Example 4. 4-layer Neural Network.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The goal is to do binary classification on the make_moons dataset from sklearn.\n",
    "\n",
    "It is utilized a neural nerwork based on relu and sigmoid activation functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from mlswarm.mlswarm import neuralnet \n",
    "\n",
    "#DATASET-------------------------------\n",
    "from sklearn.datasets import make_moons\n",
    "from sklearn.model_selection import train_test_split\n",
    "X, Y = make_moons(n_samples = 2000, noise=0.2, random_state=100)\n",
    "Y = Y.reshape((Y.shape[0],1))\n",
    "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.2, random_state = 42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Training started...\n",
      "Iteration: 00000 - Cloud mean cost: 0.25133 - Cloud variance: 0.46757\n",
      "Iteration: 00001 - Cloud mean cost: 0.25138 - Cloud variance: 0.46771\n",
      "Iteration: 00002 - Cloud mean cost: 0.25152 - Cloud variance: 0.47211\n",
      "Iteration: 00003 - Cloud mean cost: 0.25170 - Cloud variance: 0.48620\n",
      "Iteration: 00004 - Cloud mean cost: 0.25191 - Cloud variance: 0.50009\n",
      "Iteration: 00005 - Cloud mean cost: 0.25214 - Cloud variance: 0.51374\n",
      "Iteration: 00006 - Cloud mean cost: 0.25239 - Cloud variance: 0.52713\n",
      "Iteration: 00007 - Cloud mean cost: 0.25265 - Cloud variance: 0.54025\n",
      "Iteration: 00008 - Cloud mean cost: 0.25293 - Cloud variance: 0.55307\n",
      "Iteration: 00009 - Cloud mean cost: 0.25323 - Cloud variance: 0.56559\n",
      "Iteration: 00010 - Cloud mean cost: 0.25353 - Cloud variance: 0.57778\n",
      "Iteration: 00011 - Cloud mean cost: 0.25386 - Cloud variance: 0.58962\n",
      "Iteration: 00012 - Cloud mean cost: 0.25419 - Cloud variance: 0.60111\n",
      "Iteration: 00013 - Cloud mean cost: 0.25453 - Cloud variance: 0.61222\n",
      "Iteration: 00014 - Cloud mean cost: 0.25489 - Cloud variance: 0.62296\n",
      "Iteration: 00015 - Cloud mean cost: 0.25526 - Cloud variance: 0.63331\n",
      "Iteration: 00016 - Cloud mean cost: 0.25565 - Cloud variance: 0.64327\n",
      "Iteration: 00017 - Cloud mean cost: 0.25604 - Cloud variance: 0.65284\n",
      "Iteration: 00018 - Cloud mean cost: 0.25644 - Cloud variance: 0.66202\n",
      "Iteration: 00019 - Cloud mean cost: 0.25684 - Cloud variance: 0.67079\n",
      "Iteration: 00020 - Cloud mean cost: 0.25725 - Cloud variance: 0.67915\n",
      "Iteration: 00021 - Cloud mean cost: 0.25766 - Cloud variance: 0.68710\n",
      "Iteration: 00022 - Cloud mean cost: 0.25808 - Cloud variance: 0.69464\n",
      "Iteration: 00023 - Cloud mean cost: 0.25848 - Cloud variance: 0.70176\n",
      "Iteration: 00024 - Cloud mean cost: 0.25885 - Cloud variance: 0.70846\n",
      "Iteration: 00025 - Cloud mean cost: 0.25921 - Cloud variance: 0.71475\n",
      "Iteration: 00026 - Cloud mean cost: 0.25955 - Cloud variance: 0.72063\n",
      "Iteration: 00027 - Cloud mean cost: 0.25987 - Cloud variance: 0.72610\n",
      "Iteration: 00028 - Cloud mean cost: 0.26017 - Cloud variance: 0.73117\n",
      "Iteration: 00029 - Cloud mean cost: 0.26045 - Cloud variance: 0.73584\n",
      "Iteration: 00030 - Cloud mean cost: 0.26071 - Cloud variance: 0.74013\n",
      "Iteration: 00031 - Cloud mean cost: 0.26095 - Cloud variance: 0.74403\n",
      "Iteration: 00032 - Cloud mean cost: 0.26118 - Cloud variance: 0.74758\n",
      "Iteration: 00033 - Cloud mean cost: 0.26139 - Cloud variance: 0.75076\n",
      "Iteration: 00034 - Cloud mean cost: 0.26160 - Cloud variance: 0.75360\n",
      "Iteration: 00035 - Cloud mean cost: 0.26180 - Cloud variance: 0.75610\n",
      "Iteration: 00036 - Cloud mean cost: 0.26200 - Cloud variance: 0.75826\n",
      "Iteration: 00037 - Cloud mean cost: 0.26219 - Cloud variance: 0.76008\n",
      "Iteration: 00038 - Cloud mean cost: 0.26238 - Cloud variance: 0.76157\n",
      "Iteration: 00039 - Cloud mean cost: 0.26256 - Cloud variance: 0.76274\n",
      "Iteration: 00040 - Cloud mean cost: 0.26273 - Cloud variance: 0.76359\n",
      "Iteration: 00041 - Cloud mean cost: 0.26287 - Cloud variance: 0.76413\n",
      "Iteration: 00042 - Cloud mean cost: 0.26301 - Cloud variance: 0.76438\n",
      "Iteration: 00043 - Cloud mean cost: 0.26312 - Cloud variance: 0.76434\n",
      "Iteration: 00044 - Cloud mean cost: 0.26319 - Cloud variance: 0.76402\n",
      "Iteration: 00045 - Cloud mean cost: 0.26324 - Cloud variance: 0.76343\n",
      "Iteration: 00046 - Cloud mean cost: 0.26331 - Cloud variance: 0.76258\n",
      "Iteration: 00047 - Cloud mean cost: 0.26341 - Cloud variance: 0.76147\n",
      "Iteration: 00048 - Cloud mean cost: 0.26355 - Cloud variance: 0.76012\n",
      "Iteration: 00049 - Cloud mean cost: 0.26368 - Cloud variance: 0.75851\n",
      "Iteration: 00050 - Cloud mean cost: 0.26379 - Cloud variance: 0.75667\n",
      "Iteration: 00051 - Cloud mean cost: 0.26388 - Cloud variance: 0.75460\n",
      "Iteration: 00052 - Cloud mean cost: 0.26395 - Cloud variance: 0.75231\n",
      "Iteration: 00053 - Cloud mean cost: 0.26400 - Cloud variance: 0.74980\n",
      "Iteration: 00054 - Cloud mean cost: 0.26403 - Cloud variance: 0.74709\n",
      "Iteration: 00055 - Cloud mean cost: 0.26403 - Cloud variance: 0.74417\n",
      "Iteration: 00056 - Cloud mean cost: 0.26403 - Cloud variance: 0.74107\n",
      "Iteration: 00057 - Cloud mean cost: 0.26400 - Cloud variance: 0.73778\n",
      "Iteration: 00058 - Cloud mean cost: 0.26397 - Cloud variance: 0.73433\n",
      "Iteration: 00059 - Cloud mean cost: 0.26393 - Cloud variance: 0.73071\n",
      "Iteration: 00060 - Cloud mean cost: 0.26388 - Cloud variance: 0.72695\n",
      "Iteration: 00061 - Cloud mean cost: 0.26384 - Cloud variance: 0.72306\n",
      "Iteration: 00062 - Cloud mean cost: 0.26380 - Cloud variance: 0.71905\n",
      "Iteration: 00063 - Cloud mean cost: 0.26377 - Cloud variance: 0.71492\n",
      "Iteration: 00064 - Cloud mean cost: 0.26380 - Cloud variance: 0.71066\n",
      "Iteration: 00065 - Cloud mean cost: 0.26388 - Cloud variance: 0.70629\n",
      "Iteration: 00066 - Cloud mean cost: 0.26397 - Cloud variance: 0.70182\n",
      "Iteration: 00067 - Cloud mean cost: 0.26407 - Cloud variance: 0.69723\n",
      "Iteration: 00068 - Cloud mean cost: 0.26415 - Cloud variance: 0.69254\n",
      "Iteration: 00069 - Cloud mean cost: 0.26421 - Cloud variance: 0.68776\n",
      "Iteration: 00070 - Cloud mean cost: 0.26425 - Cloud variance: 0.68287\n",
      "Iteration: 00071 - Cloud mean cost: 0.26426 - Cloud variance: 0.67789\n",
      "Iteration: 00072 - Cloud mean cost: 0.26424 - Cloud variance: 0.67280\n",
      "Iteration: 00073 - Cloud mean cost: 0.26419 - Cloud variance: 0.66762\n",
      "Iteration: 00074 - Cloud mean cost: 0.26412 - Cloud variance: 0.66235\n",
      "Iteration: 00075 - Cloud mean cost: 0.26403 - Cloud variance: 0.65699\n",
      "Iteration: 00076 - Cloud mean cost: 0.26392 - Cloud variance: 0.65156\n",
      "Iteration: 00077 - Cloud mean cost: 0.26380 - Cloud variance: 0.64606\n",
      "Iteration: 00078 - Cloud mean cost: 0.26367 - Cloud variance: 0.64051\n",
      "Iteration: 00079 - Cloud mean cost: 0.26353 - Cloud variance: 0.63492\n",
      "Iteration: 00080 - Cloud mean cost: 0.26338 - Cloud variance: 0.62930\n",
      "Iteration: 00081 - Cloud mean cost: 0.26323 - Cloud variance: 0.62366\n",
      "Iteration: 00082 - Cloud mean cost: 0.26308 - Cloud variance: 0.61800\n",
      "Iteration: 00083 - Cloud mean cost: 0.26292 - Cloud variance: 0.61232\n",
      "Iteration: 00084 - Cloud mean cost: 0.26277 - Cloud variance: 0.60664\n",
      "Iteration: 00085 - Cloud mean cost: 0.26260 - Cloud variance: 0.60096\n",
      "Iteration: 00086 - Cloud mean cost: 0.26243 - Cloud variance: 0.59528\n",
      "Iteration: 00087 - Cloud mean cost: 0.26224 - Cloud variance: 0.58961\n",
      "Iteration: 00088 - Cloud mean cost: 0.26203 - Cloud variance: 0.58395\n",
      "Iteration: 00089 - Cloud mean cost: 0.26181 - Cloud variance: 0.57832\n",
      "Iteration: 00090 - Cloud mean cost: 0.26158 - Cloud variance: 0.57270\n",
      "Iteration: 00091 - Cloud mean cost: 0.26133 - Cloud variance: 0.56712\n",
      "Iteration: 00092 - Cloud mean cost: 0.26107 - Cloud variance: 0.56156\n",
      "Iteration: 00093 - Cloud mean cost: 0.26081 - Cloud variance: 0.55604\n",
      "Iteration: 00094 - Cloud mean cost: 0.26055 - Cloud variance: 0.55057\n",
      "Iteration: 00095 - Cloud mean cost: 0.26028 - Cloud variance: 0.54514\n",
      "Iteration: 00096 - Cloud mean cost: 0.26002 - Cloud variance: 0.53975\n",
      "Iteration: 00097 - Cloud mean cost: 0.25977 - Cloud variance: 0.53442\n",
      "Iteration: 00098 - Cloud mean cost: 0.25953 - Cloud variance: 0.52913\n",
      "Iteration: 00099 - Cloud mean cost: 0.25931 - Cloud variance: 0.52389\n",
      "Iteration: 00100 - Cloud mean cost: 0.25911 - Cloud variance: 0.51870\n",
      "Iteration: 00101 - Cloud mean cost: 0.25892 - Cloud variance: 0.51354\n",
      "Iteration: 00102 - Cloud mean cost: 0.25875 - Cloud variance: 0.50843\n",
      "Iteration: 00103 - Cloud mean cost: 0.25860 - Cloud variance: 0.50335\n",
      "Iteration: 00104 - Cloud mean cost: 0.25847 - Cloud variance: 0.49831\n",
      "Iteration: 00105 - Cloud mean cost: 0.25834 - Cloud variance: 0.49330\n",
      "Iteration: 00106 - Cloud mean cost: 0.25823 - Cloud variance: 0.48831\n",
      "Iteration: 00107 - Cloud mean cost: 0.25814 - Cloud variance: 0.48336\n",
      "Iteration: 00108 - Cloud mean cost: 0.25805 - Cloud variance: 0.47842\n",
      "Iteration: 00109 - Cloud mean cost: 0.25798 - Cloud variance: 0.47350\n",
      "Iteration: 00110 - Cloud mean cost: 0.25791 - Cloud variance: 0.46861\n",
      "Iteration: 00111 - Cloud mean cost: 0.25783 - Cloud variance: 0.46375\n",
      "Iteration: 00112 - Cloud mean cost: 0.25776 - Cloud variance: 0.45891\n",
      "Iteration: 00113 - Cloud mean cost: 0.25767 - Cloud variance: 0.45410\n",
      "Iteration: 00114 - Cloud mean cost: 0.25757 - Cloud variance: 0.44932\n",
      "Iteration: 00115 - Cloud mean cost: 0.25745 - Cloud variance: 0.44457\n",
      "Iteration: 00116 - Cloud mean cost: 0.25731 - Cloud variance: 0.43985\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iteration: 00117 - Cloud mean cost: 0.25714 - Cloud variance: 0.43517\n",
      "Iteration: 00118 - Cloud mean cost: 0.25696 - Cloud variance: 0.43053\n",
      "Iteration: 00119 - Cloud mean cost: 0.25676 - Cloud variance: 0.42592\n",
      "Iteration: 00120 - Cloud mean cost: 0.25654 - Cloud variance: 0.42136\n",
      "Iteration: 00121 - Cloud mean cost: 0.25630 - Cloud variance: 0.41685\n",
      "Iteration: 00122 - Cloud mean cost: 0.25605 - Cloud variance: 0.41239\n",
      "Iteration: 00123 - Cloud mean cost: 0.25580 - Cloud variance: 0.40797\n",
      "Iteration: 00124 - Cloud mean cost: 0.25553 - Cloud variance: 0.40361\n",
      "Iteration: 00125 - Cloud mean cost: 0.25526 - Cloud variance: 0.39931\n",
      "Iteration: 00126 - Cloud mean cost: 0.25500 - Cloud variance: 0.39506\n",
      "Iteration: 00127 - Cloud mean cost: 0.25473 - Cloud variance: 0.39086\n",
      "Iteration: 00128 - Cloud mean cost: 0.25447 - Cloud variance: 0.38672\n",
      "Iteration: 00129 - Cloud mean cost: 0.25419 - Cloud variance: 0.38263\n",
      "Iteration: 00130 - Cloud mean cost: 0.25391 - Cloud variance: 0.37859\n",
      "Iteration: 00131 - Cloud mean cost: 0.25364 - Cloud variance: 0.37458\n",
      "Iteration: 00132 - Cloud mean cost: 0.25338 - Cloud variance: 0.37062\n",
      "Iteration: 00133 - Cloud mean cost: 0.25314 - Cloud variance: 0.36670\n",
      "Iteration: 00134 - Cloud mean cost: 0.25291 - Cloud variance: 0.36282\n",
      "Iteration: 00135 - Cloud mean cost: 0.25270 - Cloud variance: 0.35899\n",
      "Iteration: 00136 - Cloud mean cost: 0.25251 - Cloud variance: 0.35520\n",
      "Iteration: 00137 - Cloud mean cost: 0.25233 - Cloud variance: 0.35146\n",
      "Iteration: 00138 - Cloud mean cost: 0.25217 - Cloud variance: 0.34775\n",
      "Iteration: 00139 - Cloud mean cost: 0.25203 - Cloud variance: 0.34409\n",
      "Iteration: 00140 - Cloud mean cost: 0.25189 - Cloud variance: 0.34047\n",
      "Iteration: 00141 - Cloud mean cost: 0.25178 - Cloud variance: 0.33689\n",
      "Iteration: 00142 - Cloud mean cost: 0.25167 - Cloud variance: 0.33334\n",
      "Iteration: 00143 - Cloud mean cost: 0.25157 - Cloud variance: 0.32983\n",
      "Iteration: 00144 - Cloud mean cost: 0.25149 - Cloud variance: 0.32635\n",
      "Iteration: 00145 - Cloud mean cost: 0.25141 - Cloud variance: 0.32290\n",
      "Iteration: 00146 - Cloud mean cost: 0.25134 - Cloud variance: 0.31948\n",
      "Iteration: 00147 - Cloud mean cost: 0.25127 - Cloud variance: 0.31609\n",
      "Iteration: 00148 - Cloud mean cost: 0.25118 - Cloud variance: 0.31274\n",
      "Iteration: 00149 - Cloud mean cost: 0.25097 - Cloud variance: 0.30941\n",
      "Iteration: 00150 - Cloud mean cost: 0.25042 - Cloud variance: 0.30610\n",
      "Iteration: 00151 - Cloud mean cost: 0.24950 - Cloud variance: 0.30282\n",
      "Iteration: 00152 - Cloud mean cost: 0.24813 - Cloud variance: 0.29955\n",
      "Iteration: 00153 - Cloud mean cost: 0.24620 - Cloud variance: 0.29631\n",
      "Iteration: 00154 - Cloud mean cost: 0.24298 - Cloud variance: 0.29308\n",
      "Iteration: 00155 - Cloud mean cost: 0.23369 - Cloud variance: 0.28988\n",
      "Iteration: 00156 - Cloud mean cost: 0.22401 - Cloud variance: 0.28668\n",
      "Iteration: 00157 - Cloud mean cost: 0.21508 - Cloud variance: 0.28350\n",
      "Iteration: 00158 - Cloud mean cost: 0.20732 - Cloud variance: 0.28030\n",
      "Iteration: 00159 - Cloud mean cost: 0.20069 - Cloud variance: 0.27709\n",
      "Iteration: 00160 - Cloud mean cost: 0.19524 - Cloud variance: 0.27385\n",
      "Iteration: 00161 - Cloud mean cost: 0.19076 - Cloud variance: 0.27058\n",
      "Iteration: 00162 - Cloud mean cost: 0.18703 - Cloud variance: 0.26729\n",
      "Iteration: 00163 - Cloud mean cost: 0.18397 - Cloud variance: 0.26396\n",
      "Iteration: 00164 - Cloud mean cost: 0.18129 - Cloud variance: 0.26062\n",
      "Iteration: 00165 - Cloud mean cost: 0.17877 - Cloud variance: 0.25726\n",
      "Iteration: 00166 - Cloud mean cost: 0.17628 - Cloud variance: 0.25388\n",
      "Iteration: 00167 - Cloud mean cost: 0.17391 - Cloud variance: 0.25048\n",
      "Iteration: 00168 - Cloud mean cost: 0.17166 - Cloud variance: 0.24705\n",
      "Iteration: 00169 - Cloud mean cost: 0.16943 - Cloud variance: 0.24359\n",
      "Iteration: 00170 - Cloud mean cost: 0.16735 - Cloud variance: 0.24009\n",
      "Iteration: 00171 - Cloud mean cost: 0.16533 - Cloud variance: 0.23655\n",
      "Iteration: 00172 - Cloud mean cost: 0.16334 - Cloud variance: 0.23298\n",
      "Iteration: 00173 - Cloud mean cost: 0.16138 - Cloud variance: 0.22940\n",
      "Iteration: 00174 - Cloud mean cost: 0.15920 - Cloud variance: 0.22578\n",
      "Iteration: 00175 - Cloud mean cost: 0.15697 - Cloud variance: 0.22215\n",
      "Iteration: 00176 - Cloud mean cost: 0.15448 - Cloud variance: 0.21848\n",
      "Iteration: 00177 - Cloud mean cost: 0.15187 - Cloud variance: 0.21480\n",
      "Iteration: 00178 - Cloud mean cost: 0.14944 - Cloud variance: 0.21109\n",
      "Iteration: 00179 - Cloud mean cost: 0.14730 - Cloud variance: 0.20737\n",
      "Iteration: 00180 - Cloud mean cost: 0.14531 - Cloud variance: 0.20364\n",
      "Iteration: 00181 - Cloud mean cost: 0.14350 - Cloud variance: 0.19992\n",
      "Iteration: 00182 - Cloud mean cost: 0.14182 - Cloud variance: 0.19620\n",
      "Iteration: 00183 - Cloud mean cost: 0.14027 - Cloud variance: 0.19250\n",
      "Iteration: 00184 - Cloud mean cost: 0.13872 - Cloud variance: 0.18882\n",
      "Iteration: 00185 - Cloud mean cost: 0.13729 - Cloud variance: 0.18517\n",
      "Iteration: 00186 - Cloud mean cost: 0.13592 - Cloud variance: 0.18156\n",
      "Iteration: 00187 - Cloud mean cost: 0.13478 - Cloud variance: 0.17799\n",
      "Iteration: 00188 - Cloud mean cost: 0.13352 - Cloud variance: 0.17447\n",
      "Iteration: 00189 - Cloud mean cost: 0.13226 - Cloud variance: 0.17101\n",
      "Iteration: 00190 - Cloud mean cost: 0.13095 - Cloud variance: 0.16760\n",
      "Iteration: 00191 - Cloud mean cost: 0.12968 - Cloud variance: 0.16426\n",
      "Iteration: 00192 - Cloud mean cost: 0.12846 - Cloud variance: 0.16098\n",
      "Iteration: 00193 - Cloud mean cost: 0.12738 - Cloud variance: 0.15777\n",
      "Iteration: 00194 - Cloud mean cost: 0.12649 - Cloud variance: 0.15463\n",
      "Iteration: 00195 - Cloud mean cost: 0.12567 - Cloud variance: 0.15155\n",
      "Iteration: 00196 - Cloud mean cost: 0.12480 - Cloud variance: 0.14854\n",
      "Iteration: 00197 - Cloud mean cost: 0.12391 - Cloud variance: 0.14560\n",
      "Iteration: 00198 - Cloud mean cost: 0.12294 - Cloud variance: 0.14273\n",
      "Iteration: 00199 - Cloud mean cost: 0.12204 - Cloud variance: 0.13993\n",
      "Iteration: 00200 - Cloud mean cost: 0.12122 - Cloud variance: 0.13719\n",
      "Iteration: 00201 - Cloud mean cost: 0.12041 - Cloud variance: 0.13451\n",
      "Iteration: 00202 - Cloud mean cost: 0.11956 - Cloud variance: 0.13190\n",
      "Iteration: 00203 - Cloud mean cost: 0.11870 - Cloud variance: 0.12936\n",
      "Iteration: 00204 - Cloud mean cost: 0.11765 - Cloud variance: 0.12689\n",
      "Iteration: 00205 - Cloud mean cost: 0.11616 - Cloud variance: 0.12448\n",
      "Iteration: 00206 - Cloud mean cost: 0.11442 - Cloud variance: 0.12213\n",
      "Iteration: 00207 - Cloud mean cost: 0.11255 - Cloud variance: 0.11985\n",
      "Iteration: 00208 - Cloud mean cost: 0.11066 - Cloud variance: 0.11763\n",
      "Iteration: 00209 - Cloud mean cost: 0.10886 - Cloud variance: 0.11548\n",
      "Iteration: 00210 - Cloud mean cost: 0.10723 - Cloud variance: 0.11338\n",
      "Iteration: 00211 - Cloud mean cost: 0.10569 - Cloud variance: 0.11135\n",
      "Iteration: 00212 - Cloud mean cost: 0.10430 - Cloud variance: 0.10937\n",
      "Iteration: 00213 - Cloud mean cost: 0.10306 - Cloud variance: 0.10744\n",
      "Iteration: 00214 - Cloud mean cost: 0.10198 - Cloud variance: 0.10557\n",
      "Iteration: 00215 - Cloud mean cost: 0.10102 - Cloud variance: 0.10376\n",
      "Iteration: 00216 - Cloud mean cost: 0.10018 - Cloud variance: 0.10199\n",
      "Iteration: 00217 - Cloud mean cost: 0.09945 - Cloud variance: 0.10027\n",
      "Iteration: 00218 - Cloud mean cost: 0.09874 - Cloud variance: 0.09860\n",
      "Iteration: 00219 - Cloud mean cost: 0.09806 - Cloud variance: 0.09698\n",
      "Iteration: 00220 - Cloud mean cost: 0.09738 - Cloud variance: 0.09540\n",
      "Iteration: 00221 - Cloud mean cost: 0.09677 - Cloud variance: 0.09387\n",
      "Iteration: 00222 - Cloud mean cost: 0.09622 - Cloud variance: 0.09238\n",
      "Iteration: 00223 - Cloud mean cost: 0.09567 - Cloud variance: 0.09093\n",
      "Iteration: 00224 - Cloud mean cost: 0.09512 - Cloud variance: 0.08951\n",
      "Iteration: 00225 - Cloud mean cost: 0.09461 - Cloud variance: 0.08814\n",
      "Iteration: 00226 - Cloud mean cost: 0.09412 - Cloud variance: 0.08680\n",
      "Iteration: 00227 - Cloud mean cost: 0.09365 - Cloud variance: 0.08551\n",
      "Iteration: 00228 - Cloud mean cost: 0.09315 - Cloud variance: 0.08424\n",
      "Iteration: 00229 - Cloud mean cost: 0.09266 - Cloud variance: 0.08301\n",
      "Iteration: 00230 - Cloud mean cost: 0.09219 - Cloud variance: 0.08182\n",
      "Iteration: 00231 - Cloud mean cost: 0.09175 - Cloud variance: 0.08065\n",
      "Iteration: 00232 - Cloud mean cost: 0.09133 - Cloud variance: 0.07952\n",
      "Iteration: 00233 - Cloud mean cost: 0.09094 - Cloud variance: 0.07841\n",
      "Iteration: 00234 - Cloud mean cost: 0.09055 - Cloud variance: 0.07734\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iteration: 00235 - Cloud mean cost: 0.09017 - Cloud variance: 0.07629\n",
      "Iteration: 00236 - Cloud mean cost: 0.08980 - Cloud variance: 0.07527\n",
      "Iteration: 00237 - Cloud mean cost: 0.08942 - Cloud variance: 0.07427\n",
      "Iteration: 00238 - Cloud mean cost: 0.08904 - Cloud variance: 0.07330\n",
      "Iteration: 00239 - Cloud mean cost: 0.08866 - Cloud variance: 0.07235\n",
      "Iteration: 00240 - Cloud mean cost: 0.08828 - Cloud variance: 0.07142\n",
      "Iteration: 00241 - Cloud mean cost: 0.08789 - Cloud variance: 0.07051\n",
      "Iteration: 00242 - Cloud mean cost: 0.08750 - Cloud variance: 0.06963\n",
      "Iteration: 00243 - Cloud mean cost: 0.08710 - Cloud variance: 0.06876\n",
      "Iteration: 00244 - Cloud mean cost: 0.08670 - Cloud variance: 0.06791\n",
      "Iteration: 00245 - Cloud mean cost: 0.08630 - Cloud variance: 0.06709\n",
      "Iteration: 00246 - Cloud mean cost: 0.08589 - Cloud variance: 0.06627\n",
      "Iteration: 00247 - Cloud mean cost: 0.08548 - Cloud variance: 0.06548\n",
      "Iteration: 00248 - Cloud mean cost: 0.08508 - Cloud variance: 0.06470\n",
      "Iteration: 00249 - Cloud mean cost: 0.08468 - Cloud variance: 0.06394\n",
      "Iteration: 00250 - Cloud mean cost: 0.08427 - Cloud variance: 0.06319\n",
      "Iteration: 00251 - Cloud mean cost: 0.08386 - Cloud variance: 0.06246\n",
      "Iteration: 00252 - Cloud mean cost: 0.08344 - Cloud variance: 0.06174\n",
      "Iteration: 00253 - Cloud mean cost: 0.08301 - Cloud variance: 0.06103\n",
      "Iteration: 00254 - Cloud mean cost: 0.08256 - Cloud variance: 0.06034\n",
      "Iteration: 00255 - Cloud mean cost: 0.08211 - Cloud variance: 0.05966\n",
      "Iteration: 00256 - Cloud mean cost: 0.08167 - Cloud variance: 0.05900\n",
      "Iteration: 00257 - Cloud mean cost: 0.08125 - Cloud variance: 0.05834\n",
      "Iteration: 00258 - Cloud mean cost: 0.08082 - Cloud variance: 0.05770\n",
      "Iteration: 00259 - Cloud mean cost: 0.08039 - Cloud variance: 0.05707\n",
      "Iteration: 00260 - Cloud mean cost: 0.07995 - Cloud variance: 0.05645\n",
      "Iteration: 00261 - Cloud mean cost: 0.07952 - Cloud variance: 0.05585\n",
      "Iteration: 00262 - Cloud mean cost: 0.07910 - Cloud variance: 0.05525\n",
      "Iteration: 00263 - Cloud mean cost: 0.07869 - Cloud variance: 0.05466\n",
      "Iteration: 00264 - Cloud mean cost: 0.07830 - Cloud variance: 0.05409\n",
      "Iteration: 00265 - Cloud mean cost: 0.07791 - Cloud variance: 0.05352\n",
      "Iteration: 00266 - Cloud mean cost: 0.07753 - Cloud variance: 0.05296\n",
      "Iteration: 00267 - Cloud mean cost: 0.07715 - Cloud variance: 0.05242\n",
      "Iteration: 00268 - Cloud mean cost: 0.07678 - Cloud variance: 0.05188\n",
      "Iteration: 00269 - Cloud mean cost: 0.07643 - Cloud variance: 0.05135\n",
      "Iteration: 00270 - Cloud mean cost: 0.07608 - Cloud variance: 0.05083\n",
      "Iteration: 00271 - Cloud mean cost: 0.07574 - Cloud variance: 0.05032\n",
      "Iteration: 00272 - Cloud mean cost: 0.07541 - Cloud variance: 0.04982\n",
      "Iteration: 00273 - Cloud mean cost: 0.07509 - Cloud variance: 0.04932\n",
      "Iteration: 00274 - Cloud mean cost: 0.07476 - Cloud variance: 0.04884\n",
      "Iteration: 00275 - Cloud mean cost: 0.07445 - Cloud variance: 0.04836\n",
      "Iteration: 00276 - Cloud mean cost: 0.07414 - Cloud variance: 0.04789\n",
      "Iteration: 00277 - Cloud mean cost: 0.07384 - Cloud variance: 0.04742\n",
      "Iteration: 00278 - Cloud mean cost: 0.07354 - Cloud variance: 0.04697\n",
      "Iteration: 00279 - Cloud mean cost: 0.07325 - Cloud variance: 0.04652\n",
      "Iteration: 00280 - Cloud mean cost: 0.07296 - Cloud variance: 0.04608\n",
      "Iteration: 00281 - Cloud mean cost: 0.07267 - Cloud variance: 0.04565\n",
      "Iteration: 00282 - Cloud mean cost: 0.07237 - Cloud variance: 0.04522\n",
      "Iteration: 00283 - Cloud mean cost: 0.07208 - Cloud variance: 0.04481\n",
      "Iteration: 00284 - Cloud mean cost: 0.07180 - Cloud variance: 0.04440\n",
      "Iteration: 00285 - Cloud mean cost: 0.07153 - Cloud variance: 0.04399\n",
      "Iteration: 00286 - Cloud mean cost: 0.07125 - Cloud variance: 0.04360\n",
      "Iteration: 00287 - Cloud mean cost: 0.07098 - Cloud variance: 0.04321\n",
      "Iteration: 00288 - Cloud mean cost: 0.07070 - Cloud variance: 0.04283\n",
      "Iteration: 00289 - Cloud mean cost: 0.07043 - Cloud variance: 0.04245\n",
      "Iteration: 00290 - Cloud mean cost: 0.07017 - Cloud variance: 0.04208\n",
      "Iteration: 00291 - Cloud mean cost: 0.06991 - Cloud variance: 0.04172\n",
      "Iteration: 00292 - Cloud mean cost: 0.06964 - Cloud variance: 0.04137\n",
      "Iteration: 00293 - Cloud mean cost: 0.06938 - Cloud variance: 0.04102\n",
      "Iteration: 00294 - Cloud mean cost: 0.06912 - Cloud variance: 0.04068\n",
      "Iteration: 00295 - Cloud mean cost: 0.06887 - Cloud variance: 0.04034\n",
      "Iteration: 00296 - Cloud mean cost: 0.06863 - Cloud variance: 0.04001\n",
      "Iteration: 00297 - Cloud mean cost: 0.06839 - Cloud variance: 0.03969\n",
      "Iteration: 00298 - Cloud mean cost: 0.06817 - Cloud variance: 0.03937\n",
      "Iteration: 00299 - Cloud mean cost: 0.06794 - Cloud variance: 0.03905\n",
      "Iteration: 00300 - Cloud mean cost: 0.06773 - Cloud variance: 0.03875\n",
      "Iteration: 00301 - Cloud mean cost: 0.06751 - Cloud variance: 0.03845\n",
      "Iteration: 00302 - Cloud mean cost: 0.06730 - Cloud variance: 0.03815\n",
      "Iteration: 00303 - Cloud mean cost: 0.06709 - Cloud variance: 0.03786\n",
      "Iteration: 00304 - Cloud mean cost: 0.06689 - Cloud variance: 0.03757\n",
      "Iteration: 00305 - Cloud mean cost: 0.06669 - Cloud variance: 0.03729\n",
      "Iteration: 00306 - Cloud mean cost: 0.06650 - Cloud variance: 0.03701\n",
      "Iteration: 00307 - Cloud mean cost: 0.06631 - Cloud variance: 0.03673\n",
      "Iteration: 00308 - Cloud mean cost: 0.06612 - Cloud variance: 0.03646\n",
      "Iteration: 00309 - Cloud mean cost: 0.06593 - Cloud variance: 0.03620\n",
      "Iteration: 00310 - Cloud mean cost: 0.06574 - Cloud variance: 0.03594\n",
      "Iteration: 00311 - Cloud mean cost: 0.06555 - Cloud variance: 0.03568\n",
      "Iteration: 00312 - Cloud mean cost: 0.06536 - Cloud variance: 0.03543\n",
      "Iteration: 00313 - Cloud mean cost: 0.06517 - Cloud variance: 0.03517\n",
      "Iteration: 00314 - Cloud mean cost: 0.06498 - Cloud variance: 0.03493\n",
      "Iteration: 00315 - Cloud mean cost: 0.06479 - Cloud variance: 0.03468\n",
      "Iteration: 00316 - Cloud mean cost: 0.06460 - Cloud variance: 0.03444\n",
      "Iteration: 00317 - Cloud mean cost: 0.06442 - Cloud variance: 0.03420\n",
      "Iteration: 00318 - Cloud mean cost: 0.06424 - Cloud variance: 0.03397\n",
      "Iteration: 00319 - Cloud mean cost: 0.06406 - Cloud variance: 0.03374\n",
      "Iteration: 00320 - Cloud mean cost: 0.06389 - Cloud variance: 0.03351\n",
      "Iteration: 00321 - Cloud mean cost: 0.06372 - Cloud variance: 0.03328\n",
      "Iteration: 00322 - Cloud mean cost: 0.06356 - Cloud variance: 0.03306\n",
      "Iteration: 00323 - Cloud mean cost: 0.06339 - Cloud variance: 0.03284\n",
      "Iteration: 00324 - Cloud mean cost: 0.06323 - Cloud variance: 0.03262\n",
      "Iteration: 00325 - Cloud mean cost: 0.06307 - Cloud variance: 0.03240\n",
      "Iteration: 00326 - Cloud mean cost: 0.06291 - Cloud variance: 0.03219\n",
      "Iteration: 00327 - Cloud mean cost: 0.06274 - Cloud variance: 0.03197\n",
      "Iteration: 00328 - Cloud mean cost: 0.06256 - Cloud variance: 0.03176\n",
      "Iteration: 00329 - Cloud mean cost: 0.06239 - Cloud variance: 0.03155\n",
      "Iteration: 00330 - Cloud mean cost: 0.06222 - Cloud variance: 0.03135\n",
      "Iteration: 00331 - Cloud mean cost: 0.06204 - Cloud variance: 0.03114\n",
      "Iteration: 00332 - Cloud mean cost: 0.06186 - Cloud variance: 0.03094\n",
      "Iteration: 00333 - Cloud mean cost: 0.06167 - Cloud variance: 0.03074\n",
      "Iteration: 00334 - Cloud mean cost: 0.06149 - Cloud variance: 0.03054\n",
      "Iteration: 00335 - Cloud mean cost: 0.06130 - Cloud variance: 0.03034\n",
      "Iteration: 00336 - Cloud mean cost: 0.06112 - Cloud variance: 0.03015\n",
      "Iteration: 00337 - Cloud mean cost: 0.06094 - Cloud variance: 0.02995\n",
      "Iteration: 00338 - Cloud mean cost: 0.06076 - Cloud variance: 0.02976\n",
      "Iteration: 00339 - Cloud mean cost: 0.06058 - Cloud variance: 0.02957\n",
      "Iteration: 00340 - Cloud mean cost: 0.06040 - Cloud variance: 0.02938\n",
      "Iteration: 00341 - Cloud mean cost: 0.06022 - Cloud variance: 0.02919\n",
      "Iteration: 00342 - Cloud mean cost: 0.06002 - Cloud variance: 0.02900\n",
      "Iteration: 00343 - Cloud mean cost: 0.05982 - Cloud variance: 0.02881\n",
      "Iteration: 00344 - Cloud mean cost: 0.05962 - Cloud variance: 0.02863\n",
      "Iteration: 00345 - Cloud mean cost: 0.05942 - Cloud variance: 0.02844\n",
      "Iteration: 00346 - Cloud mean cost: 0.05922 - Cloud variance: 0.02826\n",
      "Iteration: 00347 - Cloud mean cost: 0.05902 - Cloud variance: 0.02808\n",
      "Iteration: 00348 - Cloud mean cost: 0.05882 - Cloud variance: 0.02790\n",
      "Iteration: 00349 - Cloud mean cost: 0.05861 - Cloud variance: 0.02772\n",
      "Iteration: 00350 - Cloud mean cost: 0.05839 - Cloud variance: 0.02754\n",
      "Iteration: 00351 - Cloud mean cost: 0.05817 - Cloud variance: 0.02736\n",
      "Iteration: 00352 - Cloud mean cost: 0.05796 - Cloud variance: 0.02718\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iteration: 00353 - Cloud mean cost: 0.05775 - Cloud variance: 0.02701\n",
      "Iteration: 00354 - Cloud mean cost: 0.05753 - Cloud variance: 0.02683\n",
      "Iteration: 00355 - Cloud mean cost: 0.05731 - Cloud variance: 0.02666\n",
      "Iteration: 00356 - Cloud mean cost: 0.05709 - Cloud variance: 0.02648\n",
      "Iteration: 00357 - Cloud mean cost: 0.05687 - Cloud variance: 0.02631\n",
      "Iteration: 00358 - Cloud mean cost: 0.05665 - Cloud variance: 0.02614\n",
      "Iteration: 00359 - Cloud mean cost: 0.05642 - Cloud variance: 0.02596\n",
      "Iteration: 00360 - Cloud mean cost: 0.05620 - Cloud variance: 0.02579\n",
      "Iteration: 00361 - Cloud mean cost: 0.05597 - Cloud variance: 0.02562\n",
      "Iteration: 00362 - Cloud mean cost: 0.05574 - Cloud variance: 0.02545\n",
      "Iteration: 00363 - Cloud mean cost: 0.05551 - Cloud variance: 0.02528\n",
      "Iteration: 00364 - Cloud mean cost: 0.05528 - Cloud variance: 0.02512\n",
      "Iteration: 00365 - Cloud mean cost: 0.05506 - Cloud variance: 0.02495\n",
      "Iteration: 00366 - Cloud mean cost: 0.05484 - Cloud variance: 0.02478\n",
      "Iteration: 00367 - Cloud mean cost: 0.05462 - Cloud variance: 0.02462\n",
      "Iteration: 00368 - Cloud mean cost: 0.05442 - Cloud variance: 0.02445\n",
      "Iteration: 00369 - Cloud mean cost: 0.05421 - Cloud variance: 0.02429\n",
      "Iteration: 00370 - Cloud mean cost: 0.05401 - Cloud variance: 0.02412\n",
      "Iteration: 00371 - Cloud mean cost: 0.05382 - Cloud variance: 0.02396\n",
      "Iteration: 00372 - Cloud mean cost: 0.05363 - Cloud variance: 0.02380\n",
      "Iteration: 00373 - Cloud mean cost: 0.05343 - Cloud variance: 0.02364\n",
      "Iteration: 00374 - Cloud mean cost: 0.05324 - Cloud variance: 0.02348\n",
      "Iteration: 00375 - Cloud mean cost: 0.05305 - Cloud variance: 0.02333\n",
      "Iteration: 00376 - Cloud mean cost: 0.05286 - Cloud variance: 0.02317\n",
      "Iteration: 00377 - Cloud mean cost: 0.05267 - Cloud variance: 0.02301\n",
      "Iteration: 00378 - Cloud mean cost: 0.05249 - Cloud variance: 0.02286\n",
      "Iteration: 00379 - Cloud mean cost: 0.05231 - Cloud variance: 0.02271\n",
      "Iteration: 00380 - Cloud mean cost: 0.05212 - Cloud variance: 0.02255\n",
      "Iteration: 00381 - Cloud mean cost: 0.05194 - Cloud variance: 0.02240\n",
      "Iteration: 00382 - Cloud mean cost: 0.05176 - Cloud variance: 0.02225\n",
      "Iteration: 00383 - Cloud mean cost: 0.05159 - Cloud variance: 0.02210\n",
      "Iteration: 00384 - Cloud mean cost: 0.05141 - Cloud variance: 0.02196\n",
      "Iteration: 00385 - Cloud mean cost: 0.05124 - Cloud variance: 0.02181\n",
      "Iteration: 00386 - Cloud mean cost: 0.05107 - Cloud variance: 0.02167\n",
      "Iteration: 00387 - Cloud mean cost: 0.05090 - Cloud variance: 0.02152\n",
      "Iteration: 00388 - Cloud mean cost: 0.05074 - Cloud variance: 0.02138\n",
      "Iteration: 00389 - Cloud mean cost: 0.05058 - Cloud variance: 0.02124\n",
      "Iteration: 00390 - Cloud mean cost: 0.05041 - Cloud variance: 0.02110\n",
      "Iteration: 00391 - Cloud mean cost: 0.05025 - Cloud variance: 0.02097\n",
      "Iteration: 00392 - Cloud mean cost: 0.05009 - Cloud variance: 0.02083\n",
      "Iteration: 00393 - Cloud mean cost: 0.04994 - Cloud variance: 0.02070\n",
      "Iteration: 00394 - Cloud mean cost: 0.04978 - Cloud variance: 0.02057\n",
      "Iteration: 00395 - Cloud mean cost: 0.04963 - Cloud variance: 0.02044\n",
      "Iteration: 00396 - Cloud mean cost: 0.04948 - Cloud variance: 0.02031\n",
      "Iteration: 00397 - Cloud mean cost: 0.04934 - Cloud variance: 0.02018\n",
      "Iteration: 00398 - Cloud mean cost: 0.04920 - Cloud variance: 0.02005\n",
      "Iteration: 00399 - Cloud mean cost: 0.04907 - Cloud variance: 0.01993\n",
      "Iteration: 00400 - Cloud mean cost: 0.04893 - Cloud variance: 0.01980\n",
      "Iteration: 00401 - Cloud mean cost: 0.04879 - Cloud variance: 0.01968\n",
      "Iteration: 00402 - Cloud mean cost: 0.04865 - Cloud variance: 0.01956\n",
      "Iteration: 00403 - Cloud mean cost: 0.04852 - Cloud variance: 0.01944\n",
      "Iteration: 00404 - Cloud mean cost: 0.04839 - Cloud variance: 0.01933\n",
      "Iteration: 00405 - Cloud mean cost: 0.04826 - Cloud variance: 0.01921\n",
      "Iteration: 00406 - Cloud mean cost: 0.04813 - Cloud variance: 0.01910\n",
      "Iteration: 00407 - Cloud mean cost: 0.04800 - Cloud variance: 0.01898\n",
      "Iteration: 00408 - Cloud mean cost: 0.04787 - Cloud variance: 0.01887\n",
      "Iteration: 00409 - Cloud mean cost: 0.04774 - Cloud variance: 0.01876\n",
      "Iteration: 00410 - Cloud mean cost: 0.04761 - Cloud variance: 0.01865\n",
      "Iteration: 00411 - Cloud mean cost: 0.04747 - Cloud variance: 0.01854\n",
      "Iteration: 00412 - Cloud mean cost: 0.04734 - Cloud variance: 0.01844\n",
      "Iteration: 00413 - Cloud mean cost: 0.04721 - Cloud variance: 0.01833\n",
      "Iteration: 00414 - Cloud mean cost: 0.04708 - Cloud variance: 0.01823\n",
      "Iteration: 00415 - Cloud mean cost: 0.04695 - Cloud variance: 0.01812\n",
      "Iteration: 00416 - Cloud mean cost: 0.04683 - Cloud variance: 0.01802\n",
      "Iteration: 00417 - Cloud mean cost: 0.04670 - Cloud variance: 0.01792\n",
      "Iteration: 00418 - Cloud mean cost: 0.04658 - Cloud variance: 0.01782\n",
      "Iteration: 00419 - Cloud mean cost: 0.04646 - Cloud variance: 0.01773\n",
      "Iteration: 00420 - Cloud mean cost: 0.04634 - Cloud variance: 0.01763\n",
      "Iteration: 00421 - Cloud mean cost: 0.04622 - Cloud variance: 0.01753\n",
      "Iteration: 00422 - Cloud mean cost: 0.04610 - Cloud variance: 0.01744\n",
      "Iteration: 00423 - Cloud mean cost: 0.04598 - Cloud variance: 0.01734\n",
      "Iteration: 00424 - Cloud mean cost: 0.04586 - Cloud variance: 0.01725\n",
      "Iteration: 00425 - Cloud mean cost: 0.04575 - Cloud variance: 0.01716\n",
      "Iteration: 00426 - Cloud mean cost: 0.04563 - Cloud variance: 0.01707\n",
      "Iteration: 00427 - Cloud mean cost: 0.04552 - Cloud variance: 0.01698\n",
      "Iteration: 00428 - Cloud mean cost: 0.04540 - Cloud variance: 0.01689\n",
      "Iteration: 00429 - Cloud mean cost: 0.04529 - Cloud variance: 0.01680\n",
      "Iteration: 00430 - Cloud mean cost: 0.04518 - Cloud variance: 0.01672\n",
      "Iteration: 00431 - Cloud mean cost: 0.04507 - Cloud variance: 0.01663\n",
      "Iteration: 00432 - Cloud mean cost: 0.04496 - Cloud variance: 0.01654\n",
      "Iteration: 00433 - Cloud mean cost: 0.04485 - Cloud variance: 0.01646\n",
      "Iteration: 00434 - Cloud mean cost: 0.04475 - Cloud variance: 0.01638\n",
      "Iteration: 00435 - Cloud mean cost: 0.04464 - Cloud variance: 0.01629\n",
      "Iteration: 00436 - Cloud mean cost: 0.04454 - Cloud variance: 0.01621\n",
      "Iteration: 00437 - Cloud mean cost: 0.04444 - Cloud variance: 0.01613\n",
      "Iteration: 00438 - Cloud mean cost: 0.04434 - Cloud variance: 0.01605\n",
      "Iteration: 00439 - Cloud mean cost: 0.04424 - Cloud variance: 0.01597\n",
      "Iteration: 00440 - Cloud mean cost: 0.04414 - Cloud variance: 0.01589\n",
      "Iteration: 00441 - Cloud mean cost: 0.04404 - Cloud variance: 0.01581\n",
      "Iteration: 00442 - Cloud mean cost: 0.04394 - Cloud variance: 0.01574\n",
      "Iteration: 00443 - Cloud mean cost: 0.04384 - Cloud variance: 0.01566\n",
      "Iteration: 00444 - Cloud mean cost: 0.04374 - Cloud variance: 0.01559\n",
      "Iteration: 00445 - Cloud mean cost: 0.04363 - Cloud variance: 0.01551\n",
      "Iteration: 00446 - Cloud mean cost: 0.04353 - Cloud variance: 0.01544\n",
      "Iteration: 00447 - Cloud mean cost: 0.04342 - Cloud variance: 0.01536\n",
      "Iteration: 00448 - Cloud mean cost: 0.04332 - Cloud variance: 0.01529\n",
      "Iteration: 00449 - Cloud mean cost: 0.04322 - Cloud variance: 0.01522\n",
      "Iteration: 00450 - Cloud mean cost: 0.04313 - Cloud variance: 0.01515\n",
      "Iteration: 00451 - Cloud mean cost: 0.04303 - Cloud variance: 0.01508\n",
      "Iteration: 00452 - Cloud mean cost: 0.04294 - Cloud variance: 0.01500\n",
      "Iteration: 00453 - Cloud mean cost: 0.04284 - Cloud variance: 0.01494\n",
      "Iteration: 00454 - Cloud mean cost: 0.04275 - Cloud variance: 0.01487\n",
      "Iteration: 00455 - Cloud mean cost: 0.04265 - Cloud variance: 0.01480\n",
      "Iteration: 00456 - Cloud mean cost: 0.04256 - Cloud variance: 0.01473\n",
      "Iteration: 00457 - Cloud mean cost: 0.04247 - Cloud variance: 0.01466\n",
      "Iteration: 00458 - Cloud mean cost: 0.04239 - Cloud variance: 0.01460\n",
      "Iteration: 00459 - Cloud mean cost: 0.04230 - Cloud variance: 0.01453\n",
      "Iteration: 00460 - Cloud mean cost: 0.04222 - Cloud variance: 0.01447\n",
      "Iteration: 00461 - Cloud mean cost: 0.04214 - Cloud variance: 0.01440\n",
      "Iteration: 00462 - Cloud mean cost: 0.04206 - Cloud variance: 0.01434\n",
      "Iteration: 00463 - Cloud mean cost: 0.04199 - Cloud variance: 0.01428\n",
      "Iteration: 00464 - Cloud mean cost: 0.04191 - Cloud variance: 0.01422\n",
      "Iteration: 00465 - Cloud mean cost: 0.04184 - Cloud variance: 0.01415\n",
      "Iteration: 00466 - Cloud mean cost: 0.04177 - Cloud variance: 0.01409\n",
      "Iteration: 00467 - Cloud mean cost: 0.04170 - Cloud variance: 0.01403\n",
      "Iteration: 00468 - Cloud mean cost: 0.04163 - Cloud variance: 0.01398\n",
      "Iteration: 00469 - Cloud mean cost: 0.04157 - Cloud variance: 0.01392\n",
      "Iteration: 00470 - Cloud mean cost: 0.04151 - Cloud variance: 0.01386\n",
      "Iteration: 00471 - Cloud mean cost: 0.04143 - Cloud variance: 0.01380\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iteration: 00472 - Cloud mean cost: 0.04137 - Cloud variance: 0.01375\n",
      "Iteration: 00473 - Cloud mean cost: 0.04130 - Cloud variance: 0.01369\n",
      "Iteration: 00474 - Cloud mean cost: 0.04124 - Cloud variance: 0.01363\n",
      "Iteration: 00475 - Cloud mean cost: 0.04118 - Cloud variance: 0.01358\n",
      "Iteration: 00476 - Cloud mean cost: 0.04113 - Cloud variance: 0.01353\n",
      "Iteration: 00477 - Cloud mean cost: 0.04108 - Cloud variance: 0.01347\n",
      "Iteration: 00478 - Cloud mean cost: 0.04104 - Cloud variance: 0.01342\n",
      "Iteration: 00479 - Cloud mean cost: 0.04099 - Cloud variance: 0.01337\n",
      "Iteration: 00480 - Cloud mean cost: 0.04095 - Cloud variance: 0.01332\n",
      "Iteration: 00481 - Cloud mean cost: 0.04091 - Cloud variance: 0.01327\n",
      "Iteration: 00482 - Cloud mean cost: 0.04086 - Cloud variance: 0.01322\n",
      "Iteration: 00483 - Cloud mean cost: 0.04082 - Cloud variance: 0.01317\n",
      "Iteration: 00484 - Cloud mean cost: 0.04078 - Cloud variance: 0.01312\n",
      "Iteration: 00485 - Cloud mean cost: 0.04074 - Cloud variance: 0.01308\n",
      "Iteration: 00486 - Cloud mean cost: 0.04070 - Cloud variance: 0.01303\n",
      "Iteration: 00487 - Cloud mean cost: 0.04064 - Cloud variance: 0.01299\n",
      "Iteration: 00488 - Cloud mean cost: 0.04058 - Cloud variance: 0.01294\n",
      "Iteration: 00489 - Cloud mean cost: 0.04052 - Cloud variance: 0.01290\n",
      "Iteration: 00490 - Cloud mean cost: 0.04045 - Cloud variance: 0.01286\n",
      "Iteration: 00491 - Cloud mean cost: 0.04037 - Cloud variance: 0.01281\n",
      "Iteration: 00492 - Cloud mean cost: 0.04032 - Cloud variance: 0.01277\n",
      "Iteration: 00493 - Cloud mean cost: 0.04028 - Cloud variance: 0.01273\n",
      "Iteration: 00494 - Cloud mean cost: 0.04025 - Cloud variance: 0.01269\n",
      "Iteration: 00495 - Cloud mean cost: 0.04022 - Cloud variance: 0.01265\n",
      "Iteration: 00496 - Cloud mean cost: 0.04020 - Cloud variance: 0.01261\n",
      "Iteration: 00497 - Cloud mean cost: 0.04018 - Cloud variance: 0.01257\n",
      "Iteration: 00498 - Cloud mean cost: 0.04017 - Cloud variance: 0.01254\n",
      "Iteration: 00499 - Cloud mean cost: 0.04015 - Cloud variance: 0.01250\n",
      "Iteration: 00500 - Cloud mean cost: 0.04014 - Cloud variance: 0.01246\n",
      "Iteration: 00501 - Cloud mean cost: 0.04013 - Cloud variance: 0.01243\n",
      "Iteration: 00502 - Cloud mean cost: 0.04010 - Cloud variance: 0.01239\n",
      "Iteration: 00503 - Cloud mean cost: 0.04007 - Cloud variance: 0.01236\n",
      "Iteration: 00504 - Cloud mean cost: 0.04003 - Cloud variance: 0.01232\n",
      "Iteration: 00505 - Cloud mean cost: 0.04000 - Cloud variance: 0.01229\n",
      "Iteration: 00506 - Cloud mean cost: 0.03997 - Cloud variance: 0.01225\n",
      "Iteration: 00507 - Cloud mean cost: 0.03994 - Cloud variance: 0.01222\n",
      "Iteration: 00508 - Cloud mean cost: 0.03990 - Cloud variance: 0.01219\n",
      "Iteration: 00509 - Cloud mean cost: 0.03987 - Cloud variance: 0.01216\n",
      "Iteration: 00510 - Cloud mean cost: 0.03984 - Cloud variance: 0.01212\n",
      "Iteration: 00511 - Cloud mean cost: 0.03981 - Cloud variance: 0.01209\n",
      "Iteration: 00512 - Cloud mean cost: 0.03978 - Cloud variance: 0.01206\n",
      "Iteration: 00513 - Cloud mean cost: 0.03976 - Cloud variance: 0.01203\n",
      "Iteration: 00514 - Cloud mean cost: 0.03973 - Cloud variance: 0.01200\n",
      "Iteration: 00515 - Cloud mean cost: 0.03970 - Cloud variance: 0.01197\n",
      "Iteration: 00516 - Cloud mean cost: 0.03968 - Cloud variance: 0.01195\n",
      "Iteration: 00517 - Cloud mean cost: 0.03967 - Cloud variance: 0.01192\n",
      "Iteration: 00518 - Cloud mean cost: 0.03966 - Cloud variance: 0.01189\n",
      "Iteration: 00519 - Cloud mean cost: 0.03964 - Cloud variance: 0.01186\n",
      "Iteration: 00520 - Cloud mean cost: 0.03963 - Cloud variance: 0.01184\n",
      "Iteration: 00521 - Cloud mean cost: 0.03962 - Cloud variance: 0.01181\n",
      "Iteration: 00522 - Cloud mean cost: 0.03961 - Cloud variance: 0.01178\n",
      "Iteration: 00523 - Cloud mean cost: 0.03960 - Cloud variance: 0.01176\n",
      "Iteration: 00524 - Cloud mean cost: 0.03960 - Cloud variance: 0.01173\n",
      "Iteration: 00525 - Cloud mean cost: 0.03960 - Cloud variance: 0.01171\n",
      "Iteration: 00526 - Cloud mean cost: 0.03959 - Cloud variance: 0.01168\n",
      "Iteration: 00527 - Cloud mean cost: 0.03959 - Cloud variance: 0.01166\n",
      "Iteration: 00528 - Cloud mean cost: 0.03957 - Cloud variance: 0.01164\n",
      "Iteration: 00529 - Cloud mean cost: 0.03957 - Cloud variance: 0.01161\n",
      "Iteration: 00530 - Cloud mean cost: 0.03956 - Cloud variance: 0.01159\n",
      "Iteration: 00531 - Cloud mean cost: 0.03957 - Cloud variance: 0.01157\n",
      "Iteration: 00532 - Cloud mean cost: 0.03959 - Cloud variance: 0.01154\n",
      "Iteration: 00533 - Cloud mean cost: 0.03961 - Cloud variance: 0.01152\n",
      "Iteration: 00534 - Cloud mean cost: 0.03964 - Cloud variance: 0.01150\n",
      "Iteration: 00535 - Cloud mean cost: 0.03966 - Cloud variance: 0.01147\n",
      "Iteration: 00536 - Cloud mean cost: 0.03968 - Cloud variance: 0.01145\n",
      "Iteration: 00537 - Cloud mean cost: 0.03969 - Cloud variance: 0.01143\n",
      "Iteration: 00538 - Cloud mean cost: 0.03970 - Cloud variance: 0.01140\n",
      "Iteration: 00539 - Cloud mean cost: 0.03973 - Cloud variance: 0.01138\n",
      "Iteration: 00540 - Cloud mean cost: 0.03975 - Cloud variance: 0.01136\n",
      "Iteration: 00541 - Cloud mean cost: 0.03977 - Cloud variance: 0.01134\n",
      "Iteration: 00542 - Cloud mean cost: 0.03979 - Cloud variance: 0.01131\n",
      "Iteration: 00543 - Cloud mean cost: 0.03982 - Cloud variance: 0.01129\n",
      "Iteration: 00544 - Cloud mean cost: 0.03983 - Cloud variance: 0.01127\n",
      "Iteration: 00545 - Cloud mean cost: 0.03984 - Cloud variance: 0.01124\n",
      "Iteration: 00546 - Cloud mean cost: 0.03986 - Cloud variance: 0.01122\n",
      "Iteration: 00547 - Cloud mean cost: 0.03985 - Cloud variance: 0.01120\n",
      "Iteration: 00548 - Cloud mean cost: 0.03984 - Cloud variance: 0.01118\n",
      "Iteration: 00549 - Cloud mean cost: 0.03982 - Cloud variance: 0.01115\n",
      "Iteration: 00550 - Cloud mean cost: 0.03981 - Cloud variance: 0.01113\n",
      "Iteration: 00551 - Cloud mean cost: 0.03980 - Cloud variance: 0.01111\n",
      "Iteration: 00552 - Cloud mean cost: 0.03979 - Cloud variance: 0.01109\n",
      "Iteration: 00553 - Cloud mean cost: 0.03980 - Cloud variance: 0.01106\n",
      "Iteration: 00554 - Cloud mean cost: 0.03980 - Cloud variance: 0.01104\n",
      "Iteration: 00555 - Cloud mean cost: 0.03979 - Cloud variance: 0.01102\n",
      "Iteration: 00556 - Cloud mean cost: 0.03979 - Cloud variance: 0.01099\n",
      "Iteration: 00557 - Cloud mean cost: 0.03978 - Cloud variance: 0.01097\n",
      "Iteration: 00558 - Cloud mean cost: 0.03977 - Cloud variance: 0.01095\n",
      "Iteration: 00559 - Cloud mean cost: 0.03978 - Cloud variance: 0.01092\n",
      "Iteration: 00560 - Cloud mean cost: 0.03978 - Cloud variance: 0.01090\n",
      "Iteration: 00561 - Cloud mean cost: 0.03978 - Cloud variance: 0.01088\n",
      "Iteration: 00562 - Cloud mean cost: 0.03977 - Cloud variance: 0.01085\n",
      "Iteration: 00563 - Cloud mean cost: 0.03977 - Cloud variance: 0.01083\n",
      "Iteration: 00564 - Cloud mean cost: 0.03976 - Cloud variance: 0.01081\n",
      "Iteration: 00565 - Cloud mean cost: 0.03974 - Cloud variance: 0.01078\n",
      "Iteration: 00566 - Cloud mean cost: 0.03973 - Cloud variance: 0.01076\n",
      "Iteration: 00567 - Cloud mean cost: 0.03970 - Cloud variance: 0.01074\n",
      "Iteration: 00568 - Cloud mean cost: 0.03968 - Cloud variance: 0.01071\n",
      "Iteration: 00569 - Cloud mean cost: 0.03964 - Cloud variance: 0.01069\n",
      "Iteration: 00570 - Cloud mean cost: 0.03960 - Cloud variance: 0.01067\n",
      "Iteration: 00571 - Cloud mean cost: 0.03957 - Cloud variance: 0.01064\n",
      "Iteration: 00572 - Cloud mean cost: 0.03953 - Cloud variance: 0.01062\n",
      "Iteration: 00573 - Cloud mean cost: 0.03949 - Cloud variance: 0.01060\n",
      "Iteration: 00574 - Cloud mean cost: 0.03946 - Cloud variance: 0.01057\n",
      "Iteration: 00575 - Cloud mean cost: 0.03942 - Cloud variance: 0.01055\n",
      "Iteration: 00576 - Cloud mean cost: 0.03938 - Cloud variance: 0.01053\n",
      "Iteration: 00577 - Cloud mean cost: 0.03934 - Cloud variance: 0.01050\n",
      "Iteration: 00578 - Cloud mean cost: 0.03930 - Cloud variance: 0.01048\n",
      "Iteration: 00579 - Cloud mean cost: 0.03927 - Cloud variance: 0.01046\n",
      "Iteration: 00580 - Cloud mean cost: 0.03924 - Cloud variance: 0.01043\n",
      "Iteration: 00581 - Cloud mean cost: 0.03920 - Cloud variance: 0.01041\n",
      "Iteration: 00582 - Cloud mean cost: 0.03917 - Cloud variance: 0.01038\n",
      "Iteration: 00583 - Cloud mean cost: 0.03914 - Cloud variance: 0.01036\n",
      "Iteration: 00584 - Cloud mean cost: 0.03910 - Cloud variance: 0.01034\n",
      "Iteration: 00585 - Cloud mean cost: 0.03907 - Cloud variance: 0.01031\n",
      "Iteration: 00586 - Cloud mean cost: 0.03906 - Cloud variance: 0.01029\n",
      "Iteration: 00587 - Cloud mean cost: 0.03904 - Cloud variance: 0.01027\n",
      "Iteration: 00588 - Cloud mean cost: 0.03902 - Cloud variance: 0.01024\n",
      "Iteration: 00589 - Cloud mean cost: 0.03899 - Cloud variance: 0.01022\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iteration: 00590 - Cloud mean cost: 0.03898 - Cloud variance: 0.01020\n",
      "Iteration: 00591 - Cloud mean cost: 0.03897 - Cloud variance: 0.01017\n",
      "Iteration: 00592 - Cloud mean cost: 0.03895 - Cloud variance: 0.01015\n",
      "Iteration: 00593 - Cloud mean cost: 0.03894 - Cloud variance: 0.01013\n",
      "Iteration: 00594 - Cloud mean cost: 0.03893 - Cloud variance: 0.01010\n",
      "Iteration: 00595 - Cloud mean cost: 0.03892 - Cloud variance: 0.01008\n",
      "Iteration: 00596 - Cloud mean cost: 0.03891 - Cloud variance: 0.01005\n",
      "Iteration: 00597 - Cloud mean cost: 0.03890 - Cloud variance: 0.01003\n",
      "Iteration: 00598 - Cloud mean cost: 0.03888 - Cloud variance: 0.01000\n",
      "Iteration: 00599 - Cloud mean cost: 0.03886 - Cloud variance: 0.00998\n",
      "Iteration: 00600 - Cloud mean cost: 0.03885 - Cloud variance: 0.00995\n",
      "Iteration: 00601 - Cloud mean cost: 0.03883 - Cloud variance: 0.00993\n",
      "Iteration: 00602 - Cloud mean cost: 0.03881 - Cloud variance: 0.00990\n",
      "Iteration: 00603 - Cloud mean cost: 0.03880 - Cloud variance: 0.00987\n",
      "Iteration: 00604 - Cloud mean cost: 0.03879 - Cloud variance: 0.00985\n",
      "Iteration: 00605 - Cloud mean cost: 0.03877 - Cloud variance: 0.00982\n",
      "Iteration: 00606 - Cloud mean cost: 0.03876 - Cloud variance: 0.00980\n",
      "Iteration: 00607 - Cloud mean cost: 0.03875 - Cloud variance: 0.00977\n",
      "Iteration: 00608 - Cloud mean cost: 0.03874 - Cloud variance: 0.00975\n",
      "Iteration: 00609 - Cloud mean cost: 0.03872 - Cloud variance: 0.00972\n",
      "Iteration: 00610 - Cloud mean cost: 0.03871 - Cloud variance: 0.00970\n",
      "Iteration: 00611 - Cloud mean cost: 0.03869 - Cloud variance: 0.00967\n",
      "Iteration: 00612 - Cloud mean cost: 0.03868 - Cloud variance: 0.00964\n",
      "Iteration: 00613 - Cloud mean cost: 0.03867 - Cloud variance: 0.00962\n",
      "Iteration: 00614 - Cloud mean cost: 0.03865 - Cloud variance: 0.00959\n",
      "Iteration: 00615 - Cloud mean cost: 0.03864 - Cloud variance: 0.00957\n",
      "Iteration: 00616 - Cloud mean cost: 0.03862 - Cloud variance: 0.00954\n",
      "Iteration: 00617 - Cloud mean cost: 0.03859 - Cloud variance: 0.00952\n",
      "Iteration: 00618 - Cloud mean cost: 0.03856 - Cloud variance: 0.00949\n",
      "Iteration: 00619 - Cloud mean cost: 0.03852 - Cloud variance: 0.00947\n",
      "Iteration: 00620 - Cloud mean cost: 0.03848 - Cloud variance: 0.00944\n",
      "Iteration: 00621 - Cloud mean cost: 0.03845 - Cloud variance: 0.00941\n",
      "Iteration: 00622 - Cloud mean cost: 0.03841 - Cloud variance: 0.00939\n",
      "Iteration: 00623 - Cloud mean cost: 0.03837 - Cloud variance: 0.00936\n",
      "Iteration: 00624 - Cloud mean cost: 0.03833 - Cloud variance: 0.00933\n",
      "Iteration: 00625 - Cloud mean cost: 0.03829 - Cloud variance: 0.00931\n",
      "Iteration: 00626 - Cloud mean cost: 0.03825 - Cloud variance: 0.00928\n",
      "Iteration: 00627 - Cloud mean cost: 0.03821 - Cloud variance: 0.00925\n",
      "Iteration: 00628 - Cloud mean cost: 0.03817 - Cloud variance: 0.00922\n",
      "Iteration: 00629 - Cloud mean cost: 0.03813 - Cloud variance: 0.00920\n",
      "Iteration: 00630 - Cloud mean cost: 0.03809 - Cloud variance: 0.00917\n",
      "Iteration: 00631 - Cloud mean cost: 0.03805 - Cloud variance: 0.00914\n",
      "Iteration: 00632 - Cloud mean cost: 0.03800 - Cloud variance: 0.00911\n",
      "Iteration: 00633 - Cloud mean cost: 0.03796 - Cloud variance: 0.00908\n",
      "Iteration: 00634 - Cloud mean cost: 0.03793 - Cloud variance: 0.00905\n",
      "Iteration: 00635 - Cloud mean cost: 0.03789 - Cloud variance: 0.00902\n",
      "Iteration: 00636 - Cloud mean cost: 0.03785 - Cloud variance: 0.00900\n",
      "Iteration: 00637 - Cloud mean cost: 0.03782 - Cloud variance: 0.00897\n",
      "Iteration: 00638 - Cloud mean cost: 0.03777 - Cloud variance: 0.00894\n",
      "Iteration: 00639 - Cloud mean cost: 0.03771 - Cloud variance: 0.00891\n",
      "Iteration: 00640 - Cloud mean cost: 0.03763 - Cloud variance: 0.00888\n",
      "Iteration: 00641 - Cloud mean cost: 0.03756 - Cloud variance: 0.00885\n",
      "Iteration: 00642 - Cloud mean cost: 0.03750 - Cloud variance: 0.00883\n",
      "Iteration: 00643 - Cloud mean cost: 0.03743 - Cloud variance: 0.00880\n",
      "Iteration: 00644 - Cloud mean cost: 0.03736 - Cloud variance: 0.00877\n",
      "Iteration: 00645 - Cloud mean cost: 0.03730 - Cloud variance: 0.00874\n",
      "Iteration: 00646 - Cloud mean cost: 0.03723 - Cloud variance: 0.00871\n",
      "Iteration: 00647 - Cloud mean cost: 0.03716 - Cloud variance: 0.00868\n"
     ]
    }
   ],
   "source": [
    "nn_architecture = [\n",
    "    {\"input_dim\": 2, \"output_dim\": 5, \"activation\": \"relu\"},\n",
    "    {\"input_dim\": 5, \"output_dim\": 10, \"activation\": \"relu\"},\n",
    "    {\"input_dim\": 10, \"output_dim\": 5, \"activation\": \"relu\"},\n",
    "    {\"input_dim\": 5, \"output_dim\": 1, \"activation\": \"sigmoid\"},\n",
    "]\n",
    "\n",
    "#CREATE NEURAL NETWORK\n",
    "nn = neuralnet(nn_architecture)\n",
    "\n",
    "#INITIALIZE CLOUD WITH 150 PARTICLES ACCORDING TO XAVIER INITIALIZATION\n",
    "nn.init_cloud(N = 150)\n",
    "\n",
    "#Start training using default parameters\n",
    "nn.train(X_train, Y_train, {'cost_type': 'rmse',\n",
    "                            'algorithm': 'euler_adaptive',\n",
    "                            'max_epochs': 1000,\n",
    "                            'kernel_a': 0.01,\n",
    "                            'lerning_rate': 0.1,\n",
    "                            'alpha_rate': 1,\n",
    "                            'verbose': True})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test set accuracy using neuralnet 0: 0.97000\n",
      "Test set accuracy using best neuralnet: 0.97250\n"
     ]
    }
   ],
   "source": [
    "#PLOT COST HISTORY\n",
    "nn.plot_cost_history()\n",
    "\n",
    "#ACCURACY ON TEST SET USING PARTICLE 0 WEIGHTS\n",
    "nn.prediction_accuracy_particle(X_test, Y_test, acc_type = \"sigmoid_accuracy\", particle = 0)\n",
    "\n",
    "#ACCURACY ON TEST SET USING CLOUD MEAN WEIGHTS\n",
    "nn.prediction_accuracy_best_particle(X_test, Y_test, acc_type = \"sigmoid_accuracy\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x22a80358ac8>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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",
    "from mlswarm.utils import convert_prob_into_class\n",
    "%matplotlib inline \n",
    "\n",
    "#CLASSIFICATION VISUALIZED\n",
    "Y_test_hat = nn.forward_propagation(X_test, cloud = nn.cloud_mean)\n",
    "plt.scatter(X_test[:,0],X_test[:,1], c = convert_prob_into_class(Y_test_hat).flatten())"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}