04_ConvolutionalNeuralNetwork-Hoda-Keras.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<center><div style=\"direction:rtl;font-family:B Lotus, B Nazanin, Tahoma\">به نام خدا</div></center>\n",
    "<h1><center><div style=\"direction:rtl;font-family:B Lotus, B Nazanin, Tahoma\">مقدمه ای بر شبکه‌های عصبی کانولوشنالی<br>Convolutionl Neural Networks - CNN</div></center></h1>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div style=\"direction:rtl;text-align:right;font-family:Tahoma\">\n",
    "اگر روی گوگل کولب اجرا میکنید این خطوط را از حالت کامنت خارج نمائید.\n",
    "</div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#!wget https://raw.githubusercontent.com/Alireza-Akhavan/SRU-deeplearning-workshop/master/dataset.py\n",
    "#!mkdir dataset\n",
    "#!wget https://github.com/Alireza-Akhavan/SRU-deeplearning-workshop/raw/master/dataset/Data_hoda_full.mat -P dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## <div style=\"direction:rtl;text-align:right;font-family:B Lotus, B Nazanin, Tahoma\">کد یک شبکه کانولوشنالی و آموزش آن از ابتدا تا انتها بر روی مجموعه داده هدی</div>\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "11/11 [==============================] - 22s 186ms/step - loss: 2.2148 - accuracy: 0.2327 - val_loss: 1.5786 - val_accuracy: 0.6943\n",
      "Epoch 2/200\n",
      "11/11 [==============================] - 1s 116ms/step - loss: 1.4473 - accuracy: 0.5336 - val_loss: 0.7262 - val_accuracy: 0.7857\n",
      "Epoch 3/200\n",
      "11/11 [==============================] - 1s 132ms/step - loss: 0.8870 - accuracy: 0.7002 - val_loss: 0.4640 - val_accuracy: 0.8571\n",
      "Epoch 4/200\n",
      "11/11 [==============================] - 1s 121ms/step - loss: 0.6153 - accuracy: 0.7918 - val_loss: 0.3371 - val_accuracy: 0.8829\n",
      "Epoch 5/200\n",
      "11/11 [==============================] - 1s 122ms/step - loss: 0.4766 - accuracy: 0.8443 - val_loss: 0.2959 - val_accuracy: 0.8943\n",
      "Epoch 6/200\n",
      "11/11 [==============================] - 1s 119ms/step - loss: 0.4094 - accuracy: 0.8622 - val_loss: 0.2437 - val_accuracy: 0.9129\n",
      "Epoch 7/200\n",
      "11/11 [==============================] - 1s 127ms/step - loss: 0.3429 - accuracy: 0.8748 - val_loss: 0.2273 - val_accuracy: 0.9100\n",
      "Epoch 8/200\n",
      "11/11 [==============================] - 1s 117ms/step - loss: 0.3164 - accuracy: 0.8964 - val_loss: 0.1886 - val_accuracy: 0.9329\n",
      "Epoch 9/200\n",
      "11/11 [==============================] - 1s 113ms/step - loss: 0.2715 - accuracy: 0.9138 - val_loss: 0.1885 - val_accuracy: 0.9286\n",
      "Epoch 10/200\n",
      "11/11 [==============================] - 1s 117ms/step - loss: 0.2167 - accuracy: 0.9329 - val_loss: 0.1548 - val_accuracy: 0.9429\n",
      "Epoch 11/200\n",
      "11/11 [==============================] - 1s 112ms/step - loss: 0.2235 - accuracy: 0.9268 - val_loss: 0.1608 - val_accuracy: 0.9357\n",
      "Epoch 12/200\n",
      "11/11 [==============================] - 1s 116ms/step - loss: 0.1942 - accuracy: 0.9354 - val_loss: 0.1639 - val_accuracy: 0.9400\n",
      "Epoch 13/200\n",
      "11/11 [==============================] - 1s 115ms/step - loss: 0.2138 - accuracy: 0.9313 - val_loss: 0.1511 - val_accuracy: 0.9400\n",
      "Epoch 14/200\n",
      "11/11 [==============================] - 1s 119ms/step - loss: 0.1753 - accuracy: 0.9415 - val_loss: 0.1113 - val_accuracy: 0.9629\n",
      "Epoch 15/200\n",
      "11/11 [==============================] - 1s 113ms/step - loss: 0.1689 - accuracy: 0.9391 - val_loss: 0.1233 - val_accuracy: 0.9557\n",
      "Epoch 16/200\n",
      "11/11 [==============================] - 1s 115ms/step - loss: 0.1587 - accuracy: 0.9413 - val_loss: 0.1311 - val_accuracy: 0.9486\n",
      "Epoch 17/200\n",
      "11/11 [==============================] - 1s 120ms/step - loss: 0.1541 - accuracy: 0.9473 - val_loss: 0.1078 - val_accuracy: 0.9571\n",
      "Epoch 18/200\n",
      "11/11 [==============================] - 1s 130ms/step - loss: 0.1374 - accuracy: 0.9563 - val_loss: 0.1063 - val_accuracy: 0.9629\n",
      "Epoch 19/200\n",
      "11/11 [==============================] - 1s 112ms/step - loss: 0.1163 - accuracy: 0.9610 - val_loss: 0.1169 - val_accuracy: 0.9571\n",
      "Epoch 20/200\n",
      "11/11 [==============================] - 1s 116ms/step - loss: 0.1177 - accuracy: 0.9579 - val_loss: 0.1010 - val_accuracy: 0.9557\n",
      "Epoch 21/200\n",
      "11/11 [==============================] - 1s 114ms/step - loss: 0.1072 - accuracy: 0.9648 - val_loss: 0.1025 - val_accuracy: 0.9600\n",
      "Epoch 22/200\n",
      "11/11 [==============================] - 1s 112ms/step - loss: 0.1180 - accuracy: 0.9646 - val_loss: 0.1093 - val_accuracy: 0.9643\n",
      "Epoch 23/200\n",
      "11/11 [==============================] - 1s 114ms/step - loss: 0.1072 - accuracy: 0.9650 - val_loss: 0.1104 - val_accuracy: 0.9629\n",
      "Epoch 24/200\n",
      "11/11 [==============================] - 1s 114ms/step - loss: 0.0776 - accuracy: 0.9746 - val_loss: 0.1071 - val_accuracy: 0.9614\n",
      "Epoch 25/200\n",
      "11/11 [==============================] - 1s 117ms/step - loss: 0.0886 - accuracy: 0.9710 - val_loss: 0.1095 - val_accuracy: 0.9571\n",
      "Epoch 26/200\n",
      "11/11 [==============================] - 1s 111ms/step - loss: 0.0992 - accuracy: 0.9645 - val_loss: 0.0967 - val_accuracy: 0.9671\n",
      "Epoch 27/200\n",
      "11/11 [==============================] - 1s 114ms/step - loss: 0.0834 - accuracy: 0.9739 - val_loss: 0.0932 - val_accuracy: 0.9686\n",
      "Epoch 28/200\n",
      "11/11 [==============================] - 1s 114ms/step - loss: 0.0706 - accuracy: 0.9743 - val_loss: 0.1098 - val_accuracy: 0.9657\n",
      "Epoch 29/200\n",
      "11/11 [==============================] - 1s 116ms/step - loss: 0.0928 - accuracy: 0.9690 - val_loss: 0.0900 - val_accuracy: 0.9714\n",
      "Epoch 30/200\n",
      "11/11 [==============================] - 1s 110ms/step - loss: 0.0627 - accuracy: 0.9797 - val_loss: 0.0957 - val_accuracy: 0.9700\n",
      "Epoch 31/200\n",
      "11/11 [==============================] - 1s 124ms/step - loss: 0.0644 - accuracy: 0.9780 - val_loss: 0.0961 - val_accuracy: 0.9686\n",
      "Epoch 32/200\n",
      "11/11 [==============================] - 1s 115ms/step - loss: 0.0676 - accuracy: 0.9769 - val_loss: 0.1037 - val_accuracy: 0.9629\n",
      "Epoch 33/200\n",
      "11/11 [==============================] - 1s 111ms/step - loss: 0.0688 - accuracy: 0.9752 - val_loss: 0.1067 - val_accuracy: 0.9571\n",
      "Epoch 34/200\n",
      "11/11 [==============================] - 1s 121ms/step - loss: 0.0638 - accuracy: 0.9807 - val_loss: 0.1098 - val_accuracy: 0.9657\n",
      "Epoch 35/200\n",
      "11/11 [==============================] - 1s 111ms/step - loss: 0.0612 - accuracy: 0.9778 - val_loss: 0.0968 - val_accuracy: 0.9671\n",
      "Epoch 36/200\n",
      "11/11 [==============================] - 1s 113ms/step - loss: 0.0674 - accuracy: 0.9790 - val_loss: 0.0860 - val_accuracy: 0.9714\n",
      "Epoch 37/200\n",
      "11/11 [==============================] - 1s 111ms/step - loss: 0.0521 - accuracy: 0.9854 - val_loss: 0.0991 - val_accuracy: 0.9657\n",
      "Epoch 38/200\n",
      "11/11 [==============================] - 1s 111ms/step - loss: 0.0520 - accuracy: 0.9840 - val_loss: 0.1003 - val_accuracy: 0.9686\n",
      "Epoch 39/200\n",
      "11/11 [==============================] - 1s 109ms/step - loss: 0.0521 - accuracy: 0.9820 - val_loss: 0.0974 - val_accuracy: 0.9586\n",
      "Epoch 40/200\n",
      "11/11 [==============================] - 1s 111ms/step - loss: 0.0459 - accuracy: 0.9859 - val_loss: 0.1000 - val_accuracy: 0.9671\n",
      "Epoch 41/200\n",
      "11/11 [==============================] - 1s 111ms/step - loss: 0.0468 - accuracy: 0.9813 - val_loss: 0.0954 - val_accuracy: 0.9671\n",
      "Epoch 42/200\n",
      "11/11 [==============================] - 1s 116ms/step - loss: 0.0350 - accuracy: 0.9880 - val_loss: 0.0903 - val_accuracy: 0.9686\n",
      "Epoch 43/200\n",
      "11/11 [==============================] - 1s 131ms/step - loss: 0.0368 - accuracy: 0.9899 - val_loss: 0.0778 - val_accuracy: 0.9700\n",
      "Epoch 44/200\n",
      "11/11 [==============================] - 1s 113ms/step - loss: 0.0482 - accuracy: 0.9822 - val_loss: 0.1009 - val_accuracy: 0.9700\n",
      "Epoch 45/200\n",
      "11/11 [==============================] - 1s 124ms/step - loss: 0.0371 - accuracy: 0.9891 - val_loss: 0.1118 - val_accuracy: 0.9657\n",
      "Epoch 46/200\n",
      "11/11 [==============================] - 1s 107ms/step - loss: 0.0363 - accuracy: 0.9878 - val_loss: 0.1087 - val_accuracy: 0.9671\n",
      "Epoch 47/200\n",
      "11/11 [==============================] - 1s 109ms/step - loss: 0.0291 - accuracy: 0.9908 - val_loss: 0.1090 - val_accuracy: 0.9671\n",
      "Epoch 48/200\n",
      "11/11 [==============================] - 1s 111ms/step - loss: 0.0330 - accuracy: 0.9863 - val_loss: 0.1095 - val_accuracy: 0.9671\n",
      "Epoch 49/200\n",
      "11/11 [==============================] - 1s 111ms/step - loss: 0.0379 - accuracy: 0.9851 - val_loss: 0.1166 - val_accuracy: 0.9643\n",
      "Epoch 50/200\n",
      "11/11 [==============================] - 1s 112ms/step - loss: 0.0456 - accuracy: 0.9853 - val_loss: 0.0970 - val_accuracy: 0.9700\n",
      "Epoch 51/200\n",
      "11/11 [==============================] - 1s 133ms/step - loss: 0.0331 - accuracy: 0.9886 - val_loss: 0.0862 - val_accuracy: 0.9700\n",
      "Epoch 52/200\n",
      "11/11 [==============================] - 1s 115ms/step - loss: 0.0311 - accuracy: 0.9903 - val_loss: 0.1065 - val_accuracy: 0.9629\n",
      "Epoch 53/200\n",
      "11/11 [==============================] - 1s 110ms/step - loss: 0.0338 - accuracy: 0.9862 - val_loss: 0.0941 - val_accuracy: 0.9700\n",
      "Epoch 54/200\n",
      "11/11 [==============================] - 1s 108ms/step - loss: 0.0392 - accuracy: 0.9860 - val_loss: 0.1134 - val_accuracy: 0.9714\n",
      "Epoch 55/200\n",
      "11/11 [==============================] - 1s 108ms/step - loss: 0.0375 - accuracy: 0.9864 - val_loss: 0.1014 - val_accuracy: 0.9686\n",
      "Epoch 56/200\n",
      "11/11 [==============================] - 1s 109ms/step - loss: 0.0337 - accuracy: 0.9880 - val_loss: 0.0962 - val_accuracy: 0.9629\n",
      "Epoch 57/200\n",
      "11/11 [==============================] - 1s 109ms/step - loss: 0.0326 - accuracy: 0.9892 - val_loss: 0.1027 - val_accuracy: 0.9671\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 58/200\n",
      "11/11 [==============================] - 1s 115ms/step - loss: 0.0318 - accuracy: 0.9917 - val_loss: 0.1105 - val_accuracy: 0.9686\n",
      "Epoch 59/200\n",
      "11/11 [==============================] - 1s 111ms/step - loss: 0.0239 - accuracy: 0.9916 - val_loss: 0.1138 - val_accuracy: 0.9643\n",
      "Epoch 60/200\n",
      "11/11 [==============================] - 1s 113ms/step - loss: 0.0303 - accuracy: 0.9903 - val_loss: 0.1175 - val_accuracy: 0.9686\n",
      "Epoch 61/200\n",
      "11/11 [==============================] - 1s 114ms/step - loss: 0.0275 - accuracy: 0.9892 - val_loss: 0.1019 - val_accuracy: 0.9714\n",
      "Epoch 62/200\n",
      "11/11 [==============================] - 1s 113ms/step - loss: 0.0399 - accuracy: 0.9875 - val_loss: 0.1001 - val_accuracy: 0.9700\n",
      "Epoch 63/200\n",
      "11/11 [==============================] - 1s 112ms/step - loss: 0.0320 - accuracy: 0.9891 - val_loss: 0.1175 - val_accuracy: 0.9729\n",
      "Epoch 64/200\n",
      "11/11 [==============================] - 1s 109ms/step - loss: 0.0238 - accuracy: 0.9942 - val_loss: 0.1202 - val_accuracy: 0.9671\n",
      "Epoch 65/200\n",
      "11/11 [==============================] - 1s 110ms/step - loss: 0.0175 - accuracy: 0.9943 - val_loss: 0.1141 - val_accuracy: 0.9714\n",
      "Epoch 66/200\n",
      "11/11 [==============================] - 1s 109ms/step - loss: 0.0184 - accuracy: 0.9947 - val_loss: 0.1054 - val_accuracy: 0.9700\n",
      "Epoch 67/200\n",
      "11/11 [==============================] - 1s 109ms/step - loss: 0.0157 - accuracy: 0.9937 - val_loss: 0.0962 - val_accuracy: 0.9729\n",
      "Epoch 68/200\n",
      "11/11 [==============================] - 1s 112ms/step - loss: 0.0187 - accuracy: 0.9939 - val_loss: 0.1127 - val_accuracy: 0.9700\n",
      "Epoch 69/200\n",
      "11/11 [==============================] - 1s 110ms/step - loss: 0.0202 - accuracy: 0.9952 - val_loss: 0.1005 - val_accuracy: 0.9686\n",
      "Epoch 70/200\n",
      "11/11 [==============================] - 1s 119ms/step - loss: 0.0265 - accuracy: 0.9919 - val_loss: 0.1241 - val_accuracy: 0.9686\n",
      "Epoch 71/200\n",
      "11/11 [==============================] - 1s 114ms/step - loss: 0.0236 - accuracy: 0.9909 - val_loss: 0.1158 - val_accuracy: 0.9714\n",
      "Epoch 72/200\n",
      "11/11 [==============================] - 1s 128ms/step - loss: 0.0271 - accuracy: 0.9900 - val_loss: 0.1008 - val_accuracy: 0.9714\n",
      "Epoch 73/200\n",
      "11/11 [==============================] - 1s 118ms/step - loss: 0.0242 - accuracy: 0.9914 - val_loss: 0.1038 - val_accuracy: 0.9686\n",
      "Epoch 74/200\n",
      "11/11 [==============================] - 1s 119ms/step - loss: 0.0185 - accuracy: 0.9929 - val_loss: 0.1237 - val_accuracy: 0.9686\n",
      "Epoch 75/200\n",
      "11/11 [==============================] - 1s 111ms/step - loss: 0.0253 - accuracy: 0.9935 - val_loss: 0.1119 - val_accuracy: 0.9714\n",
      "Epoch 76/200\n",
      "11/11 [==============================] - 1s 116ms/step - loss: 0.0146 - accuracy: 0.9958 - val_loss: 0.0951 - val_accuracy: 0.9757\n",
      "Epoch 77/200\n",
      "11/11 [==============================] - 1s 115ms/step - loss: 0.0268 - accuracy: 0.9907 - val_loss: 0.1129 - val_accuracy: 0.9671\n",
      "Epoch 78/200\n",
      "11/11 [==============================] - 1s 117ms/step - loss: 0.0177 - accuracy: 0.9941 - val_loss: 0.1000 - val_accuracy: 0.9729\n",
      "Epoch 79/200\n",
      "11/11 [==============================] - 1s 115ms/step - loss: 0.0157 - accuracy: 0.9944 - val_loss: 0.1100 - val_accuracy: 0.9686\n",
      "Epoch 80/200\n",
      "11/11 [==============================] - 1s 112ms/step - loss: 0.0171 - accuracy: 0.9949 - val_loss: 0.1326 - val_accuracy: 0.9729\n",
      "Epoch 81/200\n",
      "11/11 [==============================] - 1s 120ms/step - loss: 0.0142 - accuracy: 0.9949 - val_loss: 0.0992 - val_accuracy: 0.9671\n",
      "Epoch 82/200\n",
      "11/11 [==============================] - 1s 132ms/step - loss: 0.0201 - accuracy: 0.9919 - val_loss: 0.1358 - val_accuracy: 0.9729\n",
      "Epoch 83/200\n",
      "11/11 [==============================] - 1s 126ms/step - loss: 0.0131 - accuracy: 0.9960 - val_loss: 0.1114 - val_accuracy: 0.9671\n",
      "Epoch 84/200\n",
      "11/11 [==============================] - 1s 114ms/step - loss: 0.0164 - accuracy: 0.9935 - val_loss: 0.1478 - val_accuracy: 0.9614\n",
      "Epoch 85/200\n",
      "11/11 [==============================] - 1s 109ms/step - loss: 0.0249 - accuracy: 0.9898 - val_loss: 0.1420 - val_accuracy: 0.9671\n",
      "Epoch 86/200\n",
      "11/11 [==============================] - 1s 110ms/step - loss: 0.0133 - accuracy: 0.9958 - val_loss: 0.1386 - val_accuracy: 0.9643\n",
      "Epoch 87/200\n",
      "11/11 [==============================] - 1s 109ms/step - loss: 0.0235 - accuracy: 0.9921 - val_loss: 0.1191 - val_accuracy: 0.9671\n",
      "Epoch 88/200\n",
      "11/11 [==============================] - 1s 106ms/step - loss: 0.0183 - accuracy: 0.9918 - val_loss: 0.1124 - val_accuracy: 0.9686\n",
      "Epoch 89/200\n",
      "11/11 [==============================] - 1s 108ms/step - loss: 0.0130 - accuracy: 0.9953 - val_loss: 0.1247 - val_accuracy: 0.9714\n",
      "Epoch 90/200\n",
      "11/11 [==============================] - 1s 108ms/step - loss: 0.0151 - accuracy: 0.9955 - val_loss: 0.1241 - val_accuracy: 0.9714\n",
      "Epoch 91/200\n",
      "11/11 [==============================] - 1s 109ms/step - loss: 0.0154 - accuracy: 0.9951 - val_loss: 0.1068 - val_accuracy: 0.9714\n",
      "Epoch 92/200\n",
      "11/11 [==============================] - 1s 108ms/step - loss: 0.0140 - accuracy: 0.9945 - val_loss: 0.1178 - val_accuracy: 0.9714\n",
      "Epoch 93/200\n",
      "11/11 [==============================] - 1s 111ms/step - loss: 0.0203 - accuracy: 0.9925 - val_loss: 0.1030 - val_accuracy: 0.9743\n",
      "Epoch 94/200\n",
      "11/11 [==============================] - 1s 109ms/step - loss: 0.0157 - accuracy: 0.9940 - val_loss: 0.1359 - val_accuracy: 0.9671\n",
      "Epoch 95/200\n",
      "11/11 [==============================] - 1s 114ms/step - loss: 0.0158 - accuracy: 0.9938 - val_loss: 0.1125 - val_accuracy: 0.9671\n",
      "Epoch 96/200\n",
      "11/11 [==============================] - 1s 117ms/step - loss: 0.0143 - accuracy: 0.9940 - val_loss: 0.1482 - val_accuracy: 0.9629\n",
      "Epoch 97/200\n",
      "11/11 [==============================] - 1s 112ms/step - loss: 0.0112 - accuracy: 0.9957 - val_loss: 0.1355 - val_accuracy: 0.9700\n",
      "Epoch 98/200\n",
      "11/11 [==============================] - 1s 113ms/step - loss: 0.0181 - accuracy: 0.9957 - val_loss: 0.1203 - val_accuracy: 0.9671\n",
      "Epoch 99/200\n",
      "11/11 [==============================] - 1s 114ms/step - loss: 0.0173 - accuracy: 0.9923 - val_loss: 0.1576 - val_accuracy: 0.9643\n",
      "Epoch 100/200\n",
      "11/11 [==============================] - 1s 115ms/step - loss: 0.0246 - accuracy: 0.9914 - val_loss: 0.1221 - val_accuracy: 0.9700\n",
      "Epoch 101/200\n",
      "11/11 [==============================] - 1s 115ms/step - loss: 0.0146 - accuracy: 0.9932 - val_loss: 0.1438 - val_accuracy: 0.9657\n",
      "Epoch 102/200\n",
      "11/11 [==============================] - 1s 131ms/step - loss: 0.0195 - accuracy: 0.9935 - val_loss: 0.1253 - val_accuracy: 0.9671\n",
      "Epoch 103/200\n",
      "11/11 [==============================] - 1s 112ms/step - loss: 0.0126 - accuracy: 0.9962 - val_loss: 0.1347 - val_accuracy: 0.9643\n",
      "Epoch 104/200\n",
      "11/11 [==============================] - 1s 118ms/step - loss: 0.0138 - accuracy: 0.9962 - val_loss: 0.1194 - val_accuracy: 0.9700\n",
      "Epoch 105/200\n",
      "11/11 [==============================] - 1s 118ms/step - loss: 0.0145 - accuracy: 0.9952 - val_loss: 0.1162 - val_accuracy: 0.9743\n",
      "Epoch 106/200\n",
      "11/11 [==============================] - 1s 105ms/step - loss: 0.0151 - accuracy: 0.9950 - val_loss: 0.1225 - val_accuracy: 0.9743\n",
      "Epoch 107/200\n",
      "11/11 [==============================] - 1s 129ms/step - loss: 0.0177 - accuracy: 0.9946 - val_loss: 0.1441 - val_accuracy: 0.9629\n",
      "Epoch 108/200\n",
      "11/11 [==============================] - 1s 116ms/step - loss: 0.0125 - accuracy: 0.9960 - val_loss: 0.1230 - val_accuracy: 0.9671\n",
      "Epoch 109/200\n",
      "11/11 [==============================] - 1s 106ms/step - loss: 0.0141 - accuracy: 0.9969 - val_loss: 0.1324 - val_accuracy: 0.9643\n",
      "Epoch 110/200\n",
      "11/11 [==============================] - 1s 117ms/step - loss: 0.0102 - accuracy: 0.9969 - val_loss: 0.1411 - val_accuracy: 0.9671\n",
      "Epoch 111/200\n",
      "11/11 [==============================] - 1s 125ms/step - loss: 0.0111 - accuracy: 0.9962 - val_loss: 0.1295 - val_accuracy: 0.9714\n",
      "Epoch 112/200\n",
      "11/11 [==============================] - 1s 119ms/step - loss: 0.0119 - accuracy: 0.9954 - val_loss: 0.1285 - val_accuracy: 0.9686\n",
      "Epoch 113/200\n",
      "11/11 [==============================] - 1s 134ms/step - loss: 0.0171 - accuracy: 0.9923 - val_loss: 0.1257 - val_accuracy: 0.9643\n",
      "Epoch 114/200\n",
      "11/11 [==============================] - 1s 108ms/step - loss: 0.0173 - accuracy: 0.9932 - val_loss: 0.1313 - val_accuracy: 0.9686\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 115/200\n",
      "11/11 [==============================] - 1s 106ms/step - loss: 0.0115 - accuracy: 0.9947 - val_loss: 0.1185 - val_accuracy: 0.9700\n",
      "Epoch 116/200\n",
      "11/11 [==============================] - 1s 106ms/step - loss: 0.0132 - accuracy: 0.9959 - val_loss: 0.1286 - val_accuracy: 0.9700\n",
      "Epoch 117/200\n",
      "11/11 [==============================] - 1s 127ms/step - loss: 0.0119 - accuracy: 0.9953 - val_loss: 0.1414 - val_accuracy: 0.9686\n",
      "Epoch 118/200\n",
      "11/11 [==============================] - 1s 113ms/step - loss: 0.0078 - accuracy: 0.9977 - val_loss: 0.1622 - val_accuracy: 0.9686\n",
      "Epoch 119/200\n",
      "11/11 [==============================] - 1s 114ms/step - loss: 0.0093 - accuracy: 0.9970 - val_loss: 0.1688 - val_accuracy: 0.9700\n",
      "Epoch 120/200\n",
      "11/11 [==============================] - 1s 116ms/step - loss: 0.0125 - accuracy: 0.9951 - val_loss: 0.1816 - val_accuracy: 0.9614\n",
      "Epoch 121/200\n",
      "11/11 [==============================] - 1s 116ms/step - loss: 0.0195 - accuracy: 0.9912 - val_loss: 0.1508 - val_accuracy: 0.9629\n",
      "Epoch 122/200\n",
      "11/11 [==============================] - 1s 111ms/step - loss: 0.0127 - accuracy: 0.9967 - val_loss: 0.1234 - val_accuracy: 0.9700\n",
      "Epoch 123/200\n",
      "11/11 [==============================] - 1s 119ms/step - loss: 0.0099 - accuracy: 0.9962 - val_loss: 0.1194 - val_accuracy: 0.9671\n",
      "Epoch 124/200\n",
      "11/11 [==============================] - 1s 122ms/step - loss: 0.0080 - accuracy: 0.9972 - val_loss: 0.1270 - val_accuracy: 0.9729\n",
      "Epoch 125/200\n",
      "11/11 [==============================] - 1s 120ms/step - loss: 0.0113 - accuracy: 0.9972 - val_loss: 0.1429 - val_accuracy: 0.9657\n",
      "Epoch 126/200\n",
      "11/11 [==============================] - 1s 124ms/step - loss: 0.0082 - accuracy: 0.9977 - val_loss: 0.1561 - val_accuracy: 0.9714\n",
      "Epoch 127/200\n",
      "11/11 [==============================] - 1s 133ms/step - loss: 0.0075 - accuracy: 0.9979 - val_loss: 0.1421 - val_accuracy: 0.9714\n",
      "Epoch 128/200\n",
      "11/11 [==============================] - 1s 127ms/step - loss: 0.0074 - accuracy: 0.9983 - val_loss: 0.1360 - val_accuracy: 0.9743\n",
      "Epoch 129/200\n",
      "11/11 [==============================] - 1s 123ms/step - loss: 0.0133 - accuracy: 0.9935 - val_loss: 0.1536 - val_accuracy: 0.9714\n",
      "Epoch 130/200\n",
      "11/11 [==============================] - 1s 109ms/step - loss: 0.0127 - accuracy: 0.9948 - val_loss: 0.1760 - val_accuracy: 0.9657\n",
      "Epoch 131/200\n",
      "11/11 [==============================] - 1s 120ms/step - loss: 0.0158 - accuracy: 0.9949 - val_loss: 0.1527 - val_accuracy: 0.9700\n",
      "Epoch 132/200\n",
      "11/11 [==============================] - 1s 121ms/step - loss: 0.0061 - accuracy: 0.9988 - val_loss: 0.1602 - val_accuracy: 0.9714\n",
      "Epoch 133/200\n",
      "11/11 [==============================] - 1s 120ms/step - loss: 0.0087 - accuracy: 0.9975 - val_loss: 0.1666 - val_accuracy: 0.9743\n",
      "Epoch 134/200\n",
      "11/11 [==============================] - 1s 135ms/step - loss: 0.0122 - accuracy: 0.9948 - val_loss: 0.1287 - val_accuracy: 0.9757\n",
      "Epoch 135/200\n",
      "11/11 [==============================] - 1s 116ms/step - loss: 0.0190 - accuracy: 0.9939 - val_loss: 0.1486 - val_accuracy: 0.9671\n",
      "Epoch 136/200\n",
      "11/11 [==============================] - 1s 116ms/step - loss: 0.0157 - accuracy: 0.9970 - val_loss: 0.1359 - val_accuracy: 0.9714\n",
      "Epoch 137/200\n",
      "11/11 [==============================] - 1s 122ms/step - loss: 0.0129 - accuracy: 0.9963 - val_loss: 0.1442 - val_accuracy: 0.9714\n",
      "Epoch 138/200\n",
      "11/11 [==============================] - 1s 111ms/step - loss: 0.0083 - accuracy: 0.9976 - val_loss: 0.1306 - val_accuracy: 0.9729\n",
      "Epoch 139/200\n",
      "11/11 [==============================] - 1s 119ms/step - loss: 0.0092 - accuracy: 0.9972 - val_loss: 0.1064 - val_accuracy: 0.9757\n",
      "Epoch 140/200\n",
      "11/11 [==============================] - 1s 111ms/step - loss: 0.0167 - accuracy: 0.9933 - val_loss: 0.1399 - val_accuracy: 0.9757\n",
      "Epoch 141/200\n",
      "11/11 [==============================] - 1s 117ms/step - loss: 0.0095 - accuracy: 0.9978 - val_loss: 0.1444 - val_accuracy: 0.9743\n",
      "Epoch 142/200\n",
      "11/11 [==============================] - 1s 118ms/step - loss: 0.0099 - accuracy: 0.9967 - val_loss: 0.1336 - val_accuracy: 0.9714\n",
      "Epoch 143/200\n",
      "11/11 [==============================] - 1s 113ms/step - loss: 0.0065 - accuracy: 0.9990 - val_loss: 0.1683 - val_accuracy: 0.9700\n",
      "Epoch 144/200\n",
      "11/11 [==============================] - 1s 112ms/step - loss: 0.0132 - accuracy: 0.9959 - val_loss: 0.1515 - val_accuracy: 0.9700\n",
      "Epoch 145/200\n",
      "11/11 [==============================] - 1s 119ms/step - loss: 0.0075 - accuracy: 0.9978 - val_loss: 0.1444 - val_accuracy: 0.9729\n",
      "Epoch 146/200\n",
      "11/11 [==============================] - 1s 123ms/step - loss: 0.0140 - accuracy: 0.9950 - val_loss: 0.1696 - val_accuracy: 0.9700\n",
      "Epoch 147/200\n",
      "11/11 [==============================] - 1s 117ms/step - loss: 0.0151 - accuracy: 0.9942 - val_loss: 0.1437 - val_accuracy: 0.9671\n",
      "Epoch 148/200\n",
      "11/11 [==============================] - 1s 118ms/step - loss: 0.0099 - accuracy: 0.9965 - val_loss: 0.1538 - val_accuracy: 0.9700\n",
      "Epoch 149/200\n",
      "11/11 [==============================] - 1s 116ms/step - loss: 0.0076 - accuracy: 0.9981 - val_loss: 0.1273 - val_accuracy: 0.9729\n",
      "Epoch 150/200\n",
      "11/11 [==============================] - 1s 110ms/step - loss: 0.0084 - accuracy: 0.9969 - val_loss: 0.1536 - val_accuracy: 0.9729\n",
      "Epoch 151/200\n",
      "11/11 [==============================] - 1s 109ms/step - loss: 0.0078 - accuracy: 0.9970 - val_loss: 0.1611 - val_accuracy: 0.9714\n",
      "Epoch 152/200\n",
      "11/11 [==============================] - 1s 130ms/step - loss: 0.0062 - accuracy: 0.9971 - val_loss: 0.1284 - val_accuracy: 0.9757\n",
      "Epoch 153/200\n",
      "11/11 [==============================] - 1s 114ms/step - loss: 0.0116 - accuracy: 0.9959 - val_loss: 0.1313 - val_accuracy: 0.9743\n",
      "Epoch 154/200\n",
      "11/11 [==============================] - 1s 113ms/step - loss: 0.0099 - accuracy: 0.9979 - val_loss: 0.1160 - val_accuracy: 0.9771\n",
      "Epoch 155/200\n",
      "11/11 [==============================] - 1s 115ms/step - loss: 0.0184 - accuracy: 0.9948 - val_loss: 0.1354 - val_accuracy: 0.9700\n",
      "Epoch 156/200\n",
      "11/11 [==============================] - 1s 113ms/step - loss: 0.0091 - accuracy: 0.9947 - val_loss: 0.1338 - val_accuracy: 0.9700\n",
      "Epoch 157/200\n",
      "11/11 [==============================] - 1s 111ms/step - loss: 0.0110 - accuracy: 0.9940 - val_loss: 0.1324 - val_accuracy: 0.9757\n",
      "Epoch 158/200\n",
      "11/11 [==============================] - 1s 122ms/step - loss: 0.0063 - accuracy: 0.9987 - val_loss: 0.1444 - val_accuracy: 0.9757\n",
      "Epoch 159/200\n",
      "11/11 [==============================] - 1s 115ms/step - loss: 0.0123 - accuracy: 0.9962 - val_loss: 0.1242 - val_accuracy: 0.9786\n",
      "Epoch 160/200\n",
      "11/11 [==============================] - 1s 111ms/step - loss: 0.0096 - accuracy: 0.9953 - val_loss: 0.1283 - val_accuracy: 0.9757\n",
      "Epoch 161/200\n",
      "11/11 [==============================] - 1s 111ms/step - loss: 0.0061 - accuracy: 0.9982 - val_loss: 0.1391 - val_accuracy: 0.9771\n",
      "Epoch 162/200\n",
      "11/11 [==============================] - 1s 116ms/step - loss: 0.0068 - accuracy: 0.9980 - val_loss: 0.1216 - val_accuracy: 0.9743\n",
      "Epoch 163/200\n",
      "11/11 [==============================] - 1s 112ms/step - loss: 0.0073 - accuracy: 0.9969 - val_loss: 0.1746 - val_accuracy: 0.9686\n",
      "Epoch 164/200\n",
      "11/11 [==============================] - 1s 114ms/step - loss: 0.0105 - accuracy: 0.9949 - val_loss: 0.1351 - val_accuracy: 0.9743\n",
      "Epoch 165/200\n",
      "11/11 [==============================] - 1s 110ms/step - loss: 0.0095 - accuracy: 0.9954 - val_loss: 0.1298 - val_accuracy: 0.9700\n",
      "Epoch 166/200\n",
      "11/11 [==============================] - 1s 110ms/step - loss: 0.0070 - accuracy: 0.9978 - val_loss: 0.1334 - val_accuracy: 0.9714\n",
      "Epoch 167/200\n",
      "11/11 [==============================] - 1s 115ms/step - loss: 0.0089 - accuracy: 0.9964 - val_loss: 0.1553 - val_accuracy: 0.9714\n",
      "Epoch 168/200\n",
      "11/11 [==============================] - 1s 123ms/step - loss: 0.0040 - accuracy: 0.9987 - val_loss: 0.1415 - val_accuracy: 0.9714\n",
      "Epoch 169/200\n",
      "11/11 [==============================] - 1s 120ms/step - loss: 0.0092 - accuracy: 0.9969 - val_loss: 0.1421 - val_accuracy: 0.9686\n",
      "Epoch 170/200\n",
      "11/11 [==============================] - 1s 114ms/step - loss: 0.0056 - accuracy: 0.9973 - val_loss: 0.1158 - val_accuracy: 0.9771\n",
      "Epoch 171/200\n",
      "11/11 [==============================] - 1s 107ms/step - loss: 0.0062 - accuracy: 0.9973 - val_loss: 0.1402 - val_accuracy: 0.9729\n",
      "Epoch 172/200\n",
      "11/11 [==============================] - 1s 105ms/step - loss: 0.0060 - accuracy: 0.9979 - val_loss: 0.1422 - val_accuracy: 0.9686\n",
      "Epoch 173/200\n",
      "11/11 [==============================] - 1s 107ms/step - loss: 0.0025 - accuracy: 1.0000 - val_loss: 0.1358 - val_accuracy: 0.9743\n",
      "Epoch 174/200\n",
      "11/11 [==============================] - 1s 106ms/step - loss: 0.0117 - accuracy: 0.9975 - val_loss: 0.1832 - val_accuracy: 0.9686\n",
      "Epoch 175/200\n",
      "11/11 [==============================] - 1s 109ms/step - loss: 0.0054 - accuracy: 0.9994 - val_loss: 0.1377 - val_accuracy: 0.9743\n",
      "Epoch 176/200\n",
      "11/11 [==============================] - 1s 112ms/step - loss: 0.0053 - accuracy: 0.9983 - val_loss: 0.1389 - val_accuracy: 0.9743\n",
      "Epoch 177/200\n",
      "11/11 [==============================] - 1s 116ms/step - loss: 0.0049 - accuracy: 0.9989 - val_loss: 0.1605 - val_accuracy: 0.9757\n",
      "Epoch 178/200\n",
      "11/11 [==============================] - 1s 123ms/step - loss: 0.0028 - accuracy: 0.9992 - val_loss: 0.1509 - val_accuracy: 0.9729\n",
      "Epoch 179/200\n",
      "11/11 [==============================] - 1s 110ms/step - loss: 0.0096 - accuracy: 0.9959 - val_loss: 0.1327 - val_accuracy: 0.9786\n",
      "Epoch 180/200\n",
      "11/11 [==============================] - 1s 123ms/step - loss: 0.0039 - accuracy: 0.9991 - val_loss: 0.1572 - val_accuracy: 0.9771\n",
      "Epoch 181/200\n",
      "11/11 [==============================] - 1s 110ms/step - loss: 0.0064 - accuracy: 0.9984 - val_loss: 0.1332 - val_accuracy: 0.9786\n",
      "Epoch 182/200\n",
      "11/11 [==============================] - 1s 109ms/step - loss: 0.0095 - accuracy: 0.9964 - val_loss: 0.1273 - val_accuracy: 0.9771\n",
      "Epoch 183/200\n",
      "11/11 [==============================] - 1s 122ms/step - loss: 0.0150 - accuracy: 0.9955 - val_loss: 0.1329 - val_accuracy: 0.9743\n",
      "Epoch 184/200\n",
      "11/11 [==============================] - 2s 135ms/step - loss: 0.0075 - accuracy: 0.9971 - val_loss: 0.1432 - val_accuracy: 0.9729\n",
      "Epoch 185/200\n",
      "11/11 [==============================] - 1s 114ms/step - loss: 0.0060 - accuracy: 0.9981 - val_loss: 0.1380 - val_accuracy: 0.9729\n",
      "Epoch 186/200\n",
      "11/11 [==============================] - 1s 108ms/step - loss: 0.0086 - accuracy: 0.9970 - val_loss: 0.1620 - val_accuracy: 0.9686\n",
      "Epoch 187/200\n",
      "11/11 [==============================] - 1s 117ms/step - loss: 0.0064 - accuracy: 0.9984 - val_loss: 0.1619 - val_accuracy: 0.9686\n",
      "Epoch 188/200\n",
      "11/11 [==============================] - 2s 151ms/step - loss: 0.0088 - accuracy: 0.9969 - val_loss: 0.1536 - val_accuracy: 0.9700\n",
      "Epoch 189/200\n",
      "11/11 [==============================] - 2s 144ms/step - loss: 0.0059 - accuracy: 0.9984 - val_loss: 0.1706 - val_accuracy: 0.9700\n",
      "Epoch 190/200\n",
      "11/11 [==============================] - 1s 115ms/step - loss: 0.0273 - accuracy: 0.9952 - val_loss: 0.1558 - val_accuracy: 0.9729\n",
      "Epoch 191/200\n",
      "11/11 [==============================] - 1s 106ms/step - loss: 0.0067 - accuracy: 0.9985 - val_loss: 0.1298 - val_accuracy: 0.9700\n",
      "Epoch 192/200\n",
      "11/11 [==============================] - 1s 106ms/step - loss: 0.0066 - accuracy: 0.9971 - val_loss: 0.1772 - val_accuracy: 0.9714\n",
      "Epoch 193/200\n",
      "11/11 [==============================] - 1s 108ms/step - loss: 0.0089 - accuracy: 0.9963 - val_loss: 0.1522 - val_accuracy: 0.9700\n",
      "Epoch 194/200\n",
      "11/11 [==============================] - 1s 120ms/step - loss: 0.0101 - accuracy: 0.9967 - val_loss: 0.1425 - val_accuracy: 0.9729\n",
      "Epoch 195/200\n",
      "11/11 [==============================] - 1s 110ms/step - loss: 0.0041 - accuracy: 0.9983 - val_loss: 0.1523 - val_accuracy: 0.9743\n",
      "Epoch 196/200\n",
      "11/11 [==============================] - 1s 107ms/step - loss: 0.0042 - accuracy: 0.9993 - val_loss: 0.1610 - val_accuracy: 0.9743\n",
      "Epoch 197/200\n",
      "11/11 [==============================] - 1s 130ms/step - loss: 0.0062 - accuracy: 0.9978 - val_loss: 0.1912 - val_accuracy: 0.9757\n",
      "Epoch 198/200\n",
      "11/11 [==============================] - 2s 147ms/step - loss: 0.0039 - accuracy: 0.9983 - val_loss: 0.1590 - val_accuracy: 0.9743\n",
      "Epoch 199/200\n",
      "11/11 [==============================] - 2s 169ms/step - loss: 0.0085 - accuracy: 0.9968 - val_loss: 0.1870 - val_accuracy: 0.9714\n",
      "Epoch 200/200\n",
      "11/11 [==============================] - 2s 157ms/step - loss: 0.0049 - accuracy: 0.9977 - val_loss: 0.1708 - val_accuracy: 0.9686\n"
     ]
    }
   ],
   "source": [
    "# 1. Import libraries and modules\n",
    "from tensorflow import keras\n",
    "from keras import layers\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout\n",
    "import numpy as np\n",
    "from dataset import load_hoda\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Load pre-shuffled HODA data into train and test sets\n",
    "X_train, y_train, X_test, y_test = load_hoda(\n",
    "                                            training_sample_size=3500,\n",
    "                                            test_sample_size=400,size=28)\n",
    "\n",
    "# Preprocess input data\n",
    "'''normalize our data values to the range [0, 1]'''\n",
    "X_train = X_train.astype('float32')\n",
    "X_test = X_test.astype('float32')\n",
    "X_train /= 255\n",
    "X_test /= 255\n",
    "\n",
    "# Reshape to original image shape (n x 784)  ==> (n x 28 x 28 x 1)\n",
    "X_train = X_train.reshape(-1,28,28,1)\n",
    "X_test = X_test.reshape(-1,28,28,1)\n",
    "\n",
    "\n",
    "# 4. Preprocess class labels\n",
    "y_train = keras.utils.to_categorical(y_train)\n",
    "y_test = keras.utils.to_categorical(y_test)\n",
    "\n",
    "\n",
    "# 5. Define model architecture\n",
    "model = Sequential()\n",
    "model.add(layers.Conv2D(32, (3, 3), activation='relu',\n",
    "                        input_shape=(28, 28, 1)))\n",
    "model.add(layers.MaxPooling2D((2, 2)))\n",
    "model.add(layers.Conv2D(64, (3, 3), activation='relu'))\n",
    "model.add(layers.MaxPooling2D((2, 2)))\n",
    "model.add(layers.Conv2D(64, (3, 3), activation='relu'))\n",
    "model.add(layers.Flatten())\n",
    "model.add(layers.Dense(64, activation='relu'))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(layers.Dense(10, activation='softmax'))\n",
    "\n",
    "\n",
    "# 6. Compile model\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer='adam',\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "\n",
    "# 7. Fit model on training data\n",
    "history = model.fit(X_train, y_train,\n",
    "          epochs=200, batch_size=256, validation_split=0.2)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "acc = history.history['accuracy']\n",
    "val_acc = history.history['val_accuracy']\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "epochs = range(len(acc))\n",
    "\n",
    "plt.plot(epochs, acc, 'bo', label='Training acc')\n",
    "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\n",
    "\n",
    "plt.figure()\n",
    "\n",
    "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
    "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "tf2.5",
   "language": "python",
   "name": "tensorflow_25"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.10"
  },
  "nbpresent": {
   "slides": {
    "300ee14f-a043-486e-b274-7ff253907cd7": {
     "id": "300ee14f-a043-486e-b274-7ff253907cd7",
     "prev": "cb74e0bc-4513-4d13-b7f1-14c3078a7927",
     "regions": {
      "26dc3f39-a230-447c-af4c-f5e5b2fb7835": {
       "attrs": {
        "height": 0.8,
        "width": 0.8,
        "x": 0.1,
        "y": 0.1
       },
       "content": {
        "cell": "c58440a5-3f8f-4f37-9c79-6bf766209406",
        "part": "whole"
       },
       "id": "26dc3f39-a230-447c-af4c-f5e5b2fb7835"
      }
     }
    },
    "878aa53a-1444-4100-8f50-7a408191c579": {
     "id": "878aa53a-1444-4100-8f50-7a408191c579",
     "prev": null,
     "regions": {
      "a6c6843a-5ea6-4fbc-b890-3b4b8ae475b3": {
       "attrs": {
        "height": 0.8,
        "width": 0.8,
        "x": 0.1,
        "y": 0.1
       },
       "content": {
        "cell": "588ee1fa-64b5-453b-ade7-8e6b2515821c",
        "part": "whole"
       },
       "id": "a6c6843a-5ea6-4fbc-b890-3b4b8ae475b3"
      }
     }
    },
    "96ffe88e-7b50-43de-afdd-942e564f4e3e": {
     "id": "96ffe88e-7b50-43de-afdd-942e564f4e3e",
     "prev": "878aa53a-1444-4100-8f50-7a408191c579",
     "regions": {
      "b7e52e12-489a-468d-b10c-af2024fd2856": {
       "attrs": {
        "height": 0.8,
        "width": 0.8,
        "x": 0.1,
        "y": 0.1
       },
       "content": {
        "cell": "de829a92-1fb6-44ad-a2c6-fc1001e1f6e1",
        "part": "whole"
       },
       "id": "b7e52e12-489a-468d-b10c-af2024fd2856"
      }
     }
    },
    "cb74e0bc-4513-4d13-b7f1-14c3078a7927": {
     "id": "cb74e0bc-4513-4d13-b7f1-14c3078a7927",
     "prev": "96ffe88e-7b50-43de-afdd-942e564f4e3e",
     "regions": {
      "444878ee-68f3-4abb-acff-a7079b21e86d": {
       "attrs": {
        "height": 0.8,
        "width": 0.8,
        "x": 0.1,
        "y": 0.1
       },
       "content": {
        "cell": "25f3f538-1ee8-4d98-a6bb-14cbeb7a702d",
        "part": "whole"
       },
       "id": "444878ee-68f3-4abb-acff-a7079b21e86d"
      }
     }
    }
   },
   "themes": {}
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}