Training & Testing.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "# Handwritten Digit Detector using Neural Networks\n",
        "\n",
        "\n",
        "# Training & Testing\n",
        "\n",
        "## This notebook contains:\n",
        "## 1. Importing the Dataset\n",
        "## 2. Pre-processing the Dataset\n",
        "## 3. Splitting the Dataset into training, validation & testing parts\n",
        "## 4. Defining the deep learning model using neural networks\n",
        "## 5. Training the model using the dataset\n",
        "## 6. Analyzing model accuracy metrics\n",
        "## 7. Saving the model for further use"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "mH6FXZQERvJl"
      },
      "outputs": [],
      "source": [
        "# Importing necessary python libs & modules\n",
        "\n",
        "import numpy as np\n",
        "import seaborn as sns\n",
        "import tensorflow as tf\n",
        "from matplotlib import pyplot as plt\n",
        "from sklearn.metrics import classification_report"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "id": "Kl6s1wkQVJ_n"
      },
      "outputs": [],
      "source": [
        "# Importing the Dataset & Splitting it into training & testing parts\n",
        "\n",
        "objects =  tf.keras.datasets.mnist\n",
        "(x_train, y_train), (x_test, y_test) = objects.load_data()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 268
        },
        "id": "oSP_V4BbVKGa",
        "outputId": "c55f38eb-9c74-4b6b-cf54-24a5f6238be8"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "(60000, 28, 28)\n"
          ]
        },
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAU4AAAD7CAYAAAAFI30bAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAABisUlEQVR4nO39d5hk6V3niX7e99jwJiN9ZmV517baSi3XLSHUEgIhDY9AewckYEcLC7uww+4g2N3LzO6dGZi7w4U7D7CrWRjEfXACARLI2251t9RqW13d5V1WehcZGT7inPO+94+Isl1dXVmVmZHmfOqpJzPDnd+Jb5xfvOZnhNaakJCQkJCbR3bagJCQkJCNRug4Q0JCQpZJ6DhDQkJClknoOENCQkKWSeg4Q0JCQpZJ6DhDQkJClsltOU4hxONCiBNCiNNCiE+tlFEhnSXUdfMSarsyiFuN4xRCGMBJ4L3AOPAc8DGt9dGVMy9krQl13byE2q4c5m089yHgtNb6LIAQ4q+ADwFvKIItHO0Su41DbmxKLM5rrbs7bcebEOq6TDaIrrBMbUNd31jX23Gcg8DYFX+PAw/f6AkuMR4W77mNQ25svqH/drTTNtwEoa7LZIPoCsvUNtT1jXW9Hcd5UwghPgl8EsAlutqHC1kjQl03J6GuN8ftbA5NAMNX/D3Uvu0qtNaf1lo/oLV+wMK5jcOFrBGhrpuXN9U21PXmuB3H+RywRwixQwhhAz8FfGFlzArpIKGum5dQ2xXilqfqWmtfCPHLwFcBA/gTrfVrK2ZZSEcIdd28hNquHLe1xqm1/hLwpRWyJWSdEOq6eQm1XRnCzKGQkJCQZbLqu+ohIWuFMM3Wf9dBRCJgmuhIa4ND1Brg++haDVWrQxCgfb/DFodsVELHGbI5kAbG8CBBNs7i/gSLBwTNXMC+va1N4xMnBrEXDLJHNaljJYzFEv6FCVBBhw0P2YhsbccpBAiJkOKa2yU6CFoXlRAIwwAh4drHKQ1aoZUOL8AOIwwDlYrRyEUobZMY+4vc3zvN7478AwC/an6Y47O9lEpJIvMRnCBASIFWnbU7ZBW4eF0bBkiB9vwVvz43t+MUAmFarambEK+7W+0aprQngZagTFCGwEsIlAm5Vxo4r5wn2D3I7KEYXkJQ2RagHQUK0ILImEnqrCI642F//xiqWl37c9ziCMfB6OlGZeKc/7E0wcEyu3sneaz7BCP2PAnRWsb/qb4fMJrN8Y3MAc7d24X7fJbBM044Xd9kGF1Z9FAvjd4YU4/Y+BHN4JM+sVem0KUSQWFpRY6zeR1ne6QobAsZi7a+ha6hvCvOzIMCbWqUo8DUpHtLRG2PhaCXgQtJCnti1N5dZk/PHL8z8nccsKOUVZ26DvifJh7nye/fQeKMw8BhF0LHueYI2yboTVPvjZB9ZJpP7/9zsjIga1wM3rYA+FBsHmLzfCx5mPwuk4+oX0S4DlQqnTM+ZMURyQTlHQkKu0w+/KGnuCM6wb9b+knc6RQGwFZ2nDIaRZhma+osJMJ10Jkk2jbxMi6BLVGWRJvQSBpUewXaeP3rVIcDenfOYkmFY/oYQpF1qpgy4Knd3dhv72NxP7x1aJT9sRlsoVhSNc55ktkgzclCN86cxCloCMKp+loio1FkKknQ18XUO5JUezXvz42RkgFue5QZaE1Dt0aUHpergKWlTzRWR/f3YFoWwWJhVaZzmx5pYHR3ta4/ywTDQBTL+NMz0KHuuTriUMsaeDEYr6dbZnpcd+B0O2w8xykEMpVEJ2JgGmgp8boiLO5x8eKC0s4Akj6m08B2fO7uneRfDXyFqHz9lMwVGrf9hkqgrjUnvBRzfhL1oOTISD+P9l/gt/q/SkxIKhpmAsXTtb2cqPYxeT5H/2lFdKaJbnpr/EZsbWQqSXNvP4u7XQ78xHE+0PUKb4mM0mtEAFAoPAIKSuEhqCqTAEGv0aTXcBhIFinvHSKSdDBP+qhSGd1UHbvgNyLStvB39tPM2nhRie9KEmNJzPk82mt2xKYg6VIeEjSzilOFbs7LLsxVmFRsHMcpBDISQTgOzb39VHsdlAHKFDQTgsqwxo8qIoNlMrEacbtB3GpwKHWB3ZbGEe6bH0PVOd4Y4FStl/Fymlrd4kI5wzeqO5EoZvwU5cDl2YXtzJbjOLMG7qKHWWq0NpNCVo/2gr90HYRtEQx0sbjbpbxNcCA+zV57hvQVUckSiacVeWVTUBGO1IdZ8qP8UOJVuo2AkXiep/eOUE9H6Wr2IwsV9Mw8qlTq3DluNAwDL2lRTxt4cYEfFbgFE/PaTdQ1RBuSIKJRrsIxAgypYBXM2TCOU0YisHMbze4oZ35G8BP3PEvKrJExK7jCo8ssYwmfLlnBFT5SaCwUCalwRORNX1+hGPdN/j8vvwfjbASzKnAbMGHF+R13G8IHuwjS09hFTbSu6R4rY5waRzebqGZnvmG3CtJxwLIQw/00e+LMPOSy70dPsj8xw0+kXmDA0ESlfdVzlpTm+7VdHKsM8MXn78GZMzjxnl7+y/av8Wu93+CdP3uSp4t7+Pq3DhGdSjLwbRcOH+vQGW48hOtQHLEoD4OXUZD0QLtETRPdaHTAIIEfM/GyPm62zr70DADfc4ZW/FAbxnEiBNox8KMGfX15fjr7PbqlT864nlO8/ml5OkChCNrrXRYGlmgtfgZaU9EWetYlNg5GU3Pl7F56GqcQYDQUZtVD1n3kXAF/cXGlzzTkWqSBiMcQrkuzO0al36bap3h/7lX2ONMMGJq4fH0lnwBBQ1kUfQd7wSA6BVPVJFXlMWAYfDQ+S59Z4Cv9B6n7DkHUXo3ByeZFGngJgZcJEKkmsUQdP+KC7EBCYnszWFkCI+4TjzSIGB6+am9uBHpFl2E2jOPUvo+xUMIxJNOlKGN+GsvMk7vOps/1WFJ1nqr3suDHWQqiVJXNA9FzvNUtoLSmrhXnvX5SxwV9382DUhBcEeSnNaLeBKVa65kqaGWghKw6RirJ3I/tpTws8PZVuXvoDI+nJngkcpa0VDjCvu7zuqXgh+JHyZplno4fJLAl04Uk/2XpTg66E7zTLZGQdXpyRWaaafyY2d6DD7kZRNSltNfjwTvOMl1JUqjdxHLYKmEkEohYlKXtFj9315NEjQZP5XdzoZghPqmQF6ZW9HrdMI6TIECXKxgRB68RZ8GP02cUb/rpVa05WhtkspFmthGn6ttEZZP7nDyB1tQ15P048amA4LUTq3giIctFRCMU9kL8QJ7/dvf3+GT6dPuey25OoZDXlF6ISot9Eup6AuUotClpVC2eXxpBCsXb3SVcAT2xMoVkBGV37sLfkBgGid4yH8wd5iviLgq1vs7Z4jjoeJRGFj6WeoGSNvni9F3kl2IMFwKChfyKHm7DOE4dBOhaHVko4R7t5t/aH2BH9wLvzJ2mHDicr3YRM5v8Us+3uMO+fFqLqs5Zz+WJyiH+7+8+ijNnIButEIU/7N3OHw2+k2SiysN9FzhV7MYqh5s86wXpushcF95wDr2txruGTrPPmbzKQZZVgxebCRaCOC9UdjBRS/PD2Vf5ycQUng7IK5/z3iB23iA6o/DOuHyvvpv8nigfTx7FCCfnt4yWAgEYosORCNKgfmiE+bts/DvKxKRgzjeYL8cIlmyMxsonObyp4xRC/AnwQWBWa31n+7Ys8NfAduA88FGt9eou9mmNqlTQvk/vC02KczHO7Yhxfk+WRsXGHrfxY5oH33+OO+zLbVUmfZMnKvv5+7F72PXZJtZrF9DVGrrZxBgaoDmUpbijiy8/loSmZF+xxlYISFk3ut7IxlgUb1uO0kiEB0ZO8nNdT9FnBMDlkWFBKb5ZvIPTlW6eO7EDe9pi7i1xfmLf31HXAWN+lNP1Ptw5QWKsjlmzaYybnLD68fZoJCDRKx3m11HWUlspNAadzVsVhsHcPTaD7x/l7bkzRIVBVTkUixGsRQO5ChtVN7OK+6fA49fc9ingm1rrPcA323+vDUGAvVAjPukTHwV9Mk7ktENiFOJjkm/l9/PNWpRxvwbABT/Dt2b3MTOVxig20dUaqt5A+z66XMFaqBCbahI7aRM/bSKLtTU7lQ7zp6wnXa9AOA5GJoMe7mP2/hgLd7VCjhLCx2qPEBdVnVeaAV+v7uXvT9/Dc0d3Yk9bmGXB+fksf1bcwR/kH+T/efbH+cyxh4lNKcxCnchsk9iUjyyaNLXGFYpd8TmGsgXqGQOjK4uMbvheO3/KKmorTBMjnUIlosScJmmjimN0NnVVmZC068SNOhJJXVvoJRtnUSBrHRhxaq2fFEJsv+bmDwGPtn//DPAd4NdX0rA3tMf3Ea+eJnLcJJZMoJNxhOejiyXozvL84F5O7cvxc7u+xydT5/lW8SAXntxG14TGmJ3Er9Uu7a4F+QJiqYRzwWLb4ShohVraGnF8603XKzFyXfjDOWbvj/PoJ37AI4lTPOhM0GvYGO2h4ZFmks/Mvo3nxkfo/VOXyFiJyo4k9YxBhTj/Yf6DOLMG/U832FFoIEfPoIpFTMPAFILYgXuoasGAYfBzXU9zPNHL/7L7n5PcN4w1VUCd2yiNK1/Pamsr4zH09kEq2+PsTI2xx1ogZXV2wKFNyDkV0kYr7XnWTxA/a5A94WHki6y067zVNc5erfVU+/dpoPeNHrgaXfN0o9H67/vIegPl+6hKFdOyMCvdlCouS/4Vx7o491bXZIaoAK2CVpZDmLMMHdYVaSAMA52MUc+51LvgLfEz3O9MkJaylQ2kNZ5WTHi9nFnKUV+I4E5XEJNzOEkXZToEtgRhEJnVuBcKsFRCFZbQXvPSR8FoQqAFljDoNpqUzEWaKUW9x8EoR1oB95sri+imtL0pXU0TP+XQTEiSZoOooHPTdWkgLBNlQsKs4woPQwg8bWLUwawEsApZfbe9OaS11kK88eqw1vrTwKcBkiK7op9E3WyiguByWTetMRrQrFtUVStE5S3xMzzz8A7mz2fpeToBHcyj3Uisua5CYPZ2oxMxFh7MMfeQIr1tgYPOFFkpyStFXQsm/CRzQZK/mHyY+ed6SU8JjNklgqUS1rEL2OdcUraFjjiIWgM1O49uemj/+hePRBIXFgNmjeE7pxl1e+h7Kk3yVQl6c24U3kjbm9I1k2L+zgiVIc2+6DQJaeJcJ6V51ZEGRiqJiMdoZgLuj51juzWPRFJVDk5RYc1X0fWVX+O8Vcc5I4To11pPCSH6gdmVNOqm0fqqsmBat4LWdUNSCywUij5ziUPd43ynHEG5dmtEEwSh87w+ndNVSHQ8SpCOUu0RpIaXOJibISt9LGFSUgYLKsrZZg/jzSxji2miU4LYbICu1dBe85ZDTixhEBWCg5lplgZd6pksqc1Xq3PltLUtGhnwMgFZs3wpiURrwVpusAspENEIOhFFu4pBc5G0bKKw8bSB0dCIRnNVCvDcquP8AvBx4LfbPz+/YhbdBrpcofe5OokLNp+P38NdsXFc4fEjmcNUfIczB/aTYh/GhakVj+vaJHREV2GayGiUhbf0UtgH5oElPr77+wxYiyjgqGfw3x/7GDOTacx5C6cgcPOa5GgTq+ihG7ef7mogGHAK9CeLjEWyt39S648V01a5Js2MwkrXScj2JmwtS3k2RrrImlUKk/EYC49uo7RNsm/3KCNmlaoWHG7CsUo/VlW1Zh2rUHP1ZsKR/pLWonJOCDEO/BatN/+zQoifB0aBj664ZbeAqlaxfnCCdDTK4t49fGPHQR5OneNnUycoZF/lPwwfxKzHSRYSsMUd57rRVYhWObJohMUDMPLwOD/Sd4RPpk7iEZAP4Gyzh8UXuul/VZMYrWCOzbcqY9kWoukRrECdACkEObPEcGyRUXv77Z9XB1ltbZVtEqR8cskqSVkn0Jr5egxr0cQu69ZewhogYq2Ii/iBPD/a9wr9RpTzfpXjjX4ulDMYNYVuNFelAM/N7Kp/7A3ues8K23L7tKfuotEgeV7zzEv7mNib4kOJ1+gyy1TvqNPIOtTTA8T292DWAsxyE1msoc6NdawUVidYL7oK08LozqFyKbwej0dyZ9nvTGEIwUsNh/84/j5OzveQOgPxsRrGQhldrYEhEQ2ztVQTVqa6ik5om69FcfICuxSgV3gZTJitWp8ynYJsChV3qfVGqOUM7D1FHuofZY89DcCRZh//ZextnD/fw76lKrpeX1dT9XWLbjQImk1y350kdSrN6AcGOL8rzrBZ4N89/PfM+Un+5t77GJ9Po+ccItMu8bE42bk8weLWcZzrBRlx8bblqPa77Ns5zq9mn8MSEjD4cvEezv7NHjITAcnnxgimpgnafZ6uIlyv7igKRX4pRs+oIjKz8o5KOA7CdfB39bO4L0q1T6DvL9KfLvKbI9/iAWeatDQBm28tHWDmO4PkJjXG5AL+KpUJ3HSOE2iNPMtVzAUTdz7GPxYOsd2dZ9haYMSe40BmGsf0GbfTVJwIWphktvVhxmOtsJV643KztpDVxXGoDLqUBwwOxArEpcOSqjMT+JyvdhGZV0RmGuhqdcX7AxlCIxGoLZErtkbcIAVLWDbCbtcXkBIhBCIWBdNsLb2YxuXXMA2CpIsyJZ5roCxJedCkPCxoZgJ6YzUyTpWErOEKcSkNt+S7OEvgFBXaW73i4pvTcQJqcRFRKtH7gzhfir6V2mDAv3r3P/GO6Gl+qefb0APVHRZF5fLX8w/zjHM37myG3udKGFN51FIxLGq7FvRkmXxfwO4dk/xo9iUAXm0m+HLxbr53dgf7XllEX5gkqKxsPyctwEBjCAlaYSCQnc653sDYtk8zLghcAymun5BodGUIBnPtotQC3zEo7I3QTAm8GPhRDRK0BD8ZcOjgObbFFrFE0PovAxzhs+DFOFHqRQpNXVtUtcbAxxIGk5UUyfM+kZkaehWrl21ax6l9H+37mPMlUucigMGZeg/7nSlGzCJZw8AiwBJ1ptPH+dbgAZRpUs+5RKvxVhhDudx+sfCCWnHaHUiDmE2mp8TDXecZNAuAwWyQ4FylC1W0EEtlgjX6AvO00arfGMp9Q4TWrW+eNhKJa3t4CUEzZeJ2ZV4f6SAFqidDvS+KFgIE+BFJZUDQTCuCRICM+W2fqulOVfho7/PssWcI2mm2hSBKQUVR9FNsuNR8i4py8DQEQqNQNAITq+wjy/VV7cqwaR3nJWbmST+viV9I8tXmW/l87i0MPjDJB/pf5cHIOd7merwtcp5feeTrHKv08/XU3UQmu+h+OUnsJYmu11espWjIZczeHpq7+1ncH+HtA8/zw8kjDBhNFA7PlXfy3PEdxEbNFQk1uh5CtwodB1qh0NR1wGuVQZ6fHsYO5b4xSoMvaHgmTW1gCMEndz/FV1J3cmEpzfgPj3C9RCK7q86O7klMqZBCY4qA+6MFkmYdQygsETDbTHC+0oWvJH88/na8wGBqMUmzYmMsWETmBGYVorOKSkbw5Z+9m/19M0S1T6A1SzWX/okCzC2gVrEP2KZ3nEGxCMUixlSMgdp2ml0u59K9POtWSBk13uqOss2M8EvpM1yIv8rpA92cS+UozzpET8da33WiGI46VxidiFEedqn2Ce6MTXDQquCK1sdxrJbBmbSIzGlYxb7ngb68vqmAyWqK0lKErppuZaOFXBehQQQCX0k8DCSSD8ZP8Ej0DGN+mld3DONdp63sQ9EzvCtSRSKR7VFkQ/soFAXlU9WC481unhD7GatlePHCMEHJIjJukc5DYtwnfjyPqNZRC3lSO4Y585Ec9V4Dj9brNH0DFpdWfbCz6R3nRXTTw5gr4FYjdD/bzdGJvbwwsoPP7ZzizvQk/2P3kySk4Ef7X+FYsp+veXfQyPaRPh0Q+1KhMz1UNjNSoo1WcQZXNLGEpKR86trnxHwP6ZOa2JS3cu97Ow9eHNxFdVuCpb0BMakoqwbPNVKcam7nlSPbybwiSZ+qvX7nPuQSxmyB3qdjVHvS/Ib4MH/bN06PUyJnlVjyo8w0ktd9nqcNJv05ZrwULyxto9h0mavEaHgWlUIEUTIx6gKrLJBNSC+0UqgjeQ+zEmDNVyG/BPEowaG9lLa5HEy9Sp/RoK4FZ31o1O01GeRsHcfpNfHHJ0Aa5GYX6I5GqN0xwPTBbZzcP8B/9b7vc6ct+MX0KbzUcT4TH+fruw9y7Kmd7PqmQxA6zpVFCgJLoEyNKz0sDPIK5oIYhakkB15YQBQr+CuUZywsE2Hb5O9OM/eQYmT/NFEhyCvFl5bu4cjiAD3fF2S/chJdq6PCGcYb4o+OkRybJNPdxVSwm6e2pWGoxvbehRs+73w5y/fZwYV8BvFKAqsMsUlFohSw7cwi6uyF1gOv+NK6NPLXqjX71xojvZO5e6JUhjT3JUfpNyK81vQ53uwjqJqtpYRVZss4zkuooDWK0Qp7rkZi3KDebVDXJhAgkVgCtttz3Jma5HB2BHpzGLbVClVaxanjVsVoT5cDBE1tIDyBqNZbwcu3ixBIx0HsGMZPRynuEKRHCuxNzbKkNOf9FM/OjjAzk2a4oNC1equnVMiNUQHU6sQnAkQgqZWjnJ29udYjVlGQuKCxqorInIdZ8RDF8s0noFyRK3+xjNwZr5tnSrsxisaazBa2nuMEVK0GtRri6GmSZxx85yDTfhrPnsURrRWYh50F7rLneXr3Tor39BGZTmK94rXWTENWhbo2KKkIRk2i5hZQ9cbtxdIKgbBtZFeWqUdzlEfgzkdO8W+3fZ6GNjjvp/jK0t0sPdNL7oImdmaeoFoN17NvkqBUIv6NoyRsC0yzleFzU08MWl+Kql2kRyn8ZXxZ+UmX2q4mgwN5Bs1FFIqvLN7FN4/tJzkq0d7qD262juNsh78IQyJcBywbYRqt2oIRgSWufrMtIbF0EMb2rREKRSFIMO2nkE1aH/5bdZrtGo0y4iLSKYJckmo/+AMNdsQWiEnFaDPDN4sH+d7sDtw5TXTeR5RrodNcDlp3JtZZCgwnIGY1L123c/U4YtHCKq9sG+A3Yss4Tuk4iG2DqITL4v4EtR5JI6Np5AKywwvst+dwxOWpxgnP5GhjhHPj3ew/sgALi6jqygZhh1ymrgP+YfF+vje5neiUuK3pltHdhe7rojoUZ/phk0Yu4P0PvsTbkqc4Xe/lf596H984up/BfzJx8x6p0Smo1lBh2NmGxNMBx6Z76XpZkBxthCPOFaHdqF5EIgTZGM2UTWmbpDYYYPdUuad3loPJKdJXJDsoFPkgzngzC2UT5vME+cVwNLKKXAwHWlqM0VW9xXAgIUBIRCxKoytKud9E7a2wM1fgn2Wf4z67xO/V+zg8P0DkjEPia0dQpdKKt1UIWXuaVYvorI9VqKPDNc5bx8hkEPEY3lAXi/uiNFKC4n4fI+mxd+A8e5Oz9NtL7HBm6TOXiLaLsbbaMwTM+gnOVXPImmwVLQid5soj2v8BR0gezZ0gajZ55fgBxDILCQvTRN93gNpAhIUDJo27quTSc3xi6AgJo86T5f38ox/hC9+7n+4fSAYu1FelMnjI1uBm6nEOA39Gq0eJBj6ttf799dZK9lpEPEbQm6awN8rc2z0SuQq/vvdJDroT7DHL5IzINc9oFR8ItCZAsxTEmKknMBqsSXjDWtNxXa8oBmEIhStM3hE9yU57lucy++EN8p3f8OVsm8KeGIV9gsyDM/zR/r8gJT0GTIeZoMG/HP1xjs700fu0IPmX3wM2Z2Zlx3XdItzMiNMHfk1r/aIQIgG8IIT4OvAJWu1Gf1sI8Sla7UbXvCMiANJAug7CcWCghyDhML8nRnlIUBsKuGP3BNtii+xxpumWVdxrLsqyajAWSOaCGF8q3MNYLcPzZ0ewLjhkjmn0ChTKXYesG10D3dIjLZsE5hJ6qE7jPfdgLTWxxhda73/TgyBApJLoeIQgFaHW66IsQTMmCVzI3xOQHChxKDdBRVsU/AiHm1HONnp46fwwcsLFzW/6UKN1o+taYKCRSAwnoJm0sIv2TfU8v11uppDxFDDV/r0khDgGDLJOWskCSNtCZjPoVJyZt2Wp9gnc+/P8/K4fsNOZ5RF3BguBI0wMYV0qQXWROaX5avkOjpQG+f637yAxCruPVDFffa3VEG4l4gnXGR3XVevWeOiKYV+vYZKVHh/Y9yr/+JF7caZi9D9jYRU9zGIdPJ/qzizlfpPyiMC6d5HueIX3dJ+h3ypwX+Q8280mk77JBT/DiUY/X5s5yMRiiuT3IyRHfSJn82zmYoEd17UDGEKQiNWp9kSxSxauYaxqqi4sc42z3av5EPAsK9ludLlcTJ+zLYRtIzIpartzNNImpR3gdTc5lJvhoDvBgLlEStpXOcslVaeqNZN+hAt+lmO1Qb48cZD5QpzEuCA27WPOl7dMzOZ60fVi8sHOyBy5gSXmjSSFPTZmxcIuu8gASkMGtR5Nc6DJPd0z9LlFDkVHyRplYsKnrjUTQYpXatt4pTjI6fO9GIsmmTmFk28galtnXXO96LoWCKFRBmj5xvVAV5KbdpxCiDjwOeBXtdZFccUa1W23G10OQrRagkYjqFyKeneUwh4b3p9nezrPx7tfY5c9w4BRos9oxWNe6TQ9HfBUvZejtUH++ux9NA5niMxB90tVMqU6Mj/X6pq4wvUf1yvrRldaIwcDg59IvMojd5xicl+Gpw/toei7LDRiNJXJu9Nj3B0ZI21U6DPKuCIgIQWB1jxRG+ZEvZ9n5ndyaqIHa9Rh7xcrGAt5KJRala62yIbQetJ1tQkubtxesdm42tyU4xRCWLRE+HOt9d+1b167VrLtMBMZcRG2DV1pVMyl0Ruj0mdRGYAfGzrJvbELvCsySq8RARygtUve0B4BmpIKKCnJ4eo2jhQHKE4l6D6ric34mEfOEpRKWypHuaO6Ko0MWlV2Ksqmrn0sWiXKug2HbgN2mjP0GUtUtM2cn6SuLe5zL7DPakVABNpAIajrgIqGE/V+XioMc34+izHtEJ0SGKfGt1xH045fr2vIxVqdhtR4JlynKNOqcDO76gL4Y+CY1vp3r7hrTVrJCstGxiLQ38Pke7up5zSNngAj4ZFJFdmRXuDR2AL/LP0cXbLR7j1ymVG/yTO1HbxWHeTzJ+7GKzgkTptEZxQ7pj3csTyiUttyqXYd13WpROpMHGVE+NLC3RhC86B7gRHTvvQYVxjstOo0dY3t5hIBgqwEMKgqj7xSnPdT/MXcWzlTzDH75ACZEwH9FYVVrGEW6qhieTXMX7d0WtdOYAjBjvQCz49ksJcMYoax6hETNzPifBvw08ARIcTL7dt+kzVqJSssE+G6NHNxlu5tMjy4wNt7znBHdIL99hR32xe/YiRwbYgRzAURXiyP8OL8MM6LMbJTiszLeTg3hm56BFuos+U1dFRXXa1hziwRzdicWczxSnSYXdYsI+bl4E1HWDjCuvS3uqI6bl0rFpTDmWYv358YoTobY9fTdYxvv3jF47ckHdW1U/S5RYxsAy8RvWHfo5XiZnbVn+KNVw5Wtt2oNDB7u9GxCLVdXVT6LLyEoN4Fja6ARw8e5WB8koPuBH1GkW6jybXOsqE9PK34anWQ75d38d3JnZQPd+EsCnKvNrEXG8iFAkHTW9XS+uudNdX1OqhGA1koEr3gkH8yx+e6u/j2wT080DPGo6ljfDA6x7Xr/IHWfK2W5Xvl3RwpDHD8Qh8sWaROGCQKCnu8sKl3zG+GTuu61lwMR1pr1lXmkLQtgqFuGl0uE+80Sd8zz770Ah/oeoU+c4n7nAIJ2ZrKSQyudZoKRV0HlJTmr6Yf4vDREVJHTXb/wwV0uYIqldBBgL+FpuTrFd1oEDQaiFKZbcUqKhll8t3dfGVnltlDcT6w45+AqxesFIrPLxziO6/tI3rGZteTVczCEoxOoCpVgrD48JZlrYvxrCvHiWVRz7lUe028Ho97cpPsiMyz3Z4nLeu4olWmf1HVqSjNTNAKJ1LtAOqiivDs0k6makmOHRsicdokMRG0WsvWaq0RZug01xdBgK7UkEBiPAXa4AVzJz8VfBhTKKRoOUOlJb6WvHZiiNg5i/i4wspXW+vTt1t+LmRDIZsBajHKZDRJUbnA2sdZryvHKRybxb0WpZ2KR+88wW/1fxVXiEsO0xIGng442kxwptnDF+fu4vDoEFqJVte9skn/E5A4X+XA4hwUiuhGk6BcDh3mOkX7PsFCHrEoSOYLJG2LIddFRSJcb/X5YG2y1cDNa6IqrRYXYXHprYVcqpI8kaBSSXF+fzcqMorSaxSH1GZdOU4A6YFRF4xX0jxbH8AQV0+/6sriufIOxmoZXpvqx5hwQbUaSJllQeJ8CXl2Al2pbsqMn02JCtCKLZNwEHJ7iEYTd1ERRCTfyu8nbVR4JT+IX7C3Zj1OtVSi/2tT6HiE5td6+L3Ux677OKOuEIFmpOxjlAuX3ijhBTA1i6rU0P6mz0kOCdmSqJk5ur4LWddh5vBO/lNsD3bBZ1+xipEvr1ifqhuxrhyn9poEp88BLcNuxrhwOyAkZGuh6nXU6BgA9gm4GPmrYc1qq679Pn5ISEjIBid0nCEhISHLJHScISEhIcskdJwhISEhyyR0nCEhISHLROg1DAwXQswBFWB+zQ66cuS4fbtHtNbdK2HMeiLUNdR1HbKquq6p4wQQQjyvtX5gTQ+6AmxUu9eKjfr+bFS714qN+v6stt3hVD0kJCRkmYSOMyQkJGSZdMJxfroDx1wJNqrda8VGfX82qt1rxUZ9f1bV7jVf4wwJCQnZ6IRT9ZCQkJBlEjrOkJCQkGWyZo5TCPG4EOKEEOK0EOJTa3Xc5SKEGBZCfFsIcVQI8ZoQ4lfat2eFEF8XQpxq/8x02tb1wkbQNtR1+YS63uC4a7HGKYQwgJPAe4Fx4DngY1rro6t+8GXS7jndr7V+UQiRAF4Afhz4BJDXWv92+0OU0Vr/eucsXR9sFG1DXZdHqOuNWasR50PAaa31Wa11E/gr4ENrdOxlobWe0lq/2P69BBwDBmnZ+5n2wz5DS5yQDaJtqOuyCXW9AbflOJcxlB8Exq74e7x927pGCLEdOAQ8C/Rqrafad00DvZ2ya7VZ5hRtw2m7VXWFzX3NrqWut+w420P5PwDeDxwEPiaEOLhShnUaIUQc+Bzwq1rrq5rh6Nb6xqaM4wp13Zy6wubWdq11veU1TiHEW4F/rbV+X/vv3wDQWv/7N3qshf3DLrHbMHdjU2Jxfr0Xg1iOrhcfb2E/E+q6vnWF5V+zoa5vrOvt9By63lD+4WsfJIT4JPBJ4C4Dk4fFe27jkBubb+i/He20DTfBcnUl1HVD6Ao3oW2o62VupOuqbw5prT/drlLyYQtntQ8XskZc1FVr/UCo6+Yh1PXmuB3HOQEMX/H3UPu266K1/tJtHCtk7ViWriEbilDbFeJ2HOdzwB4hxA4hhA38FPCFlTErpIOEum5eQm1XiFte49Ra+0KIXwa+ChjAn2itX1sxy0I6Qqjr5iXUduW4nc2hi9PvcAq+yQh13byE2q4MYZGPkJCQkGUSOs6QkJCQZRI6zpCQkJBlcltrnJsSaQAgDAOkuOou7fmgFYRV80NCOocQIK4Z863xdRk6ziuQsRjNh/dT67Yo7JXUhj0QgNCYCxbD3/RwJ0owOUuwuNhpc0NCtgTCshGWiTBNsExEKkl9exeBIwlciZaQPF5AnTyHDgJQwarbFDrOKxARl4U7HEo7FO97x4v8/sDTSASGkPxZMcf/MfNRslaSWLkKoeMMCVkThG0hXAdh2+A6eP1p8gcdvBh4CY02wKwliZxruTMdOs61QSYSiKE+mn0JivsCcjvy3B1rpfQqNMaVD9aACqfqISGrgbBsZCyCiMVo7uzFj5nUswZeVODFBH4MvKQmGKph2T5Rx0MIzWy5ix59B5GpCrx6Eu37q2pn6DgBmcsy+5YclUHBh9/6fX6u62myMkC1c3U9DQqJCP1lSMiqImMR6O2mPpTiwvtsgp4mu4an2Zec5VB8lEciZ7GFwhKtnW0D8IB/mflxXhjeSeZwit6Tdug41wJtGvgxgR/TdFkVug2FIwwCrTGEePMXCFkzhOMgHQciLiRiYJko1wZToiyJNi5vGhh1H2O+CJ4Pvo/WGuE6aNtC1JvoUgnt+ahqNdzw6xDCcRCmiUyn0LEIfi5OeThCrVvCUJWBbJG70xMcjE6yx5mm29CAwGtfm1HRmg/uis9zoreHRiaNSMSRSqEajVXTNXScALZFIw1eJqDXWiIqDGQYqbX+EAJjsB+vP01pxCV/h8BLKBLbimSiFQ5mphlyLq89f3N2H2PfHMIugl3SSE9THpTUuzXRSUH34TpWoY48NYqqVDp4YlsTYZoYA32oZJSZh9IU9mnkYI0P7/8+/XaBu9wx0rJGVjZxBeSVwavNBBVtUwhiuMLjYXeSbsPhw6kX2Lt/mn9XeT/+zn7MuRhibBLdaKyK7VvbcQqBMC1UxMKPa0TUJyYbSCQKRYCmrBQFBaONHNIDEehwdNIBhGUjDEmQjVPtc6j0S/yRGulklceHj7HNWeCR6Bl2tj/REknOKvHvt/XhFQy8JYH0BNUhhdFXpapj1CcsEGBbW/syWGuEabZGmhEXvztJM+NQGRBYI2XuHxzj57NPk5YQF1Z7xucQaM2ob3CkPkxV2eT9GAmjzkF7mm4Deo0a0h0jHq8TRKMYroMwjFUr57+lPzHm4ADecI6Fu6PsfniU+zJj3OVMAgZnfTjvdfF/T76TE0/uIDIn6P1BCXO+hFoqvulrh6wc0nVR9+6l3uUw+Q6T7N1z7E8u8o7MabJmmT32NK4IONro51uVFGmjStYoE5MNPvHIUzSUSdGP4GnJsLtIv1XgiwN38XJiJ5Fxl+1nk1BY6vRpbnqEZSNcB0YGmX5nlkYGmgdrdGcXeSw3wf3x82y35+iWAktIDCEItOa0r5gLYvzG8Y9Qe6Ib0d40b2Q08Q/X+cX0KbKGgSsaZGNVmskkRtXBMIwbG3QbbGnHqZMxqoMulUH4aP/zPOiOMmC0vqPmghinG328em6QPf9YxsiXYaGAajRQ9dUZ/odcH2HbVAcilAcMUnfO8x/3f5Zuo8aQ0RqRSCRV7fONcobnl0bIOWUGnQIj9jz/ffZ5otICuGr5JW1Uma0mmJA5dNTt1KltKYQhEa5LozdG4W6fRF+JX9v3bd4RPUNWQkLa7Udal56jaDnNU40+5s9k2feVPCLQaMugMhLn9Pt6IX0KV5i4hknCbjDvCLQtEau4P7H1HKcQmL096ESM+Ye6mHurT/dQnj32NCkZIIWBR8A/Fg7x1bMHcE85mHPz6FIZXauhA9XKUghZXYRARqPI7i78nhRTb5U4O5f4kaFj9BlVCsrmG800o81u/mHqHubKMSpnU7hzEuVA4Gj8gQaj932Xfe4UDzjT9BqRSy8/7aeYmk9h5w1Ew+vgiW4dZF8P9Z3d5A843LH/HPemx7nDmSAtwbomE6iufSZ9zUwQ518d/Wcsns2Se1EgF4pgGuhUZ3shvanjFEL8CfBBYFZrfWf7tizw18B24DzwUa31hogIF4ZBMJij3hdl7qGA33nsswyaixy06jii9Y3X0D5PjO/GfiZB6qyPmppB1esdtnxlWde6CoEwDGQyQWNHjtKwwzve8Sq/0f8V0hJS0mGs7vDFxXt4cXYY/8s5UlMBQ6/No8enWoHSpknlLTv4q/T93Ns7wXDvN+m9YuY21Uwjph3cOQHNzeU416u2Xn+GhTsclg76/NG2f+JuO2ivYdqve2xDK054vbxaG8L7do79X11A5Jfwp2cwEolWREUHuZmt4z8FHr/mtk8B39Ra7wG+2f57Y2AYNHIRSkMmVqbBoLlIt1HDEq0NoVHf54gXpTAXJz6hcOcaqx4T1iH+lHWqq5HNIHeO0Ng/wMyDLgt3aw7GJ0kIzUIgONyELy7dw9ePH2DxeJb4ZEBkuo4oV9GNBsKx0b1d1LoMhtMFdkXniAofhaKqmyyqOuerXUSmJdFZhfY2l+NknWrrJyxqfRorUychm1eF+lWVx1zQ4Kzn8b26w5cqO/iDC4/x5yceJDalEMUKulZrbcwaBipu40Ulplz9LKHr8aYjTq31k+1G71fyIeDR9u+fAb4D/PpKGrZaCNMkv9+meG+Dd+84w512AwsLSxgsqTpfLt/FkfIgqVdsUl95Fd1sojah41zPuqodA8wfSlDYp/kfP/B57nUvsNOqExUmX61u41uL+3nqxQPs+f/VMPJLMDuPrjfwmx6ogGAwx8I9SfJ3av63wW9zn5O/FO836WsmgyTPjw8z8lQZc7ZIsMk2htartuV+i+Q9C9zbPUFKBlzpfsYCyVmvh1drQ3x3bjdnZ3Lk/sllaKyBffYCwcxsKw8dENEI1f4I1R5Jxqyu5Slc4lbXOHu11lPt36eB3jd64JXtRl2it3i4FUIIhG3jJSGbK7EtkscV5qXwo5LSnKj2cnyxB6egCYpbbve8o7oaySQ4DpW+KJVBAQM1HomeYbcpmVeauUDzfGkHL88M4k4bmOML6GKJoFy5qrBDELep9gp0tkmfUSTV3nQItGbMT/F8dSeNgouxuIguVSDozKhljbkpbVfzepW+ptSwmKknOO0lyasqJWVT1xZH6sOcrvZyqtjN6aluxLRDfKKBPZ5HFUtXz/qkJHAEygZLrNMR55uhtdZCvHEyotb608CnAZIi27kASGlgpJLQncU7UOX3Dvw9w+YSYFHVTeaDgJcbA3zt+btInjTpPbW1A6LXWlfpupTfvZ/CLpPKoRq/cO832WnPMWwo5pXP/z71Pl7N97H0TC89L3i4M0uo+QVUe5R5JYt7XYbfO8o9mQl6DQ8ups4S8Dvn38/4s4P0HAdm5lCV2mZdinlDbqTtal6v2RcXsGpZxgZS/PxbPo5hKPSFKFZR4CyCW1BYFcWu+SZGpYgYn0FVa6+PYrFMmgmBFwdXdmaZ5VYd54wQol9rPSWE6AdmV9Ko1UBYJiIWJUi49GaLPBapAw4KRUMrCspm0ssQmTRJnfcxF8psiXHI1XRGV2kgXIfSgElpl89bd5znl9LHsISBp00mg4BX831Mj2XpP6WIPnsGXW+8fsOuXaexkRI83vsae5xpou36qp4OaGjF2FyGzAlIjDVaTtNrrskprgM6f81OzZKsN7GXctR6XLQB2aOayIKPM1PFmF9CV6sE+UXUDZJMtCEJbIGy9YYbcX4B+Djw2+2fn18xi1YY6bqIVBI10M35x1PU+gP+64Hnr3rMcS/G/zH2OKdmu+k6HhA/ugALGyJIYKVZc12NrizVh3ZR7TGpvKPCR/Ye4ZHEaTwCRj2fJ6p7eKG0neJ3exk4FZA8UUBXqleNEoXZqtUoRobwu2JUhxQHnQkGzSIWBvmgwe8tvJ3Di4M4L8XIHl5EFsr4/qbbFLoRHb9mVa2OpIB7RjNIN1qAM11BVuvoShVVrYHnvWlmno46VAc0fn+TbrMzy2k3E470l7QWlXNCiHHgt2i9+Z8VQvw8MAp8dDWNvB2E60A2RWV7nK5Hp3i8/ygfTB7myiDbs80ejpwYxp2wSB5bIDh5pnMGrxHrRVeRiDN/j0V1OOC/ufMZ/mX2eHt02FqP/MrcnRyf6WHw2Qb2D06i642rR4lCtJxmJEJ9OEV50MborbLHWiQtJZYwKCjJF8/dQf1cgqEjHurwMTZzJO560fZadKNB0GhAYQnz/AUAVPv/clARi2a3Ty5XImuUV9zOm+FmdtU/9gZ3vWeFbVkVRDxObThFud/gvtQcByMTJIQPWFzwa5zxMnxt4Q7ipy2iUxpR7swu3VrTcV2lgbBMVCpGZcSndyTPDqc1exz1fY40+/ly/m5e/cFOIrMCeyaPbjYv76xaNjIeQ2RSLB3qpZGSLO0BL+fztpFRYlKQV4pnGjFeqt5B41SS9GlwZ6urlr+8Xui4treBketCDfe1ql2Z7WjJKxOANBT2RolkS/QnisRka/1z3G8wHUQ5l8/SO9PEmq+u6tr1ps8cUrkUCwdsytsV78se4bHIHK5obRYcafbx2dkH+d7xXez7VhFjYp4gvyWn6GuOsExkIk69N8Zb7z7FJ3qfYr+9iMLhufo2Pjv9IEdf28b+/zyLnphG1epXbQLJeAwGeijtSbH0z0u8ZWCUH8ke5j5nmqgQJKTNcc/ij8Yf4+R0N4NP+MRevNDaoe3geYfcGDXcx+S7UwQOKBu01OgrHKfQ0OgJeNvAOHtis2RlnUCbvNAY5KniXqrnk9hHz6IKS+jm6q1fb1rHKWMxRDRKrS9GdVAjeuv0GKVLa151Dd8r388L48PYUxZyKY+uVLdKaErHEUK0AplNQdau0meWiLYDos80ejk+0YczYyCqdXQQYPbkwLHRURfl2jTTDpU+m/KQ5ED3DPcmLrDHmqPXcC4doxBEuVBI4+VdrGITVSqv6sUUsgzaG3nSbdXjFNk0Kh5laV+C8kiAdhRYGiE1lyqItz1oMl1lV2yOfrsAQFl7vFDZwRPju3HnJDQaLZ1XsYrZ5nScQsCeEUo7E8w8JPnZx7/FPneKfVYRD4MnasO8Whvib555mO2fD7AXiujJmdeNakJWESkRpomyBQNOgRFTY7XXnf/+3N0M/qWFs1gFQyKGB5h/ay/VHkFle0BssERPYp53dl2g3y7w7thxuqVPQl79cT5WH8R7KUPXuMaeLOCHBYvXB9JoOUzHQQ/34iVdph+JUrmjzsjAFP/vkW+TkDXs9o65FAoDRYBEaUlUNugzWlP0fGBxpBnlsz94kOEvC6ITJVSltuo6bz7H2a6x2cxGKPcbeL0N3h0/yoBZIyFNPK2Y9lNcqGWw5w0ip2fQpQpB6DTXHiHQQmBJHwvjUgpes2lilzyEFxBkkwRxm/KgoN6r6Nq+yKMDp9jlzvKO6GkSMiArTSxxeaQZaI1CkfdjOHmILgRQq4dOs9O0axAIx0FmM+iIQ6MvTiNlUBkKuHP7JG/JnOPdkWkcYeK1AwJdYeIIi0C3quS2MsUjeDpgLoCSimAUTaITZYyFEv4azBo3l+OUBkZXFhGPcu7dNm//oVe4NzF2KV2vFReoGKtnOVXoximAzi+GI81OoBT4PkZTMdHIMB54dEtBVFr8T3d9nb/4Xx+iERgEWmAbZR5LT9PnLLHbmWG7NQ9AQTmUlCYvfKQIGDACEtKmrD0WAsGRwgC5V+s45+ZRi4XOnu9WRhoIKVobPz0ZSntSTPywwko1GMjO0+3UeCw1yZ2RcdJGlVHfoKBcnijvp6ps3p98hTvtEq4wcMTVJecMBK7wYKDO7ANxkqMukZk5dGN1a0xsKscppEDEowTpOHp3hf9z+In2PVfXW8x7MQrlKNGqRpUrWy5zZD2gtUYHChFAyXOZCyIkRJUo8DPJCX4m+feXHquuE7Ay6Tc44XURIDDQWMInK5dIAHWtyasI89UY3aN5/HboS0hnELK1nq0TMRo9MZZ2GfyLt36HR2KnOGiXSEmbkmpS1ZqCMpkLYpxvdvO9hR1UPZu97jQjZhFkgCWuLk4sRUv7TKrC0rCLVTGJ2u1mbUEQ9hy6GWQ0SvHePsoDBju6L6BQr+sdVNKKZ0Z3YL4cJznqocNWv50hCNDVKu5UmSeeupOnhnfyi3c+yU8mXiUqDaLCpqqbjPutkeUz1T1MNDIczg8ymU/hNUx02UTEfT5w4DUOxCbpM8p0G/Bio4/PLxxibjxNd3Oy02e6dZFGqzzgyCAqFWX6wSSLD3j0DszycOw0Udngj/IPcqGW5Uwxx2wxTr3sIBcsZFNgFQXagM+/914GBxcZMRdJXHE5S2S7H1GRD217he9GdnMqNkRibAdWvoY8N4GqVlshbCvsQDeV4xTRCPN3GXj7avx47iyB1iCuHq2UlIE4EWPomyXMmQJ+OEXvCNr30b6PMTrJjn90qfZG+bv0IR7de4IcHlEDSirgteYg5xo9/NnJh6jOxcgcNth2pIZRbyJLRaq7M3z3X+yEQbjXHQUCflDeyRMn9hAdtdCNcBe9I1ysqRpxaYxkKQ/alB+t8Mwjf0hUGrjC5IQX8NnTh2icT5A4J8mN+0Qma8hXjwEgU0l0NsWRHUMczm7DivrstC6n2RpC0N2OoviX2Vf41ezL/ELkcY6cPUhs2iazUEQ0m63BkV7Z63xTOE5hmsh4DJ1N0cwq+rqW6LVapcKC9jdNVXuc9W2ONkawigKjUEVXN1dx4o2IbnqYCzWiwLlXe/kl/TGilkfCqlPyXKaWktTrFsbZCPGCID7pYy1UQEq0Y+JHJD2JMrujM6RlAzApBw66amI0CKv1d4J29X52DOOlXebvcqgMKQ72z5KQJnUdcNrTvFzfTm0iTuKCJDYV4M40MJZqaEAmE1TuHabWbZLtWWDEnictawQa5lWT5+oDeNq41Ftq2KzSazjclZjg6QN7qfcYCDWEm+/DmSwiFout7g1XDpSC4JaX6jaF45TxGGrXENXhGF17Fvjpbc/yYOTcVY+ZCST/18xjvJbvI3UuQJ0bu5SFEtI5VK2GOHEW0zLZdzIJERctTeoigaU1I0G5tZHUaLZGqY0mqt7AGOyjtq+H8oDBP+9/kY8lT18qETjfiGPPG9gFDUHoONeUdlSL6O9h7P1ZqgOKdz9ymJ/vfpJeo4YlHE54kk/PvYuX5obof0qQemkaimX0UhFsG5lK0tg3wOTPNLl3+Cwf73uat7utwh8NDd+tjfBvXvwgftOgv7dAd6TCT/X9gA/HZ/nZ9Mv80Pte41Szl//88DuYXExhP9NN5kQao6Ewaj4XMyBk00eeGb+leqybwnGKSITqUIxyv8H2RIFhe4GE9AALj4CqCpjwM7yW72NmOs2OJX8rVcVZ32iN9ppor4mqLiPdVWu0KQjsVuO1lIy0w1U0Zd/BKgusqgoTGtYYGYkgM2m8niS1PoXRX+X+xHnutgPyCo41Fc/VdvLS3BBzMyl2LniQL6Db7UtEIk4w0EVlwGZHzziPZM7QZ7QKeUwHcNbP8nRxD2oygtkUTAZZ5mNxnoruZcBaJCY0adlk2Frg3sw4KbvGKwNxjHprBmLWLWiv4FlVRXwycksdTjeF42zsH6D880sc6pngv+r+HvusJWLt5k+jvuCJyh18efYOgr/tZte5Bs7xScJ99A2OlASORLc/wRedpkJxcqab4e/XsOarrVCzkNWnnQmkD+xg/J0pKoOKDz/2LG+Jn2GPPUte+fyH2cf4p1fvxpi26XlBsTPv4x6dICiWMQb7CHJJZu9KsPBog57ueX5x23fYZ80yp6I8VXf5nTPvZ/bZPiKzgl3Pl5F1Hy/tErgOT++6n68N3ofX63HHrgn6I0XekTrBu5LHOfuBU8x7CTxt4CtJ0Y8wUU1xdq4Le7EPOTW97NPdFI6zmTL5yR0v8v7EEQaMgKi43PypoFxOVPsYzWcYPF7FPHqeoFLroLUhK0I7WP5iHrO6IgO9WbWxxxbQpUq4HLNGCMMAw6CZcSntCIgMlvnx9Ascsn3mlU9embxW6Cdy0iE6rUm9PAfFMqpUBilQ8Sj17giVIcFj+05yR3ySu+xpUlJwyotyttnD2EQXg4cVkZkG8uWTqEYDO5EAx8EqDRKZdylutzkZ7aGYdfmR7GH2WPO8xZ0jfkX850zQ5LnGIF+N3MXZ+IHrtIp7czaF41SGoNdaos8IcK+J83q2uosvvnIX7piNOd9Kq9Rbqw7jpkTnF0m8ZuK7ORaCODDfaZO2LkIg9u+mNpJg5n6Lxx56hb2xGbKyzlSg+J/Hf4yXxoeQR+P0HvaQnqY5mEaNZCiO2HhJQWm7wh0ss7f7Ah/qepGSivBvJj/AVDXF6VP9uNMmvacUyaMLiHKNwPdBa1Stjmg2sUbnSBfixCYiVM5HKGej/A/3/xTRTI23DI5yV2L8krnPLO7i+Rf2EJmRjFyYv6WC5TdTj3MY+DNaPUo08Gmt9e+vh3ajF9ESus0iGemiUJd20gGOl/tJHLWJTSpYLKIbjRu80tZhI+h6I4LCEhSWiPckWPI73MtqHdERXYWkti3B3N0W3FXi3wx8hbQ0KSjFuB/h2WM7ybxgkTrnET0ygcokKe1PUc9IFh5pMtC/yMeHXuajyVdw25Wtnqo3ePr0LowJl+HvB8SPzUKhSDBzdeH61vo4qPEJGG8lYyaATF8vztIOqj0pvnP/HiYGU5eec+JsPzv+0cOZrqDHprgVbmbE6QO/prV+UQiRAF4QQnwd+AStdqO/LYT4FK12o2vaNc/YvYPa7hz5AwZdsoJEAPKq2M0lz8Wd17h5H8IMoStZt7reDDKRQOayFPpsEka4jnkFa66rkILygEn9jhr390/iCoFEYgBR6dE7UGC2kqPWY+Hu2o4Xh+pQgI43uWfnOHsSs+xzW8874UV4onyA7y/uwD0aITapiU7UEMUyehnr1bpaI36hjl20gQjnzm+7dF96WuBM55GFEuoWW0PfTCHjKWCq/XtJCHEMGGQdtBtdOtTD1Ac99gxNMWTWMES8Hbd3Ob1gtpogdaaGNVNs9WUOAda3rjeDzGUp3dPL0k6DrNmZKuDrkY7oahgU9sF/fPhvGDAXibbXEy0hSAifT2z/Hsd7+3Gkjys9+q3CpQItrhBYCCwhsYTNE+UD/Odn3kVk3GT75+fRY1Ot1s/+m7fUuJKgWMR49igRKYh+22ytwbbRvo+qN1Ba3XJG0bLWONu9mg8Bz9LBdqPCcRC2TSMlyeVKbE8sYLc3Cy7urF6sCD1XjDNS8xD1JiqM6bsu60XX5aAtEy8qCRywRDiTuB5rqqts5YzbKC6WbG8V4NAM2wsESGKyQULWickGAYKCMskHUeraYtZPkA/ifHHyDiLjJtFpjShWCCrVWy7AczHkcDWW527acQoh4sDngF/VWheFuFyWeU3bjUoDY7CfoCtBcTf8d9t/wE5nhqgw2r1qPKo64A/n38lXzh9AvJzAmBlFLRXDHdbrsG50XSY66lDtljSz6lL7hJDLrKmuSmPUBSfqA+DCbquAJQwcTBwBDzsL3GPPY9BKkzzuxfj03LuYrCV5dWIAr2gTPW+RuKCILARsPzuDqNYJZubWbdWym3KcQgiLlgh/rrX+u/bNHWk3KqRARxz8mIUfVex0Zhg0li4V8/C0oq41J0q91CbjpBc0ul5vd88LR5xXsp50XS5aCJCt1goGGom4HJL0xu3gtwSd0NWoCU5Ue5FCscdawBU+Fi1HeXGztq6hqiTnmzleXexnvhxDjUeJFASZUwGJk0vIYpVgaqZVBGYd70nczK66AP4YOKa1/t0r7upYu1HlmvgxAx0L2GMtkJZgCBtPB4z6BhNBlmMvj7DzH5rYc5VWywTPDwvZXsF61HU5GIUyqbMxlGVSCKIYogH6egXothad0FX7HkPfqvDixL081XeI39/7HqyIx0C2iGt6lJoOdc8kP5EmOmpiVSA6o8g2FM5CHaPmYeTL6MUlVLPZvlbXt5I3M+J8G/DTwBEhxMvt236TDrYbbTWkl0inean4rUTi4bGgYow1u4iNSYzvvHhLMVpbhHWn63LQ1RqRmTrVnhh1fTm4+doygluQtddVa4yXT9F9zCHYM8RcIU4z6XBhWwRtK2TZwGgIel6D3FMTUG9cWjrTzSZa6w2XyXczu+pPcXWDzitZ9+1GQ67PRtdVlyuY0wUi/RGOVIY45p6nz4CotDDtAD+XwDAkorCEXqfrZKtBp3S92ATPnCnQddTEdw3qYybKNDEaGsNTRMer6GIJPB/t+atSJ3Ot2BSZQyFbD1WpoCoVoj0pXlvq53B8ENsdIyrBcT0a3TFcDcKQ6DBRbNW5WF9VjVaRF8axhXx9KqO+OjllI7Mh5zWy7mGVfFTZ4qjnMu43rtteIWTzI+sep6e7+cL8vZz3W9khA8kii3ssijtjyFwXRjIJ0niTVwpZMbRu7YZf+3+TOE3YgI5TBwFyroA7tkT0gsmfLzzCVyoHqKpwWLEVkfkSiaejPP/Efr5ZvAOAjw48T/ZHJph6TFHf14feNoCMhWmZISvHhpyq62YTUa1jF+C5mW1MJxPEZANLBLxSHWailsYqb55vt5A3RnsekQWFsiXnq13MBQ3SRpW7MpPM5BJUe+IIFcWecaFc3lSjnpDOsfEcp9aoYhlRqdL/NYvm0QwFO8kfx3YCYDQVwtf0nZracDt1IctHLxVJvzRPfCzO9+7axX+Kvp0DkUk+3vU0g06B/+v978CYcthZ6UNWKq1d3HUcHxiyMdh4jpPLFVE4fQ7j9DkMwLrmMeGlsTVQTQ9jZg6r6WHODXB4cZB+u8BuK6AQPcsTw3s4ZfbgJW1c22rt5IaOM+Q22ZCOMyTkElqhanUkBQa+28vMhRF+/65BDt87TNaq8P6e1xiMLvHKwN243V3I+UWCsLRgyG0SOs6QjY3W6EaDoNEg+sQxYo6Ns7SH7zh72T40z3+790n2OZP8oOsegkwMs7yMvkYhIW9A6DhDNg0Xg7ATYw3qL0aYHh3k/fO/hFc32Xbab6X1haUFQ1aA0HGGbBp0o4FuNDCeeY3+5y2QEmFboDSqWiXYADnQIRuD0HGGbDouthsOCVkthF7DuDYhxBxQYWN21spx+3aPaK27V8KY9USoa6jrOmRVdV1TxwkghHhea/3Amh50Bdiodq8VG/X92ah2rxUb9f1Zbbs3XMplSEhISKcJHWdISEjIMumE4/x0B465EmxUu9eKjfr+bFS714qN+v6sqt1rvsYZEhISstEJp+ohISEhyyR0nCEhISHLZM0cpxDicSHECSHEaSHEp9bquMtFCDEshPi2EOKoEOI1IcSvtG/PCiG+LoQ41f6Z6bSt64WNoG2o6/IJdb3BcddijVMIYQAngfcC48BzwMe01kdX/eDLpN1zul9r/aIQIgG8APw48Akgr7X+7faHKKO1/vXOWbo+2Cjahrouj1DXG7NWI86HgNNa67Na6ybwV8CH1ujYy0JrPaW1frH9ewk4BgzSsvcz7Yd9hpY4IRtE21DXZRPqegNuy3EuYyg/CIxd8fd4+7Z1jRBiO3AIeBbo1VpPte+aBno7Zddqs8wp2obTdqvqCpv7ml1LXW/ZcbaH8n8AvB84CHxMCHFwpQzrNEKIOPA54Fe11sUr79Ot9Y1NGccV6ro5dYXNre1a63rLa5xCiLcC/1pr/b72378BoLX+92/0WAv7h11it2HuxqbE4vx6LwaxHF0vPt7CfibUdX3rCsu/ZkNd31jX2ykrd72h/MPXPkgI8Ungk8BdBiYPi/fcxiE3Nt/QfzvaaRtuguXqSqjrhtAVbkLbUNfL3EjXVd8c0lp/ul2l5MMWzmofLmSNuKir1vqBUNfNQ6jrzXE7jnMCGL7i76H2bddFa/2l2zhWyNqxLF1DNhShtivE7TjO54A9QogdQggb+CngCytjVkgHCXXdvITarhC3vMaptfaFEL8MfBUwgD/RWr+2YpaFdIRQ181LqO3KcVs9h9rT73AKvskIdd28hNquDGGRj5CQkJBlsmm7XMpEAhmLgmWhoy4AwvNBKXRhCVWuoIMAwnqkISEhy2RzOk5pUHvnfmbvtagN+xw6eA5fG5zLZ6mUXPr/aYD0sxPoYplgcbHT1oaEhGwwNqXjFIZBuc+kvrfO/Tsv8Mc7/hFPK77at42XK9v42otvJRWLIBph7+0NiRDtnxIhBRjGrb2O0qAV2vdXzraQ20eIlrZtXbXvrbuZ4eZynNLAyKQQ8RilnfD4waPcHz+PhYElDO51xnGFxxdyb6E+kMD1A5iZXXeihFyDNDC6sgjXASlACFQ6Tr0vRq3LYO5+UEkftLjpjGSjaBCZlkTmNd3fuEAwPRMu3XSStrM0sml0XzdeLsrcIZfAhqFvLKFfWF+b/5vKcQopEIk4Kh3HG2zyC93fIS19LBEBYK9l4IpJvIyi3mVhL7odtjjkZhCGAZkkQcIFKVGmpNbvUthtUNkW8H8+/if8cNQj0Ap1k57zi9UUf3DhMU6f66XrpSTMzbdHoMEqn03IdRESYZmIZILqSILiNpPE+6bpiZaYObuT+AudNvBqNpXj1EpDo4moNmApyTPVXexxpsnJGpa4ZjonOmNjyJsjo1FELIaIR2kOZvDjFoXdFs0EaAlIaGQUxkCZka4lBs0iEFnWMYbNPI/kzuIryexbBkgM3E3s+CzB5DTa80GFDnQtEZaJdBxqu3OMP2agck0ezE5hCcV4VCJjMXTTQ3vrY3ltUzlOtEKVKwiliEz38rmp+3h77gwPOC9gcYvrYCFrjkwlCQZzVLbFmH5I4nf5PHrXER5InsdAIYVm0Mpzlz2PKwQpaS9rtAlwwIKd2ed4LH6MX/vgTzA6nWLbF3qJFUvoWh1Vra7iGYZci3QcRCLO3D02f/KRP2TQKDMRxDneGOC76fuRuWx7Mzd0nKuD50HTxGjCUt2l6IfT8Y2GziQpb49RHjBQQzVy6Qr3JMa4w5lACgVAt6ySk/almcRFp6lQN3UMQwiiwqLbqLAvO8dxLains8TjsdaUPXSca8vFtWsTBo0y/YZNXddIGxWUDdqxwVw/7mr9WLISaI1qNBBaYxc1C/k4Y6kMSutwar6BKNyVZfbHGgx1L/Kvd3yDAXORAaOBKwRGW0hLSAxx+6IOmfC/Dn6REz09/PrenyY+0YMzZkGhEG4UdRBHmOwwIWCGZkrjd8Wxmh7MddqyFpvLcULrw640KNBK4KswOWqj4UcEuUyJfekZHnCm6TYcuKLEWaA1HgGloAFwyYEG1zg6QwgsBFIILIzrOlpXmOy0wBAzBFFNEJFgW6t3ciE3hSEkUWGTEGW0CdqWaHP9LLdtLscpBDISQUQjNLoEQ32L7E/MIFdgZBKydkQWAsZO5fiBZ1LqkXRfcb3kgwZ5ZfDd6m4+N3UfDd9EivY0XV+tc9at8kBmlB6ryHujJxkww/qSISvD5nKcgLAthOvixWB7coFBZ/HS9C5kY2CVfCIzLkvpKA192WsGWlPSgrkgxgul7Zw+1Y/wLgbD87oYzvGkh68l22KLPOSeY+BmDh7OztcVxjq9dDeX4xQSLBsskyCi2RFdoM9cQl5Ry8QSIFJNKn0R7GKUaC4HjQZBqRSuaa0T7IkCuVdylOcdPtz4ZWT0cmaPqhvgSZxZk95TCnFN1JC4QsJmzOFczw6OZ7fT+8NFEukfkJKChLQvPWZJNflObYDXakNEJyWRiSVEIfwsrBcC3QpBU6YEKVuB8utAmzd1nEKIPwE+CMxqre9s35YF/hrYDpwHPqq17njSt5ACYVto1yaIKfa60wyai1etbUkgm65QGHCxl0yiPVlEuYqoVrdU6t161jU4M0rkwgSxeIze73ahr0ipFEEAgUKUKgTz+Va2zxsgbQuRSBDs6ufrd+3n4dgZ9lgLJK5Y9p4LJH83dz9H53pJng/g9AWCpreap7fqrGdtb5ZAKwzREkobGmVJMNbPfsXNWPKnwOPX3PYp4Jta6z3AN9t/dxytNLrRQNQaWAXJk0t7eaUxjHdFNogrBAe6pjF3lCkPC+qDCYJs8tbznTcuf8p61VUrtOeja3VEsYJcKl/6L4oVRLGMrtVawdAqeOP/hgHZFM20Q9xuEJUNrGumfp6WFBoR6nULo6HRvn9DZ7xB+FPWq7bLoOU8IYgoGikDFbXf/ElrxJuOOLXWT7YbvV/Jh4BH279/BvgO8OsradgtoQKCfAFRKpM+PsjXMndxdl+Oj+w9hSNaO6UJafP/GvwSpX6DX0p9jNnGAIkLJpkzDkGj0eETWDvWta66lfqoajVUvdEq5HHxLqUv/vKmLyMzaQr35CgPSh5LzTBslEmIq8cKdW0yW47jFVzMWoDeBJ+Bda3tMrEAK1unuD2OU4hgC7ku0mJvdY2zV2s91f59Guh9owde2W7UJXqLh1sGKkB74BYC3GmT6f4E3jVrIr2GQ68BuUiZGRcCR7QCcEPWl65tB3oTPvIqhGUjIy4qnaDWJWhkNBmziit4XYRFRdsUS1HMgoFRXx9ZKavETWm75tfrRZQGrREaAgQKjURgCIHj+PgRUPb6marf9uaQ1loLId5wtVZr/Wng0wBJkV2bVV0VEH/+ArGzKc7JLqr3rclRNxXrUtebRO7cRulgF+VBg9IjVfqyRR6OnSEhzas2CgGONwZIf8cl+1oV88wUnR/LrD430rZTumqtEUGA9GA6iJEQZXJGBAnk4hVGuxM0E5L1kgd4qy58RgjRD9D+ObtyJq0M/tQ0wWsncOehqS+fpmz/C7ku617X6yIESANhmgjHwe9qpWtWBjS7++a4LzdGn1HEElcHwQdas+jHSIz5mCfGUPlC585h9Vnf2iqFDhRCQSGIUtWt9FkDQcxqot0AZa6fWeGtjji/AHwc+O32z8+vmEUrjBZgiNaw/8oiEBJ5KXA65BIbRteLyEQCdWA7Xtph4YBNrVfj9TcZGZzkQGyJx7uO0Gcu0Wt4XJl9NBM0ONrs4sXCMEY9AM+/qXXTDcy61lY3GqA1Tl7zj/lDTCbO85HESax1GoN9M+FIf0lrUTknhBgHfovWm/9ZIcTPA6PAR1fTyNvFQLdCG5ZZQWczsxl0BZCxKIt7YlT7JNH3zPKzI89z0JngXqeAIySuuPgRvzpraC6web66g7Fihkwj2Aw76ZfYiNpq30f7Pk5R8fLcIEoL3hs7QXad7j3czK76x97grvessC2rRoBYdtmxzc561lU4DjLiIuJx/IEsKmLSyFgE19kcaKQEC4cCrEyN9/aeZb8zRZ9ZwhHyqoyxuvZpaMXXq9t4trST1wr9nL3Qgz1h0Z2fQ/n+5R37Dc561nazsLkyh66D0BBo8TqnqVCvy20OWR/IeAxyWWojaaYfsmlmFMk9iwwki697bH+kyM93P8mIWcMVEku01rCvXctcUgH5wOL3Tr2H8nM5YpOavUcqGEtL6LHJTRGGFLJ2bHrHeS03W68xZO0RjoMwTfRQL6XdSSq9BrVtHlaywcHcDNujC697Tr9dYMSskTNaFeCvp69HwHTgMOlnWFiIk53QxGYCjPkSolJDeVsnY2wjYgqFMDR6He3pbjnHGbJOkQbGQB9BOs7Y+1Pse/wUO2ILvCt5nLRRpVtWccXrnaIlICVvnFFSVQFfLD7A4cIQiRdder94Ft1sosoVUHrdtGMIeT1SCDJOFSfWJHBshBTrYg8vdJwh6wZtW6ioRaNL8eM9L7Hdnuceu0ZU2Fy7uXMt8lLQ9OuHJQFQ9h0qvo1V0Zdz3MO+QhuChFkn4jTXVTjSOhr8hoS8npspCdiKz5TIN3hsWpr8ZPZZfmH4CRYPaoJH7sDYtxPklqtPsOFwhckH0y/zM7uepdqnWxXQ1gHrw4qQkNsgeJO5myMsDtmSH4rOYPTXKA05eLl4q+1wyLpEtb8ETQzushf5QPw1vOT6iXrYElP1KwPgr5zKjUTzvDDgYVasVh3PkM6hFeSXsJoeuZf6+dfOP0Nmmtw3cgFbBpwu5CjXHZQS6OtEQwihL90+lC1wX3aMve40H4mfJS7Dyu8bBavkMz6Z4lWg1Lt+W5hsesepxRsHwO+JzNC3Lc/8Ug9iHfUz2ZJoTTA3B3OQrdbIvJKlNpzg8Fv3goDMUU33QoBsKmRw4xHm3N3DfPbeXoaHFnh0/2ni7e9KpfV1nW7I+sEqNolcSDAvUhQORID1GfGw6R3njRiwFrkzO803ujLoTBKjUiUoV8JNg05TqyOXKriOReJcHCTEpjysxToiCBDBjads8SmbZtJm3MxQ2WeGIWgbCOEpjDqIhqSpDcDHEAJDa7SlkMk4ulJF1esdtXPTO06hW0U+rpc59A53nvv6v0rJd5jauZuolMjRCVSp1CFrQwCCUglRrSJm5ug53YrP1I0GOgjQ8KYZPsnZFMlXkkz9UC/Tb0uwzwp7pG8UZMPDzWsaWUldW0ADC4EjNEQC6OtGLpXR0zMd7diw+R2nrxnzU8TkIllpYonLU/KotIgCfW6R82mJnYlgTdsQ+s3OovWl3GVuZWTRaCAaTaSvUeH+58YiaPWREgGXtJNCYGhaQfCWsS6W1Tb9pyo+HfCpox/hN8c/yJh//Snbwegksw9rJt8ege7sGlsYsuIM9bH0wADlEUjIWqetCVkOWiMDjVCC4IpUIUMIhGw5zvXQX33TjzjNcsDidJITUlPqv/4uXZdRxuqpUfeiKDfcXV9zLuaV3273QiFASIKkS7VH4qUUrvBhnZYmC7kOWiNUa8QZXFMKUshW0zZpGB2P53zTowshhoUQ3xZCHBVCvCaE+JX27VkhxNeFEKfaPzOrb+7yMWsBVt4kX4hRVWFYykU6rmu7+LCRTCLv2oe4/w7Mvl6E49xSYLowTeTd+/F+6BBjPxTH/tE53vPgq/QaWyudsuO63i7zi6RfKZA+pXhyaT8vNKCkAqLCYP/ADJPvjLF4fw4Zi3TUzJtx2z7wa1rrg8BbgF8SQhxkg3TNkw0fuyBQJau92BzSpuO6CsNAxKJUdiQp7o6ju9KtKu63EJguTJPKjgTzd9lYDyzyXw7+Gf9b/1fJvkke+yak47reDsHiIurV46ROV3it0MdrjUGqWhARNm/JnqN+d5WlXRKczg6CbqYe5xQw1f69JIQ4BgyyQbrmGfky6VMxhDI5Uh9ij3WYrJRXBUUPWws8MnKOF50h6n1xopkMqlrd1KXGOqWrkckgknGC7hTlkRj1tKRwUKMsjV1KEp2PQbny5oU3pNHqm+460JNDpaLM32nS2Ffj0b4xElLhisvl5TwdUNcKpQQyAKH0pqz4vtGv14uIps/EQopnErvZbs/Ra5R5ZmEn7uEoiQsKahsoHKndcvQQ8CzL6IjYSfzzY8Qnpog8eIAnPrCXEXueB52JS0HRAPfY8J+Gvs43unL8mx0/Q/RCD3Jmfsu0C14zXYWAgR4q25Is3GFhv2ueHek8vzzwLQIEvzL537DtZLzl1CqVG76UtC1EKgnZFPn7uqj1SLa/5zz/88g/0WdU6TYufzEGWlPVASUtUJ5EehrxJkH0m4GNeL1eRFQb6NE039W72BubYbv5EsdODnLwz0fR5TJBh0MGb9pxCiHiwOeAX9VaF8UVhWJv1DWvY+1GL6ICdCPAqPsU6hFmvDRVe/qqh5gYONIiIWtoA7QhEOukmMBqs6a6CkmzO0Zxu0l1QHFP1yz749OMmEU8BI20pjGcwUpFMROxG76USkRodEVppk2WdkkaXYoDyWkGjCqJdruFuvbJBwElbXKksY1JL40xb2MvNZHlJmqTVHy/Hhv2er3SlgCCQOLp1tKN8CW6XEbV6re/kXib3JTjFEJYtET4c63137VvnhFC9Gutp27UNW+9tJEVTZ+p+RTfje9mnzPJXuvyUF/RmrYF2mIr9W9ba12FYTDzkMvIB87xeGqCn8r8gITw6TZMSsqn6+45Rq0csmlj1JI3fK1Gd0Bm+yLdsQr/j76XGbYX2G/N021cbgE87sNnlx7mVLmH77+8F3faYPg5D/f50+imt2kzxDbD9XotAQIU6KaHXgeFp2+mWZsA/hg4prX+3SvuWtdd865F+ApVMZmuJCmpCIrL2SQSueX6EXVK12ZC80j2LAcjE+w0wRKtKbUhAran8uQHYviegdeUrUIDb0C6p8T7ho6zzVngQ/ET5IwICodAazwCykox4ac5sjTAucUuomMGsQmNO1kiKCyt5CmtKzbL9XotBhokCNtCBK1ZZCe5mRHn24CfBo4IIV5u3/abrPOuedciihXSh3PMzPfxg56dfDD2+jYMW4x1pWtUGPyLvic4k+3F0wYNdeMIiF5riT32NAnhkZCtj/FM0KCkDL5b3c2Ti3t5YXyYxNdjxBYV8dEisliDufxanE4nWVe63i5CQMqokTMMIn1lqo/sxZ2tIY6c6ujm7c3sqj/FG0cQb5iuebpWI3nBx/AMJmrpTpvTcTqpq7rOYS1h8Ha3ztvd0WW9lmxXhlcoSspgLojxUnkbz1/YhjwZo/cb46jZeVS9QbBJp+ZXslmuV+DSWTjSIy4cuhMVSkNJtIwSPWGvb8e5afB8nHwDZbucWericBO6ZYMBMwyKXzO0InNC88dPvZN4f5mXdh5mm73AB2KnyRo3p8Nc0GA+sAgQ1LVJSUU4XNvGvBfnm+N7WVyIY0/aJM9DdNZHl8roZnNThh5tZnTERm2r88D2UbZbczS0z1LNJbagsAtN8LyO2rdlHKdqNDBni0Q0nJlP8kx1D/dFzjNgdlaArYQOArLPzRObSpE/kOYvag+yrSfP/btGyd5kzPu4H+F4s5+KcigHLmP1LN8a3UOt4NL1fYvdp+pYkzOocxfQSm+JUeZmREVtHtg+yi/0fYdhs0hVC8oVl+7pBuZ8GdXhcLIt4zhRGlGpYVomzolu/tB8F6nEQwwmWhsFUijOF7LEpgNkuR52PlwNtEaUq9jzFvFJk/rJKOcWHf4o+Sh3x8d5S+QMO00fS0gsYVBSTU57LhVtc77ZzVIQ5ZnFnZzJ52j6Bo2GRVAzsSctYmVBfNLDmq8iSpWOlhwLuX1EoJmuJDne6CdA0G1UCJoGwmuC6vzsYcs4Tu018WfmEPN5tv/nJUTEBSlpGpfjBfuDBnr+GEG12uqCGLLi+FMziPkFEqMuyReS+L1pnr5wiG/03sP73vEyv9D9HbLSp9swOO25/H+nfojxcpqxM91YBYP0Ceg+XUNWPYzCIvgBeB46UOhqFdX0Qu02AbLaZHQ8x9/I+7k3M852dwGWLGSjivD8ji+9bBnHCbSC4VVAMHPdELaQtaCdkBB4PqJWxwRik3GkL3l2ehtDzj1kzTJZs8wr1W0cnh6gWnSJTJg4i5AYb2BdmEdXq/j5xY4HQoesEk0Pc85i1Omi7puciXRjFySi6be+LDvM1nKcIesHFaCbCjUzR+5J0I5N84UUX028q5W9JQVGQzGw5CF9D6NYgqYHS2VUqdSaiodOc9OiJ2fY/RcOQcxGOWmKVpYdk3n0hUnUxSLXHSR0nCGdQ2tUvY4aHQPAOAlvtEfU+TFGyFqiKhU4fAxB6zNhwLrqHLU1ErJDQkJCVpDQcYaEhIQsk9BxhoSEhCyT0HGGhISELJPQcYaEhIQsE6HXMKRDCDEHVID5NTvoypHj9u0e0Vp3r4Qx64lQ11DXdciq6rqmjhNACPG81vqBNT3oCrBR7V4rNur7s1HtXis26vuz2naHU/WQkJCQZRI6zpCQkJBl0gnH+ekOHHMl2Kh2rxUb9f3ZqHavFRv1/VlVu9d8jTMkJCRkoxNO1UNCQkKWyZo5TiHE40KIE0KI00KIT63VcZeLEGJYCPFtIcRRIcRrQohfad+eFUJ8XQhxqv0z02lb1wsbQdtQ1+UT6nqD467FVF0IYQAngfcC48BzwMe01kdX/eDLpN1zul9r/aIQIgG8APw48Akgr7X+7faHKKO1/vXOWbo+2Cjahrouj1DXG7NWI86HgNNa67Na6ybwV8CH1ujYy0JrPaW1frH9ewk4BgzSsvcz7Yd9hpY4IRtE21DXZRPqegPWynEOAmNX/D3evm1dI4TYDhwCngV6tdZT7bumgd5O2bXO2HDahrreFKGuNyDcHHoDhBBx4HPAr2qti1fep1vrG2E4wgYk1HVzsta6rpXjnACGr/h7qH3bukQIYdES4c+11n/XvnmmvZ5ycV0lbFzUYsNoG+q6LEJdb8BaOc7ngD1CiB1CCBv4KeALa3TsZSGEEMAfA8e01r97xV1fAD7e/v3jwOfX2rZ1yobQNtR12YS63ui4axUAL4T4APB7tNqH/InW+t+uyYGXiRDi7cB3gSNcbnPym7TWTT4LbANGgY9qrfMdMXKdsRG0DXVdPqGuNzhumDkUEhISsjzCzaGQkJCQZRI6zpCQkJBlEjrOkJCQkGUSOs6QkJCQZRI6zpCQkJBlEjrOkJCQkGUSOs6QkJCQZRI6zpCQkJBl8v8HEW1N42ehji0AAAAASUVORK5CYII=",
            "text/plain": [
              "<Figure size 432x288 with 9 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "# Printing sample images from dataset\n",
        "\n",
        "print(x_train.shape)\n",
        "\n",
        "for i in range(9):\n",
        "\tplt.subplot(330 + 1 + i)\n",
        "\tplt.imshow(x_train[i])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "O6Pw9GCiVKMZ"
      },
      "outputs": [],
      "source": [
        "# Reshaping the training & testing data to make it suitable for neural network imlementation\n",
        "\n",
        "x_train  = x_train.reshape(-1, 28 * 28) / 255.0\n",
        "x_test = x_test.reshape(-1, 28 * 28) / 255.0 "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "id": "QhN6hbbDkhnd"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "(50000, 784) (10000, 784)\n"
          ]
        }
      ],
      "source": [
        "# Reserving sample images for validation\n",
        "\n",
        "x_validate = x_train[-10000:]\n",
        "y_validate = y_train[-10000:]\n",
        "x_train = x_train[:-10000]\n",
        "y_train = y_train[:-10000]\n",
        "\n",
        "print(x_train.shape , x_validate.shape)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "id": "dTsxwl7kVzbw"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense (Dense)                (None, 128)               100480    \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 64)                8256      \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 10)                650       \n",
            "=================================================================\n",
            "Total params: 109,386\n",
            "Trainable params: 109,386\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "# Defining the deep learning model using neural networks & printing its summary\n",
        "\n",
        "model = tf.keras.models.Sequential([tf.keras.layers.Dense(128, activation='relu', input_shape=(784,)), tf.keras.layers.Dense(64, activation='relu'), \n",
        "tf.keras.layers.Dense(10, activation=tf.nn.softmax)])\n",
        "\n",
        "model.compile(optimizer = 'adam', loss = 'sparse_categorical_crossentropy', metrics=['accuracy'])\n",
        "\n",
        "model.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_5phYxm2VzhI",
        "outputId": "abb8d3be-022a-44e4-a5e8-c6809717e3af"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.3724 - accuracy: 0.8959 - val_loss: 0.1773 - val_accuracy: 0.9502\n",
            "Epoch 2/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.1556 - accuracy: 0.9547 - val_loss: 0.1287 - val_accuracy: 0.9642\n",
            "Epoch 3/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.1091 - accuracy: 0.9673 - val_loss: 0.1086 - val_accuracy: 0.9699\n",
            "Epoch 4/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0828 - accuracy: 0.9753 - val_loss: 0.0966 - val_accuracy: 0.9715\n",
            "Epoch 5/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0646 - accuracy: 0.9805 - val_loss: 0.0885 - val_accuracy: 0.9735\n",
            "Epoch 6/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0515 - accuracy: 0.9842 - val_loss: 0.0860 - val_accuracy: 0.9750\n",
            "Epoch 7/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0422 - accuracy: 0.9877 - val_loss: 0.0832 - val_accuracy: 0.9760\n",
            "Epoch 8/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0339 - accuracy: 0.9901 - val_loss: 0.0844 - val_accuracy: 0.9752\n",
            "Epoch 9/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0279 - accuracy: 0.9919 - val_loss: 0.0878 - val_accuracy: 0.9755\n",
            "Epoch 10/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0242 - accuracy: 0.9930 - val_loss: 0.0860 - val_accuracy: 0.9759\n",
            "Epoch 11/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0199 - accuracy: 0.9941 - val_loss: 0.0926 - val_accuracy: 0.9756\n",
            "Epoch 12/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0181 - accuracy: 0.9946 - val_loss: 0.1105 - val_accuracy: 0.9733\n",
            "Epoch 13/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0147 - accuracy: 0.9956 - val_loss: 0.1033 - val_accuracy: 0.9742\n",
            "Epoch 14/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0121 - accuracy: 0.9965 - val_loss: 0.0940 - val_accuracy: 0.9768\n",
            "Epoch 15/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0110 - accuracy: 0.9966 - val_loss: 0.0925 - val_accuracy: 0.9777\n",
            "Epoch 16/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0150 - accuracy: 0.9950 - val_loss: 0.0999 - val_accuracy: 0.9752\n",
            "Epoch 17/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0088 - accuracy: 0.9975 - val_loss: 0.1024 - val_accuracy: 0.9762\n",
            "Epoch 18/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0066 - accuracy: 0.9982 - val_loss: 0.1004 - val_accuracy: 0.9789\n",
            "Epoch 19/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0057 - accuracy: 0.9985 - val_loss: 0.1113 - val_accuracy: 0.9778\n",
            "Epoch 20/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0075 - accuracy: 0.9977 - val_loss: 0.1190 - val_accuracy: 0.9740\n",
            "Epoch 21/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0131 - accuracy: 0.9957 - val_loss: 0.1170 - val_accuracy: 0.9762\n",
            "Epoch 22/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0080 - accuracy: 0.9976 - val_loss: 0.1224 - val_accuracy: 0.9751\n",
            "Epoch 23/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0063 - accuracy: 0.9979 - val_loss: 0.1206 - val_accuracy: 0.9760\n",
            "Epoch 24/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0020 - accuracy: 0.9997 - val_loss: 0.1122 - val_accuracy: 0.9796\n",
            "Epoch 25/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0013 - accuracy: 0.9998 - val_loss: 0.1080 - val_accuracy: 0.9806\n",
            "Epoch 26/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0081 - accuracy: 0.9974 - val_loss: 0.1270 - val_accuracy: 0.9747\n",
            "Epoch 27/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0109 - accuracy: 0.9962 - val_loss: 0.1267 - val_accuracy: 0.9773\n",
            "Epoch 28/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0074 - accuracy: 0.9974 - val_loss: 0.1218 - val_accuracy: 0.9789\n",
            "Epoch 29/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0047 - accuracy: 0.9985 - val_loss: 0.1360 - val_accuracy: 0.9747\n",
            "Epoch 30/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 0.0052 - accuracy: 0.9985 - val_loss: 0.1241 - val_accuracy: 0.9780\n",
            "Epoch 31/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 9.0177e-04 - accuracy: 0.9998 - val_loss: 0.1213 - val_accuracy: 0.9790\n",
            "Epoch 32/32\n",
            "391/391 [==============================] - 1s 2ms/step - loss: 4.3892e-04 - accuracy: 1.0000 - val_loss: 0.1233 - val_accuracy: 0.9804\n"
          ]
        }
      ],
      "source": [
        "# Fitting the training data on the defined model\n",
        "\n",
        "history = model.fit(x_train, y_train,batch_size=128, epochs=32 , validation_data=(x_validate, y_validate))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 460
        },
        "id": "AsQYs9LWVzuf",
        "outputId": "98a4124d-1c0d-4aca-f2b2-71cb4b393136"
      },
      "outputs": [],
      "source": [
        "# Predicting the test data using the trained neural network\n",
        "\n",
        "prediction = model.predict(x_test)\n",
        "predict = [np.argmax(i) for i in prediction]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "text/plain": [
              "Text(69.0, 0.5, 'Truth')"
            ]
          },
          "execution_count": 9,
          "metadata": {},
          "output_type": "execute_result"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 720x504 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "# Printing Confusion Matrix\n",
        "\n",
        "cm = tf.math.confusion_matrix(labels=y_test,predictions=predict)\n",
        "\n",
        "plt.figure(figsize = (10,7))\n",
        "sns.heatmap(cm, annot=True, fmt='d')\n",
        "plt.xlabel('Predicted')\n",
        "plt.ylabel('Truth')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SZzLnXE0Y_6p",
        "outputId": "3cbe0807-abde-4260-dcb3-57a64d6db348"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.99      0.99      0.99       980\n",
            "           1       0.99      0.99      0.99      1135\n",
            "           2       0.98      0.98      0.98      1032\n",
            "           3       0.97      0.98      0.97      1010\n",
            "           4       0.98      0.97      0.98       982\n",
            "           5       0.98      0.98      0.98       892\n",
            "           6       0.99      0.99      0.99       958\n",
            "           7       0.98      0.98      0.98      1028\n",
            "           8       0.98      0.97      0.97       974\n",
            "           9       0.97      0.98      0.97      1009\n",
            "\n",
            "    accuracy                           0.98     10000\n",
            "   macro avg       0.98      0.98      0.98     10000\n",
            "weighted avg       0.98      0.98      0.98     10000\n",
            "\n"
          ]
        }
      ],
      "source": [
        "# Printing Classification Report\n",
        "\n",
        "print(classification_report(y_test, predict))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "id": "BUhJHFvmY_9Y"
      },
      "outputs": [],
      "source": [
        "# Saving the model for further use\n",
        "\n",
        "model.save(\"trained_model.h5\")"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "collapsed_sections": [],
      "name": "MNIST.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}