keras_mnist.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "source": [
        "from keras.datasets import mnist\n",
        "import numpy as np\n",
        "from keras.models import Sequential\n",
        "from keras.datasets import mnist\n",
        "from keras.optimizers import SGD\n",
        "from keras.layers import Dense, Activation\n",
        "np.random.seed(1671)"
      ],
      "metadata": {
        "id": "HPrc4f-w8wzv"
      },
      "execution_count": 31,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "num_epoch = 200\n",
        "batch_size = 128\n",
        "verbose = 1\n",
        "num_classes = 10\n",
        "optimizer = SGD()\n",
        "num_hidden = 128\n",
        "train_portion = 0.2"
      ],
      "metadata": {
        "id": "nm5evAMlMsed"
      },
      "execution_count": 32,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "(X_train, y_train), (X_test, y_test) = mnist.load_data()"
      ],
      "metadata": {
        "id": "zA9L8WPbMwH0"
      },
      "execution_count": 33,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "plt.figure(figsize = (12, 6))\n",
        "\n",
        "for i in range(1, 10):\n",
        "  plt.subplot(3, 3, i)\n",
        "  plt.imshow(X_train[i])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 525
        },
        "id": "sg4VK7zXMs3R",
        "outputId": "355ea04e-ef5f-46ce-dcf6-dd38d833db39"
      },
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1200x600 with 9 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAA0cAAAH8CAYAAAAaKNQcAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAABJGklEQVR4nO3dfXyU5Zn3/+8kJEOAZGKgTEhJalykWGnRImDU+piV2l8VhLbqbn+LDyurJrRAu7a4Fe9t3Y2VXbVg1HZXQdtbcdku0GJL1waEqgElYi2CEStCLCSINQ8GEpLMef9hnXY4JpIMk8w1M5/36zV/XN+cM3MMwmGOXLnOy+eccwIAAACANJeR6AIAAAAAwAsYjgAAAABADEcAAAAAIInhCAAAAAAkMRwBAAAAgCSGIwAAAACQxHAEAAAAAJIYjgAAAABAEsMRAAAAAEhiOAIAAAAASdKQgXrh6upqLVmyRI2NjZo0aZKWLVumqVOnHvd5oVBI+/fvV25urnw+30CVhyTlnFNbW5uKioqUkcFsj+QUa3+U6JHoHf0RqYD+iIHS5x7pBsDKlStddna2e+SRR9yrr77qbrzxRpefn++ampqO+9yGhgYniQePj3w0NDQMxF9dYMCdSH90jh7J4/gP+iOSFf2Rx2A8jtcjfc45pzibNm2apkyZovvvv1/SB5N8cXGx5s2bp29/+9sf+dyWlhbl5+frPH1BQ5QV79KQ5LrVpWf1CzU3NysQCCS6HKDfTqQ/SvRI9I7+iGRHf8RA6muPjPuv1R09elR1dXVatGhROMvIyFB5eblqa2vN+s7OTnV2doaP29ra/lRYlob4+IuNY/xplOd0OZJRf/ujRI9EP9AfkcTojxhwfeyRcf+l5EOHDqmnp0fBYDAiDwaDamxsNOurqqoUCATCj+Li4niXBACe0N/+KNEjAaQH+iO8IuFXbC5atEgtLS3hR0NDQ6JLAgDPoEcCQHT0RwyEuP9a3ahRo5SZmammpqaIvKmpSYWFhWa93++X3++PdxkA4Dn97Y8SPRJAeqA/wivifuYoOztbkydPVk1NTTgLhUKqqalRWVlZvN8OAJIG/REAoqM/wisG5D5HCxcu1Jw5c3TWWWdp6tSpuu+++9Te3q7rrrtuIN4OAJIG/REAoqM/wgsGZDi66qqr9M4772jx4sVqbGzUGWecofXr15uL7AAg3dAfASA6+iO8YEDuc3QiWltbFQgEdKFmsA0jjG7XpWe0Vi0tLcrLy0t0OcCgo0eiN/RHpDv6Iz5KX3tkwnerAwAAAAAvYDgCAAAAADEcAQAAAIAkhiMAAAAAkMRwBAAAAACSGI4AAAAAQBLDEQAAAABIYjgCAAAAAEkMRwAAAAAgSRqS6AIQm+6LJ5vswC2dJvtt2aMRx5Nq55g1RdXZJsvc+NIJVAcAAAAkH84cAQAAAIAYjgAAAABAEsMRAAAAAEhiOAIAAAAASWzIkBRCF5xpsqWP3G+ycVn2P2fomOPtZcvNmvqzekz2jyef3fcCASDNtH9pmsm+f/eDJvveV/7OZG7bjgGpCQAG0u+XlJls19/Y70ezfJkmO/+WuSbLWfNCfAqLM84cAQAAAIAYjgAAAABAEsMRAAAAAEhiOAIAAAAASWzI4Dldl55lslsf+LHJxmdlmyxktl+Q3uzqijhuCfnNmjNtpM7LppgsZ+Pv7Ht2dNgnA0gaR2ZMtdlIezFtwSO1g1FO0jh4lv3Z4vfeujwBlQDAwGhccE7E8TNX3W3WdDn7/WhULh4VDQ7OHAEAAACAGI4AAAAAQBLDEQAAAABIYjgCAAAAAElsyDBoMvPyTNZ+/gSTLbj3cZNdlPN+lFfs21y74r3Ii+lqHrB3N37u/yw12dP/+ZDJPvWTSpOd8i0u0gaS2f7zbS8Z9lfNduEjA1+LZ2XYDSpcyRGTXTL6NZPV+M4xGQAkg/eLIzf6Ksjo4+YLSY4zRwAAAAAghiMAAAAAkMRwBAAAAACSGI4AAAAAQBIbMgyatx/7uMlenFI94O/73dEvRhyvH2EvDr7urUtN9ujJvzZZ3qfejV9hADzhn7+4ymTf32V7QjrL/KtPmOy1C+wOFWe88FWTFb34uwGpCQDi6f0vTzPZT6/8wTGJz6x5qNluLvbrr5xlsuF7XzVZyCTewJkjAAAAABDDEQAAAABIYjgCAAAAAEkMRwAAAAAgiQ0ZBkz3xZMjjp84436zJkN9u9PwdXsvMdm2X59mst/dYN9j45GhEcejt9m7ur/xnr2YLutfN5osw16HByDJZfm6E12C5w35z8N9Wnfk93kDXAkAnLiOL0412R1VdpOZ8VnH/8bv0f/4vMkKdz4fW2EewZkjAAAAABDDEQAAAABIYjgCAAAAAEkxDEebN2/W5ZdfrqKiIvl8Pq1Zsybi6845LV68WGPGjFFOTo7Ky8u1e/fueNULAJ5FfwSA6OiPSBb93pChvb1dkyZN0vXXX69Zs2aZr999991aunSpHn30UZWWlur222/X9OnTtXPnTg0dOjTKKya/0AVnmmzpI5GbI4zLsn/UoSj3Br7itStNlvmldpPl/3/OZJ/6caXJxlc3RBxnNGw3a076jYnU9S89JvvpZ+zFetdf9DWTZW58yb4gkAaSoT+Gzjsj4vhzQ58dlPdNZicPf7dP64p/bfsmgA8kQ39MFwe+2mGyi3JsJmVGHM15q9ysKPxBcm++EE2/h6PLLrtMl112WdSvOed033336Tvf+Y5mzJghSXrssccUDAa1Zs0aXX311eY5nZ2d6uzsDB+3trb2tyQA8IR490eJHgkgNdAfkSzies3Rnj171NjYqPLyP0+WgUBA06ZNU21tbdTnVFVVKRAIhB/FxcXxLAkAPCGW/ijRIwGkPvojvCSuw1FjY6MkKRgMRuTBYDD8tWMtWrRILS0t4UdDQ0PUdQCQzGLpjxI9EkDqoz/CSxJ+E1i/3y+/35/oMvrMN/l0kx1aaG+sOj4r8gavdZ1miTa8/ymTvbvS/tRj5Hv2pyaBn2yxmX0LxfP2jsFM+9/p3fn25oij7f1jAcQo3j1y7xdzIo5HZw6L22ungiEnl5jsSwU/69Nzc/a8ZzKuQgIGTrJ9D5kIQ8Z+3GSvfm65ybqc7Va7uiKP990z3qwZrq2xF+dRcT1zVFhYKElqamqKyJuamsJfA4B0RH8EgOjoj/CSuA5HpaWlKiwsVE1NTThrbW3V1q1bVVZWFs+3AoCkQn8EgOjoj/CSfv9a3fvvv6833ngjfLxnzx69/PLLKigoUElJiebPn68777xTp556angrxqKiIs2cOTOedQOA59AfASA6+iOSRb+Ho23btumiiy4KHy9cuFCSNGfOHK1YsUK33nqr2tvbNXfuXDU3N+u8887T+vXr2aMeQMqjPwJAdPRHJIt+D0cXXnihnLM3IP2Qz+fTd7/7XX33u989ocK8IGOYvVC5+267h/6WCf9jsj3dRyOOF972DbPmpN/sM9no4QdN5uULeqeO2Wuytwa/DMATkqE/DhnXdtw1Ha/lD3whHtVw33CTneu3N+x+uHWsfXIz91gBepMM/THZZZ7+SZOd9fiOmF/vqv/5WsTxX/3UbgaWiuJ6zREAAAAAJCuGIwAAAAAQwxEAAAAASGI4AgAAAABJMWzIkE6OXHC6yX414YE+Pffvv74g4jh3jb2IrTu2sgBgQI3eZjcgSDaZo0aarGm2vbt7wVfejjjeNP7hKK9md8t6sHqmyUY3Pd/n+gAg3vZeYfvef4/cHmVlpkn+5veXm2z8Xb+POPbyBmHxxJkjAAAAABDDEQAAAABIYjgCAAAAAEkMRwAAAAAgiQ0ZPtJnvveyyTKizJPX7b3EZDlrXhiIkgZNls9erNcV5cbWmb7e73YNIDkdKbB9bvgJvF7oc2eazGX6Io4byv1mzdGiLpNlZNtLgv/3c8tMluUzkRp77Hvc/uaVEcd/DNnNKIZl2PcMbm0zGd0QwGD543VlJlt905IoK7NMclPDBSbrmmP7Y887+2KqLdlx5ggAAAAAxHAEAAAAAJIYjgAAAABAEsMRAAAAAEhiQ4aw5v/fXtj2neC/mSykbJPV/e+nTFai5L5TepezFyCHZC9UXr/LfvZT9dKA1ATgxHV2RF6cG4qyjcDy2+412c8qz4j5Pb818j9NlqHIHROOuKNmzf4e24fuf+dCk5X/er7J8rfbXj3mf5tM5tv7dsTxO7tyzJpgpt0Ywr34O5MBwEDIPP2TJnv+zvujrBzap9erfftkkxW/taOfVaUuzhwBAAAAgBiOAAAAAEASwxEAAAAASGI4AgAAAABJbMgQ1m2vwVUgw17QW9th7yB8ymP77evFpar4yxg2zGSv/dvEKCvrTPK3b15msglf32Myewk1AK8Y99XtEcenV1WaNcVT/hDX99x4cLzJ3vnl2Ijjka/aTQ+y178Y5dXsuvHa1qc6ovWmP3zrnIjjKf5as2bl+x/v0+sDwEB4/Tb7vVu0jbP6quQum9mtedIXZ44AAAAAQAxHAAAAACCJ4QgAAAAAJDEcAQAAAIAkNmTot3d7Rpis+823Br+QPjp2A4b6uz5t1rw2w95l+ZeHAybbXz3OZLnvbTmB6gAkWukiuwHBYBijfQl532MNO/+d4675zsbZJhuvFwaiHABQ6IIzI47vPGtNzK/11zuuNtmIbTtifr10wJkjAAAAABDDEQAAAABIYjgCAAAAAElcc9Rv33zuyyYbH+WGqYlw7O+oStLBhUcijnedZa8vuuR3V5ls+OffNFmuuL4IQPr5xFpujwhg8PzLih9FHE/M6lsP+uaB800WuOY9k8V++9j0wJkjAAAAABDDEQAAAABIYjgCAAAAAEkMRwAAAAAgiQ0Z/sxno4wos+MPznvCZNUaPxAVfaS93y0z2U//7h6Tjc/Kjjj+7AtzzJqiK3fGrzAAAADE7MzsyO8/u1zftlCoXf5Zk41+7/m41JROOHMEAAAAAGI4AgAAAABJDEcAAAAAIKmfw1FVVZWmTJmi3NxcjR49WjNnzlR9fX3Emo6ODlVUVGjkyJEaMWKEZs+eraamprgWDQBeQ38EgN7RI5Es+rUhw6ZNm1RRUaEpU6aou7tbt912my699FLt3LlTw4cPlyQtWLBATz31lFatWqVAIKDKykrNmjVLzz333IB8gLiJcvPhkEImuyDnXZPNXzHZZH+1PPK5WY1tZk3TBR8zWcFVb5tsXkmNyS4bVmeyn7UHTfZ3v/t8xPGoHw43awCcuJTuj2kk02d/Zvje+CyTFf5yMKoBUgc9MrqG/55osizfyzG91phnDpmsb1s54C/1azhav359xPGKFSs0evRo1dXV6fzzz1dLS4sefvhhPf7447r44oslScuXL9dpp52mLVu26Oyzz45f5QDgIfRHAOgdPRLJ4oSuOWppaZEkFRQUSJLq6urU1dWl8vLy8JoJEyaopKREtbW1UV+js7NTra2tEQ8ASHbx6I8SPRJAauJ7SHhVzMNRKBTS/Pnzde6552rixA9OCTY2Nio7O1v5+fkRa4PBoBobG6O+TlVVlQKBQPhRXFwca0kA4Anx6o8SPRJA6uF7SHhZzMNRRUWFduzYoZUrV55QAYsWLVJLS0v40dDQcEKvBwCJFq/+KNEjAaQevoeEl/XrmqMPVVZWat26ddq8ebPGjh0bzgsLC3X06FE1NzdHTP5NTU0qLCyM+lp+v19+vz+WMhJiqM/+ke3664dM9uznhkYc7+60n/+6wFsx1/H1/Z8z2frnzzDZqV/fEvN7AOi/ePZHKfl6ZLLrcXYjHm56AcRPOn8PGbrgTJPdd8ZPTNblIrdRaAl1mDVTfjnfZBP27oy9OIT1q+U751RZWanVq1drw4YNKi0tjfj65MmTlZWVpZqaP++uVl9fr3379qmsrCw+FQOAB9EfAaB39Egki36dOaqoqNDjjz+utWvXKjc3N/w7oIFAQDk5OQoEArrhhhu0cOFCFRQUKC8vT/PmzVNZWRm7jABIafRHAOgdPRLJol/D0YMPPihJuvDCCyPy5cuX69prr5Uk3XvvvcrIyNDs2bPV2dmp6dOn64EHHohLsQDgVfRHAOgdPRLJol/DkXNR7pR6jKFDh6q6ulrV1dUxFwUAyYb+CAC9o0ciWcS0IUMqCj5z0GTf+gf7O67fL+z9fiR/6fyhRyOOzxv6Vp+et73TXgZ2zaa5Jht/XZ3JThWbLwBAvB2ecjjRJQBIAR0F2SY7b2h7lJWZEUe/OlxiVoyf+6LJomwngxiwBw8AAAAAiOEIAAAAACQxHAEAAACAJIYjAAAAAJDEhgxhPa//3mS7v3yyyT41b57Jdn5lWUzvOeEXt5jskw/YC3/Hb7ebLwAA4i/Tx88MASCd8X8BAAAAABDDEQAAAABIYjgCAAAAAEkMRwAAAAAgiQ0ZPlL3m2+ZbNwCm12xYEpMrz9e9u7GLqZXAgDEovPXH4s47jmDe8wDGBh5LzeabN7bF5vsoeJNg1EOesGZIwAAAAAQwxEAAAAASGI4AgAAAABJXHMEAEhjhfc+H3H8hXs/a9acopcHqRoAqax7z16TvX22XfdFTR6EatAbzhwBAAAAgBiOAAAAAEASwxEAAAAASGI4AgAAAABJDEcAAAAAIInhCAAAAAAkMRwBAAAAgCSGIwAAAACQxHAEAAAAAJIYjgAAAABAEsMRAAAAAEhiOAIAAAAASQxHAAAAACBJGpLoAo7lnJMkdatLcgkuBp7TrS5Jf/57AqQbeiR6Q39EuqM/4qP0tUd6bjhqa2uTJD2rXyS4EnhZW1ubAoFAossABh09EsdDf0S6oj+iL47XI33OYz9iCoVC2r9/v3Jzc9XW1qbi4mI1NDQoLy8v0aX1W2tra1LXL3nvMzjn1NbWpqKiImVk8FuhSD8f9kjnnEpKSjzzbzMWXusv/eW1+umPSHf0R+/wYv197ZGeO3OUkZGhsWPHSpJ8Pp8kKS8vzzN/sLFI9volb30GfiKKdPZhj2xtbZXkrX+bsUr2z+Cl+umPSGf0R+/xWv196ZH8aAkAAAAAxHAEAAAAAJI8Phz5/X7dcccd8vv9iS4lJslev5QanwFIRanwbzPZP0Oy1w+kqlT4t5nsnyGZ6/fchgwAAAAAkAiePnMEAAAAAIOF4QgAAAAAxHAEAAAAAJIYjgAAAABAEsMRAAAAAEjy8HBUXV2tk08+WUOHDtW0adP0wgsvJLqkXm3evFmXX365ioqK5PP5tGbNmoivO+e0ePFijRkzRjk5OSovL9fu3bsTU2wUVVVVmjJlinJzczV69GjNnDlT9fX1EWs6OjpUUVGhkSNHasSIEZo9e7aampoSVDGAZOmR9EcAgy1Z+qNEj/QiTw5HTz75pBYuXKg77rhDL730kiZNmqTp06fr4MGDiS4tqvb2dk2aNEnV1dVRv3733Xdr6dKleuihh7R161YNHz5c06dPV0dHxyBXGt2mTZtUUVGhLVu26Omnn1ZXV5cuvfRStbe3h9csWLBAP//5z7Vq1Spt2rRJ+/fv16xZsxJYNZC+kqlH0h8BDKZk6o8SPdKTnAdNnTrVVVRUhI97enpcUVGRq6qqSmBVfSPJrV69OnwcCoVcYWGhW7JkSThrbm52fr/fPfHEEwmo8PgOHjzoJLlNmzY55z6oNysry61atSq8ZteuXU6Sq62tTVSZQNpK1h5JfwQw0JK1PzpHj/QKz505Onr0qOrq6lReXh7OMjIyVF5ertra2gRWFps9e/aosbEx4vMEAgFNmzbNs5+npaVFklRQUCBJqqurU1dXV8RnmDBhgkpKSjz7GYBUlUo9kv4IIJ5SqT9K9MhE8dxwdOjQIfX09CgYDEbkwWBQjY2NCaoqdh/WnCyfJxQKaf78+Tr33HM1ceJESR98huzsbOXn50es9epnAFJZKvVI+iOAeEql/ijRIxNlSKILgLdUVFRox44devbZZxNdCgB4Cv0RAHqXKj3Sc2eORo0apczMTLOLRVNTkwoLCxNUVew+rDkZPk9lZaXWrVunjRs3auzYseG8sLBQR48eVXNzc8R6L34GINWlUo+kPwKIp1TqjxI9MlE8NxxlZ2dr8uTJqqmpCWehUEg1NTUqKytLYGWxKS0tVWFhYcTnaW1t1datWz3zeZxzqqys1OrVq7VhwwaVlpZGfH3y5MnKysqK+Az19fXat2+fZz4DkC5SqUfSHwHEUyr1R4kemTAJ3hAiqpUrVzq/3+9WrFjhdu7c6ebOnevy8/NdY2NjokuLqq2tzW3fvt1t377dSXL33HOP2759u9u7d69zzrm77rrL5efnu7Vr17pXXnnFzZgxw5WWlrojR44kuPIP3HzzzS4QCLhnnnnGHThwIPw4fPhweM1NN93kSkpK3IYNG9y2bdtcWVmZKysrS2DVQPpKph5JfwQwmJKpPzpHj/QiTw5Hzjm3bNkyV1JS4rKzs93UqVPdli1bEl1SrzZu3OgkmcecOXOccx9sxXj77be7YDDo/H6/u+SSS1x9fX1ii/4L0WqX5JYvXx5ec+TIEXfLLbe4k046yQ0bNsxdeeWV7sCBA4krGkhzydIj6Y8ABluy9Efn6JFe5HPOuYE9NwUAAAAA3ue5a44AAAAAIBEYjgAAAABADEcAAAAAIGkAbwJbXV2tJUuWqLGxUZMmTdKyZcs0derU4z4vFApp//79ys3Nlc/nG6jykKScc2pra1NRUZEyMpjtkZxi7Y8SPRK9oz8iFdAfMVD63CMHYpeHlStXuuzsbPfII4+4V1991d14440uPz/fNTU1Hfe5DQ0Nve58wYPHh4+GhoaB+KsLDLgT6Y/O0SN5HP9Bf0Syoj/yGIzH8XrkgOxWN23aNE2ZMkX333+/pA8m+eLiYs2bN0/f/va3P/K5LS0tys/P13n6goYoK96lIcl1q0vP6hdqbm5WIBBIdDlAv51If5Tokegd/RHJjv6IgdTXHhn3X6s7evSo6urqtGjRonCWkZGh8vJy1dbWmvWdnZ3q7OwMH7e1tf2psCwN8fEXG8f40yjP6XIko/72R4keiX6gPyKJ0R8x4PrYI+P+S8mHDh1ST0+PgsFgRB4MBtXY2GjWV1VVKRAIhB/FxcXxLgkAPKG//VGiRwJID/RHeEXCr9hctGiRWlpawo+GhoZElwQAnkGPBIDo6I8YCHH/tbpRo0YpMzNTTU1NEXlTU5MKCwvNer/fL7/fH+8yAMBz+tsfJXokgPRAf4RXxP3MUXZ2tiZPnqyamppwFgqFVFNTo7Kysni/HQAkDfojAERHf4RXDMh9jhYuXKg5c+borLPO0tSpU3Xfffepvb1d11133UC8HQAkDfojAERHf4QXDMhwdNVVV+mdd97R4sWL1djYqDPOOEPr1683F9kBQLqhPwJAdPRHeMGA3OfoRLS2tioQCOhCzWAbRhjdrkvPaK1aWlqUl5eX6HKAQUePRG/oj0h39Ed8lL72yITvVgcAAAAAXsBwBAAAAABiOAIAAAAASQxHAAAAACCJ4QgAAAAAJDEcAQAAAIAkhiMAAAAAkMRwBAAAAACSGI4AAAAAQJI0JNEFAAAAAOi715dPNtme6Q+b7J4/nmKyX3/lLJP17Hw9PoWlAM4cAQAAAIAYjgAAAABAEsMRAAAAAEhiOAIAAAAASWzIAABIQZkjC0zmC+SZbN/soojjjlHOrBn3z781Wejw4ROoDgD6LvP0T5ps7UXVJutyWSarOKneZP/9mUtNlrszxuJSEGeOAAAAAEAMRwAAAAAgieEIAAAAACQxHAEAAACAJDZkAAAkkYyJE0y2e1GOya7/9PMm+8bIX8X0nqcFbzLZqdfWxfRaANBvf2g00ddev9pkT5/+08GoJuVx5ggAAAAAxHAEAAAAAJIYjgAAAABAEsMRAAAAAEhiQ4aEOjr9LJPt/duQyW7+7CaTzT/p9T69x6f/c17E8bAD9u7vzed0muwT/9fOzdm/2tan9wSA/vJN+bTJ3liQabJnzrvfZB/L9JssI8rP/p46fJLJ3uwcHXEc7W7yPz7/P0z2vSlzTOZe/J3JAOBE9TS3mGzv26fahacPQjFpgDNHAAAAACCGIwAAAACQxHAEAAAAAJIYjgAAAABAEhsyDJp3bioz2bJbq012lr/HZNEuLJ7zVrnJzgzsM9lv//4Hx60t2uufU3CNyQpiu7k8gDSX+bGPmez1H3w84vjn5zxg1pySlRXl1ezmC9Esby022ZrZ55ks5I98j4p1dkOGaH35SDDHZEP7VBkA9E9mcLTJPnda3zbmQv9x5ggAAAAAxHAEAAAAAJIYjgAAAABAEsMRAAAAAEhiQ4a48GVlm6yjfFLE8U8XLTFriobYC4tv2PvXJtv7b5802fCnXjbZxmElJtu0enxkHaf+zKyJpvXlkSYr6NMzASDSH75q7+T+6gXHbhYTbfOFvvlJtM0XZp5jsp56ewGz70xuKQ/A43KHm+gLBS/G/HIHJ/tMlv/KeJP17EzPTR84cwQAAAAAYjgCAAAAAEkMRwAAAAAgKYbhaPPmzbr88stVVFQkn8+nNWvWRHzdOafFixdrzJgxysnJUXl5uXbv3h2vegHAs+iPABAd/RHJot8bMrS3t2vSpEm6/vrrNWvWLPP1u+++W0uXLtWjjz6q0tJS3X777Zo+fbp27typoUNT8/7hByrPMtkL3zz2YmO7+cKX37jcZN2zu0w27NBWk7kodeyfO9lkW089tg7rl4dzTTbuhw22tuO+EpDe6I/RffyKt2J63n+/X2iye16/xGTBW21H7Knv2zdV7306r/+FAeg3+mPset7YY7Lv/Pwqk82+prpPr/fq3yw12ZktXzdZcZpuyNDv4eiyyy7TZZddFvVrzjndd999+s53vqMZM2ZIkh577DEFg0GtWbNGV199tXlOZ2enOjs7w8etra39LQkAPCHe/VGiRwJIDfRHJIu4XnO0Z88eNTY2qry8PJwFAgFNmzZNtbW1UZ9TVVWlQCAQfhQX2y1ZASDZxdIfJXokgNRHf4SXxHU4amxslCQFg8GIPBgMhr92rEWLFqmlpSX8aGiwv84FAMkulv4o0SMBpD76I7wk4TeB9fv98vvt9ThetXvZNJPVz1pmstAxx6c9fZNZM+Gbb5ms59C7sZamm25eG9Pz7vyXOSY7qaH3n9QAGDzJ1iOjutHW/6mKeRHHxU/3mDXDX7XfFI3aa38H3j6z7w4H7c0QASSHlOiPMfqrb26x4TWDX0cqiuuZo8LCDy6ebWpqisibmprCXwOAdER/BIDo6I/wkrgOR6WlpSosLFRNTU04a21t1datW1VWVhbPtwKApEJ/BIDo6I/wkn7/Wt3777+vN954I3y8Z88evfzyyyooKFBJSYnmz5+vO++8U6eeemp4K8aioiLNnDkznnUDgOfQHwEgOvojkkW/h6Nt27bpoosuCh8vXLhQkjRnzhytWLFCt956q9rb2zV37lw1NzfrvPPO0/r169N+j3oAqY/+CADR0R+RLHzOuWj3E02Y1tZWBQIBXagZGuLLSmgtv//3s02262p7g62WUIfJvvza30QcD531R7Mm1NbWpzoyhg832btf+ozJ1n5viclGZeZEHE9YVWHWjJsf5aI+j+p2XXpGa9XS0qK8PG7eiPTjpR6ZCvb+16cjjn977iN9et6l/2B76dB1L8SlpljRH5Hu0r0//mr/yybrcn3bsmbSg/NMVnzn8ydakqf0tUfG9ZojAAAAAEhWDEcAAAAAIIYjAAAAAJDEcAQAAAAAkmLYrS5VZQZHm+zRKx8wWUghkx27+YIkZf/13mOe1zcZZ3zKZBMf2WWyO4NLozzb3iX63Jevjjj+5P+xr3Uid5cHgMG0b/E5JuseFmVfIV+UJ0dZNuvU2uO+Z+XbF5osZ/1LfXl5ABg00TZfiPZ9Kz4aZ44AAAAAQAxHAAAAACCJ4QgAAAAAJDEcAQAAAIAkNmQI8w21mxmc5e/bVgU5X8u2r/eJ4ojj3TeNNWsuLbcX9C4Y/SOTlQzJMVm0y+t6nL0c2PfkqMg1zbujPBMABlfmMXcn75h6qlmTtajJZK9MWNan18/yZZqsL3eK33hkmMnenltiMtdtN7cBACQ/zhwBAAAAgBiOAAAAAEASwxEAAAAASGI4AgAAAABJbMgQ5jo6Tba1M8tk0/xdJlv765Umi/WOxL8+Mspku7vsRgsX5bxvsm1H7cYQ+Y8d/+7vABAvPr/d3OboBZ822YIHfhxxfFFOjVnT1GP78sYjJ5ls8eszTPbE6StMVjTE1nasoRm2x7/5lXyTnVI/1GShjo7jvj4AwNs4cwQAAAAAYjgCAAAAAEkMRwAAAAAgieEIAAAAACSxIUNYT9NBk91x89+b7N8eesBkn7H7IOgnrcURx3duusKsGb/CXrw7pKnFZKOf+KPJLireYLI5G22947XNFgcAcZAx1G5K8O5VZ5rsN/+69LivdfoT80w2dmOPyfxPvWiykWPsBjVP/Gqyyb4xcsdx64i26c4r19r6yxq+ZrLgY781Wejw4eO+JwDEQ5Yv02RR9vSKKu8c+31wuuLMEQAAAACI4QgAAAAAJDEcAQAAAIAkhiMAAAAAkMSGDB8p+1d2M4PbSqfG9Frj9UKf1rXNsK//VMlak3U5O9fmvBVlZwgAiAOf32+y1+75jM1mHH/zBUmaUT8z4nj8kjfNmmgb5QwpHmuyST/bZ7J/HLnTZC2hoyab9tNvRByPmWDfs+bTT5qs9nb7Oa+65osmO7T00yYb+q7d9OFYmc+8dNw1APCXupzdxCakUJ+eu2nSEya74uwbIoMtr8RUV7LhzBEAAAAAiOEIAAAAACQxHAEAAACAJK458pzuHDuv9vV3SEtX2N+7745PWQDSiG+I/V9D/X2TTPbaFdUme7u702RX/PBWk538yO8jjrujXF/UVW5v5Drx+9tNdsfoOpMtb/2EyX78T5ebbNz/bIk4zhw10qy58K/tDWrbr7I37F595n+YbOxSe61WNOvaI9/3R+NP6dPzAOBDEzb8vcl2XvyjmF/v9bmR17KP39LLwhTDmSMAAAAAEMMRAAAAAEhiOAIAAAAASQxHAAAAACCJDRk8J3dllKvd/n3w6wCQvhr+0d6M+rUrfmCy/VE2X/jyXf9ospPX2Bu8/vHi0ohj99Vcs+a/J9r3/Fim3eDg9JV2w4TxPzpksmH1W012rJ5D75os74lomX3ul26xG08Ev7T3uO8pSfpG/jHBq317HgD8if/1HBtePPh1JDvOHAEAAACAGI4AAAAAQBLDEQAAAABI6udwVFVVpSlTpig3N1ejR4/WzJkzVV9fH7Gmo6NDFRUVGjlypEaMGKHZs2erqakprkUDgNfQHwGgd/RIJAufc871dfHnP/95XX311ZoyZYq6u7t12223aceOHdq5c6eGDx8uSbr55pv11FNPacWKFQoEAqqsrFRGRoaee+65Pr1Ha2urAoGALtQMDfFlxfapkljb1WebbNO/27vQhxQy2ayzrzRZd8Pb8SnMI7pdl57RWrW0tCgvLy/R5QBhg9EfpcHpkf/05ssmm+bvMtkfe+yGDA+9N81kH89+z2Rz8vq4UcExTn/8ayYbt+hFk7nu7pheP5nRH+FlfA+ZGNe8tt9kf5t7oE/PzfJlRhxfdtk1Zk3ot7tiKywB+toj+7Vb3fr16yOOV6xYodGjR6uurk7nn3++Wlpa9PDDD+vxxx/XxRd/sD3G8uXLddppp2nLli06+2z7jT8ApAL6IwD0jh6JZHFC1xy1tLRIkgoKCiRJdXV16urqUnl5eXjNhAkTVFJSotra2qiv0dnZqdbW1ogHACS7ePRHiR4JIDXxPSS8KubhKBQKaf78+Tr33HM1ceJESVJjY6Oys7OVn58fsTYYDKqxsTHq61RVVSkQCIQfxcXFsZYEAJ4Qr/4o0SMBpB6+h4SXxTwcVVRUaMeOHVq5cuUJFbBo0SK1tLSEHw0NDSf0egCQaPHqjxI9EkDq4XtIeFm/rjn6UGVlpdatW6fNmzdr7Nix4bywsFBHjx5Vc3NzxOTf1NSkwsLCqK/l9/vl99s7nqerllPYXR1IZvHsj1JieuTm9yeYbJr/dyYryLR13Tbq5T69xxdfmxVxvK92rFlzyn+3mGzcq3UmS8fNF4BkxfeQg2vFvnNMds3pq/r03K4+b9mWWvr1nbhzTpWVlVq9erU2bNig0tLSiK9PnjxZWVlZqqmpCWf19fXat2+fysrK4lMxAHgQ/REAekePRLLo15mjiooKPf7441q7dq1yc3PDvwMaCASUk5OjQCCgG264QQsXLlRBQYHy8vI0b948lZWVscsIgJRGfwSA3tEjkSz6NRw9+OCDkqQLL7wwIl++fLmuvfZaSdK9996rjIwMzZ49W52dnZo+fboeeOCBuBQLAF5FfwSA3tEjkSz6NRz15X6xQ4cOVXV1taqr7Y1LASBV0R8BoHf0SCSLmDZkwMD5+KbDJsuqzDRZul4kB2DgPX9Rkcmm/e3FJmuZdNRkQ96xd6Uf/9Af7LrGgxHHJ3fYXaZCH1klAOB4OldE2cxiyeDXkUzYGg0AAAAAxHAEAAAAAJIYjgAAAABAEsMRAAAAAEhiQwbP8T33sslWtI422TW59gLnw6ePMVl2w9txqQtA+uh5948mCy593mZ9fL3uE6wHABCbk162/bz6vU+arOKk+sEoJylw5ggAAAAAxHAEAAAAAJIYjgAAAABAEsMRAAAAAEhiQ4akcO8Pv2Sya775A5ONuf0Nk73b/JnIYMsrcasLAAAA3tWz83WT/Wpins00pQ+vtisOFXkfZ44AAAAAQAxHAAAAACCJ4QgAAAAAJHHNUVL4+I/tjbmumvlFkz05bp3JLlh8TcRxwd8EzJqe5pYTqA4AAABIDZw5AgAAAAAxHAEAAACAJIYjAAAAAJDEcAQAAAAAktiQISn0HHrXZEdnjzTZaf/+DybbVf7DiOMrJtxg34AbwwIAAACcOQIAAAAAieEIAAAAACQxHAEAAACAJIYjAAAAAJDEhgxJK9omDafOsdkVmnJMwuYLAAAAQDScOQIAAAAAMRwBAAAAgCSGIwAAAACQ5MFrjpxzkqRudUkuwcXAc7rVJenPf0+AdEOPRG/oj0h39Ed8lL72SM8NR21tbZKkZ/WLBFcCL2tra1MgEEh0GcCgo0fieOiPSFf0R/TF8Xqkz3nsR0yhUEj79+9Xbm6u2traVFxcrIaGBuXl5SW6tH5rbW1N6vol730G55za2tpUVFSkjAx+KxTp58Me6ZxTSUmJZ/5txsJr/aW/vFY//RHpjv7oHV6sv6890nNnjjIyMjR27FhJks/nkyTl5eV55g82Fslev+Stz8BPRJHOPuyRra2tkrz1bzNWyf4ZvFQ//RHpjP7oPV6rvy89kh8tAQAAAIAYjgAAAABAkseHI7/frzvuuEN+vz/RpcQk2euXUuMzAKkoFf5tJvtnSPb6gVSVCv82k/0zJHP9ntuQAQAAAAASwdNnjgAAAABgsDAcAQAAAIAYjgAAAABAEsMRAAAAAEhiOAIAAAAASR4ejqqrq3XyySdr6NChmjZtml544YVEl9SrzZs36/LLL1dRUZF8Pp/WrFkT8XXnnBYvXqwxY8YoJydH5eXl2r17d2KKjaKqqkpTpkxRbm6uRo8erZkzZ6q+vj5iTUdHhyoqKjRy5EiNGDFCs2fPVlNTU4IqBpAsPZL+CGCwJUt/lOiRXuTJ4ejJJ5/UwoULdccdd+ill17SpEmTNH36dB08eDDRpUXV3t6uSZMmqbq6OurX7777bi1dulQPPfSQtm7dquHDh2v69Onq6OgY5Eqj27RpkyoqKrRlyxY9/fTT6urq0qWXXqr29vbwmgULFujnP/+5Vq1apU2bNmn//v2aNWtWAqsG0lcy9Uj6I4DBlEz9UaJHepLzoKlTp7qKiorwcU9PjysqKnJVVVUJrKpvJLnVq1eHj0OhkCssLHRLliwJZ83Nzc7v97snnngiARUe38GDB50kt2nTJufcB/VmZWW5VatWhdfs2rXLSXK1tbWJKhNIW8naI+mPAAZasvZH5+iRXuG5M0dHjx5VXV2dysvLw1lGRobKy8tVW1ubwMpis2fPHjU2NkZ8nkAgoGnTpnn287S0tEiSCgoKJEl1dXXq6uqK+AwTJkxQSUmJZz8DkKpSqUfSHwHEUyr1R4kemSieG44OHTqknp4eBYPBiDwYDKqxsTFBVcXuw5qT5fOEQiHNnz9f5557riZOnCjpg8+QnZ2t/Pz8iLVe/QxAKkulHkl/BBBPqdQfJXpkogxJdAHwloqKCu3YsUPPPvtsoksBAE+hPwJA71KlR3ruzNGoUaOUmZlpdrFoampSYWFhgqqK3Yc1J8Pnqays1Lp167Rx40aNHTs2nBcWFuro0aNqbm6OWO/FzwCkulTqkfRHAPGUSv1RokcmiueGo+zsbE2ePFk1NTXhLBQKqaamRmVlZQmsLDalpaUqLCyM+Dytra3aunWrZz6Pc06VlZVavXq1NmzYoNLS0oivT548WVlZWRGfob6+Xvv27fPMZwDSRSr1SPojgHhKpf4o0SMTJsEbQkS1cuVK5/f73YoVK9zOnTvd3LlzXX5+vmtsbEx0aVG1tbW57du3u+3btztJ7p577nHbt293e/fudc45d9ddd7n8/Hy3du1a98orr7gZM2a40tJSd+TIkQRX/oGbb77ZBQIB98wzz7gDBw6EH4cPHw6vuemmm1xJSYnbsGGD27ZtmysrK3NlZWUJrBpIX8nUI+mPAAZTMvVH5+iRXuTJ4cg555YtW+ZKSkpcdna2mzp1qtuyZUuiS+rVxo0bnSTzmDNnjnPug60Yb7/9dhcMBp3f73eXXHKJq6+vT2zRfyFa7ZLc8uXLw2uOHDnibrnlFnfSSSe5YcOGuSuvvNIdOHAgcUUDaS5ZeiT9EcBgS5b+6Bw90ot8zjk3sOemAAAAAMD7PHfNEQAAAAAkAsMRAAAAAIjhCAAAAAAkDeBNYKurq7VkyRI1NjZq0qRJWrZsmaZOnXrc54VCIe3fv1+5ubny+XwDVR6SlHNObW1tKioqUkYGsz2SU6z9UaJHonf0R6QC+iMGSp975EDs8rBy5UqXnZ3tHnnkEffqq6+6G2+80eXn57umpqbjPrehoaHXnS948Pjw0dDQMBB/dYEBdyL90Tl6JI/jP+iPSFb0Rx6D8ThejxyQ3eqmTZumKVOm6P7775f0wSRfXFysefPm6dvf/vZHPrelpUX5+fk6T1/QEGXFuzQkuW516Vn9Qs3NzQoEAokuB+i3E+mPEj0SvaM/ItnRHzGQ+toj4/5rdUePHlVdXZ0WLVoUzjIyMlReXq7a2lqzvrOzU52dneHjtra2PxWWpSE+/mLjGH8a5TldjmTU3/4o0SPRD/RHJDH6IwZcH3tk3H8p+dChQ+rp6VEwGIzIg8GgGhsbzfqqqioFAoHwo7i4ON4lAYAn9Lc/SvRIAOmB/givSPgVm4sWLVJLS0v40dDQkOiSAMAz6JEAEB39EQMh7r9WN2rUKGVmZqqpqSkib2pqUmFhoVnv9/vl9/vjXQYAeE5/+6NEjwSQHuiP8Iq4nznKzs7W5MmTVVNTE85CoZBqampUVlYW77cDgKRBfwSA6OiP8IoBuc/RwoULNWfOHJ111lmaOnWq7rvvPrW3t+u6664biLcDgKRBfwSA6OiP8IIBGY6uuuoqvfPOO1q8eLEaGxt1xhlnaP369eYiOwBIN/RHAIiO/ggvGJD7HJ2I1tZWBQIBXagZbMMIo9t16RmtVUtLi/Ly8hJdDjDo6JHoDf0R6Y7+iI/S1x6Z8N3qAAAAAMALGI4AAAAAQAxHAAAAACCJ4QgAAAAAJDEcAQAAAIAkhiMAAAAAkMRwBAAAAACSGI4AAAAAQBLDEQAAAABIkoYkugAAAAAA6WXkcyeZLMPnTPbOOc2DUM1f1DCo7wYAAAAAHsVwBAAAAABiOAIAAAAASQxHAAAAACCJDRkGjW/y6SYLZds//j9cONxkr857wGRdric+hfXikh1fMtnwGQdMFuroGNA6AKQvn99vssOXTTLZZ/7ptybbPaVzQGoCAPTf6w+fZbIXS35gsrLfVJjsFL08ECX1ijNHAAAAACCGIwAAAACQxHAEAAAAAJIYjgAAAABAEhsyxIUrsxcI7742O+L43oufMGuyfN0mK89pM1mXszNsSKH+lNhvT0/8L5Od8ePrTVZ6836T9Rx6d0BqApBeMj82ymQbqx8y2W867P/KlpRebrLuPXvjUxgAoFevPzjVZC9eeq/J2kLOZHmbcgakpv7gzBEAAAAAiOEIAAAAACQxHAEAAACAJIYjAAAAAJDEhgxx4e78o8lem/A/CahkYL18ziMmmz7tFpP5n2JDBgCD53ND7eY2/1JSYLIMNmQAgAF34Zm7TJabkW2yW/Z+3mSjflg7IDX1B2eOAAAAAEAMRwAAAAAgieEIAAAAACQxHAEAAACAJDZkiIs/PFNswwnHf15th99k1//iRrvQF+XJ9qbCUZ392dcjjpef/L99eyIAJIlMHz/nA5BejsyYarJR39hjss6rMk3WfaAxbnUcvOUck30/eK/JftL6CZO9t6jEZBlK/KZe/B8FAAAAAMRwBAAAAACSGI4AAAAAQBLDEQAAAABIYkOGuCi5a5vJrvyva477PN/RLpOdumdrXGr6UPOokRHHv96Sa9aU57T16bUu/t1VJsvb+KrJQn2sDQDiocfZrtM1zP7vzW6BAwDJ6at3rTPZdXkNJiuffLPJhq6L34YMcyp+YbIz/Lbb3vi9K01W8JvauNURT5w5AgAAAAAxHAEAAACAJIYjAAAAAJAUw3C0efNmXX755SoqKpLP59OaNWsivu6c0+LFizVmzBjl5OSovLxcu3fvjle9AOBZ9EcAiI7+iGTR7w0Z2tvbNWnSJF1//fWaNWuW+frdd9+tpUuX6tFHH1Vpaaluv/12TZ8+XTt37tTQoUPjUrTXuK6jJuupfyMBlVhNs8ZHHH86e22UVX27THn//gKTjTj8ZixlASmJ/ugdBydnmaz4lwkoBIAk+mO8HTiab7KQ9pqsO8cXt/cMXXCmyWaMWGayLpdj6xgavzoGWr+Ho8suu0yXXXZZ1K8553TffffpO9/5jmbMmCFJeuyxxxQMBrVmzRpdffXV5jmdnZ3q7OwMH7e2tva3JADwhHj3R4keCSA10B+RLOJ6zdGePXvU2Nio8vLycBYIBDRt2jTV1kbfrq+qqkqBQCD8KC4ujmdJAOAJsfRHiR4JIPXRH+ElcR2OGhs/2Dc9GAxG5MFgMPy1Yy1atEgtLS3hR0OD3aMdAJJdLP1RokcCSH30R3hJwm8C6/f75Y9ysyj03zs3l5lswldfizgOZsb+Z33arXtM1hPzqwHoi3Tuka7L3ij79a4Ok43PstcjHCm114ICSC3p0h93L51mstUj7bU+DzaPN1n+lj+YrLuP75uZH4g4PvTNdrOmaIj981+w/xyTBR+uM5nrYx2DLa5njgoLCyVJTU1NEXlTU1P4awCQjuiPABAd/RFeEtfhqLS0VIWFhaqpqQlnra2t2rp1q8rK7FkNAEgX9EcAiI7+CC/p96/Vvf/++3rjjT9vU71nzx69/PLLKigoUElJiebPn68777xTp556angrxqKiIs2cOTOedQOA59AfASA6+iOSRb+Ho23btumiiy4KHy9cuFCSNGfOHK1YsUK33nqr2tvbNXfuXDU3N+u8887T+vXr2aMeQMqjPwJAdPRHJAufc85T10O1trYqEAjoQs3QEJ+9iV86OlhpL2ybc/MvTPbVvFdNlpuRHdN7fu+dz5qs7nP5Jgu1tcX0+rHqdl16RmvV0tKivLy8QX1vwAvSvUeGauxWvesn2Jtbj1s/12Tjb9g2IDV5Bf0R6S5V+mPmJ8dFHP/TL//LrPlk1hGTzZq3wGQ5a16IuY43Hz8j4njHBf9h1vz6SK7Jlo6bEPN7DqS+9si4XnMEAAAAAMmK4QgAAAAAxHAEAAAAAJIYjgAAAABAUgy71cHKPP2TJnv9upMiji84b0fMr7+u2N4FOaRQlJXH33zhjS57X+SrHvyGyUpWN5ks1Pb7474+AAAA+sade4bJrn54XcTxWf4es2bC+q+bbPwJbL7w1p32flLbzr/nmMSODd/6z+tN9nE9H3MdXsCZIwAAAAAQwxEAAAAASGI4AgAAAABJDEcAAAAAIIkNGfot2oVz1y5fbbIZww/F8V3jN8N+7Y2rTPbx79sL5+ylfwCQPEYUHE50CQDSmC/LbpJ1oPIsk237pt10K8uXGXHc5ez3gbPOeMlkP/u+3VRh3D//1mQZhaNNdsUXtpgsU76I4zOet5svlNyV3JsvRMOZIwAAAAAQwxEAAAAASGI4AgAAAABJDEcAAAAAIIkNGeIiU85kGXGcO4+9ME+Suuxb9sn60+zmEZ/72wqTBf6vvTAPAJLFTz/7Hyabp3MTUAmAdNR4k9184YVv/sBkoSjPPfZ7vMdaP27W/GvhVpt91Wa3lU8z2V8Hfmmyi3LeN9nWzqERxyVf/p1Zk4o4cwQAAAAAYjgCAAAAAEkMRwAAAAAgieEIAAAAACSxIUO/+Z572WQPz/y8yb597ciI45JfHTVrMo90x60uSdp9Q5bJXvv8g3F9DwBIpIZni204YfDrAIAPvXNTmcme/9Z9JmsLdZlsZ9dwk/3TN/8h4njou/Z7yJp/fctky0/+X5NF27gh2qZh0TaGOCs78n0XvLHLrPnB7Fn2tX5r1yUTzhwBAAAAgBiOAAAAAEASwxEAAAAASGI4AgAAAABJbMgQFz07XzfZKbcOfh2n7f6YDe1eEQCQtEY0uOMvkpTrs+syPzXeZNH6NwD0x6f+zm5A8LP2oMn+9UfXmGzMvz9vsmGymygc691vfMZkC5Z9zmT3Fv3muK/Vm0yfL+L4H38326wp+u3OmF/fqzhzBAAAAABiOAIAAAAASQxHAAAAACCJa45SStOscYkuAQAGVEYf75197O/KS1Iox94oGwBOVN2vPmWyP64cZbIx9fb6olgdCQ412byPbYiy0va9s79babJRv20/7nsWv/EHk/Uc91nJhzNHAAAAACCGIwAAAACQxHAEAAAAAJIYjgAAAABAEhsyhPn8fpM1f/lMk5209lWThdraBqSmj3LgG+eYbO3X7o6y0n4uAEhWJ62oNdlDt37CZDcF9pps94Jsk437anzqApC+Sv7ZbrQQz40KMj/2MZO9PdvuTjMuy37P93/bxphs1A9tH+2LVNx8IRrOHAEAAACAGI4AAAAAQBLDEQAAAABI6udwVFVVpSlTpig3N1ejR4/WzJkzVV9fH7Gmo6NDFRUVGjlypEaMGKHZs2erqakprkUDgNfQHwGgd/RIJIt+bciwadMmVVRUaMqUKeru7tZtt92mSy+9VDt37tTw4cMlSQsWLNBTTz2lVatWKRAIqLKyUrNmzdJzzz03IB8gFh2XTzVZ4Jv7TLZp3DKTXfniNfYF6+O3IcOQMYUm+8OXTjHZk/P+zWRFQ46/+UJTT6fJso64PlYHoDep0h+T0b9tmW6yz19yn8nG/8PrJgsNREEADHpk7HZ/Y5zJdl2y1GS1nVkm+68rPhflFX8fj7JSVr+Go/Xr10ccr1ixQqNHj1ZdXZ3OP/98tbS06OGHH9bjjz+uiy++WJK0fPlynXbaadqyZYvOPvvs+FUOAB5CfwSA3tEjkSxO6JqjlpYWSVJBQYEkqa6uTl1dXSovLw+vmTBhgkpKSlRbG33bwM7OTrW2tkY8ACDZxaM/SvRIAKmJ7yHhVTEPR6FQSPPnz9e5556riRMnSpIaGxuVnZ2t/Pz8iLXBYFCNjY1RX6eqqkqBQCD8KC4ujrUkAPCEePVHiR4JIPXwPSS8LObhqKKiQjt27NDKlStPqIBFixappaUl/GhoaDih1wOARItXf5TokQBSD99Dwsv6dc3RhyorK7Vu3Tpt3rxZY8eODeeFhYU6evSompubIyb/pqYmFRbajQYkye/3y+8//kYC8TT9XzaZ7Bsjd/Tpua/dlmfD96edaElhV59jTx2vGf2UyUKyF91FM+etyAuV31j+SbNm5P/EdqdkAFY8+6OUmB6ZCnrkM1noSEcCKgHwl5L9e8jBkPmp8RHH37vSDpE9zm6mdd3PbjLZuNe3xK+wNNGvM0fOOVVWVmr16tXasGGDSktLI74+efJkZWVlqaamJpzV19dr3759Kisri0/FAOBB9EcA6B09EsmiX2eOKioq9Pjjj2vt2rXKzc0N/w5oIBBQTk6OAoGAbrjhBi1cuFAFBQXKy8vTvHnzVFZWxi4jAFIa/REAekePRLLo13D04IMPSpIuvPDCiHz58uW69tprJUn33nuvMjIyNHv2bHV2dmr69Ol64IEH4lIsAHgV/REAekePRLLo13Dkovx+47GGDh2q6upqVVdXx1wUACQb+iMA9I4eiWQR04YM6WxX+Q8T8K720rDaDnsB4o1b/85k427cHXE8sp3NFwCkvr8akmOyd6+barKRD9MTAXjLV/7nmYjjK0ccNGs+u+U6k42bz+YL8XBCN4EFAAAAgFTBcAQAAAAAYjgCAAAAAEkMRwAAAAAgKU03ZNjwtXNN9tgt9kLd3577yIDW8ZPWYpMd6Mo32SMv2XrH/UePyU557mWThWKqDACSx/ILbK9+L3TEZKNeed9kx98/CwAG17+snR1xfM1Xl5o1Ob/IG6xy0g5njgAAAABADEcAAAAAIInhCAAAAAAkMRwBAAAAgKQ03ZAh85mXTFb6wjCTTf7a10326D/cZ7KJ2T6TXfy7qyKOW54pNGs+8eQfTNa9Z6/JTlWdyQAAH/jHXV8y2Zc+sd1kGe2dJrNb2wBAYp3yrdqI4yu+NcWsGalakyE+OHMEAAAAAGI4AgAAAABJDEcAAAAAIClNrzmKJnT4sMk+ftfzJrvtLnuz2GhG6M2PPJak7j7WBgDoXcEXXzfZBg2PstKuAwDgL3HmCAAAAADEcAQAAAAAkhiOAAAAAEASwxEAAAAASGI4AgAAAABJDEcAAAAAIInhCAAAAAAkMRwBAAAAgCSGIwAAAACQxHAEAAAAAJIYjgAAAABAEsMRAAAAAEhiOAIAAAAASdKQRBdwLOecJKlbXZJLcDHwnG51Sfrz3xMg3dAj0Rv6I9Id/REfpa890nPDUVtbmyTpWf0iwZXAy9ra2hQIBBJdBjDo6JE4Hvoj0hX9EX1xvB7pcx77EVMoFNL+/fuVm5urtrY2FRcXq6GhQXl5eYkurd9aW1uTun7Je5/BOae2tjYVFRUpI4PfCkX6+bBHOudUUlLimX+bsfBaf+kvr9VPf0S6oz96hxfr72uP9NyZo4yMDI0dO1aS5PP5JEl5eXme+YONRbLXL3nrM/ATUaSzD3tka2urJG/924xVsn8GL9VPf0Q6oz96j9fq70uP5EdLAAAAACCGIwAAAACQ5PHhyO/364477pDf7090KTFJ9vql1PgMQCpKhX+byf4Zkr1+IFWlwr/NZP8MyVy/5zZkAAAAAIBE8PSZIwAAAAAYLAxHAAAAACCGIwAAAACQxHAEAAAAAJIYjgAAAABAkoeHo+rqap188skaOnSopk2bphdeeCHRJfVq8+bNuvzyy1VUVCSfz6c1a9ZEfN05p8WLF2vMmDHKyclReXm5du/enZhio6iqqtKUKVOUm5ur0aNHa+bMmaqvr49Y09HRoYqKCo0cOVIjRozQ7Nmz1dTUlKCKASRLj6Q/AhhsydIfJXqkF3lyOHryySe1cOFC3XHHHXrppZc0adIkTZ8+XQcPHkx0aVG1t7dr0qRJqq6ujvr1u+++W0uXLtVDDz2krVu3avjw4Zo+fbo6OjoGudLoNm3apIqKCm3ZskVPP/20urq6dOmll6q9vT28ZsGCBfr5z3+uVatWadOmTdq/f79mzZqVwKqB9JVMPZL+CGAwJVN/lOiRnuQ8aOrUqa6ioiJ83NPT44qKilxVVVUCq+obSW716tXh41Ao5AoLC92SJUvCWXNzs/P7/e6JJ55IQIXHd/DgQSfJbdq0yTn3Qb1ZWVlu1apV4TW7du1yklxtbW2iygTSVrL2SPojgIGWrP3ROXqkV3juzNHRo0dVV1en8vLycJaRkaHy8nLV1tYmsLLY7NmzR42NjRGfJxAIaNq0aZ79PC0tLZKkgoICSVJdXZ26uroiPsOECRNUUlLi2c8ApKpU6pH0RwDxlEr9UaJHJornhqNDhw6pp6dHwWAwIg8Gg2psbExQVbH7sOZk+TyhUEjz58/Xueeeq4kTJ0r64DNkZ2crPz8/Yq1XPwOQylKpR9IfAcRTKvVHiR6ZKEMSXQC8paKiQjt27NCzzz6b6FIAwFPojwDQu1TpkZ47czRq1ChlZmaaXSyamppUWFiYoKpi92HNyfB5KisrtW7dOm3cuFFjx44N54WFhTp69Kiam5sj1nvxMwCpLpV6JP0RQDylUn+U6JGJ4rnhKDs7W5MnT1ZNTU04C4VCqqmpUVlZWQIri01paakKCwsjPk9ra6u2bt3qmc/jnFNlZaVWr16tDRs2qLS0NOLrkydPVlZWVsRnqK+v1759+zzzGYB0kUo9kv4IIJ5SqT9K9MiESfCGEFGtXLnS+f1+t2LFCrdz5043d+5cl5+f7xobGxNdWlRtbW1u+/btbvv27U6Su+eee9z27dvd3r17nXPO3XXXXS4/P9+tXbvWvfLKK27GjBmutLTUHTlyJMGVf+Dmm292gUDAPfPMM+7AgQPhx+HDh8NrbrrpJldSUuI2bNjgtm3b5srKylxZWVkCqwbSVzL1SPojgMGUTP3ROXqkF3lyOHLOuWXLlrmSkhKXnZ3tpk6d6rZs2ZLoknq1ceNGJ8k85syZ45z7YCvG22+/3QWDQef3+90ll1zi6uvrE1v0X4hWuyS3fPny8JojR464W265xZ100klu2LBh7sorr3QHDhxIXNFAmkuWHkl/BDDYkqU/OkeP9CKfc84N7LkpAAAAAPA+z11zBAAAAACJwHAEAAAAAGI4AgAAAABJDEcAAAAAIInhCAAAAAAkMRwBAAAAgCSGIwAAAACQxHAEAAAAAJIYjgAAAABAEsMRAAAAAEhiOAIAAAAASdL/A6cQWkHMmET1AAAAAElFTkSuQmCC\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Temos que o dataset é com posto por 60000 itens 28x28 (60000 x 28 x 28). Vamos aplicar o conceito de *flattening* e torná-lo 60000 x 784"
      ],
      "metadata": {
        "id": "YIgSWvKaxJll"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "new_shape = 784"
      ],
      "metadata": {
        "id": "SBScT8-COR8Y"
      },
      "execution_count": 35,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_train = X_train.reshape(60000, new_shape)\n",
        "X_test = X_test.reshape(10000, new_shape)"
      ],
      "metadata": {
        "id": "e46eyKGKOb2T"
      },
      "execution_count": 36,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_train = X_train.astype('float32')\n",
        "X_test = X_test.astype('float32')"
      ],
      "metadata": {
        "id": "Tbhw_26fOxuW"
      },
      "execution_count": 37,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Normalização dos valores do intervalo [ 0, 255 ] para [ 0, 1 ]."
      ],
      "metadata": {
        "id": "i5KGo968x1Ur"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X_train /= 255\n",
        "X_test /= 255"
      ],
      "metadata": {
        "id": "ajZ8gB3lPwov"
      },
      "execution_count": 38,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(X_train.shape[0], 'train samples')\n",
        "print(X_test.shape[0], 'test samples')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Z-Tdd834P6wa",
        "outputId": "d245100c-bd93-4f60-f191-9033268fee69"
      },
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "60000 train samples\n",
            "10000 test samples\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Convertemos os valores de Y em matrizes binárias com as linhas compostas por zeros e apenas um número 1 na posição que representa o valor do número."
      ],
      "metadata": {
        "id": "pVXScejM6ZJA"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from keras.utils import to_categorical"
      ],
      "metadata": {
        "id": "Xzh-MkCTVkGl"
      },
      "execution_count": 40,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "Y_train = to_categorical(y_train, num_classes)\n",
        "Y_test = to_categorical(y_test, num_classes)"
      ],
      "metadata": {
        "id": "-PVHrUpVSqph"
      },
      "execution_count": 41,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "função SoftMax:\n",
        "\n",
        "$S(y)i = exp(yi)/∑jexp(yj)$"
      ],
      "metadata": {
        "id": "FEV8z6tmqgmv"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model = Sequential()\n",
        "model.add(Dense(num_classes, input_shape = (new_shape,)))\n",
        "model.add(Activation('softmax'))\n",
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3d7asEchTBs4",
        "outputId": "9f1aba15-1c4e-4be8-8f45-5c54d889523b"
      },
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " dense_1 (Dense)             (None, 10)                7850      \n",
            "                                                                 \n",
            " activation_1 (Activation)   (None, 10)                0         \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 7850 (30.66 KB)\n",
            "Trainable params: 7850 (30.66 KB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "$H(y,y^​)=−∑i​yi​⋅log(y^​i​)$"
      ],
      "metadata": {
        "id": "DV47pN4MmDqt"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model.compile(loss = 'categorical_crossentropy', optimizer = optimizer, metrics = ['accuracy'])"
      ],
      "metadata": {
        "id": "Ci6ukGAHy44q"
      },
      "execution_count": 43,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Acessando os hiperparâmetros:"
      ],
      "metadata": {
        "id": "_eit-ye1s82r"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model_config = model.get_config()\n",
        "print(model_config)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hjQUBamatB2_",
        "outputId": "2268a156-3583-4d7e-befa-3c082d2e63db"
      },
      "execution_count": 44,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "{'name': 'sequential_1', 'layers': [{'module': 'keras.layers', 'class_name': 'InputLayer', 'config': {'batch_input_shape': (None, 784), 'dtype': 'float32', 'sparse': False, 'ragged': False, 'name': 'dense_1_input'}, 'registered_name': None}, {'module': 'keras.layers', 'class_name': 'Dense', 'config': {'name': 'dense_1', 'trainable': True, 'dtype': 'float32', 'batch_input_shape': (None, 784), 'units': 10, 'activation': 'linear', 'use_bias': True, 'kernel_initializer': {'module': 'keras.initializers', 'class_name': 'GlorotUniform', 'config': {'seed': None}, 'registered_name': None}, 'bias_initializer': {'module': 'keras.initializers', 'class_name': 'Zeros', 'config': {}, 'registered_name': None}, 'kernel_regularizer': None, 'bias_regularizer': None, 'activity_regularizer': None, 'kernel_constraint': None, 'bias_constraint': None}, 'registered_name': None, 'build_config': {'input_shape': (None, 784)}}, {'module': 'keras.layers', 'class_name': 'Activation', 'config': {'name': 'activation_1', 'trainable': True, 'dtype': 'float32', 'activation': 'softmax'}, 'registered_name': None, 'build_config': {'input_shape': (None, 10)}}]}\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.summary()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wzXNAM5M0A9c",
        "outputId": "a6f53885-b18d-42b3-9a27-71d2611cae32"
      },
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " dense_1 (Dense)             (None, 10)                7850      \n",
            "                                                                 \n",
            " activation_1 (Activation)   (None, 10)                0         \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 7850 (30.66 KB)\n",
            "Trainable params: 7850 (30.66 KB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "history = model.fit(\n",
        "                    X_train, Y_train,\n",
        "                    batch_size = batch_size,\n",
        "                    epochs = num_epoch,\n",
        "                    verbose = verbose,\n",
        "                    validation_split = train_portion,\n",
        "                   )\n",
        "score = model.evaluate(X_test, Y_test, verbose = verbose)\n",
        "\n",
        "print(\"Test score:\", score[0])\n",
        "print(\"Test accuracy:\", score[1])\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lTJuFQst7hvN",
        "outputId": "63081be7-0750-4882-be34-8e0a1663b458"
      },
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/200\n",
            "375/375 [==============================] - 2s 3ms/step - loss: 1.3582 - accuracy: 0.6793 - val_loss: 0.8821 - val_accuracy: 0.8257\n",
            "Epoch 2/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.7862 - accuracy: 0.8295 - val_loss: 0.6501 - val_accuracy: 0.8608\n",
            "Epoch 3/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.6395 - accuracy: 0.8523 - val_loss: 0.5570 - val_accuracy: 0.8730\n",
            "Epoch 4/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.5683 - accuracy: 0.8621 - val_loss: 0.5052 - val_accuracy: 0.8802\n",
            "Epoch 5/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.5250 - accuracy: 0.8691 - val_loss: 0.4720 - val_accuracy: 0.8834\n",
            "Epoch 6/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.4952 - accuracy: 0.8746 - val_loss: 0.4483 - val_accuracy: 0.8881\n",
            "Epoch 7/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.4731 - accuracy: 0.8778 - val_loss: 0.4304 - val_accuracy: 0.8905\n",
            "Epoch 8/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.4559 - accuracy: 0.8817 - val_loss: 0.4164 - val_accuracy: 0.8935\n",
            "Epoch 9/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.4420 - accuracy: 0.8842 - val_loss: 0.4055 - val_accuracy: 0.8951\n",
            "Epoch 10/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.4306 - accuracy: 0.8862 - val_loss: 0.3957 - val_accuracy: 0.8966\n",
            "Epoch 11/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.4209 - accuracy: 0.8881 - val_loss: 0.3878 - val_accuracy: 0.8988\n",
            "Epoch 12/200\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.4125 - accuracy: 0.8896 - val_loss: 0.3809 - val_accuracy: 0.8997\n",
            "Epoch 13/200\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.4052 - accuracy: 0.8905 - val_loss: 0.3749 - val_accuracy: 0.9013\n",
            "Epoch 14/200\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.3988 - accuracy: 0.8919 - val_loss: 0.3697 - val_accuracy: 0.9020\n",
            "Epoch 15/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3930 - accuracy: 0.8930 - val_loss: 0.3650 - val_accuracy: 0.9021\n",
            "Epoch 16/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3879 - accuracy: 0.8941 - val_loss: 0.3607 - val_accuracy: 0.9038\n",
            "Epoch 17/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3832 - accuracy: 0.8953 - val_loss: 0.3569 - val_accuracy: 0.9038\n",
            "Epoch 18/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3790 - accuracy: 0.8964 - val_loss: 0.3532 - val_accuracy: 0.9052\n",
            "Epoch 19/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3751 - accuracy: 0.8974 - val_loss: 0.3501 - val_accuracy: 0.9059\n",
            "Epoch 20/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3715 - accuracy: 0.8981 - val_loss: 0.3475 - val_accuracy: 0.9062\n",
            "Epoch 21/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3682 - accuracy: 0.8990 - val_loss: 0.3445 - val_accuracy: 0.9074\n",
            "Epoch 22/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3651 - accuracy: 0.8996 - val_loss: 0.3420 - val_accuracy: 0.9075\n",
            "Epoch 23/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3622 - accuracy: 0.9003 - val_loss: 0.3396 - val_accuracy: 0.9080\n",
            "Epoch 24/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3595 - accuracy: 0.9004 - val_loss: 0.3374 - val_accuracy: 0.9081\n",
            "Epoch 25/200\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.3570 - accuracy: 0.9012 - val_loss: 0.3354 - val_accuracy: 0.9088\n",
            "Epoch 26/200\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.3546 - accuracy: 0.9016 - val_loss: 0.3334 - val_accuracy: 0.9090\n",
            "Epoch 27/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3523 - accuracy: 0.9020 - val_loss: 0.3315 - val_accuracy: 0.9097\n",
            "Epoch 28/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3502 - accuracy: 0.9025 - val_loss: 0.3299 - val_accuracy: 0.9091\n",
            "Epoch 29/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3482 - accuracy: 0.9031 - val_loss: 0.3283 - val_accuracy: 0.9099\n",
            "Epoch 30/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3463 - accuracy: 0.9031 - val_loss: 0.3267 - val_accuracy: 0.9105\n",
            "Epoch 31/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3445 - accuracy: 0.9036 - val_loss: 0.3252 - val_accuracy: 0.9109\n",
            "Epoch 32/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3427 - accuracy: 0.9043 - val_loss: 0.3239 - val_accuracy: 0.9110\n",
            "Epoch 33/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3411 - accuracy: 0.9048 - val_loss: 0.3226 - val_accuracy: 0.9118\n",
            "Epoch 34/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3395 - accuracy: 0.9052 - val_loss: 0.3213 - val_accuracy: 0.9115\n",
            "Epoch 35/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3380 - accuracy: 0.9056 - val_loss: 0.3201 - val_accuracy: 0.9122\n",
            "Epoch 36/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3366 - accuracy: 0.9059 - val_loss: 0.3189 - val_accuracy: 0.9123\n",
            "Epoch 37/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3352 - accuracy: 0.9066 - val_loss: 0.3180 - val_accuracy: 0.9127\n",
            "Epoch 38/200\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.3339 - accuracy: 0.9070 - val_loss: 0.3167 - val_accuracy: 0.9129\n",
            "Epoch 39/200\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.3326 - accuracy: 0.9075 - val_loss: 0.3158 - val_accuracy: 0.9125\n",
            "Epoch 40/200\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.3314 - accuracy: 0.9076 - val_loss: 0.3148 - val_accuracy: 0.9131\n",
            "Epoch 41/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3302 - accuracy: 0.9079 - val_loss: 0.3138 - val_accuracy: 0.9133\n",
            "Epoch 42/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3290 - accuracy: 0.9086 - val_loss: 0.3129 - val_accuracy: 0.9133\n",
            "Epoch 43/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3279 - accuracy: 0.9085 - val_loss: 0.3121 - val_accuracy: 0.9133\n",
            "Epoch 44/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3268 - accuracy: 0.9091 - val_loss: 0.3113 - val_accuracy: 0.9133\n",
            "Epoch 45/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3258 - accuracy: 0.9094 - val_loss: 0.3105 - val_accuracy: 0.9135\n",
            "Epoch 46/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3248 - accuracy: 0.9097 - val_loss: 0.3096 - val_accuracy: 0.9137\n",
            "Epoch 47/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3239 - accuracy: 0.9098 - val_loss: 0.3088 - val_accuracy: 0.9137\n",
            "Epoch 48/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3229 - accuracy: 0.9101 - val_loss: 0.3081 - val_accuracy: 0.9135\n",
            "Epoch 49/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3220 - accuracy: 0.9103 - val_loss: 0.3074 - val_accuracy: 0.9137\n",
            "Epoch 50/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3211 - accuracy: 0.9109 - val_loss: 0.3067 - val_accuracy: 0.9143\n",
            "Epoch 51/200\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.3203 - accuracy: 0.9111 - val_loss: 0.3060 - val_accuracy: 0.9140\n",
            "Epoch 52/200\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.3194 - accuracy: 0.9116 - val_loss: 0.3056 - val_accuracy: 0.9142\n",
            "Epoch 53/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3186 - accuracy: 0.9118 - val_loss: 0.3048 - val_accuracy: 0.9146\n",
            "Epoch 54/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3179 - accuracy: 0.9120 - val_loss: 0.3041 - val_accuracy: 0.9142\n",
            "Epoch 55/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3171 - accuracy: 0.9123 - val_loss: 0.3036 - val_accuracy: 0.9149\n",
            "Epoch 56/200\n",
            "375/375 [==============================] - 2s 4ms/step - loss: 0.3163 - accuracy: 0.9123 - val_loss: 0.3030 - val_accuracy: 0.9150\n",
            "Epoch 57/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3156 - accuracy: 0.9125 - val_loss: 0.3024 - val_accuracy: 0.9153\n",
            "Epoch 58/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3149 - accuracy: 0.9130 - val_loss: 0.3019 - val_accuracy: 0.9154\n",
            "Epoch 59/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3142 - accuracy: 0.9129 - val_loss: 0.3014 - val_accuracy: 0.9152\n",
            "Epoch 60/200\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.3135 - accuracy: 0.9134 - val_loss: 0.3009 - val_accuracy: 0.9159\n",
            "Epoch 61/200\n",
            "375/375 [==============================] - 2s 5ms/step - loss: 0.3129 - accuracy: 0.9133 - val_loss: 0.3004 - val_accuracy: 0.9152\n",
            "Epoch 62/200\n",
            "375/375 [==============================] - 3s 9ms/step - loss: 0.3123 - accuracy: 0.9134 - val_loss: 0.2998 - val_accuracy: 0.9155\n",
            "Epoch 63/200\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.3116 - accuracy: 0.9135 - val_loss: 0.2993 - val_accuracy: 0.9158\n",
            "Epoch 64/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3110 - accuracy: 0.9139 - val_loss: 0.2989 - val_accuracy: 0.9163\n",
            "Epoch 65/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3104 - accuracy: 0.9139 - val_loss: 0.2985 - val_accuracy: 0.9165\n",
            "Epoch 66/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3098 - accuracy: 0.9141 - val_loss: 0.2980 - val_accuracy: 0.9162\n",
            "Epoch 67/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3092 - accuracy: 0.9141 - val_loss: 0.2977 - val_accuracy: 0.9165\n",
            "Epoch 68/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3087 - accuracy: 0.9142 - val_loss: 0.2972 - val_accuracy: 0.9167\n",
            "Epoch 69/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3081 - accuracy: 0.9145 - val_loss: 0.2967 - val_accuracy: 0.9168\n",
            "Epoch 70/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3076 - accuracy: 0.9146 - val_loss: 0.2963 - val_accuracy: 0.9169\n",
            "Epoch 71/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3071 - accuracy: 0.9150 - val_loss: 0.2959 - val_accuracy: 0.9169\n",
            "Epoch 72/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3065 - accuracy: 0.9149 - val_loss: 0.2955 - val_accuracy: 0.9169\n",
            "Epoch 73/200\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.3061 - accuracy: 0.9150 - val_loss: 0.2952 - val_accuracy: 0.9170\n",
            "Epoch 74/200\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.3056 - accuracy: 0.9154 - val_loss: 0.2949 - val_accuracy: 0.9173\n",
            "Epoch 75/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3051 - accuracy: 0.9150 - val_loss: 0.2945 - val_accuracy: 0.9169\n",
            "Epoch 76/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3046 - accuracy: 0.9154 - val_loss: 0.2941 - val_accuracy: 0.9170\n",
            "Epoch 77/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3041 - accuracy: 0.9154 - val_loss: 0.2937 - val_accuracy: 0.9178\n",
            "Epoch 78/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3037 - accuracy: 0.9156 - val_loss: 0.2934 - val_accuracy: 0.9174\n",
            "Epoch 79/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3032 - accuracy: 0.9159 - val_loss: 0.2930 - val_accuracy: 0.9177\n",
            "Epoch 80/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3028 - accuracy: 0.9160 - val_loss: 0.2928 - val_accuracy: 0.9176\n",
            "Epoch 81/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3023 - accuracy: 0.9159 - val_loss: 0.2925 - val_accuracy: 0.9178\n",
            "Epoch 82/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3019 - accuracy: 0.9163 - val_loss: 0.2922 - val_accuracy: 0.9178\n",
            "Epoch 83/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3015 - accuracy: 0.9161 - val_loss: 0.2918 - val_accuracy: 0.9180\n",
            "Epoch 84/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3011 - accuracy: 0.9164 - val_loss: 0.2916 - val_accuracy: 0.9181\n",
            "Epoch 85/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.3007 - accuracy: 0.9162 - val_loss: 0.2912 - val_accuracy: 0.9182\n",
            "Epoch 86/200\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.3003 - accuracy: 0.9167 - val_loss: 0.2909 - val_accuracy: 0.9181\n",
            "Epoch 87/200\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.2999 - accuracy: 0.9168 - val_loss: 0.2906 - val_accuracy: 0.9185\n",
            "Epoch 88/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2995 - accuracy: 0.9166 - val_loss: 0.2904 - val_accuracy: 0.9182\n",
            "Epoch 89/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2991 - accuracy: 0.9171 - val_loss: 0.2901 - val_accuracy: 0.9185\n",
            "Epoch 90/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2988 - accuracy: 0.9170 - val_loss: 0.2898 - val_accuracy: 0.9186\n",
            "Epoch 91/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2984 - accuracy: 0.9169 - val_loss: 0.2896 - val_accuracy: 0.9184\n",
            "Epoch 92/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2980 - accuracy: 0.9172 - val_loss: 0.2893 - val_accuracy: 0.9187\n",
            "Epoch 93/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2977 - accuracy: 0.9171 - val_loss: 0.2891 - val_accuracy: 0.9187\n",
            "Epoch 94/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2973 - accuracy: 0.9173 - val_loss: 0.2888 - val_accuracy: 0.9185\n",
            "Epoch 95/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2970 - accuracy: 0.9174 - val_loss: 0.2886 - val_accuracy: 0.9183\n",
            "Epoch 96/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2966 - accuracy: 0.9178 - val_loss: 0.2883 - val_accuracy: 0.9190\n",
            "Epoch 97/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2963 - accuracy: 0.9177 - val_loss: 0.2881 - val_accuracy: 0.9187\n",
            "Epoch 98/200\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2959 - accuracy: 0.9178 - val_loss: 0.2879 - val_accuracy: 0.9189\n",
            "Epoch 99/200\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.2957 - accuracy: 0.9177 - val_loss: 0.2876 - val_accuracy: 0.9188\n",
            "Epoch 100/200\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2953 - accuracy: 0.9179 - val_loss: 0.2874 - val_accuracy: 0.9192\n",
            "Epoch 101/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2950 - accuracy: 0.9181 - val_loss: 0.2872 - val_accuracy: 0.9189\n",
            "Epoch 102/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2947 - accuracy: 0.9181 - val_loss: 0.2869 - val_accuracy: 0.9188\n",
            "Epoch 103/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2944 - accuracy: 0.9183 - val_loss: 0.2867 - val_accuracy: 0.9193\n",
            "Epoch 104/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2941 - accuracy: 0.9184 - val_loss: 0.2865 - val_accuracy: 0.9193\n",
            "Epoch 105/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2938 - accuracy: 0.9183 - val_loss: 0.2864 - val_accuracy: 0.9192\n",
            "Epoch 106/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2935 - accuracy: 0.9184 - val_loss: 0.2862 - val_accuracy: 0.9197\n",
            "Epoch 107/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2932 - accuracy: 0.9184 - val_loss: 0.2859 - val_accuracy: 0.9198\n",
            "Epoch 108/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2929 - accuracy: 0.9185 - val_loss: 0.2858 - val_accuracy: 0.9193\n",
            "Epoch 109/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2927 - accuracy: 0.9185 - val_loss: 0.2855 - val_accuracy: 0.9195\n",
            "Epoch 110/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2923 - accuracy: 0.9186 - val_loss: 0.2854 - val_accuracy: 0.9193\n",
            "Epoch 111/200\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2921 - accuracy: 0.9187 - val_loss: 0.2852 - val_accuracy: 0.9195\n",
            "Epoch 112/200\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.2918 - accuracy: 0.9188 - val_loss: 0.2849 - val_accuracy: 0.9193\n",
            "Epoch 113/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2915 - accuracy: 0.9188 - val_loss: 0.2848 - val_accuracy: 0.9196\n",
            "Epoch 114/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2913 - accuracy: 0.9187 - val_loss: 0.2846 - val_accuracy: 0.9197\n",
            "Epoch 115/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2910 - accuracy: 0.9189 - val_loss: 0.2845 - val_accuracy: 0.9194\n",
            "Epoch 116/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2908 - accuracy: 0.9188 - val_loss: 0.2842 - val_accuracy: 0.9200\n",
            "Epoch 117/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2905 - accuracy: 0.9192 - val_loss: 0.2841 - val_accuracy: 0.9196\n",
            "Epoch 118/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2902 - accuracy: 0.9191 - val_loss: 0.2840 - val_accuracy: 0.9197\n",
            "Epoch 119/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2900 - accuracy: 0.9194 - val_loss: 0.2838 - val_accuracy: 0.9199\n",
            "Epoch 120/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2898 - accuracy: 0.9193 - val_loss: 0.2837 - val_accuracy: 0.9197\n",
            "Epoch 121/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2895 - accuracy: 0.9191 - val_loss: 0.2834 - val_accuracy: 0.9206\n",
            "Epoch 122/200\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2893 - accuracy: 0.9195 - val_loss: 0.2832 - val_accuracy: 0.9202\n",
            "Epoch 123/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2890 - accuracy: 0.9194 - val_loss: 0.2831 - val_accuracy: 0.9196\n",
            "Epoch 124/200\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.2888 - accuracy: 0.9195 - val_loss: 0.2828 - val_accuracy: 0.9202\n",
            "Epoch 125/200\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2886 - accuracy: 0.9197 - val_loss: 0.2828 - val_accuracy: 0.9198\n",
            "Epoch 126/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2883 - accuracy: 0.9195 - val_loss: 0.2826 - val_accuracy: 0.9204\n",
            "Epoch 127/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2881 - accuracy: 0.9198 - val_loss: 0.2825 - val_accuracy: 0.9201\n",
            "Epoch 128/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2879 - accuracy: 0.9197 - val_loss: 0.2823 - val_accuracy: 0.9202\n",
            "Epoch 129/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2877 - accuracy: 0.9196 - val_loss: 0.2822 - val_accuracy: 0.9202\n",
            "Epoch 130/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2874 - accuracy: 0.9197 - val_loss: 0.2821 - val_accuracy: 0.9202\n",
            "Epoch 131/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2872 - accuracy: 0.9201 - val_loss: 0.2819 - val_accuracy: 0.9206\n",
            "Epoch 132/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2870 - accuracy: 0.9197 - val_loss: 0.2818 - val_accuracy: 0.9210\n",
            "Epoch 133/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2868 - accuracy: 0.9198 - val_loss: 0.2816 - val_accuracy: 0.9211\n",
            "Epoch 134/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2866 - accuracy: 0.9200 - val_loss: 0.2815 - val_accuracy: 0.9206\n",
            "Epoch 135/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2864 - accuracy: 0.9204 - val_loss: 0.2813 - val_accuracy: 0.9206\n",
            "Epoch 136/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2862 - accuracy: 0.9203 - val_loss: 0.2812 - val_accuracy: 0.9209\n",
            "Epoch 137/200\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.2859 - accuracy: 0.9202 - val_loss: 0.2810 - val_accuracy: 0.9207\n",
            "Epoch 138/200\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2858 - accuracy: 0.9203 - val_loss: 0.2809 - val_accuracy: 0.9208\n",
            "Epoch 139/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2855 - accuracy: 0.9204 - val_loss: 0.2808 - val_accuracy: 0.9208\n",
            "Epoch 140/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2854 - accuracy: 0.9204 - val_loss: 0.2807 - val_accuracy: 0.9202\n",
            "Epoch 141/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2852 - accuracy: 0.9204 - val_loss: 0.2805 - val_accuracy: 0.9206\n",
            "Epoch 142/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2850 - accuracy: 0.9202 - val_loss: 0.2804 - val_accuracy: 0.9208\n",
            "Epoch 143/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2848 - accuracy: 0.9205 - val_loss: 0.2803 - val_accuracy: 0.9207\n",
            "Epoch 144/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2846 - accuracy: 0.9204 - val_loss: 0.2801 - val_accuracy: 0.9210\n",
            "Epoch 145/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2844 - accuracy: 0.9206 - val_loss: 0.2800 - val_accuracy: 0.9208\n",
            "Epoch 146/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2842 - accuracy: 0.9204 - val_loss: 0.2800 - val_accuracy: 0.9212\n",
            "Epoch 147/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2840 - accuracy: 0.9205 - val_loss: 0.2799 - val_accuracy: 0.9212\n",
            "Epoch 148/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2838 - accuracy: 0.9207 - val_loss: 0.2798 - val_accuracy: 0.9212\n",
            "Epoch 149/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2837 - accuracy: 0.9208 - val_loss: 0.2795 - val_accuracy: 0.9210\n",
            "Epoch 150/200\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.2834 - accuracy: 0.9208 - val_loss: 0.2794 - val_accuracy: 0.9209\n",
            "Epoch 151/200\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2833 - accuracy: 0.9211 - val_loss: 0.2793 - val_accuracy: 0.9209\n",
            "Epoch 152/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2831 - accuracy: 0.9209 - val_loss: 0.2792 - val_accuracy: 0.9211\n",
            "Epoch 153/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2830 - accuracy: 0.9208 - val_loss: 0.2791 - val_accuracy: 0.9213\n",
            "Epoch 154/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2828 - accuracy: 0.9208 - val_loss: 0.2790 - val_accuracy: 0.9216\n",
            "Epoch 155/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2826 - accuracy: 0.9211 - val_loss: 0.2789 - val_accuracy: 0.9212\n",
            "Epoch 156/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2824 - accuracy: 0.9210 - val_loss: 0.2788 - val_accuracy: 0.9212\n",
            "Epoch 157/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2823 - accuracy: 0.9212 - val_loss: 0.2787 - val_accuracy: 0.9215\n",
            "Epoch 158/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2821 - accuracy: 0.9214 - val_loss: 0.2786 - val_accuracy: 0.9212\n",
            "Epoch 159/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2819 - accuracy: 0.9214 - val_loss: 0.2786 - val_accuracy: 0.9212\n",
            "Epoch 160/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2817 - accuracy: 0.9214 - val_loss: 0.2784 - val_accuracy: 0.9212\n",
            "Epoch 161/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2816 - accuracy: 0.9214 - val_loss: 0.2783 - val_accuracy: 0.9218\n",
            "Epoch 162/200\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2814 - accuracy: 0.9213 - val_loss: 0.2782 - val_accuracy: 0.9214\n",
            "Epoch 163/200\n",
            "375/375 [==============================] - 2s 4ms/step - loss: 0.2813 - accuracy: 0.9214 - val_loss: 0.2781 - val_accuracy: 0.9218\n",
            "Epoch 164/200\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2811 - accuracy: 0.9217 - val_loss: 0.2780 - val_accuracy: 0.9216\n",
            "Epoch 165/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2809 - accuracy: 0.9215 - val_loss: 0.2779 - val_accuracy: 0.9215\n",
            "Epoch 166/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2808 - accuracy: 0.9215 - val_loss: 0.2777 - val_accuracy: 0.9216\n",
            "Epoch 167/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2806 - accuracy: 0.9218 - val_loss: 0.2777 - val_accuracy: 0.9214\n",
            "Epoch 168/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2805 - accuracy: 0.9215 - val_loss: 0.2776 - val_accuracy: 0.9218\n",
            "Epoch 169/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2803 - accuracy: 0.9214 - val_loss: 0.2775 - val_accuracy: 0.9218\n",
            "Epoch 170/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2802 - accuracy: 0.9216 - val_loss: 0.2774 - val_accuracy: 0.9214\n",
            "Epoch 171/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2800 - accuracy: 0.9215 - val_loss: 0.2773 - val_accuracy: 0.9217\n",
            "Epoch 172/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2799 - accuracy: 0.9217 - val_loss: 0.2773 - val_accuracy: 0.9215\n",
            "Epoch 173/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2797 - accuracy: 0.9216 - val_loss: 0.2771 - val_accuracy: 0.9216\n",
            "Epoch 174/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2796 - accuracy: 0.9214 - val_loss: 0.2770 - val_accuracy: 0.9218\n",
            "Epoch 175/200\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2794 - accuracy: 0.9219 - val_loss: 0.2770 - val_accuracy: 0.9221\n",
            "Epoch 176/200\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.2793 - accuracy: 0.9217 - val_loss: 0.2769 - val_accuracy: 0.9218\n",
            "Epoch 177/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2792 - accuracy: 0.9218 - val_loss: 0.2768 - val_accuracy: 0.9217\n",
            "Epoch 178/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2790 - accuracy: 0.9219 - val_loss: 0.2768 - val_accuracy: 0.9215\n",
            "Epoch 179/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2789 - accuracy: 0.9215 - val_loss: 0.2766 - val_accuracy: 0.9222\n",
            "Epoch 180/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2787 - accuracy: 0.9218 - val_loss: 0.2767 - val_accuracy: 0.9221\n",
            "Epoch 181/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2786 - accuracy: 0.9218 - val_loss: 0.2765 - val_accuracy: 0.9219\n",
            "Epoch 182/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2785 - accuracy: 0.9219 - val_loss: 0.2764 - val_accuracy: 0.9220\n",
            "Epoch 183/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2783 - accuracy: 0.9220 - val_loss: 0.2763 - val_accuracy: 0.9222\n",
            "Epoch 184/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2782 - accuracy: 0.9220 - val_loss: 0.2762 - val_accuracy: 0.9216\n",
            "Epoch 185/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2781 - accuracy: 0.9220 - val_loss: 0.2761 - val_accuracy: 0.9222\n",
            "Epoch 186/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2779 - accuracy: 0.9222 - val_loss: 0.2760 - val_accuracy: 0.9220\n",
            "Epoch 187/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2778 - accuracy: 0.9222 - val_loss: 0.2760 - val_accuracy: 0.9222\n",
            "Epoch 188/200\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.2776 - accuracy: 0.9221 - val_loss: 0.2760 - val_accuracy: 0.9225\n",
            "Epoch 189/200\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.2775 - accuracy: 0.9220 - val_loss: 0.2759 - val_accuracy: 0.9221\n",
            "Epoch 190/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2774 - accuracy: 0.9221 - val_loss: 0.2758 - val_accuracy: 0.9223\n",
            "Epoch 191/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2772 - accuracy: 0.9223 - val_loss: 0.2757 - val_accuracy: 0.9224\n",
            "Epoch 192/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2771 - accuracy: 0.9222 - val_loss: 0.2757 - val_accuracy: 0.9224\n",
            "Epoch 193/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2770 - accuracy: 0.9226 - val_loss: 0.2756 - val_accuracy: 0.9223\n",
            "Epoch 194/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2769 - accuracy: 0.9221 - val_loss: 0.2755 - val_accuracy: 0.9226\n",
            "Epoch 195/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2768 - accuracy: 0.9225 - val_loss: 0.2754 - val_accuracy: 0.9224\n",
            "Epoch 196/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2766 - accuracy: 0.9224 - val_loss: 0.2753 - val_accuracy: 0.9226\n",
            "Epoch 197/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2765 - accuracy: 0.9227 - val_loss: 0.2753 - val_accuracy: 0.9224\n",
            "Epoch 198/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2764 - accuracy: 0.9224 - val_loss: 0.2752 - val_accuracy: 0.9224\n",
            "Epoch 199/200\n",
            "375/375 [==============================] - 1s 2ms/step - loss: 0.2762 - accuracy: 0.9226 - val_loss: 0.2751 - val_accuracy: 0.9222\n",
            "Epoch 200/200\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2761 - accuracy: 0.9226 - val_loss: 0.2750 - val_accuracy: 0.9226\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.2772 - accuracy: 0.9225\n",
            "Test score: 0.27721282839775085\n",
            "Test accuracy: 0.9225000143051147\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "plt.plot(history.history['loss'], label='Training Loss')\n",
        "plt.plot(history.history['val_loss'], label='Validation Loss')\n",
        "plt.title('Loss vs Epoch')\n",
        "plt.xlabel('Epoch')\n",
        "plt.ylabel('Loss')\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "id": "pWCXyEcFLeYt",
        "outputId": "6b9b1821-b79b-448d-a0b2-1f9a6661fc41"
      },
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Acuracidade:"
      ],
      "metadata": {
        "id": "wTuKkCSnnwPa"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "plt.plot(history.history['accuracy'], label = \"Training Accuracy\")\n",
        "plt.plot(history.history['val_accuracy'], label = \"Validation Accuracy\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.xlabel(\"Epoch\")\n",
        "plt.title(\"Model Accuracy vs Epoch\")\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "id": "OUxB2-aXMiWm",
        "outputId": "8def2f20-de63-40e8-eced-b96bf07ad8c7"
      },
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CsIS8AGLn54F",
        "outputId": "0adb4377-7acf-4144-8a65-6c0d14dea916"
      },
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " dense_1 (Dense)             (None, 10)                7850      \n",
            "                                                                 \n",
            " activation_1 (Activation)   (None, 10)                0         \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 7850 (30.66 KB)\n",
            "Trainable params: 7850 (30.66 KB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.optimizer.get_config()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pPca92pH2oCw",
        "outputId": "957f1e04-55fe-4837-f7a7-eb40a35238c8"
      },
      "execution_count": 51,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'name': 'SGD',\n",
              " 'weight_decay': None,\n",
              " 'clipnorm': None,\n",
              " 'global_clipnorm': None,\n",
              " 'clipvalue': None,\n",
              " 'use_ema': False,\n",
              " 'ema_momentum': 0.99,\n",
              " 'ema_overwrite_frequency': None,\n",
              " 'jit_compile': False,\n",
              " 'is_legacy_optimizer': False,\n",
              " 'learning_rate': 0.01,\n",
              " 'momentum': 0.0,\n",
              " 'nesterov': False}"
            ]
          },
          "metadata": {},
          "execution_count": 51
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "oa-wfbYE5wBr"
      },
      "execution_count": null,
      "outputs": []
    }
  ],
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}