRvR6vXKssLCQuro62tra8Hq9q2ZLyjtyIQRarZaysjJ27NhBSUkJ2dnZygRmIBBgdHQUj8fD9PQ04XAYv9+vfF6j0WC32yksLMRkMmE0Glm/fj1Wq5XS0lKCwSATExPE4/EkXuWtodfrycjIQAhBKBRiaGgIl8ul9sQXQaPRKD0pvV6P0WgE5tpbIpMrFAoRjUbx+XzE43GklMroLxwOs5oZX6tBYuJu+/btbN26lT179mCz2TAYDABIKQmFQmi1WjQaDU6nU/lsogP19ttvJ8v8Vcdms+FwOBYc0+l0SvrvapLyjtxsNlNUVMSePXs4cOAAVquVSCTC5OQkb7/9Nj/4wQ8YGRnB6/UyOztLOBwmGAwqn09kc/zKr/wKe/bs4d577yUjIwOn08mnP/1pTpw4wcjIiJLVko4YDAYyMzORUuLxeDh16hR9fcu+InxN4XA4KCws5PHHH6e2tpbdu3djNBoVRx6NRjl27Bjt7e1897vfxePx4PV62bhxIxaLhfPnzxMIBFZ1+LySaLVaHnjgAbZu3coXvvAFpdMz3yG53W46OjqoqKggPz9/wecrKyspLi7mmWeeWW3TU4qjR4/y9NNPL+hMrgYp78gTqYZmsxmbzYZWq8Xn89He3k5bWxsdHR1MT08TDAYJBAJEo9EFN1fixjxz5gwZGRkUFhZSXFyMwWCgqqoKl8uF0+lkYmIi7Ry5RqPBYDCQn59PWVkZoVCIqakpXC7Xqg7r0gUhhDKBV1NTw9atW9m6dSvr1q2jqKgIg8GgtJdYLMaGDRvQ6XTU1tYSDAaJRqPU19djMpmYmppiYmKC0dHRZF/WLWM0GrFYLGzevJna2lrlPksQjUbp7u5mYGCAkydPsm3bNmpqaigpKVFGMjqdDq1Wu+BztyOhUCgp917KO3KtVktGRgYWiwWLxQLA9PQ0L730EqdOneLChcUXdEkpmZqa4vnnn2d4eJhAIMBjjz3G+vXr2bFjB+FwmI0bN3LhwgVmZmZW45KWDYPBgMPhYMuWLRw4cIDm5ma6u7vp6elZME+gModWq8VisVBfX8+hQ4f41Kc+RXZ2tuKMEvFfmHP6W7Zsobi4mGAwiMlkwuFwUFdXh1arxePx0NzcvCYcucPhoKioiIcffpjNmzdf4Yz9fj8/+MEPOHfuHD/96U95+OGHueuuu3jyySfJy8tLktUq80l5Ry6EQK/XK8vNW1paaGlp4d1332V4ePiGvqu7u5uf/OQn1NXVkZubi91ux2azsWXLFlwuV9rlmOfl5fHAAw9QXl6uHLvRuK3BYKCwsJCMjAxMJhMwl+UzNjaG3+9f6dWgq4YQgvz8fNatW8cXvvAF1q1bh81mIxwO43a7eeONN/D7/Qvy7RN/7+rqYtOmTdTV1WE0GgkGg8zOzqb9RLLRaMThcLB37162bNlCUVGR0gZCoRA+n4/XXnuN9vZ23nrrLVwuF9FolKamJqanp5menmb9+vV85CMfwWQyYTAY+PjHP05FRQX/8R//kfb6LEZilG+xWNi/f7/SGdi8eTMf+MAHOH78+KreOynvyBM9cp1Oh5SSrq4umpubaWxsvOH4ZCK/emhoCI/Hg81mw2KxsGHDBk6dOrVCV7ByZGdns2/fPgoKCgCU1K/rOfPEEFij0WCxWCgvLycrKwu73Q6gpHROTU0RCASIxWJpPRmcmDDPz8+npqaGD3/4w8qCqfHxcUZHR3n11Vdxu93K/IpGo6GgoEB53/r168nNzUUIoTjy+XMx6YjRaMTpdFJXV8edd95JTk4ORqMRKSXBYJCpqSlef/11jh07Rl9fH9FoFCklPT09DAwMMD4+Tn19Pffee68S5jtw4ABlZWW88MILymTxWqW7u5ucnBxlfgWgvLyc3bt309jYqDry+ZSXl/Mbv/Eb1NbWEo/H6e7upqur66YzBqSUtLe3k5ubS0FBAVarldraWrKzs5fZ8pXH4XBwxx13UFBQQCwW49133+Xs2bPXdboPPPAAe/bsYevWrTgcDnJycjAYDEpjjMfjuN1uenp6eOONNzh27Nh1Q1ipTHZ2Nrm5ufzhH/4htbW16PV6ZVXwj3/8Y1paWjhx4gTBYHBBaCU3N5eqqip+67d+i3Xr1mGxWDh8+DDNzc2cPHmSycnJJF/ZrZHoxOzYsYN9+/ZhsVgQQhCLxejs7OTtt9/m/PnzDAwMKE4c5tpHJBKhu7ubWCzGK6+8wrZt29i7dy8FBQVkZmbyO7/zO5w5c4bnn38+yVe5MgghePDBBzl06BCZmZnJNie1HbnBYMBms1FeXo7ZbCYajTI9PY3H47ml1K+pqSllMZBOpyMzM3NBLmg6IITAaDRis9nQ6XSEQiF6enro7e29wpELIcjIyMBqtVJQUMC2bdvYsWMHW7duxWq1Kr2wWCyGwWBAq9VSUFCAyWTC5XJx8eJFOjs7iUQiaZly53Q6WbduHTU1NRQXF+PxeBgcHOTChQs0NzfT1tbG5OTkgoUsQgiysrLQaDSUl5eTk5ODVqtlZGSEnp6eBb33dCUjI4OCggIlR1yj0RCNRpmdnWVoaIjm5mYmJyevOt+S6LVPT0/T3NxMdnY2W7ZsUeazampqcLlcSbiq1cPhcJCbu9RNn1aWlHXkiaFwYmmwRqPB6/XS39/P4ODgLTmUrq4uTCaTEkJINxJOPJGv6vV6CQQCvPPOOzQ3Ny9w5Ikhb01NDfv37+cP/uAPcDgcWCwWdDod0WiUmZkZRkZG6O3tpbKykuzsbPLz86mqqiIrK4uenh46OztxuVxpmW63d+9eDh06RHl5OUIITpw4wTvvvMNLL73E4OAgs7OzV7QnIQTFxcVUVVWxfv16TCaTMpo7e/ZsWupwOQ6HY0FoLh6P4/P5aGho4K233uL73//+dVdpjo2N8c1vfhO/309dXR1lZWUYjUbWrVtHd3f3alyGCinsyHU6HTU1NVRVVWG32/F4PExOTiqxuVtx5Ika3emKTqdj48aNVFVVYbFYGB4eVupeJK5Lr9djMBiorKwkPz+f+++/nw0bNpCdnY1Go8Hv99Pf38/ExATNzc2MjY0xNDREbm4u2dnZfOhDH1Jq2FitViwWCxMTE2nlwHJzc9m6dSu7d+9m48aNStz/yJEjNDU1MTY2tiCckiAxwV5fX09dXR16vR6fz8f09DSDg4MMDQ2l/TJ0o9FIVlYWRUVFmM1m5XggEODChQsMDw8vOb4djUYZGxujpaUFu91Obm6uUg9fZXVIWUduMBjYsmULmzdvJicnh4GBAdrb2+np6bnhbJW1hsFgYOfOndTW1mK1WpmcnKS1tXXBUN9gMGC329m9ezf19fV89rOfVdI3Z2Zm8Hg8HD16lPb2dl588UXcbreSfpmdnU1BQQGbNm1i586dZGVlYbFY0q4UbnFxMZ/85CfZt28fmzdvVlbx/vSnP2VkZISJiYmrfi4xitm3bx/btm1Dq9Xidrvp6uqis7Mz7RdbJeoS2e12ysvLsVqtyt98Ph9nz56lv7//hr5zbGyM06dPs3nzZjUlMQmktCPft28f69evx+fzcf78eV5//fU1kw53K+h0OsrLy5XVdYnQyPzecnV1Nffddx8HDx5k3bp1mEwmPB4Pvb29vPfee5w7d05JI7u8Uls8HsflcilD7nTFYrGwfv167Ha7srx8dnZWCadcjhACjUbD7t272bJlCzt37lTCMR0dHTz//POMjIwk4UqWF61WS3l5ORUVFeTl5WEwGIjH4/T29nLhwgXOnDlzS/nxifpGWVlZZGdn4/P51HUNK0zKOvLEhJvdbicYDDI0NERHR0faTzAtB4n6MYkediLnNxFWSWi3detWNm3aRHFxMYFAgImJCdrb2zl9+jTvv/8+AwMDV73B4vE4fr8fv9+vVIycn7WQLuj1eux2u1IrJBaLEYlE8Hq9V1y3VqtFr9crBdZ27txJYWEhWVlZxONxxsfHaW5uXhMdCa1Wi9PpVDZGgDlthoeH6evrY2Rk5KoPuqWSmMNJTHyGw+HbxpEnOgurHbpNWUeewO/3MzU1RUtLC42NjWs6L3WpaDQacnJysNlsV/zNYDBQVFTE1q1beeihh3A4HAQCAV555RVOnz7NM888g9frVZacL4bX66WtrU0phbBWb8bExHplZSX3338/999/Pzt27FB2oUrExjs6OtbEIhedTkd1dTUlJSXKsWg0yg9/+EPOnj3L9PT0Ld1nGo2G3NxcSkpKWLduHbFY7LYpGaHWWpmHRqNRqszBXCO7vIbK7Uxig2WdTocQQnnBXEpZ4ia1Wq243W4mJyc5ceIEFy5cYGJi4rqLhhI1SYQQyvZ4oVAo7SaII5EIHo9HqdKn0+nIzs7m3nvvXVBXJ+HIS0pK2L59OyUlJcoEYCgUwuVyMTU1pZRJTncS6aiJkUpi1fTQ0BDDw8O3fI2JBVg6nW7BfXw7oNZamUci4+J2awRLJXEjJhbwzD9us9n44Ac/yNatWzGbzZw5c4ampiaeeuqpJZe1TXy/Xq9XHqLp6MCCwSCDg4PK5JvZbGbDhg387d/+7YIHWaIOd2ZmplKWNPH3QCBAU1MTfX19q97LWikSHYFE6YsEXV1dt5wyqGarJIeUdOSJxRv5+fno9fplj0uWl5cr6WiJbavSqbeZKO50ea9aSolWq8Vmsykbb5hMJqxW65JvrszMTGXZcWZmJhMTE2kbUhkYGOA///M/8Xq9RKNRKioqMBqNFBUVLdAuEdMdGhri+eefp7y8XNnI2uv1cuzYMXp6epJ4JcuLVquluLh4RRazJOq2p9t8SrqTko68qKiIyspKHA7HsjtYjUZDcXExlZWVyorIdAwbJLiaU0+MZOLxuLIFVWLF5vV61larldzcXDZs2EAkEqGrqyvtyvsmGB8f580336S4uJi8vDxyc3OVvPgECe0SmTqvvPKKkrJZWlqKz+ejqalpTaW8JjaFsNlsSrtZTsd7rY6GysqRko58pUjkR+/atYtdu3ah1+tpa2vj3/7t32hubk62ecuCy+XiH/7hH/j4xz+ubARrtVr51Kc+RVtbG++///6im2h8+tOf5s4778RoNNLR0cGzzz5Le3v7Kl/F8hCPxwmFQrzwwgu89dZb7Nq1i9zcXCoqKpQRyujoKG63m/Pnz5Obm8udd97JXXfdRW1tLV6vV9kOMN1KHCcTKSWBQACfz8fMzEzajujSidvCkWs0GqXSW2Klo81mIxAIMD4+TmtrK1NTU8k2c0lotVolvnm1Iv7hcJje3l7a29tpamqiqqoKo9FIfX09RqORqakp+vv7GRsbIx6Po9VqMZvNWCwWsrKyqK2tZf369UxOTjI4OEhXV9ctpaIlGyklExMTTE1NodPpyMnJYWpqSnHkLpeL6elpGhoa2Lp1K+vXr6egoICsrCxluzyfz5e2o5LVIiMjQ6meGIvFlEqjHo9nzSYpBAIBvF4vZrNZaU82m42CggImJiZWNcPuuo5cCFEKPAPkAxL4hpTyH4UQ2cD3gQqgF/iElHJ65Uy9eYxGIzU1NezYsYP77ruPdevWYTAYaGpq4vz585w6deqmG9tq65NYZJGTk7MgRJAgFovh8Xj4yU9+wtGjR/njP/5j9u3bxyc+8Qk8Hg8PPvgg//7v/85Pf/pT/H4/ZrOZffv2sW/fPh544AFqamrQ6XQ888wznD59mpaWlpsOO6VS24nH47S3tyu1VhLMD60UFRXx6KOPKmGpnp4eLl68qOwFu9ykkj63SkVFBQcPHiQ3Nxe/38+zzz7L2bNn6ejouKn2kw7adHd3c+HCBbZv365kAO3du5dwOMy3v/3tVS0atpQeeRT4ipTyrBDCCpwRQrwO/BrwppTyb4QQXwW+CvzRchjl9/vxer3LcvMIIbBYLOzYsYMdO3YoZUynp6c5evQoTU1NRCKRW4mRr6o+oVAIv9/PzMwMOTk5CCFwOByUlJRgMpmUOHiinvTPf/5zenp62L59O36/n66uLlwuF5mZmcoOOPfffz8VFRUUFxczMDDAxMQER44coaen51bnDla97SzGfKc9n0QbMZvN6HQ6ZSPhxCYmKzh/khR9wuEwJ0+eJBqNUl1drcyf7Nu3D6PRyKlTp274mvV6PRaLhXg8jtfrpbGxkYsXL6bNfXUzXC1JIrGwbLUzd67ryKWUI8DIpZ9nhRCtQDHwKHDg0tu+DRxmmQT1eDxMT08TCATQarW3JIpGo8Fms3HfffdRV1fHtm3bmJycZGhoiFdeeYWurq5bulFXW5+EEx8fHycnJweY2ymouroaq9WKXq8nFosp+d8//OEPycjI4O677yYcDjM0NEQ0GsVqtXLo0CHq6ur40Ic+hJSScDjMkSNHOHfuHK+//vot58Mmo+3cDIn0w8QIJxqN4vf7OXXqFA0NDSuWepksfYLBIK+//jqhUIiHHnpImQw/dOgQeXl5S6ppfzkJRx6JRJienubkyZO3VM4gHdpOKqVZ3lCMXAhRAWwHTgD5l8QGGGVuCLQsJHYjGR8fx2azKZkXVqt1wVL0q5GZmUleXh6bNm2iqqqKkpISCgoKuPfee5WVkD/+8Y85efIkTU1NyzqJtVr6+Hw+nn76aXbv3o3dbsdut7Nnzx5+//d/n+7ubt555x2ltC3MxS9ramrIzs4mLy+PvLw8HA4HZWVlZGRkMDo6SktLCydPnuTtt9+mp6dH+exysVra3Awmk4mPfOQj7NmzB4D+/n46Ozvp6OhgaGhoVbIvVlOfaDRKf38/p06d4lvf+hYPPvgg27dvZ9u2beh0OtatW8f4+PiSNs7Izc3lYx/7GPfddx82m43W1lY6Ozuv2DbvVki1tpOTk0NhYSF1dXVs3LgxJfYyWLIjF0JYgB8CX5ZSeuY/jaSUUghx1dYuhPg88PkbMWp2dlZZSWexWDCZTGRnZ1NYWMjMzMyiDcRqtVJZWcm2bduoq6ujqqqK7OxsioqKlHrLicJA4+PjyzYRczP63Iw2MDc0bmxsxGQy0dPTw7p16ygoKGDPnj0UFhYyOTnJzMwMXq9XyZEuKyujsLCQTZs2UVJSQm5urrIKraOjg+bmZmUnoLGxsVuR4gpWs+3cKFqtFpPJRG1tLRUVFcDcxiMDAwNMTU2tyiq91dZHSsns7CwjIyOcO3eOzZs3s3HjRpxOJ0VFReTn5xMIBK7ryPV6PQ6Hgz179lBZWYler2d8fJz+/v5l24QkFduOzWajqqqK/Px8ZQFZslmSIxdC6JkT83tSyhcuHXYJIQqllCNCiELgqne/lPIbwDcufc8N/8+azWYKCgr49V//dT7wgQ8wNTW1qPN1OBzU1NQoRaW0Wi1SSqXyX1tbG6dOnaKjo2M5eww3pc/NahOPx5mYmODdd9+lu7ubz33uc3z4wx+mvLycqqoq9uzZo8SCE8v3E1kuOp1OiaO/8847tLS08PTTTzM9PX1dbW+GZLadpVBQUEBFRQX33nsvRUVFAPT09HDixIlVqauSTH1mZmZoaGigrKwMKSV33303drud7du3E4vFFi3Xq9FoqK2tpb6+ngMHDmAymZiamuLll1/myJEjy7IKNlXbTl1dHZ/73OcWbHqebJaStSKAp4BWKeU/zPvTS8BngL+59O+Pl9OwaDRKR0cHMNfLtlqtVFVVkZeXt2jMMuH4E0vMI5EIfr+fzs5OWlpaOHXqFKOjo4RCoeXqMSRFn0SFwuHhYc6dO4fZbGbjxo1YrVYyMjKw2WxkZ2czNjZGIBBYcK1TU1NMT0/z3nvvKfXdA4HAsuf7JkubG8FqtSoZQInVsOFwmEAgsOKLxJKtTyQSwe12c/HiRcxms7KJRl1dnTJZ3t/fr4xKEhuhZ2ZmYjab2b17N7W1tdhsNmZnZ3G5XAwNDeFyuZajXkvKtR2dTofNZqOoqIiKiooFe3UmCvudP3+e9vb2Va/SupQe+Z3Ap4EmIcT5S8f+hDkhnxNC/DrQB3xiOQ3z+/288MILbN26FYvFQklJCcXFxUv+fKJH6vP5GBoa4uWXX+b48eO88cYby2kmJEkfQJlYeu655/jpT3/KAw88QEFBAcXFxWzZsoV9+/Zx/PhxpWeVcObHjx+nubmZvr6+lW5wSdNmqSR65IkJv8TepeFweDVi40nVJ1EQ7N1336WtrY0777yTjRs38rGPfYzq6mpKS0v5zne+Q1tbGzA311JYWEhFRQUlJSV88YtfpKKigqysLLq6ujh8+DCtra3LVbM95dpORkYGmzZtora2lrq6ugV/a2lp4X/8j/+hzKusNkvJWjkCXGt69oHlNecXRCIRWlpalCH/oUOH2Lt3L06nc9F9Nqempmhvb2dqagqPx8Pp06cZHR2ls7PzlorlX4tk6TOfhNM5c+YMZrMZs9nMkSNHeOGFF+jv779iQc/o6CiTk5MrvlAjFbS5HiaTSal0mOiFdnR00NjYuOJFslJFH5/Ph8vl4tvf/jbV1dUcOnQIo9HIPffcg8lkor+/n4aGBvLy8rj33nvJycnB4XBQXFyMlJKWlhbOnj3L0aNHmZ5enpTuVNFmMRI7KR09epSenh46OjqWXJhuuUnZlZ2JGN3o6Ci9vb0UFBRQWlqKwWC4ourffMbGxjh79ixDQ0OMjo7yk5/8ZEmz7+lMLBYjFotx8eLFZJuSdiRCBUIIQqEQ4+Pj9PX13VYbByfqDb366qu0t7dTXV1NZWUlNTU15OTkMDk5idFopKSkhEcffRSz2UxGRgaxWAy3201HR4eyX0A6rwJeClJKZXOSpqYmzp49y9NPP530eutiNQvb3MykQ6I2dqLIj9FoXDR/MxwOMzs7SyQSUWKAy5gHfEZKuWu5vmw+KzWZt4qsmDawMvoIIfiLv/gLDhw4wM6dO+nr6+PFF1/kpZde4vjx48t9upTXR6vVkpmZSXl5Obm5uRQUFPDAAw9QWVlJeXk5U1NTNDY2otFokFIqo92WlhbcbreyIcXNzC1IKVcsKXu52o5Go8FkMuFwOHA6nbjdbvx+PxMTEys9n3LdtpOyPfIEiSfgyMjImtgvUSW1MBqNmEwmhBBEIhGmpqZu2+0EY7EYPp+Pnp4eJicnGRsbo6CggFAohMFgwOVy0djYqMw/nTt3ThnBXG+zkrVAYs7N5/MxODiYbHMWkPKOXEVlpZBSMjY2xuDgIBs2bEi2OSlBwln5/X5GR0fp6OhQNqVOTATPf69arjY1UB25ym2Ny+Wit7cXl8ulvJZ7VWs6knDOa7Vy4VpDdeQqtzUDAwOYzWbq6+vp6+u7apaPikqqozpylduapqYmenp6OH/+PIFAgMHBwVVZ0amispyojlzltmZycpLJyclFl6OrqKQ6q+3IJwDfpX9THSdX2rmSxRXSSRu4Up+VLjyh6rM4XiBd9uRT287i3LA+q5pHDiCEOL2S+bTLRTLsTBdtQNXneqy2rao2qXfOm+VmbNWslDEqKioqKquD6shVVFRU0pxkOPJvJOGcN0My7EwXbUDV53qstq2qNql3zpvlhm1d9Ri5ioqKisryooZWVFRUVNKcVXPkQogPCiHahRAXhRBfXa3zLgUhRKkQ4m0hRIsQ4oIQ4ncvHf9zIcSQEOL8pdfDK2hDSuqjarM4qj6Lo+pzbZZVm0TRm5V8AVqgC6gCDEADsHk1zr1E+wqBHZd+tgIdwGbgz4E/uJ31UbVR9VH1SX1tVqtHvge4KKXsllKGgWeBR1fp3NdFSjkipTx76edZoBVY+r5yt07K6qNqsziqPouj6nNtllOb1XLkxcDAvN8HWd3/zCUjhKgAtgMnLh36khCiUQjxtBDCsUKnTQt9VG0WR9VncVR9rs2taqNOds5DCGEBfgh8WUrpAf4FWAdsA0aAv0+edclF1WZxVH0WR9Xn2iyHNqvlyIeA0nm/l1w6ljIIIfTMifk9KeULAFJKl5QyJqWMA99kbpi2EqS0Pqo2i6PqsziqPtdmubRZLUd+CqgWQlQKIQzAE8BLq3Tu6yLmNgF9CmiVUv7DvOOF8972UaB5hUxIWX1UbRZH1WdxVH2uzXJqsyrVD6WUUSHEl4DXmJtFflpKeWE1zr1E7gQ+DTQJIc5fOvYnwKeEENsACfQC/30lTp7i+qjaLI6qz+Ko+lybZdNGXdmpoqKikuaok50qKioqaY7qyFVUVFTSHNWRq6ioqKQ5qiNXUVFRSXNUR66ioqKS5qiOXEVFRSXNUR25ioqKSpqjOnIVFRWVNGdNOHIhxGEhRFAI4b30ak+2TamCEGKTEOItIcTMpcL6H022TanCvPaSeMWEEP+UbLtSCSHEd4UQI0IIjxCiQwjxG8m2KdUQQlRf8j/fTZYNa8KRX+JLUkrLpdeGZBuTCgghdMCPgZeBbODzwHeFEDVJNSxFmNdeLEABEACeT7JZqcb/AiqklFnAR4C/EkLsTLJNqcbXmavpkjTWkiNXuZKNQBHwtUvV1N4CjjJX30FlIR8DxoD3km1IKiGlvCClDCV+vfRal0STUgohxBOAG3gzmXasJUf+v4QQE0KIo0KIA8k2JoURQF2yjUhBPgM8I9XiQ1cghPhnIYQfaGOuPvbPkmxSSiCEyAL+Avj9ZNuyVhz5HzG3J18x8A3gJ0IItdcA7cz1Mv9QCKEXQnwAuBfITK5ZqYUQopw5Xb6dbFtSESnlbzK3p+TdwAtAaPFP3Db8JfCUlHIw2YasCUcupTwhpZyVUoaklN9mLnywYrtypwtSygjwGPAhYBT4CvAcc9tdqfyCTwNHpJQ9yTYkVbkUmjvC3MYMX0y2PcnmUpnZB4GvJdkUYJXqkScByVwI4bZHStnIXG8TACHEMdSe5+X8KvA3yTYiTdChxsgBDgAVQP/c/hBYAK0QYrOUcsdqG5P2PXIhhF0IcVAIkSGE0Akhfhm4B3g12balAkKILZe0yRRC/AFQCHwryWalDEKI/cyF5NRslcsQQuQJIZ4QQliEEFohxEHgUyR5Yi9F+Aa/2FdzG/CvwE+Bg8kwZi30yPXAXzGXoRFjbkLmMSllR1KtSh0+DfwGczq9Bzw0LwtBZW6S8wUp5WyyDUlBJHNhlH9lrtPXx9wGwSmxVVoykVL6AX/idyGEFwhKKceTYY+6Q5CKiopKmpP2oRUVFRWV2x3VkauoqKikObfkyIUQHxRCtF+q4fHV5TJqraDqc21UbRZH1WdxVH0WctMxciGEFugAHmIuL/kU8CkpZcvymZe+qPpcG1WbxVH1WRxVnyu5lR75HuCilLJbShkGngUeXR6z1gSqPtdG1WZxVH0WR9XnMm4l/bAYGJj3+yCwd7EPCCHSPUVmQkqZu8T33pA+qjZq25nHbaePlPJGFvCp99ZlrHgeuRDi88yVT10L9C3nl6naLI6qz+KsMX2WlTWmzXXbzq048iGgdN7vJZeOLUBK+Q3mVkGthSfjjXBdfVRtFNS2sxBVn8VR763LuJUY+SmgWghRKYQwAE8At/2Kr3mo+lwbVZvFUfVZHFWfy7jpHrmUMiqE+BLwGqAFnpZSXlg2y9IcVZ9ro2qzOKo+i6PqcyWrukR/DQxxzkgpd63EF6vaLI6qz+Kkuz43ONl5Q6S7Niyh7aR10SydTofBYMBsNmM0GrFYLOh0OoxG44L3xWIxgsEg0WiUaDTK9PQ0wWCQUEitHaVyeyGEQKPRoNPpyMjIICsrCwApJTMzM4TDYfW+SEPS2pE7HA6Ki4vZt28fFRUV3HvvvTidTgoKCtBofhH+9/v9dHZ2Mjk5icvl4sUXX6StrY3e3l5isVgSr0BFZXUxGo0YDAby8vJYt24dDz30EEIIotEor7zyCoODg3R1daEW00svUt6RZ2VlUVZWhtlsJiMjg9zcXPR6PQB5eXmUlpZSXl5OTk4OJSUlmM1mMjMzGR4eZnJykoGBAaampmhqasLr9TIzM0Nvby8zMzNqY1W5bdBoNGi1WtavX690foqKiti4cSMA8XiczMxMRkdHOXv2LP39/Vy8eJFIJEI8Hk+y9elJVlYWd9xxB16vl4mJCYaGhvB6vStyrpR35NnZ2ezfv5+CggJycnLYunUrZrMZmHPkJSUlVzjkeDxOd3c3jY2NvPvuu/T393Pq1KnbskFe2r1kAeoD7PZDq9WSkZHBjh072L59O5/97GexWCwL3nPXXXcxNTXFW2+9xZtvvsno6Cizs7OEw+EkWZ3eOJ1OfvmXf5mhoSGam5vxer23ryOvrKzki1/8IhkZGej1ejIzM4lGo4yNjdHd3c25c+eYmJjA7/fj8/mYnJykra2N6elpZmZmmJqawu/33zZO3GQyUVVVRVFREZWVlWzcuBGn00lWVhaBQIDR0VG6u7sZGBjg8OHDuN3uZJussgpUV1ezd+9eHn/8cTZt2oTJZLriPRqNhqysLO666y4qKiq47777ePHFF2ltbaW9vZ1oNJoEy9MTIQSZmZnU19djMpmYmJggIyNjxc6X8o7cYrFQU1ODXq9Hq9Xi9/txu91MTEwwPDxMX18fIyMjzM7O4vF4GBkZ4ezZs8k2e1VJTGBlZWVhs9moqalh3bp1bN68me3bt1NQUEB2djY+n4+BgQHy8/NxOp2cP3+eQCCQ1pNbQggMBgMmk4nMzMxrvk9KSTgcVh7oUkri8TjBYJBYLIZOp1OOR6PRNffgz8rKorKykurqaiorK4FfJAHEYjHi8ThCCLRaLQUFBdjtdioqKuju7iYWizE2NobX6yUYDKojuiWQmEx2Op1MTk6SmZmJVqtdufOt2DcvE+Pj47z11lvU1tZSWlrK0NAQDQ0N/N7v/R6BQIBYLKY0RCnlbTl5aTabsdvt/NEf/RG1tbVs3rwZg8GAXq/H6/USDocZHBxEq9WSn59PcXExDzzwAFNTUzQ0NPDuu++mreOyWCzs3LmTD37wg3zqU59Co9EsmOgGiEajhEIhTpw4oQxtZ2dnmZ6e5vXXX2dkZISNGzcSj8fx+Xz09fXhcrmScTkrhtfrZWRkhGAwqBwbHBzkrbfewuVy4Xa70Wq1FBYW8uSTTypO6Itf/CJTU1N873vfo7GxkVdffZVwOHxb3mdLRQhBeXk569evJzs7+6qjn+Um5R359PQ0J06cICsrC6fTiVarVXrmPp+PSCSSbBOThhACvV7P+vXrqampoba2lvLycjIzM5XJlZGREbxeL4FAAIPBgM1mY8OGDRQUFFBSUsLk5CQajYbMzExsNhs+n49wOEwgEEiLnpdWq8Vut5Obm0tRUREajeaKeYFYLEYkEmF2dha/f26bRb/fj9frJR6PMzExQUVFBVJK/H4/PT09DA0NMTExQTAYxO/3E4lEiEQiSuch3TCbzeTl5WE0GonH44yOjnLx4kWOHz/O5OQkPp8PjUbDyMgIlZWVSu/darWi1+vZvn07BoMBl8tFT08Po6OjadE+bhWDwYDRaKSqqgopJV1dXYTD4UX9jhACnU6HTqdDo9EQiUTweDwrGppKeUfe09PDv/3bv2G1WikvL8dqtZKbm4vT6QS4rWO8er0eq9XKRz7yET760Y+yfv16dDodY2NjvPXWWzz33HN0dHQwNTWF1+vFaDTidDr5/Oc/z8MPP6xkLPzgBz+guLiYrVu30t3dzdTUFP39/WkRE9Xr9TidTqxW6zWHromH//bt26/424c//GGklAtCK729vQwNDXH48GFGRkbo7e3F7XYzNTXF0NAQPp9vRa9pJSgoKGDPnj3Y7XaCwSBnzpzhvffe46mnnlrgkC0WCxcvXuTRRx+lsLAQi8WCxWLhwx/+MNu3b6esrIzvfe97jI+Pp0X7uFVsNht5eXl84QtfIBqN8o//+I9MTk4yPT295O/weDz09/crnYiVIOUdeTQaZXZ2lra2No4fP84999yD2WzGYrEwO3t7b3yem5vLXXfdRX19PUVFRYTDYXp6evj6179OV1cX7e3tuN1ugsEg8XiccDjM9PQ0hw8fVjISZmdnqamp4c477+TDH/4wHo8Hl8vFX/7lXzI9PZ3yvU+v18vx48eRUhKJRMjKykKv1+PxeBQHpdVq0el0VFVVXTHMtdlsZGRkoNPpEEIghMDpdGIymcjIyMDn8zE7O0sgEMDn83Hs2DG6u7s5duxYWs0tZGdnU1tbi8ViIRwO8/bbb9PY2HhFrzoUCtHd3c3bb79NNBrlzjvvpKysjOLiYux2O7t27cLtduN0Onn11VeZnJxc0z3znJwcKisr2bx5M9FolNLSUuLx+HUdeXZ2Nna7HSEEoVAIt9u9otGDlHfkiQmpoaEhWltbufPOOzGZTMpqzsvjoVLKNd2wEmi1WpxOJ3v27KGqqors7GyGh4fp7u7mP//zP/H7/Vc44Vgshtfr5fz583R1dRGLxcjMzGTTpk3s2rWLgwcPEo/HGR4e5mtf+xozMzMp78gDgQBtbW3A3EO/oKAAo9GIy+VS4v6J1b7hcFhZyQhzQ+CSkhLF+Sfaktlsxmq1UlhYqLw3EZ4xm82cO3eO06dPp5Ujz8rKorR0rmBgog10d3df8b5IJMLIyAhSSqanp7FarWg0GvLz85XEA5/PR1ZWFidPnmRmZmZNhzdtNhvFxcWUl5cTjUbJz89ncnJy0c8IIbDZbNjtdgDC4bAaWplPIjvDaDRSVlampCPOx+/3MzY2RiQSSXkndLMYjUbuuOMO7rjjDp544gkMBgMTExN89atfpaGh4apOfD5utxufz0dBQQFVVVX88R//MUVFRQC0tLTQ3t6O1+tNm6FzNBrl4sWLjIyMLIhLJkj0tF966aUrHvwVFRXk5OSwbds2DAYDAJs2baKyspKKigql3INWq0Wj0XD33XdTWlrK97//fSKRyILJw1TG5XJx+vRp1q9fj5SSYDC4aH74xMQEs7OzfP/73+fs2bOUlpaSk5NDZmYmGzZsoKioiKNHj5Kdnc3JkyfTdrL8epSWlrJ9+3YyMzOZnJxkYmLiurngWq2WAwcOsHXrViXlt6Ojg0AgsGJ2prwjT6REZWVlkZOTo/SuNmzYQG5uLiUlJQve73a76enpwe12Kwn4a82hJ1LECgsLycvLw+PxMDs7S0dHh9LTXoxElk8i7a6srEzpqXo8HiWkkk4jm2AweF2nOjY2dsUxn8+HzWYjHo8rjjwYDOJ2u/F4PEq7S4wCs7OzCYVCSihreHh4Ra5nuUl0cMrKygAW/P8mVkNnZWUhpcTtdit/83q9TE5OMj4+rnScMjMz0el05OXl4XA4rrrobC2g0WhwOBwUFBQQCASYnp7G7XZf1yELIXA4HNjtdmKxGOFw+Lqdq1sl5R25Xq8nKyuLHTt28PDDDyvD4D/6oz8CfrFKUUqJEIKRkRFOnjzJmTNnaGtr47333mNmZiaZl7Ds6HQ6ysvLlZoyU1NTStmBpQ73E0Pn6enpBQ47kbGyVntYlzM6Osro6CidnZ2KQzIYDBgMBpxOJ6WlpTz55JNs2bKF3bt3YzQayc7O5rHHHuP06dM899xzSb6CpeHxeOjt7WXz5s1XzBPU1dVRX1/PAw88QDQa5eWXX1acjhCCjIwMjh49qmQ7JY7n5+dTUFCwJh25VqvFZDJRWVnJ1q1buXDhAk1NTVy4cGFJ91higj2xXmGlO5Mp7cgzMjIoLy/n/vvvZ+vWrdhsNiUml5io83g8wJxzSwwbS0pKMBgMbNy4kbq6OoaGhnjjjTfw+XzKkzGdepuXI4TAbrcrS6z7+/s5c+bMDQ3dhBBKr35+fDgRZ19ro5hrkWgH8683Ho8vCM0k0l93796tLKLp7u5Om944zPXIXS4XoVAIs9mM0+nE6/USiUTYunUr9913H5s2bVLWY8x/kGu1WoqLi8nPz1eOaTQaqqurCYfDrF+/XukUrJXFVAaDgZycHOU+m5ycZGxsTBnNXovEQiC73U5GRoZS62mlSVlHrtFoMJvNbN68md/+7d8mPz+frKwsZfl9X18fg4ODyoSN2WzmscceUyoi1tfXK0PFlpYWuru7GRkZweVyKavZ0hWtVkt2djZWqxWA9vZ2Dh8+fENZPBqNRskXNhqNSu+hu7ubCxcurOkJrOsRj8eJx+PKMPrtt9+mvLwcmBuxeL1ezp49S1/fsm/DuWLMT4FzOp3k5+crPcs77riDxx9/XEnB3Lp163W/T6PRUF9fj9ls5vTp03R1dREKhfD5fGvCkWdkZCj1nbKyshgbG2N0dPS615YIPzmdTsxmM6dOnbpqSG+5SVlHnlhlVlJSQnFxMSMjIzQ2NvKv//qvDA4O4vP5CAaDSi9Uq9Vy8uRJMjMzsVqt1NbWUllZSW1tLUajkb/+679mcnKS0dFRnn/+edra2hZkNqQLOp1OWc2YcC5+v5/JycklT05qNBoyMjJ45JFH2LZtG0ajESmlMjEzODiY1g+65cRsNrN7924qKiqAuYnmrKws9uzZg8lkSpsVoKOjoxw/fhy73U5OTg6nTp3C7/ej0+mIRqOEw+Grroq9Fok48KZNm/jKV76Cx+PB7Xbz0ksv0dbWRnNzs7IHQDpitVrZtGkT2dnZxONxLl68yMWLF6/rL9atW8eGDRvIz89HCEFrayujo6Mrbm/KOnKYc1qJgve9vb10dnZy7NgxBgYGrvr+7u5u9Ho9FosFt9vN+Pg4QgiKi4vZsWMHgUCAmZkZLly4gN/vV2LK6eS0NBoNer2e/Px8HA4HgJI9sdRwUaLXUF1dzfr169FqtQQCAWZnZ3G73czMzKTdA24lSKx4Xb9+Pbm5ucAvVu1ZrdZFa7ukGj6fj+HhYc6ePYvVamVkZASYS0v0eDxMTk5it9uV2G4i934xjEYjRqMRh8OhpGcODQ0hhGBqakpZRJWOoczMzEzKysqwWCxEo1E8Ho8yMQ4oD7xErZ8E69atY9OmTZjNZmKxGNPT0yu6EChByjrySCTChQsX6Ozs5Ac/+IGyu8/1wgeRSAS3282xY8c4efIkzz33HLm5uXzoQx9Sikh94QtfwOPx8Md//Mf09fWl1RA5wdWWol+PRAZQRUUFVVVVVFZWkpeXhxCC7u5u3n//fVpbWxkbG7vtHbkQgry8PGpra/mN3/gN5aEZCoWYnJzklVdeob+/P8lWLp1IJEI0GuXkyZNoNBolrOLxePiv//ovGhoauPvuu8nJycFisVBWVkZ1dfWSvz/h/D/zmc/wS7/0S3R0dHD69Gn+7d/+Tanlki7o9XrKysr4+Mc/jsPhwOv1YrVa2bx5M7/6q7+qrHY1Go1kZGRQX1+v7JFgNBqVzuTIyAhGo/G6D8TlIGUdOaDUt7jRJdFSSkKhkBKzi8ViNDY24vf7CQQC3HHHHUqd85ycHCKRCDMzM2mz9DoWixEIBAiHwxgMBiwWC7m5uYs+kIxGo1JvY+fOndTW1mKz2ZSexfT0NG1tbXg8ntveicMvsjIKCwux2+1Kpsfo6Ch9fX2r1tNaThIVIOcTi8VwuVzKSC8rK4vMzEwKCgro7OxU3mcwGJSOQzQaVUaxiYQEm82Gw+FQNoCpqKhgdnaW7du309LSQjweV2rbpDKJzo7JZCI3N1dZR7B9+3Z8Ph+bN29W1irE43HlAZa4jzIyMjCZTMp3bNq0iampKbq7uxkfH1+xdQcp7ciXC4/Hw5tvvsmxY8ewWCz89V//NXfddRe/8zu/Q1NTEwANDQ1XXemWasTjcaLRKBMTE4oDLykpUVKkrjZi0Wg02Gw2ZTu8j3zkI+zfv3/BewYGBnj77beZmJhYrUtJaTQaDZs3b6aurg6TyaT0qs6fP8+5c+eYmppKm8VA12NgYICBgQFOnjypLLpLFIuCud62w+FQJsS9Xq/yECsuLuazn/0sW7duZe/evcrEeXFxMQaDAZ1Ox49+9CMAOjs7V3RRzHKRCD0mRqsAn/vc55RV41NTU7hcLqW0b2Njo/LZDRs2UFpaikajITs7myeeeAKn00koFOLtt99mcHBwRWy+LRx5gkTGweHDh5mcnOTgwYPY7XaeeOIJSkpKOHfuHOfPn1+xXTyWg1gshsfj4d///d/ZvXs3v/qrv0ptbS0OhwOLxcLQ0BCdnZ3KAg+9Xo/RaGTdunXk5+crW32FQiEMBgNSSmUR0PT09G2drZLAYrFgt9u5//772bJlC1qtVnmAtra2cvr06TWlUyJvPhHrHh0dXVCqNrGnZ8KphcNhZRJzZGSEF198kcbGRs6cOcOBAwcoKytTsjY2bNjAvffei81m41vf+lZaOPJoNEowGGR8fJyMjAwlkWJ4eJiWlhY8Hg9TU1MEAoErJnN/+Zd/mezsbKV2Tzwep7CwkDvvvJOGhgbVkS8HiZDEmTNnmJqaYsuWLRQVFXHgwAHi///2zjUmziu947/DHWYwBAw214FgcGywZUhMiG+J4kSpmkirxsqqVbTqhyRbKV2ljlqpq/0SK4qURnJW6adKrjaXSps0jdZSN5aSOHHcCBSb8Q1sbgMDATzj4Q7DMMAweE4/wHuCtQaGy9zM+UnI8ILnPfrPeZ855znPJRAgISGBzs7OqDbkxoro3LlzzM7O8uKLL1JQUEBJSQlpaWk4HA7q6+tV1TZjq1dVVUV6erqqnWG4ZYwPBrfbjcfjidkog83EbDaTm5tLTU0Ne/bsUXpNT0/T09ODzWZ7YAx5fHy8KkJXVFTE3Nycqmy4NAhguSSYsbExGhoasNvt2Gw2tm/fTlJSEpmZmSoLdGZmhrS0NM6ePRv15y9GT4Pp6WkGBgZUGV+r1Up7ezvnz5/H4/Hc1w0rhKC2tpYjR45gNpsRQjA9PY3JZKKioiKkh+NbypAb9Pb2MjAwwNDQELW1tZw8eZLHH3+c6upqzp8/H5a4z41g1NC2Wq28/fbbHD58mH379lFaWsquXbt49NFHla/PaHVns9lUCOeJEyc4duwYsBDNcP36dXp6era8fzwxMRGTycQzzzzDsWPHKCwsVO6FW7ducfHiRa5evfrAhGdmZ2dTWVnJs88+y8GDB8nJyWFgYIAPPviAnp4euru7g34tozbLe++9h8Vi4fTp0+Tn55OVlUVeXh7JyckcPHiQlJQUWlpaonqe+f1+rFYrr776KgkJCcTHx9Pf34/X6132GTFq/S/dGff19XHmzBmViGi320M25i1pyGdnZ/H5fHR1dZGbm8v4+DgWi4Xt27er0+doZ35+ntHRUZqbm0lLS8Pv97Nnzx71qW90TnK5XExMTNDZ2YnL5aK9vR232018fDxCCObm5lT22YNgnDZCSkoKhYWFlJeX39PXcm5ujoGBAdUf9kFpRmy4U0pKSqisrCQzM5PMzExqamqAhYPdYJPnjKiyvr4+ZmZmaG9vJxAIkJWVRXJyMhkZGZSWlqrw32jGcDe2t7erQ8zVEp2Sk5PJzs7GZDKRkJDA6Ogot2/fVkEWRuXRULGqIRdCFAH/BewAJHBGSvnvQogs4HOgBOgFfimlDL7aeoQx3iyXy8WtW7fU9nKtRFKf0dFRGhoasFqtJCUlUVFRgclkAhYM+fz8PD09PSou3GQyqYw+4/DO4/FQX18fktVCrM2doqIiXn75ZY4fP05NTY1yqbhcLq5evarCYDeLSOsTHx9PWloamZmZahFjsVg4deoUX3zxBW63m87OzjWFDvp8PgYHBzl9+jTHjx+nqqpKNUB54YUXyMvL4/z586t+OERam7t3764piu2hhx6irq6OoqIikpKSaGlp4fr169jtduWGC2UsfTBpXPPAP0sp9wJ1wD8KIfYCvwUuSCnLgQuLP8cMRr2SHTt2UFpaSkJCgqoVsUYiqk8gEFBhlnfu3KG3t5fe3l76+/txOBxMTk7i8/nw+/1IKTGZTPfsOnw+H06nM1RxvjExd4ymE1VVVdTW1pKXl0dcXBx3795leHiYc+fOcfPmTebn5zf7YYyoPh6PB5vNxo0bN2hsbMTr9aoEF6P+znoKYgkh7kmYMl7DqI8UJDExdwzi4uJUdJOUEqfTidPpVMlQoU6IWnVFLqV0Aa7F7z1CiHagAPgF8NTin30C/B/wryEZZQgwGjNYLBb27dvH1NQUg4ODazbk0aCPcUCzWhGnxMREMjIylN8XFvyB/f39ITHk0aBNMCQmJrJ//35qa2t56qmn7onOGBgY4JNPPsHpdG76wxhpfdxuN9evXycjI4OZmRnKysrUrtQ4Y1mPIU9ISGDnzp1kZWWp/y+lZHh4mJGRkaB0jLQ2ayUhIUG5VaSUOBwObt++HbaM1jX5yIUQJUA10AjsWBQbYICFLdCmY/QaNJrgbkZ4oHEo8dJLL7F//35SUlL48ssv+e677zZUOyMS+qyF7OxsDh06pEqRhpNo1aa4uJji4mJeeeUVysrK7kl6+fTTT2lubqa7uzvkyT+R1MdIBKurq2N6epqysjJqamp48803aW1txeFwcOHCBZVyfz+MHc2BAwcoKChg7969qt7IRonWubMS4a4CGbQhF0KYgT8BJ6WUk0vfICmlFELc96NHCPFr4NfrGVxycjK5ublUV1czMTHB+Pg4bW1twY4XIQTJyclqMiUkJKg6JTt27KCqqor8/HwmJyex2WxYrdZ1Z3euR5+NaLNW4uLiSE9Pp6ysTNWL8Pv9zM3NqdKloSIScycY4uLiyM/PZ/fu3Rw4cIDs7Gzg53yDpqYmmpqaQl57JtL6jI2N4fP56OzsJDU1leLiYnJycnjsscdIT0/H4XDQ39/P4OAgKSkp932NXbt2UV1dzbFjxyguLmbnzp2qBonRXGFqagqv17umuRZpbdaLUUEzXARlyIUQiSyI+Ucp5dnFy4NCiDwppUsIkQfcN2ZPSnkGOLP4OkG/g8nJyRw6dIijR49y8uRJrly5QnNzM99//31Q9X2Nji5PPPGE6n5TUlLCnj17ePjhh9m2bZtqEfbxxx9jtVrp6+tbV+TGevVZrzZrJS4ujtzcXMrLy3nuuefYtm0bgUCAjo4O2tramJqaClkkRiTmTjAkJSWRmprK66+/zvHjx8nJyVEHwDdu3KC5uZkLFy7Q398faiMecX2MM5T333+fqqoqSkpKyMjIUDs4gOeff14doN8PQ08jm3Opf92oqPnZZ5/R0dER9DMWDdrECsFErQjgD0C7lPL3S371Z+DvgX9b/Pd/N3NgRmcSI0OxsLAQn8/H008/zZ07d+5JJTe6oPv9flWNzeiCXllZqSI5srOzVVedyclJ2trasNvttLe3MzIysl4jHhF91kJcXJzKVkxPT1cZnS6XC5fLFbJtYDRrs23bNgoKClTNaWP+eL1ebDYbjY2NIU/DjyZ9AoGA6jT19ddfk5+fr5pJmEwmTCYT8fHxy4bnGjtgQFUsnZ6eVrvd7u5unE4nbrc7qBV5NGkTDElJSWRlZakdi9E+L1wEsyI/DPwKuCWEaFq89jsWhPwfIcQrQB/wy5CMkAVD9Mgjj1BRUcGhQ4cYGRnhhx9+IBAIIIQgPz8fk8nExMSE8qkbE8uYNEZpzcHBQZUA89FHHzE2Nqa6DK2TiOuzGkYXdCNtGhbcKi0tLbS0tIQiGsMgKrURQlBYWMiTTz5JXl6ecgF4vV56enr45ptvVKhhiA+rokqfubk57HY7b7zxBnv37qW2tpajR49isVjYtWsXZrNZVYFcCb/fj91ux+Fw0NTURENDA62trYyNja0lmCCqtFkNs9nM7t27ycrKUjvgnJycsLXBCyZqpQFYbjTHN3c4P+P3+1XVtPT0dPLz88nOzmZubg4hBFVVVaqwj1Gic35+Hq/Xy6VLl/D5fGpiGvUdDENuJMmMjo5ueMUVKX3WgtFtyVgtBAIBZmdnuXbtGjdu3AhZIlA0apOUlER+fj51dXWcOHGCvLw89bvh4WG+/fZb5WILdcRBNOpj4HK5uHz5Mn19fWRmZlJQUEBqaipms5nS0lJ27tyJw+FASklubi6zs7N4PB6V4WnkLxg9UT0ez5rmWTRrcz+MZi1GYbGZmZmg++duBlGb2Xn37l36+vpUVlRVVRWlpaUAZGZmcvjwYcxmMyaTCafTqVbVbrebtrY2vF4vXq+XhoYG1Xx5fHycoaEhfD5fVKcIbzZCCEwmk8pUnJ+fx+fz0dHRQVdXV8wV/d8IycnJWCwWKisrOXLkiNq1BQIBRkdHsVqtQbX0etAxiqh1dHSoSn5G/e26ujp2795Nc3MzgUCAiooKJicnGR4epqenRy2YjAXWViAuLk4ZcaPRy1qavWyUqDXkBuPj4zQ2NnLz5s17ymqazWbVmsrv96uDGKNxrtEkdWpqSq0EjC4mW8lwwc8Hx5WVlcBCXQyHw8Hs7OyWMlipqamUl5fzzjvvYLFY7gk1tNlsXLt2jfr6+qgumhYJjP6lRonbr776iosXL6qQzB9//FE9d0ZlxK32nI2NjXH58mVGR0fJzs7m0qVLdHV1he35inpDbnQFWktjYc29xMfHU1xcrLqge73etforYx4j2zAnJ4eKigrVmNvn8+HxeNTB9/j4+Jb6cAuWpXMlnC6DWGFqaoquri5mZ2fJyMhQ7qRwEfWGXLNxjBW54Q8eHh7GbrdvqQdyaaOIjIwMdcD5008/0d3dzbvvvsudO3e2fOEwzfpwOp18/vnnKsgi3DsSbci3AF6vlw8//JD09HRgoUm1w+HYUrscKaVqLm2cn0xPT1NfX09rayuDg4PapaJZN1LKiO5wtSHfArjdbt56661IDyOiSCkZGRlhaGgIt9vNzMwMQ0NDnDt3jitXrkR9wwONZiW0IddsCaSUDA0NYbVaee2111TVyO7ubiYmJrQR18Q02pBrtgwzMzPMzMxsqDCaRhONhNuQjwDexX+jne385TgtIbxfLGkDf6lPKLUBrc9qTAG2EN9js9BzZ2XWrI8Id6ynEOKqlPKxsN50HURinLGiDWh9ViPcY9XaRN8918t6xhpMhyCNRqPRRDHakGs0Gk2MEwlDfiYC91wPkRhnrGgDWp/VCPdYtTbRd8/1suaxht1HrtFoNJrNRbtWNBqNJsYJmyEXQvyVEMImhLALIX4brvsGgxCiSAhxUQjRJoRoFUL80+L1U0IIpxCiafHrr0M4hqjUR2uzMlqfldH6LM+maiOlDPkXEA90Aw8DSUAzsDcc9w5yfHlAzeL36UAnsBc4BfzLVtZHa6P10fpEvzbhWpHXAnYpZY+Ucg74b+AXYbr3qkgpXVLK64vfe4B2oCCMQ4hafbQ2K6P1WRmtz/JspjbhMuQFwO0lPzsI75sZNEKIEqAaaFy89BshxE0hxIdCiNUbFq6PmNBHa7MyWp+V0fosz0a10YedSxBCmIE/ASellJPAfwBlwAHABbwfudFFFq3Nymh9VkbrszyboU24DLkTKFryc+HitahBCJHIgph/lFKeBZBSDkop70opA8B/srBNCwVRrY/WZmW0Piuj9VmezdImXIb8ClAuhCgVQiQBfwv8OUz3XhWx0LzxD0C7lPL3S67nLfmzvwFaQjSEqNVHa7MyWp+V0fosz2ZqE5bqh1LKeSHEb4BvWDhF/lBK2RqOewfJYeBXwC0hRNPitd8BfyeEOABIoBf4h1DcPMr10dqsjNZnZbQ+y7Np2ujMTo1Go4lx9GGnRqPRxDjakGs0Gk2Mow25RqPRxDjakGs0Gk2Mow25RqPRxDjakGs0Gk2Mow25RqPRxDjakGs0Gk2M8//TcyjGeU+J9QAAAABJRU5ErkJggg==\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now, let's convert each label into categorical from:\n",
        "\n",
        "0 -> [1 0 0 0 0 0 0 0 0 0]\n",
        "\n",
        "1 -> [0 1 0 0 0 0 0 0 0 0]\n",
        "\n",
        "2 -> [0 0 1 0 0 0 0 0 0 0]\n",
        "\n",
        "3 -> [0 0 0 1 0 0 0 0 0 0]\n",
        "\n",
        "4 -> [0 0 0 0 1 0 0 0 0 0]\n",
        "\n",
        "5 -> [0 0 0 0 0 1 0 0 0 0]\n",
        "\n",
        "6 -> [0 0 0 0 0 0 1 0 0 0]\n",
        "\n",
        "7 -> [0 0 0 0 0 0 0 1 0 0]\n",
        "\n",
        "8 -> [0 0 0 0 0 0 0 0 1 0]\n",
        "\n",
        "9 -> [0 0 0 0 0 0 0 0 0 1]"
      ],
      "metadata": {
        "id": "uj48ICTP-q9r"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "num_class = 10\n",
        "y_train = keras.utils.to_categorical(y_train, num_class)\n",
        "y_test = keras.utils.to_categorical(y_test, num_class)"
      ],
      "metadata": {
        "id": "PRWIjhq60fEl"
      },
      "execution_count": 54,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "for i in range(20):\n",
        "  print(f'Labels of train dataset:\\n{ytrain[i]}->{y_train[i]}\\nLabels of test dataset:\\n{ytest[i]}->{y_test[i]}')\n",
        "\n",
        "print(f'The shape of training data is : {y_train.shape}.\\nThe shape of test data is {y_test.shape}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Q8PtOqHPAHpA",
        "outputId": "b5a7e56f-f433-4409-99f3-4b0e2e4b4fd9"
      },
      "execution_count": 55,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Labels of train dataset:\n",
            "5->[0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "7->[0. 0. 0. 0. 0. 0. 0. 1. 0. 0.]\n",
            "Labels of train dataset:\n",
            "0->[1. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "2->[0. 0. 1. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "4->[0. 0. 0. 0. 1. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "1->[0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "1->[0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "0->[1. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "9->[0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n",
            "Labels of test dataset:\n",
            "4->[0. 0. 0. 0. 1. 0. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "2->[0. 0. 1. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "1->[0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "1->[0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "4->[0. 0. 0. 0. 1. 0. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "3->[0. 0. 0. 1. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "9->[0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n",
            "Labels of train dataset:\n",
            "1->[0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "5->[0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "4->[0. 0. 0. 0. 1. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "9->[0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n",
            "Labels of train dataset:\n",
            "3->[0. 0. 0. 1. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "0->[1. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "5->[0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "6->[0. 0. 0. 0. 0. 0. 1. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "3->[0. 0. 0. 1. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "9->[0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n",
            "Labels of train dataset:\n",
            "6->[0. 0. 0. 0. 0. 0. 1. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "0->[1. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "1->[0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "1->[0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "7->[0. 0. 0. 0. 0. 0. 0. 1. 0. 0.]\n",
            "Labels of test dataset:\n",
            "5->[0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "2->[0. 0. 1. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "9->[0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n",
            "Labels of train dataset:\n",
            "8->[0. 0. 0. 0. 0. 0. 0. 0. 1. 0.]\n",
            "Labels of test dataset:\n",
            "7->[0. 0. 0. 0. 0. 0. 0. 1. 0. 0.]\n",
            "Labels of train dataset:\n",
            "6->[0. 0. 0. 0. 0. 0. 1. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "3->[0. 0. 0. 1. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "9->[0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n",
            "Labels of test dataset:\n",
            "4->[0. 0. 0. 0. 1. 0. 0. 0. 0. 0.]\n",
            "The shape of training data is : (60000, 10).\n",
            "The shape of test data is (10000, 10)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### **Build CNN architecture**"
      ],
      "metadata": {
        "id": "LNdOBquzBijh"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model_input = keras.layers.Input(shape=(x_train.shape[1], x_train.shape[2], x_train.shape[3]), # As it is the first layer, needs to enter input shape\n",
        "                                 name='conv2d_input')\n",
        "\n",
        "#######################################################\n",
        "#-------------- 1st convolutional layer --------------#\n",
        "#------------------ with 64 filters ------------------#\n",
        "#######################################################\n",
        "conv_layer_1 = keras.layers.Conv2D(64, \n",
        "                              kernel_size=(3,3), # filter/grid size\n",
        "                              activation='relu', # Rectified Linear Unit used as an activation function\n",
        "                              padding='same', # zero padding to apply filters into whole image pixels \n",
        "                              name='conv_layer_1')\n",
        "\n",
        "conv_layer_1 = conv_layer_1(model_input)\n",
        "conv_layer_1 = keras.layers.MaxPool2D((3,3), # grid size\n",
        "                                     strides=(2,2), # strides\n",
        "                                     padding='same',# applies max pooling to whole input\n",
        "                                     name='max_pool_1')(conv_layer_1)\n",
        "conv_layer_1 = keras.layers.BatchNormalization()(conv_layer_1) # standardizes/normalizes the contributions\n",
        "                                                 # to a layer for every mini-batch\n",
        "\n",
        "#######################################################\n",
        "#-------------- 2nd convolutional layer --------------#\n",
        "#------------------ with 32 filters ------------------#\n",
        "#######################################################\n",
        "conv_layer_2 = keras.layers.Conv2D(32,\n",
        "                              kernel_size=(3,3),\n",
        "                              activation='relu',\n",
        "                              padding='same',\n",
        "                              name='conv_layer_2')\n",
        "\n",
        "conv_layer_2 = conv_layer_2(conv_layer_1)\n",
        "conv_layer_2 = keras.layers.MaxPool2D((3,3),\n",
        "                                     strides=(2,2),\n",
        "                                     padding='same',\n",
        "                                     name='max_pool_2')(conv_layer_2)\n",
        "conv_layer_2 = keras.layers.BatchNormalization()(conv_layer_2)\n",
        "\n",
        "#######################################################\n",
        "#---------------------- Flatten ----------------------#\n",
        "#######################################################\n",
        "flatten = keras.layers.Flatten()(conv_layer_2)\n",
        "\n",
        "#######################################################\n",
        "#------------------ Fully Connected ------------------#\n",
        "#######################################################\n",
        "fully_connected = keras.layers.Dense(32, \n",
        "                                     activation='relu', \n",
        "                                     name='fully_connected_layer')(flatten)\n",
        "# to avoid overfitting we add dropout\n",
        "fully_connected = keras.layers.Dropout(0.3)(fully_connected)\n",
        "\n",
        "#######################################################\n",
        "#-------------------- Output Layer -------------------#\n",
        "#######################################################\n",
        "model_output = keras.layers.Dense(10, # number of classes\n",
        "                                  activation='softmax', \n",
        "                                  name='model_output')(fully_connected)\n"
      ],
      "metadata": {
        "id": "n7DpTIyUAISd"
      },
      "execution_count": 56,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Make this network a model\n",
        "conv_model = keras.Model(model_input, model_output, name='conv_model_aao')"
      ],
      "metadata": {
        "id": "tbtcIzljjUUN"
      },
      "execution_count": 57,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "conv_model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VsOuLu9SDw4k",
        "outputId": "d8900498-0453-4867-cc00-f50a7f246d7e"
      },
      "execution_count": 58,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"conv_model_aao\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " conv2d_input (InputLayer)   [(None, 28, 28, 1)]       0         \n",
            "                                                                 \n",
            " conv_layer_1 (Conv2D)       (None, 28, 28, 64)        640       \n",
            "                                                                 \n",
            " max_pool_1 (MaxPooling2D)   (None, 14, 14, 64)        0         \n",
            "                                                                 \n",
            " batch_normalization_8 (Batc  (None, 14, 14, 64)       256       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " conv_layer_2 (Conv2D)       (None, 14, 14, 32)        18464     \n",
            "                                                                 \n",
            " max_pool_2 (MaxPooling2D)   (None, 7, 7, 32)          0         \n",
            "                                                                 \n",
            " batch_normalization_9 (Batc  (None, 7, 7, 32)         128       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " flatten_4 (Flatten)         (None, 1568)              0         \n",
            "                                                                 \n",
            " fully_connected_layer (Dens  (None, 32)               50208     \n",
            " e)                                                              \n",
            "                                                                 \n",
            " dropout_4 (Dropout)         (None, 32)                0         \n",
            "                                                                 \n",
            " model_output (Dense)        (None, 10)                330       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 70,026\n",
            "Trainable params: 69,834\n",
            "Non-trainable params: 192\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Compile the model\n",
        "conv_model.compile(optimizer=keras.optimizers.Adam(learning_rate=0.001), loss=keras.losses.BinaryCrossentropy(), metrics=['accuracy'])"
      ],
      "metadata": {
        "id": "BMQsO43aEVYr"
      },
      "execution_count": 59,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Train the model\n",
        "history = conv_model.fit(x_train, y_train, batch_size=32, epochs=20, validation_data=(x_test, y_test), shuffle=True, verbose =1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "W96JqybjEmxT",
        "outputId": "6886d050-409e-4e50-f69a-59d59cddd907"
      },
      "execution_count": 60,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "1875/1875 [==============================] - 101s 53ms/step - loss: 0.0713 - accuracy: 0.8950 - val_loss: 0.0115 - val_accuracy: 0.9857\n",
            "Epoch 2/20\n",
            "1875/1875 [==============================] - 98s 52ms/step - loss: 0.0265 - accuracy: 0.9649 - val_loss: 0.0102 - val_accuracy: 0.9876\n",
            "Epoch 3/20\n",
            "1875/1875 [==============================] - 97s 52ms/step - loss: 0.0202 - accuracy: 0.9740 - val_loss: 0.0084 - val_accuracy: 0.9900\n",
            "Epoch 4/20\n",
            "1875/1875 [==============================] - 97s 52ms/step - loss: 0.0178 - accuracy: 0.9768 - val_loss: 0.0076 - val_accuracy: 0.9890\n",
            "Epoch 5/20\n",
            "1875/1875 [==============================] - 100s 53ms/step - loss: 0.0155 - accuracy: 0.9799 - val_loss: 0.0072 - val_accuracy: 0.9908\n",
            "Epoch 6/20\n",
            "1875/1875 [==============================] - 97s 52ms/step - loss: 0.0144 - accuracy: 0.9811 - val_loss: 0.0080 - val_accuracy: 0.9900\n",
            "Epoch 7/20\n",
            "1875/1875 [==============================] - 98s 52ms/step - loss: 0.0129 - accuracy: 0.9836 - val_loss: 0.0070 - val_accuracy: 0.9911\n",
            "Epoch 8/20\n",
            "1875/1875 [==============================] - 97s 52ms/step - loss: 0.0116 - accuracy: 0.9855 - val_loss: 0.0066 - val_accuracy: 0.9916\n",
            "Epoch 9/20\n",
            "1875/1875 [==============================] - 97s 52ms/step - loss: 0.0111 - accuracy: 0.9851 - val_loss: 0.0069 - val_accuracy: 0.9916\n",
            "Epoch 10/20\n",
            "1875/1875 [==============================] - 99s 53ms/step - loss: 0.0101 - accuracy: 0.9870 - val_loss: 0.0060 - val_accuracy: 0.9928\n",
            "Epoch 11/20\n",
            "1875/1875 [==============================] - 97s 52ms/step - loss: 0.0098 - accuracy: 0.9876 - val_loss: 0.0065 - val_accuracy: 0.9918\n",
            "Epoch 12/20\n",
            "1875/1875 [==============================] - 97s 52ms/step - loss: 0.0088 - accuracy: 0.9883 - val_loss: 0.0062 - val_accuracy: 0.9934\n",
            "Epoch 13/20\n",
            "1875/1875 [==============================] - 97s 52ms/step - loss: 0.0084 - accuracy: 0.9889 - val_loss: 0.0076 - val_accuracy: 0.9912\n",
            "Epoch 14/20\n",
            "1875/1875 [==============================] - 98s 52ms/step - loss: 0.0074 - accuracy: 0.9906 - val_loss: 0.0074 - val_accuracy: 0.9914\n",
            "Epoch 15/20\n",
            "1875/1875 [==============================] - 97s 52ms/step - loss: 0.0073 - accuracy: 0.9907 - val_loss: 0.0058 - val_accuracy: 0.9941\n",
            "Epoch 16/20\n",
            "1875/1875 [==============================] - 98s 52ms/step - loss: 0.0071 - accuracy: 0.9910 - val_loss: 0.0068 - val_accuracy: 0.9930\n",
            "Epoch 17/20\n",
            "1875/1875 [==============================] - 97s 52ms/step - loss: 0.0068 - accuracy: 0.9917 - val_loss: 0.0067 - val_accuracy: 0.9932\n",
            "Epoch 18/20\n",
            "1875/1875 [==============================] - 97s 52ms/step - loss: 0.0067 - accuracy: 0.9916 - val_loss: 0.0070 - val_accuracy: 0.9922\n",
            "Epoch 19/20\n",
            "1875/1875 [==============================] - 97s 52ms/step - loss: 0.0061 - accuracy: 0.9925 - val_loss: 0.0064 - val_accuracy: 0.9935\n",
            "Epoch 20/20\n",
            "1875/1875 [==============================] - 97s 52ms/step - loss: 0.0063 - accuracy: 0.9924 - val_loss: 0.0054 - val_accuracy: 0.9936\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(5,5))\n",
        "plt.plot(history.history['accuracy'], label='Train')\n",
        "plt.plot(history.history['val_accuracy'], label='Test')\n",
        "plt.xlabel('# epochs')\n",
        "plt.ylabel('Accuracy in Model')\n",
        "plt.legend()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 351
        },
        "id": "Btnl3YouFs9s",
        "outputId": "9ec64ecf-d567-4222-f651-92236049a8a1"
      },
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7fc64a5b6bb0>"
            ]
          },
          "metadata": {},
          "execution_count": 61
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 360x360 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(5,5))\n",
        "plt.plot(history.history['loss'], label='Train')\n",
        "plt.plot(history.history['val_loss'], label='Test')\n",
        "plt.xlabel('# epochs')\n",
        "plt.ylabel('Loss in Model')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 351
        },
        "id": "5pdzu8GaHaCj",
        "outputId": "1d2da670-1b8b-49ef-cf22-d85e79e506c1"
      },
      "execution_count": 62,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0, 0.5, 'Loss in Model')"
            ]
          },
          "metadata": {},
          "execution_count": 62
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 360x360 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Prediction\n",
        "predictions = conv_model.predict(x_test[:10])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZNhCwE6hHksJ",
        "outputId": "9c051d01-1758-446e-a16c-118d6fb80814"
      },
      "execution_count": 63,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 125ms/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "predictions"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wQCqMl-NdklT",
        "outputId": "1a7f4e4b-c65e-4808-ea37-c36e95c4d6a0"
      },
      "execution_count": 64,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.00000000e+00, 2.39301033e-35, 4.72205955e-32, 0.00000000e+00,\n",
              "        1.56789732e-34, 0.00000000e+00, 0.00000000e+00, 9.99999940e-01,\n",
              "        0.00000000e+00, 1.01357698e-36],\n",
              "       [1.95509166e-22, 8.52958869e-19, 9.99999940e-01, 7.62192812e-21,\n",
              "        4.26685839e-23, 1.14450393e-23, 9.05522705e-23, 9.05277643e-15,\n",
              "        6.32993545e-17, 1.37404304e-28],\n",
              "       [2.86158733e-37, 9.99999940e-01, 5.53014269e-28, 4.10609489e-36,\n",
              "        1.83467872e-21, 6.96427515e-27, 2.32967188e-27, 2.02192828e-23,\n",
              "        8.10909883e-23, 1.03255100e-29],\n",
              "       [9.99999940e-01, 2.38791598e-25, 7.78688462e-22, 5.55600590e-27,\n",
              "        1.48870352e-18, 1.53255226e-20, 1.68732920e-16, 2.37102735e-22,\n",
              "        1.28111781e-17, 4.73914868e-17],\n",
              "       [4.47058042e-34, 1.12209809e-22, 3.08892021e-25, 5.42225330e-34,\n",
              "        9.99999940e-01, 1.09693066e-26, 3.06598112e-25, 3.43059387e-25,\n",
              "        9.01112404e-22, 5.78631326e-18],\n",
              "       [5.19657859e-34, 9.99999940e-01, 2.31132182e-27, 2.83984765e-38,\n",
              "        1.72087937e-22, 2.36861529e-28, 5.00083980e-26, 6.52586564e-24,\n",
              "        2.66750507e-23, 8.69073442e-31],\n",
              "       [1.14974422e-19, 2.83255238e-14, 4.91947147e-14, 5.66620250e-24,\n",
              "        9.99999940e-01, 4.31814194e-16, 2.18706872e-15, 4.87038775e-19,\n",
              "        2.06585788e-08, 8.76175310e-09],\n",
              "       [4.45997873e-22, 3.28027875e-20, 1.10638738e-24, 2.78899051e-24,\n",
              "        5.67481264e-16, 1.22519087e-18, 8.94473471e-33, 2.06847306e-22,\n",
              "        1.43805498e-17, 9.99999940e-01],\n",
              "       [6.74368299e-16, 6.34237996e-13, 1.93446062e-13, 1.42807492e-13,\n",
              "        4.62549408e-11, 9.99873400e-01, 1.26593921e-04, 1.08730523e-13,\n",
              "        4.75159467e-08, 3.58629074e-14],\n",
              "       [1.63108995e-26, 2.98718546e-19, 1.72654673e-23, 1.71086782e-26,\n",
              "        5.19802029e-14, 7.40709542e-21, 1.70726859e-34, 4.75620704e-18,\n",
              "        2.69793539e-19, 1.00000000e+00]], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 64
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "predictions = np.argmax(predictions,axis=1)"
      ],
      "metadata": {
        "id": "JuRk_FsrdNgj"
      },
      "execution_count": 65,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "predictions"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HZekhzoDdP2x",
        "outputId": "dba6bf43-83e8-476d-a553-d11a323c34ce"
      },
      "execution_count": 66,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([7, 2, 1, 0, 4, 1, 4, 9, 5, 9])"
            ]
          },
          "metadata": {},
          "execution_count": 66
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "fig = plt.figure(figsize=(5,5))\n",
        "for k in range(len(predictions)):\n",
        "  fig.add_subplot(5,2,k+1)\n",
        "  plt.imshow(x_test[k], cmap='gray')\n",
        "  plt.title(f'prediction:{predictions[k]}')\n",
        "  fig.tight_layout(pad=0.2)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 387
        },
        "id": "xNpVIAifdgzf",
        "outputId": "398f7d8d-5e5b-4988-f0e8-5c7d395a0cfe"
      },
      "execution_count": 67,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 360x360 with 10 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "shJWNLFAkpGs"
      },
      "execution_count": 67,
      "outputs": []
    }
  ]
}