MLP_Models_Comparision.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.model_selection import StratifiedShuffleSplit, cross_val_predict\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.metrics import accuracy_score, confusion_matrix\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, LeakyReLU"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## A) Loading the Dataset & Dividing into training & testing sets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset = pd.read_csv(\"powerline_dataset.csv\")\n",
    "dataset = dataset[['Output (S)', 'Ia', 'Ib', 'Ic', 'Va', 'Vb', 'Vc']] # Since there were 2 additional noisy columns in csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "split = StratifiedShuffleSplit(n_splits=1, test_size=0.1, random_state=42)\n",
    "\n",
    "train_index, test_index = next(split.split(dataset, dataset['Output (S)']))\n",
    "train_set, test_set = dataset.iloc[train_index], dataset.iloc[test_index]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## B) Training & Evaluating SVM, Decision Tree & Random Forest classifiers"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### (i) SVM:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Normalizing dataset before sending into SVM\n",
    "\n",
    "scaler = StandardScaler()\n",
    "scaler.fit(dataset.iloc[:, 1:])\n",
    "train, test = train_set.to_numpy(), test_set.to_numpy()\n",
    "train[:, 1:], test[:, 1:] = scaler.transform(train_set.iloc[:, 1:]), scaler.transform(test_set.iloc[:, 1:])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SVC(C=1000, random_state=42)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svm_clf = SVC(C=1000, kernel=\"rbf\", random_state=42)\n",
    "svm_clf.fit(train[:, 1:], train[:, 0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAATwAAAEWCAYAAAD7MitWAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAAAlAElEQVR4nO3debxXVb3/8df7gAIOkKAQAiomaQ45m0OZQ6ZeLdTyhtNFs4sVaqVelfTnVJR6s2tm3JtdU3LGKUnNIa5oDglIOOGEM4KiiDOiwOf3x14HvxzP+Z7v+XL2+U7vp4/9+O7v2muvtQ5HPqy9195rKSIwM2sETZVugJlZV3HAM7OG4YBnZg3DAc/MGoYDnpk1DAc8M2sYDnh1TlIvSX+R9Laka1egnEMk3dGZbasESX+VNLLS7bDKcMCrEpIOljRN0nuS5qa/mF/uhKK/DQwA+kXEgeUWEhFXRMTXO6E9y5G0i6SQdEOL9M1T+uQSyzlD0uXt5YuIvSNifJnNtRrngFcFJB0HnA/8giw4rQOMA4Z3QvHrAk9HxOJOKCsvrwM7SupXkDYSeLqzKlDG/783uojwVsEN6AO8BxxYJE8PsoA4J23nAz3SsV2A2cDxwDxgLnBEOnYm8BHwcarjSOAM4PKCstcDAuievh8OPAe8CzwPHFKQfm/BeTsCU4G30+eOBccmAz8D7kvl3AGs2cbP1tz+/wFGp7RuKe00YHJB3t8ALwPvAA8BX0npe7X4OR8uaMfY1I6FwAYp7Xvp+H8D1xWUfw4wCVCl/7/wls/mf/EqbwegJ3BjkTynANsDWwCbA9sBpxYc/yxZ4BxEFtR+J2mNiDidrNd4TUSsFhEXF2uIpFWBC4C9I2J1sqA2o5V8fYFbUt5+wK+BW1r00A4GjgD6AysDJxSrG/gT8G9pf0/gcbLgXmgq2Z9BX+BK4FpJPSPithY/5+YF5xwGjAJWB15sUd7xwBclHS7pK2R/diMjRT+rPw54ldcPeCOKX3IeApwVEfMi4nWyntthBcc/Tsc/johbyXo5G5bZnqXAppJ6RcTciHi8lTz7AM9ExGURsTgirgKeBL5RkOeSiHg6IhYCE8gCVZsi4n6gr6QNyQLfn1rJc3lEzE91nkfW823v57w0Ih5P53zcorwPgEPJAvblwDERMbud8qyGOeBV3nxgTUndi+RZm+V7Jy+mtGVltAiYHwCrdbQhEfE+8B3g+8BcSbdI2qiE9jS3aVDB91fLaM9lwNHArrTS45V0vKQn0ojzW2S92jXbKfPlYgcjYgrZJbzIArPVMQe8ynsA+BDYr0ieOWSDD83W4dOXe6V6H1il4PtnCw9GxO0RsQcwkKzX9ocS2tPcplfKbFOzy4AfArem3tcy6ZLzJOBfgTUi4jNk9w/V3PQ2yix6eSppNFlPcQ5wYtktt5rggFdhEfE22c3530naT9IqklaStLekc1O2q4BTJa0lac2Uv91HMNowA9hZ0jqS+gBjmg9IGiDpm+le3iKyS+MlrZRxK/D59ChNd0nfATYGbi6zTQBExPPAV8nuWba0OrCYbES3u6TTgN4Fx18D1uvISKykzwM/J7usPQw4UdIW5bXeaoEDXhWIiF8Dx5ENRLxOdhl2NPDnlOXnwDTgEeBRYHpKK6euO4FrUlkPsXyQaiK7kT8HeJMs+PywlTLmA/umvPPJekb7RsQb5bSpRdn3RkRrvdfbgb+SParyIlmvuPBytfmh6vmSprdXT7qFcDlwTkQ8HBHPAD8FLpPUY0V+Bqte8oCUmTUK9/DMrGE44JlZw3DAM7OG4YBnZg2j2MOuFdVry6M9mlJDFky9sNJNsDL07L7sOcaydOTv6cJ/XrhCdXUG9/DMrGFUbQ/PzGpAjc245YBnZuVr6lbpFnSIA56ZlU8Vvy3XIQ54ZlY+X9KaWcNwD8/MGoZ7eGbWMNzDM7OG4VFaM2sYvqQ1s4bhS1ozaxju4ZlZw6ixgFdbrTWz6tKtW+lbCSR9RtJ1kp5MS3LuIKmvpDslPZM+1yjIP0bSLElPSdqzvfId8MysfFLpW2l+A9wWERsBmwNPACcDkyJiGDApfUfSxsAIYBNgL2CcpKKR1QHPzMqnptK39oqSegM7AxcDRMRHEfEWMBwYn7KN55M1nIcDV0fEorTE5yxgu2J1OOCZWfk60MOTNErStIJtVIvS1idbpvQSSf+U9L9pjeQBETEXIH32T/kHsfxSnbNTWps8aGFm5evAoEVEXARcVCRLd2Ar4JiIeFDSb0iXr23V3lo1xdrgHp6Zla9z7+HNBmZHxIPp+3VkAfA1SQOz6jQQmFeQf0jB+YPJFpFvkwOemZWvqVvpWzsi4lXgZUkbpqTdgZnARGBkShsJ3JT2JwIjJPWQNBQYBkwpVocvac2sfJ3/HN4xwBWSVgaeA44g65hNkHQk8BJwIEBEPC5pAllQXAyMjoglxQp3wDOz8nXyq2URMQPYppVDu7eRfywwttTyHfDMrHw19qaFA56Zlc8Bz8wahufDM7OG4emhzKxh+JLWzBqGe3hm1ijkgGdmjcIBz8wahpoc8MysQbiHZ2YNwwHPzBqGA56ZNY7aincOeGZWPvfwzKxhNDX5TQszaxDu4ZlZ46iteOeAZ2blcw/PzBqGA56ZNQy/WmZmDcM9PDNrGA54ZtYwHPDMrGHUWsCrrcekzay6qANbKcVJL0h6VNIMSdNSWl9Jd0p6Jn2uUZB/jKRZkp6StGd75TvgmVnZmpqaSt46YNeI2CIitknfTwYmRcQwYFL6jqSNgRHAJsBewDhJRdeNdMAzs7JJKnlbAcOB8Wl/PLBfQfrVEbEoIp4HZgHbFSvIAc/MytfJl7RAAHdIekjSqJQ2ICLmAqTP/il9EPBywbmzU1qbPGiRgz6r9eK/Tz+YjT83kAj4/plX8LUdvsB3D9iR1xe8B8DpF07k9ntnAnDCd7/O4cN3YMnSpRx/7nX87YEnKtl8S047dQz33D2Zvn37ccNNN1e6OVWpIz23FMBGFSRdFBEXtci2U0TMkdQfuFPSk8WKbCUtirXBAS8Hvzrx29xx/0wO/o+LWal7N1bpuTJf2+EL/Pbyuzj/sknL5d1o/c9y4J5bsdW3xzJwrT7c+j9Hs9l+Z7F0adHfm3WB4fsdwEEHH8opY06qdFOqVkcCXgpuLQNcyzxz0uc8STeSXaK+JmlgRMyVNBCYl7LPBoYUnD4YmFOs/NwvaSWtIWkTSetLNbZMeRlWX7UnX97qc1x64wMAfLx4CW+/t7DN/Pvu8kWuvX06H328mBfnzOfZl99g203X66LWWjFbb7Mtvfv0qXQzqlpn3sOTtKqk1Zv3ga8DjwETgZEp20jgprQ/ERghqYekocAwYEqxOnLp4UnqA4wGDgJWBl4HegIDJP0DGBcRd+VRd6UNHdSPNxa8x0VnHspmnx/EP594mRPOvQ6A74/YmYP33Y7pM1/i5F/fwFvvLmTQWn148NEXlp3/yrwFrN3ff8msNnTyu7QDgBtTcOwOXBkRt0maCkyQdCTwEnAgQEQ8LmkCMBNYDIyOiCXFKsirx3Ud2c3Er0TEhhHx5YjYJiKGAGcDw1PjlyNplKRpkqYtfuPxnJqWr+7du7HFRkP4w7V/Z4eDzuGDhYs44bt78Idr/87G3ziDL404m1ffeIezjzsgO6GVf/nCV7NWIzqzhxcRz0XE5mnbJCLGpvT5EbF7RAxLn28WnDM2Ij6X4sxf26sjl4AXEXtExGUR8VYrxx6KiB9HxMWtHLsoBcZtuq+5SR5Ny90rry3glXlvMfWxFwG48W8z2GKjIcx7812WLg0igj/ecB/bbLpuln/eWwz+7LLnKBnUfw3mvv52Rdpu1lFd9FhKp8ntnpoyX5J0gKT90351/NQ5em3+u8x+dQHD1s1GznfZbkOefO5VPrtm72V5hu+2OTOfnQvALZMf4cA9t2Lllbqz7tr92GCdtZj62AuVaLpZh0mlb9Ugr3t4XwfGAc8Ar6TkwcAGkn4YEXfkUW+1OO6ca7nkF4ezcvduvPDKG4w6/XLOO/FAvrjhYCKCF+e+yTE/vwqAJ557levv+Cf/vP4UFi9Zyo/PnuAR2ipx0gnHMW3qFN56awF77LYzPxh9DAd868BKN6uq1FofRpHDDSNJTwB7R8QLLdKHArdGxBfaK6PXlkf7b30NWTD1wko3wcrQs/uKrUqx4Um3l/z39Klz9qx4dMzrObzuZM/ItPQKsFJOdZpZF6uxDl5uAe+PwFRJV/PJqx9DyF70/dRghZnVpiZP8Q4R8UtJfyZ7uXcHsldAZgOHRMTMPOo0s67nHl4SEU8AfinUrI7V2qBFLo+lSPqLpG9I+tT9uvSK2VmSvptH3WbWdfxYSubfgeOA8yW9ySevlq0HPAtcGBE3tX26mdWCDk7sWXF53cN7FTgROFHSesBAYCHwdER8kEedZtb1qqXnVqrcp4dKz+K9kHc9Ztb1au0enufDM7Oy1Vi8c8Azs/LVWg8v1zuOkn5USpqZ1aZaG6XNe4hlZCtph+dcp5l1kaYmlbxVg7xmSzkIOBgYKmliwaHVgfl51GlmXa/WLmnzuod3PzAXWBM4ryD9XeCRnOo0sy5WY/Eut+fwXgReJHuP1szqlHt4gKR3aX19SAEREb1bOWZmNabG4l1uPbzV8yjXzKpLtQxGlCrX5/AkrdNaekS8lGe9ZtY1fEm7vFsK9nsCQ4GngNpckszMluOAVyAiNiv8Lmkr4Kg86zSzrlNj8a5rXy2LiOmStu3KOs0sP+7hFZB0XMHXJmArsrnxzKwO5BHvJHUDpgGvRMS+kvoC15DNp/kC8K8RsSDlHQMcCSwBjo2I24uVnferZasXbD3I7ukNz7lOM+siOb1a9iOWXx7iZGBSRAwDJqXvSNqYbGGwTYC9gHEpWLYp73t4Z+ZZvplVVlMnd/EkDQb2AcaSzZoOWSdpl7Q/HpgMnJTSr46IRcDzkmYB2wEPtFV+3pe0a5HNfLwJ2SgtABGxW571mlnX6Ei8kzQKGFWQdFFEXNQi2/lkMaPwWd4BETEXICLmSuqf0gcB/yjINzultanNgJdGVNsUEdOLHU+uILv23hf4PtnsKb6HZ1YnOjJokYJbywBXWNa+wLyIeEjSLqVU31o1xU4o1sM7r8ixAErppfWLiIsl/Sgi7gbulnR3CeeZWQ3o5BctdgK+KelfyK4Ie0u6HHhN0sDUuxsIzEv5ZwNDCs4fDMwpVkGbAS8idl2hpmc+Tp9zJe2TGjO4E8o1syrQma+WRcQYYAxA6uGdEBGHSvpPsqvDs9Nn84qHE4ErJf0aWBsYBkwpVke79/AkrUJ283CdiBglaRiwYUTcXMLP8HNJfYDjgd8CvYGflHCemdUAtXpV2enOBiZIOhJ4CTgQICIelzQBmAksBkZHxJJiBZUyaHEJ8BCwY/o+G7gWaDPgSbo0Ig6PiJsljYyI8UBn9BjNrIrkNXdAREwmG40lIuYDu7eRbyzZiG5JSnkO73MRcS7p8jQiFtL6zcJCmxfsew0LszolqeStGpTSw/tIUi/S6IekzwGL2jmn6EiJmdWHKoljJSsl4J0O3AYMkXQF2UjK4e2cM1jSBWQ9web9ZSLi2DLaamZVprMfPM5buwEvIu6UNB3YniyA/Sgi3mjntP8o2J+2Au0zsypWrxOAfhX4Mtml6krAjcUyp0EKM6tzNdbBK+mxlHHABsBVKekoSV+LiNG5tszMql7dXdKS9e42jYjmQYvxwKO5tsrMakJthbvSHkt5Cihcm2IIXlvWzKijx1Ik/YXsnl0f4AlJU9L3L5EttN2uNNXLb8nu/y0F7iUb9Ji9gu02sypQY2MWRS9pf9UJ5V8CXEl6FQQ4NKXt0Qllm1mF1c0obZrdZEWtFRGXFHy/VNKPO6FcM6sC1XKpWqp27+FJ2l7SVEnvSfpI0hJJ75RY/huSDpXULW2HAvNXrMlmVi2aVPpWDUoZtLgQOAh4BugFfC+lleK7wL8CrwJzgW+nNDOrA3UzaFEoImZJ6pamXrlEUkmDFhHxEvDNFWmgmVWv6ghjpSsl4H0gaWVghqRzyXpqqxY7QdJpRQ5HRPysA200syrVrVquVUtUyiXtYSnf0cD7ZM/hHdDOOe+3skG2fuRJZbXUzKpO3V3SRsSLafdD4EwASdcA3ylyzrL1MCStTjYn3hHA1RRfK8PMakiVxLGSlbtM4w7tZUirhR8HHEK2luRWzauFm1l9qMd3aTssLbpxANmSbJtFxHt51GNmlVVj8a6sdWlFNkVUMceTzYp8KnBKwfW7yAYterfXsAVTS33yxarBgMMuq3QTrAxvX3XYCp1fLffmSlXuurRPFis0IkoZDDGzGtetXgJeJ61La2Z1rMaeSsnnHp6ZNQYHPDNrGLV2D8/32sysbJ05eYCknpKmSHpY0uOSmp/77SvpTknPpM81Cs4ZI2mWpKck7dlue0tohNKMJ6el7+tI2q795ptZvZNK30qwCNgtIjYHtgD2krQ9cDIwKSKGAZPSdyRtDIwANgH2AsZJ6lasglJ6eOPIHjQ+KH1/F/hdSc03s7rWXSp5a09kmp/ZXSltAQwne3mB9Llf2h8OXB0RiyLieWAWULQzVkrA+1JaoezD1KgFwMolnGdmda6Te3ikeTNnAPOAOyPiQWBARMwFSJ/9U/ZBwMsFp89OaW0qJeB9nLqJzauWrUW2PoWZNbgmqeRN0ihJ0wq2US3Li4glEbEFMBjYTtKmRapvLYxGsfaWMkp7AdnC2/0ljSWbxPPUEs4zszrXkUHaiLiI7HXTUvK+JWky2b251yQNjIi5kgaS9f4g69ENKThtMDCnWLnt9vAi4grgROCXZHPh7RcR15bSaDOrb508SruWpM+k/V7A18je6poIjEzZRgI3pf2JwAhJPSQNBYYBU4rV0W4PT9I6wAfAXwrT0mzGZtbAOnkC0IHA+HQLrQmYEBE3S3oAmCDpSOAl0iqIEfG4pAnATGAxMDrNyt6mUi5pbyG7LhbQExhKtjj3JuX9TGZWLzoz3kXEI8CWraTPB3Zv45yxwNhS6yhlAtDNCr+nWVSOKrUCM6tfqrFVLTr8allETJe0bR6NMbPaUnfv0ko6ruBrE7AV8HpuLTKzmlF3AQ9YvWB/Mdk9vevzaY6Z1ZJamzygaMBLoyWrRcR/dFF7zKyGdKux6UeKTfHePSIWF5nq3cwaXD0t4jOF7H7dDEkTgWv5ZH1ZIuKGnNtmZlWuHu/h9QXmA7vxyfN4ATjgmTW4GuvgFQ14/dMI7WN8EuiaFX1B18waQ1MdPYfXDViNMmYkMLPGUE89vLkRcVaXtcTMak73GruJVyzg1dZPYmZdrp56eK2+rGtm1qxuHkuJiDe7siFmVntqLN55XVozK1+NvWjhgGdm5aubS1ozs/Y44JlZw6itcOeAZ2YroMY6eA54Zla+upoPz8ysGI/SmlnD8KCFmTUMX9KaWcPwJa2ZNYxa6+HVWoA2syqiDmztliUNkXSXpCckPS7pRym9r6Q7JT2TPtcoOGeMpFmSnpK0Z3t1OOCZWdm6SSVvJVgMHB8RXwC2B0ZL2hg4GZgUEcOASek76dgIYBNgL2BcWmmxTQ54ZlY2qfStPRExNyKmp/13gSeAQcBwYHzKNh7YL+0PB66OiEUR8TwwC9iuWB0OeGZWNnXkP2mUpGkF26g2y5XWA7YEHgQGRMRcyIIi0D9lGwS8XHDa7JTWJg9amFnZOjJmEREXARe1X6ZWA64HfhwR7xQZGOnwejsOeGZWts5etUzSSmTB7oqCta9fkzQwIuZKGgjMS+mzgSEFpw8G5hRvr5lZmTrzHp6yrtzFwBMR8euCQxOBkWl/JHBTQfoIST0kDQWGAVOK1eEenpmVrZNfLdsJOAx4VNKMlPZT4GxggqQjgZeAAwEi4nFJE4CZZCO8oyNiSbEKHPDMrGyduUpjRNxL24/stbqoWESMBcaWWocDnpmVTTU2BagDnpmVrcbeLHPA6yqnnTqGe+6eTN++/bjhppsr3Rxr4ZEL9ue9hR+zZGmwZGmwyym3Ljt2zD4b8/NDt2boqAm8+e4iVurWxPnf+xJbrt+PpRGcPH4a9z7xWgVbXznu4Vmrhu93AAcdfCinjDmp0k2xNuz78zt5891Fy6UN6rsKu242kJdef29Z2sjdNgBgx5NuZs3ePbn+pN3Y5dRbiaJPgNWnzryH1xX8WEoX2Xqbbendp0+lm2Ed9Mt/24bTrpy+3NOsGw3+DHc//ioAb7zzIW9/8BFbrt+vMg2ssCap5K0a5NrDk7QN8BVgbWAh8Bjwt4h4M896zTos4M9jdicCLpn0DJf+3zPsvfVg5rz5AY+9tGC5rI+9uIB9th7C9fe/wOB+q7L50H4M7rcq05+dX6HGV051hLHS5RLwJB0OHAs8DzwEPAX0BL4MnCTpMeD/RcRLLc4bBYwCuHDc7zny39t81c6sU339jNt4dcFC1uzdkz//dHeenvM2J+y3Gfv/4m+fynvZ5Fl8flAfJo/9F15+432mPP06i5csrUCrK69aem6lyquHtyqwU0QsbO2gpC3InopeLuAVvmv34eLi78SZdaZXF2T/q77xzofcPPVldvrCANZdazXuPWdfILuXd88v9mG3U29l3tsf8tPLpi07944z9+TZV9+tSLsrrbbCXU4BLyJ+187xGXnUa1aOVXp0p0nw3oeLWaVHd3b74kDOueFRNvj+tcvyPHLB/uxyyq28+e4ieq3cDUl8sGgxu242kMVLgqdeebuCP0EF1VjE6/JRWkmnRcRZXV1vpZ10wnFMmzqFt95awB677cwPRh/DAd86sNLNMqB/n55cftxXAejerYnr7nueSQ+3/Q76Wr17csOY3VkaMPfNDzhq3H1d1dSqU2uXtIouHkuX9FJErNNePl/S1pYBh11W6SZYGd6+6rAVilhTn3u75L+n267fp+LRMa9Bi3faOgT0yqNOM6uAioewjsnrkvYtYNuI+NTj55Je/nR2M6tFtfamRV4PHv8JWLeNY1fmVKeZdbHOnA+vK+Q1SntqkWN+t8qsTlRJHCtZLj28tABHseOSNDiPus2s60gqeasGed3D+09JTWRTMT8EvE72psUGwK5kk/mdTjYnvZnVqCqJYyXL65L2wLRI7iHAd4GBwAdk60zeCoyNiA/zqNvMuk6Nxbv8HjyOiJnAKXmVb2ZVoMYinufDM7Oy1dpjKQ54ZlY238Mzs4ZRawEv1xmPJU0qJc3MapM68F81yOtd2p7AKsCaktbgk1ubvclmPzazOuAeXuYosufvNgKmp/2HyJ7LKzpXnpnVDnVga7cs6Y+S5qUZ0ZvT+kq6U9Iz6XONgmNjJM2S9JSkPUtpby4BLyJ+ExFDgRMiYmjBtnlEXJhHnWZWAZ0Z8eBSYK8WaScDkyJiGDApfSc95zsC2CSdM05St/YqyOuS9oC0+0rB/jIRcUMe9ZpZ1+rMCUAj4p5WXksdDuyS9scDk4GTUvrVEbEIeF7SLGA74IFideQ1SvuNIscCcMAzqwNdcAtvQETMBYiIuZL6p/RBwD8K8s1OaUXl9WrZEXmUa2ZVpgMRr3BVwuSitHBXZ9Xc7uzLea9Le1pr6Y24poVZPerI4yaFqxJ2wGuSBqbe3UBgXkqfDQwpyDcYaHshkiTX5/CA9wu2JcDewHo512lmXaQLJgCdCIxM+yPJnvRoTh8hqYekoWTLvk5pr7Bce3gRcV7hd0m/ImuomdWBzryHJ+kqsgGKNSXNJptC7mxggqQjydaxPhAgIh6XNAGYCSwGRkfEkvbq6OpXy1YB1u/iOs0sJ505sWdEHNTGod3byD8WGNuROvK+h/con9xI7AasBfj+nVmdqLU3LfLu4e1bsL8YeC0iFudcp5l1kRqLd7nfw3sRID070xNYWxIR8VKe9ZpZF6mxiJf3bCnflPQM8DxwN/AC8Nc86zSzrlNrs6Xk/VjKz4DtgafTu7W7A/flXKeZdZFaW5c274D3cUTMB5okNUXEXcAWOddpZl2kSaVv1SDvQYu3JK0G3ANcIWke2eCFmdWFKolkJcprIe510u5wsuUZfwLcBjxL8YkFzKyG1NolbV49vD8DW0XE+5Kuj4hvkU3tYmZ1pEriWMnyCniFfw5+s8KsTlVLz61UeQW8aGPfzOpIZ75a1hXyCnibS3qHrKfXK+2TvkdE9M6pXjPrQrUV7vKbALTdueXNrPbVWAfPC3GbWfmq5Q2KUjngmVn5aiveOeCZWflqLN454JlZ+Tpzmcau4IBnZmWrsXiX++QBZmZVwz08MytbrfXwHPDMrGx+LMXMGoZ7eGbWMBzwzKxh+JLWzBqGe3hm1jBqLN454JnZCqixiOeAZ2Zlq7VXyxThCYm7mqRREXFRpdthpfHvq3741bLKGFXpBliH+PdVJxzwzKxhOOCZWcNwwKsM3w+qLf591QkPWphZw3APz8wahgOemTUMBzxAUkg6r+D7CZLO6MD5h0t6XdKMtP2pzHa8lz7Xk3RwkXwDJd1c8H2MpFmSnpK0Z0H63yStUU5bap2kJQW/jxmS1iujjEslfTvt/1jSKkXyXidp/bQ/VtLLzb/PgjxHSzqio+2wzuOAl1kEHCBpzRUo45qI2CJt/7aC7VkPaDPgAccBfwCQtDEwAtgE2AsYJ6l5IfTLgB+uYFtq1cKC38cWEfHCCpb3Y6DVgCdpE6BbRDyXkv4CbNdK1j8Cx65gO2wFOOBlFpONxP2k5QFJ60qaJOmR9LlOKQVK+oakByX9M/W0BqT0MySdUJDvsVZ6H2cDX0k9k0+1CfgWcFvaHw5cHRGLIuJ5YBaf/GWbCBxUSnvrnaTV0u9vuqRHJQ1P6etJeqwg36d695KOBdYG7pJ0VyvFHwLc1PwlIv4REXNbZoqID4AXJLUWDK0LOOB94nfAIZL6tEi/EPhTRHwRuAK4oI3zv1Nw+XQEcC+wfURsCVwNnNiBtpwM/D31TP6r8ICkocCCiFiUkgYBLxdkmZ3SiIgFQA9J/TpQd73oVfD7uBH4ENg/IrYCdgXOk0p7ETQiLgDmALtGxK6tZNkJeKjEdk0DvlJiXutknjwgiYh30r23Y4GFBYd2AA5I+5cB57ZRxDURcXTzF0mbAddIGgisDDzfSU0dCLxe8L21v7SFzxrNI+udzO+k+mvFwojYovmLpJWAX0jaGVhK9o/CgE6qq+XvpJh5wEadVK91kHt4yzsfOBJYtUieUh9c/C1wYURsBhwF9Ezpi1n+z71nyxPbsbDFObOBIQXfB5P1RgrLLwzgjeoQYC1g6xQIXyP7s1nR3wd8+ndSjH8fFeSAVyAi3gQmkAW9ZveTDQpA9pfm3hKL6wO8kvZHFqS/AGwFIGkrYGgr574LrN5GuU+TDWo0mwiMkNQjXe4OA6ak8gV8NtXZ6PoA8yLiY0m7Auum9NeA/pL6SeoB7NvG+cV+J08AG5TYjs8Dj7Wby3LhgPdp5wGFo7XHAkdIegQ4DPhRieWcAVwr6e/AGwXp1wN9Jc0AfkAWwFp6BFgs6eGWgxYR8T7wrKQN0vfHyYL0TLKBjNERsSRl3xr4R0QsLrHN9ewKYBtJ08j+4XoSICI+Bs4CHgRubk5vxUXAX9sYtLgF2KX5i6RzJc0GVpE0u8UgyE7A31bsR7Fy+dWyGiRpf7JLs1PbyfcbYGJETOqaljUmSb2Au4CdCv6xaS3flsBxEXFYlzXOluMeXg2KiBsp7TL1MQe7/EXEQuB00uh4EWsC/y//Fllb3MMzs4bhHp6ZNQwHPDNrGA54ZtYwHPBqUMFMII9JurbYLB4llFU4I8j/pskI2sq7i6Qdy6jjhdYmZmgrvY0yDpd0YWfUa43LAa82Nc8EsinwEfD9woMFs6V0SER8LyJmFsmyC9DhgGdWLRzwat/fgQ1S7+suSVcCj0rqJuk/JU1NM70cBdnbF5IulDRT0i1A/+aCJE2WtE3a3yvNLPJwmmVkPbLA+pPUu/yKpLUkXZ/qmCppp3RuP0l3pJlifk8H1qeXtJ2k+9O590vasODwEEm3KZv37/SCcw6VNCW16/flBnyrf548oIZJ6g7szSdTRW0HbBoRz0saBbwdEdumV6buk3QHsCWwIbAZ2cvzM8nmaSssdy2y+fZ2TmX1jYg3Jf0P8F5E/CrluxL4r4i4V9m0WbcDXyB7Ju3eiDhL0j50bF3XJ1O9iyV9DfgF2XRYy34+4ANgagrY7wPfIXvo92NJ48jepChrElarbw54talXejUNsh7exWSXmlPSnHgAXwe+2Hx/juxd0mHAzsBV6Y2AOZL+r5XytwfuaS4rvWPcmq8BGxfMstRb0uqpjgPSubdIWtCBn60PMF7SMLKJGlYqOHZnRMwHkHQD8GWyl/+3JguAAL3IZiQx+xQHvNq03NRHAOkv+/uFScAxEXF7i3z/QvszvqiEPJDdEtkhvWnQsi3lPtH+M+CuiNg/XUZPLjjWssxIbR0fEWPKrM8aiO/h1a/bgR+keeCQ9HlJqwL3kM2u0k3ZXH2tTWj5APDVNPsKkvqm9JYzhtwBFM4BuEXavYfsshJJewMdWVejcJaZw1sc20NS3/Tu6n7AfcAk4NuS+je3VdK6mLXCAa9+/S/Z/bnpyqYw/z1Zj/5G4BngUeC/gbtbnhgRr5Pdd7tB0sPANenQX4D9mwctyGaS2SYNiszkk9HiM4GdJU0nu7R+qUg7H0kzisyW9GuyCVZ/Kek+oOXgw71kk7DOAK6PiGlpVPlU4A5lM9rcSTYhp9mn+F1aM2sY7uGZWcNwwDOzhuGAZ2YNwwHPzBqGA56ZNQwHPDNrGA54ZtYw/j9LeX8TgFa1igAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_pred = svm_clf.predict(test[:, 1:])\n",
    "y_test = test[:, 0]\n",
    "\n",
    "conf_mat = confusion_matrix(y_test, y_pred)\n",
    "\n",
    "lang = ['No Fault (0)', 'Fault (1)']\n",
    "ax = plt.axes()\n",
    "sns.heatmap(conf_mat, square=True, annot=True, cmap='Blues', fmt='d', cbar=True, ax=ax, xticklabels=lang, yticklabels=lang)\n",
    "ax.set_title('Confusion Matrix')\n",
    "plt.xlabel('Predicted Label')\n",
    "plt.ylabel('True Label')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The Evaluation Metrics of the SVM Classifier (on the test set) is as follows:\n",
      "\n",
      "True Positives: 549\n",
      "True Negatives: 650\n",
      "False Positives: 1\n",
      "False Negatives: 1\n",
      "\n",
      "Precision: 0.9981818181818182\n",
      "Recall: 0.9981818181818182\n",
      "\n",
      "Accuracy: 0.9983347210657785\n"
     ]
    }
   ],
   "source": [
    "true_positive = conf_mat[1, 1]\n",
    "true_negative = conf_mat[0, 0]\n",
    "false_positive = conf_mat[0, 1]\n",
    "false_negative = conf_mat[1, 0]\n",
    "accuracy = accuracy_score(y_pred, y_test)\n",
    "\n",
    "print(\"The Evaluation Metrics of the SVM Classifier (on the test set) is as follows:\\n\")\n",
    "\n",
    "print(\"True Positives:\", true_positive)\n",
    "print(\"True Negatives:\", true_negative)\n",
    "print(\"False Positives:\", false_positive)\n",
    "print(\"False Negatives:\", false_negative, end='\\n\\n')\n",
    "\n",
    "print(\"Precision:\", true_positive/(true_positive+false_positive))\n",
    "print(\"Recall:\", true_positive/(true_positive+false_negative), end='\\n\\n')\n",
    "\n",
    "print(\"Accuracy:\", accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### (ii) Decision Tree:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeClassifier(max_depth=14, random_state=42)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Since only min_samples_leaf=1 gave the best result, Regularization was not required\n",
    "tree_clf = DecisionTreeClassifier(max_depth=14, random_state=42)\n",
    "tree_clf.fit(train_set.iloc[:, 1:], train_set.iloc[:, 0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_pred = tree_clf.predict(test_set.iloc[:, 1:])\n",
    "\n",
    "conf_mat = confusion_matrix(y_test, y_pred)\n",
    "\n",
    "lang = ['No Fault (0)', 'Fault (1)']\n",
    "ax = plt.axes()\n",
    "sns.heatmap(conf_mat, square=True, annot=True, cmap='Blues', fmt='d', cbar=True, ax=ax, xticklabels=lang, yticklabels=lang)\n",
    "ax.set_title('Confusion Matrix')\n",
    "plt.xlabel('Predicted Label')\n",
    "plt.ylabel('True Label')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The Evaluation Metrics of the Decision Tree Classifier (on the test set) is as follows:\n",
      "\n",
      "True Positives: 549\n",
      "True Negatives: 650\n",
      "False Positives: 1\n",
      "False Negatives: 1\n",
      "\n",
      "Precision: 0.9981818181818182\n",
      "Recall: 0.9981818181818182\n",
      "\n",
      "Accuracy: 0.9983347210657785\n"
     ]
    }
   ],
   "source": [
    "true_positive = conf_mat[1, 1]\n",
    "true_negative = conf_mat[0, 0]\n",
    "false_positive = conf_mat[0, 1]\n",
    "false_negative = conf_mat[1, 0]\n",
    "accuracy = accuracy_score(y_pred, y_test)\n",
    "\n",
    "print(\"The Evaluation Metrics of the Decision Tree Classifier (on the test set) is as follows:\\n\")\n",
    "\n",
    "print(\"True Positives:\", true_positive)\n",
    "print(\"True Negatives:\", true_negative)\n",
    "print(\"False Positives:\", false_positive)\n",
    "print(\"False Negatives:\", false_negative, end='\\n\\n')\n",
    "\n",
    "print(\"Precision:\", true_positive/(true_positive+false_positive))\n",
    "print(\"Recall:\", true_positive/(true_positive+false_negative), end='\\n\\n')\n",
    "\n",
    "print(\"Accuracy:\", accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### (iii) Random Forest:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(n_estimators=162, random_state=42)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest_clf = RandomForestClassifier(n_estimators=162, random_state=42)\n",
    "forest_clf.fit(train_set.iloc[:, 1:], train_set.iloc[:, 0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_pred = forest_clf.predict(test_set.iloc[:, 1:])\n",
    "\n",
    "conf_mat = confusion_matrix(y_test, y_pred)\n",
    "\n",
    "lang = ['No Fault (0)', 'Fault (1)']\n",
    "ax = plt.axes()\n",
    "sns.heatmap(conf_mat, square=True, annot=True, cmap='Blues', fmt='d', cbar=True, ax=ax, xticklabels=lang, yticklabels=lang)\n",
    "ax.set_title('Confusion Matrix')\n",
    "plt.xlabel('Predicted Label')\n",
    "plt.ylabel('True Label')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The Evaluation Metrics of the Random Forest Classifier (on the test set) is as follows:\n",
      "\n",
      "True Positives: 549\n",
      "True Negatives: 650\n",
      "False Positives: 1\n",
      "False Negatives: 1\n",
      "\n",
      "Precision: 0.9981818181818182\n",
      "Recall: 0.9981818181818182\n",
      "\n",
      "Accuracy: 0.9983347210657785\n"
     ]
    }
   ],
   "source": [
    "true_positive = conf_mat[1, 1]\n",
    "true_negative = conf_mat[0, 0]\n",
    "false_positive = conf_mat[0, 1]\n",
    "false_negative = conf_mat[1, 0]\n",
    "accuracy = accuracy_score(y_pred, y_test)\n",
    "\n",
    "print(\"The Evaluation Metrics of the Random Forest Classifier (on the test set) is as follows:\\n\")\n",
    "\n",
    "print(\"True Positives:\", true_positive)\n",
    "print(\"True Negatives:\", true_negative)\n",
    "print(\"False Positives:\", false_positive)\n",
    "print(\"False Negatives:\", false_negative, end='\\n\\n')\n",
    "\n",
    "print(\"Precision:\", true_positive/(true_positive+false_positive))\n",
    "print(\"Recall:\", true_positive/(true_positive+false_negative), end='\\n\\n')\n",
    "\n",
    "print(\"Accuracy:\", accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Comparision of the 3 models:\n",
    "<br><br>\n",
    "The observation is that all 3 models when **well tuned** gives the **exact same result**.<br>\n",
    "Their accuracy, precision, recall and all parameters are same are are over 99.8%.<br><br>\n",
    "\n",
    "The probably reasons for this are as follows:<br>\n",
    "1) The given dataset is well balanced, and has less noises.<br>\n",
    "2) The \"StratifiedShuffleSplit\" funstion of sklearn made sure that the classes are split evenly among the training anf testing set, further promoting the balance.<br><br>\n",
    "\n",
    "The following tuning were done on each model:<br>\n",
    "1) The SVM classifier performed poorly when linear, polynomial and SGD kernels were used giving an accuracy of only around 40% to 70%. But however when the \"rbf\" kernel was used the and the regularization parameter 'C' was set to 1000% the accuracy improved to more than 99.8%.<br>\n",
    "2) The Decision Tree which was expected to give the least accuracy surprisingly gave the best result, since the correaltion between the attributes in the dataset was simple and pretty consistent. It is to be noted here that, there seemed to be no overfitting in the data, since increasing the 'min_samples_leaf' parameter to value over 1, actually cause the accuracy of the test set to reduce instead of increasing. And the best result was achived either by setting min_samples_leaf=1 or max_depth=14 or above.<br>\n",
    "3) The Random Forest method gave the best accuaracy when 'n_estimators' was set at a value of 162 or above. The accuracy readuced when 'n_estimators' was set at a lower value.<br><br>\n",
    "\n",
    "Hence my take on the best classifier is as follows:<br>\n",
    "1) For this given simple dataset, \"Decision Tree\" is the best because it has the least time complexity for training compared to the other 2 methods, and gives the same accuracy as the other 2, without even requiring the dataset to be normalized.<br>\n",
    "2) But for a lot conventinal datasets, when noise in dataset is present, or when the dataset is not well balanced, but not too complex, Random Forest may give a better performance.<br>\n",
    "3) When the dataset is noisy and is complex, SVM with a proper regulaization constant, and \"rbf\" kernel may give the best result."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## C) Training & Evaluating MLP Models:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### (i) Model 1: (1 hidden layer, linear activation, SGD optimizer)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "model.add(Dense(10, input_dim=6, activation='linear'))\n",
    "model.add(Dense(1, activation='sigmoid'))\n",
    "model.compile(loss='binary_crossentropy', optimizer='SGD', metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 9180 samples, validate on 1620 samples\n",
      "Epoch 1/500\n",
      " - 0s - loss: 0.7298 - accuracy: 0.4834 - val_loss: 0.7154 - val_accuracy: 0.5049\n",
      "Epoch 2/500\n",
      " - 0s - loss: 0.7120 - accuracy: 0.4932 - val_loss: 0.7005 - val_accuracy: 0.5123\n",
      "Epoch 3/500\n",
      " - 0s - loss: 0.7006 - accuracy: 0.5037 - val_loss: 0.6911 - val_accuracy: 0.5278\n",
      "Epoch 4/500\n",
      " - 0s - loss: 0.6932 - accuracy: 0.5191 - val_loss: 0.6847 - val_accuracy: 0.5494\n",
      "Epoch 5/500\n",
      " - 0s - loss: 0.6883 - accuracy: 0.5427 - val_loss: 0.6806 - val_accuracy: 0.5753\n",
      "Epoch 6/500\n",
      " - 0s - loss: 0.6851 - accuracy: 0.5833 - val_loss: 0.6779 - val_accuracy: 0.6210\n",
      "Epoch 7/500\n",
      " - 0s - loss: 0.6831 - accuracy: 0.6688 - val_loss: 0.6760 - val_accuracy: 0.7611\n",
      "Epoch 8/500\n",
      " - 0s - loss: 0.6817 - accuracy: 0.7568 - val_loss: 0.6749 - val_accuracy: 0.7716\n",
      "Epoch 9/500\n",
      " - 0s - loss: 0.6808 - accuracy: 0.7557 - val_loss: 0.6741 - val_accuracy: 0.7691\n",
      "Epoch 10/500\n",
      " - 0s - loss: 0.6803 - accuracy: 0.7523 - val_loss: 0.6735 - val_accuracy: 0.7623\n",
      "Epoch 11/500\n",
      " - 0s - loss: 0.6798 - accuracy: 0.7484 - val_loss: 0.6731 - val_accuracy: 0.7586\n",
      "Epoch 12/500\n",
      " - 0s - loss: 0.6795 - accuracy: 0.7478 - val_loss: 0.6728 - val_accuracy: 0.7568\n",
      "Epoch 13/500\n",
      " - 0s - loss: 0.6793 - accuracy: 0.7468 - val_loss: 0.6725 - val_accuracy: 0.7556\n",
      "Epoch 14/500\n",
      " - 0s - loss: 0.6792 - accuracy: 0.7439 - val_loss: 0.6724 - val_accuracy: 0.7549\n",
      "Epoch 15/500\n",
      " - 0s - loss: 0.6791 - accuracy: 0.7418 - val_loss: 0.6723 - val_accuracy: 0.7562\n",
      "Epoch 16/500\n",
      " - 0s - loss: 0.6790 - accuracy: 0.7411 - val_loss: 0.6722 - val_accuracy: 0.7556\n",
      "Epoch 17/500\n",
      " - 0s - loss: 0.6790 - accuracy: 0.7400 - val_loss: 0.6721 - val_accuracy: 0.7543\n",
      "Epoch 18/500\n",
      " - 0s - loss: 0.6789 - accuracy: 0.7388 - val_loss: 0.6721 - val_accuracy: 0.7537\n",
      "Epoch 19/500\n",
      " - 0s - loss: 0.6789 - accuracy: 0.7398 - val_loss: 0.6720 - val_accuracy: 0.7537\n",
      "Epoch 20/500\n",
      " - 0s - loss: 0.6789 - accuracy: 0.7369 - val_loss: 0.6720 - val_accuracy: 0.7512\n",
      "Epoch 21/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7383 - val_loss: 0.6720 - val_accuracy: 0.7500\n",
      "Epoch 22/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7378 - val_loss: 0.6719 - val_accuracy: 0.7500\n",
      "Epoch 23/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7376 - val_loss: 0.6719 - val_accuracy: 0.7488\n",
      "Epoch 24/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7345 - val_loss: 0.6719 - val_accuracy: 0.7481\n",
      "Epoch 25/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7378 - val_loss: 0.6718 - val_accuracy: 0.7500\n",
      "Epoch 26/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7371 - val_loss: 0.6719 - val_accuracy: 0.7488\n",
      "Epoch 27/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7369 - val_loss: 0.6719 - val_accuracy: 0.7500\n",
      "Epoch 28/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7376 - val_loss: 0.6719 - val_accuracy: 0.7494\n",
      "Epoch 29/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7389 - val_loss: 0.6719 - val_accuracy: 0.7481\n",
      "Epoch 30/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7375 - val_loss: 0.6720 - val_accuracy: 0.7494\n",
      "Epoch 31/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7373 - val_loss: 0.6720 - val_accuracy: 0.7481\n",
      "Epoch 32/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7366 - val_loss: 0.6719 - val_accuracy: 0.7469\n",
      "Epoch 33/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7367 - val_loss: 0.6719 - val_accuracy: 0.7481\n",
      "Epoch 34/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7366 - val_loss: 0.6719 - val_accuracy: 0.7481\n",
      "Epoch 35/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7354 - val_loss: 0.6719 - val_accuracy: 0.7469\n",
      "Epoch 36/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7367 - val_loss: 0.6719 - val_accuracy: 0.7463\n",
      "Epoch 37/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7336 - val_loss: 0.6719 - val_accuracy: 0.7469\n",
      "Epoch 38/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7359 - val_loss: 0.6719 - val_accuracy: 0.7475\n",
      "Epoch 39/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7363 - val_loss: 0.6719 - val_accuracy: 0.7457\n",
      "Epoch 40/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7326 - val_loss: 0.6719 - val_accuracy: 0.7463\n",
      "Epoch 41/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7331 - val_loss: 0.6719 - val_accuracy: 0.7463\n",
      "Epoch 42/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7332 - val_loss: 0.6718 - val_accuracy: 0.7463\n",
      "Epoch 43/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7344 - val_loss: 0.6718 - val_accuracy: 0.7444\n",
      "Epoch 44/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7315 - val_loss: 0.6718 - val_accuracy: 0.7457\n",
      "Epoch 45/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7330 - val_loss: 0.6719 - val_accuracy: 0.7457\n",
      "Epoch 46/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7322 - val_loss: 0.6718 - val_accuracy: 0.7469\n",
      "Epoch 47/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7338 - val_loss: 0.6718 - val_accuracy: 0.7463\n",
      "Epoch 48/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7356 - val_loss: 0.6718 - val_accuracy: 0.7469\n",
      "Epoch 49/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7347 - val_loss: 0.6718 - val_accuracy: 0.7469\n",
      "Epoch 50/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7351 - val_loss: 0.6718 - val_accuracy: 0.7457\n",
      "Epoch 51/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7317 - val_loss: 0.6719 - val_accuracy: 0.7457\n",
      "Epoch 52/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7337 - val_loss: 0.6718 - val_accuracy: 0.7463\n",
      "Epoch 53/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7328 - val_loss: 0.6718 - val_accuracy: 0.7444\n",
      "Epoch 54/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7321 - val_loss: 0.6718 - val_accuracy: 0.7438\n",
      "Epoch 55/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7312 - val_loss: 0.6719 - val_accuracy: 0.7438\n",
      "Epoch 56/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7331 - val_loss: 0.6718 - val_accuracy: 0.7438\n",
      "Epoch 57/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7314 - val_loss: 0.6718 - val_accuracy: 0.7432\n",
      "Epoch 58/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7310 - val_loss: 0.6717 - val_accuracy: 0.7444\n",
      "Epoch 59/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7327 - val_loss: 0.6718 - val_accuracy: 0.7438\n",
      "Epoch 60/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7307 - val_loss: 0.6718 - val_accuracy: 0.7432\n",
      "Epoch 61/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7316 - val_loss: 0.6718 - val_accuracy: 0.7438\n",
      "Epoch 62/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7309 - val_loss: 0.6718 - val_accuracy: 0.7432\n",
      "Epoch 63/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7315 - val_loss: 0.6718 - val_accuracy: 0.7426\n",
      "Epoch 64/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7294 - val_loss: 0.6718 - val_accuracy: 0.7444\n",
      "Epoch 65/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7324 - val_loss: 0.6719 - val_accuracy: 0.7444\n",
      "Epoch 66/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7319 - val_loss: 0.6719 - val_accuracy: 0.7438\n",
      "Epoch 67/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7328 - val_loss: 0.6719 - val_accuracy: 0.7469\n",
      "Epoch 68/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7338 - val_loss: 0.6720 - val_accuracy: 0.7444\n",
      "Epoch 69/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7338 - val_loss: 0.6719 - val_accuracy: 0.7444\n",
      "Epoch 70/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7336 - val_loss: 0.6719 - val_accuracy: 0.7438\n",
      "Epoch 71/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7321 - val_loss: 0.6719 - val_accuracy: 0.7444\n",
      "Epoch 72/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7325 - val_loss: 0.6718 - val_accuracy: 0.7444\n",
      "Epoch 73/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7301 - val_loss: 0.6718 - val_accuracy: 0.7444\n",
      "Epoch 74/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7329 - val_loss: 0.6718 - val_accuracy: 0.7438\n",
      "Epoch 75/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7296 - val_loss: 0.6718 - val_accuracy: 0.7444\n",
      "Epoch 76/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7328 - val_loss: 0.6719 - val_accuracy: 0.7432\n",
      "Epoch 77/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7316 - val_loss: 0.6719 - val_accuracy: 0.7420\n",
      "Epoch 78/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7300 - val_loss: 0.6719 - val_accuracy: 0.7420\n",
      "Epoch 79/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7303 - val_loss: 0.6719 - val_accuracy: 0.7420\n",
      "Epoch 80/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7298 - val_loss: 0.6718 - val_accuracy: 0.7432\n",
      "Epoch 81/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7322 - val_loss: 0.6718 - val_accuracy: 0.7420\n",
      "Epoch 82/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7290 - val_loss: 0.6718 - val_accuracy: 0.7420\n",
      "Epoch 83/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7294 - val_loss: 0.6718 - val_accuracy: 0.7420\n",
      "Epoch 84/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7316 - val_loss: 0.6718 - val_accuracy: 0.7426\n",
      "Epoch 85/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7303 - val_loss: 0.6718 - val_accuracy: 0.7432\n",
      "Epoch 86/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7314 - val_loss: 0.6718 - val_accuracy: 0.7444\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 87/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7321 - val_loss: 0.6718 - val_accuracy: 0.7444\n",
      "Epoch 88/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7321 - val_loss: 0.6719 - val_accuracy: 0.7432\n",
      "Epoch 89/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7302 - val_loss: 0.6719 - val_accuracy: 0.7432\n",
      "Epoch 90/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7309 - val_loss: 0.6718 - val_accuracy: 0.7426\n",
      "Epoch 91/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7297 - val_loss: 0.6718 - val_accuracy: 0.7420\n",
      "Epoch 92/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7291 - val_loss: 0.6718 - val_accuracy: 0.7420\n",
      "Epoch 93/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7296 - val_loss: 0.6719 - val_accuracy: 0.7407\n",
      "Epoch 94/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7284 - val_loss: 0.6718 - val_accuracy: 0.7420\n",
      "Epoch 95/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7290 - val_loss: 0.6719 - val_accuracy: 0.7438\n",
      "Epoch 96/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7322 - val_loss: 0.6718 - val_accuracy: 0.7444\n",
      "Epoch 97/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7308 - val_loss: 0.6718 - val_accuracy: 0.7432\n",
      "Epoch 98/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7312 - val_loss: 0.6718 - val_accuracy: 0.7420\n",
      "Epoch 99/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7298 - val_loss: 0.6719 - val_accuracy: 0.7401\n",
      "Epoch 100/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7270 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 101/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7291 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 102/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7278 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 103/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7264 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 104/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7296 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 105/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7294 - val_loss: 0.6718 - val_accuracy: 0.7420\n",
      "Epoch 106/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7302 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 107/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7285 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 108/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7281 - val_loss: 0.6719 - val_accuracy: 0.7401\n",
      "Epoch 109/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7276 - val_loss: 0.6719 - val_accuracy: 0.7401\n",
      "Epoch 110/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7298 - val_loss: 0.6719 - val_accuracy: 0.7414\n",
      "Epoch 111/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7279 - val_loss: 0.6719 - val_accuracy: 0.7401\n",
      "Epoch 112/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7268 - val_loss: 0.6719 - val_accuracy: 0.7414\n",
      "Epoch 113/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7279 - val_loss: 0.6719 - val_accuracy: 0.7383\n",
      "Epoch 114/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7283 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 115/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7265 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 116/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7260 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 117/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7268 - val_loss: 0.6718 - val_accuracy: 0.7414\n",
      "Epoch 118/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7290 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 119/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7275 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 120/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7261 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 121/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7259 - val_loss: 0.6719 - val_accuracy: 0.7414\n",
      "Epoch 122/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7304 - val_loss: 0.6719 - val_accuracy: 0.7420\n",
      "Epoch 123/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7273 - val_loss: 0.6719 - val_accuracy: 0.7407\n",
      "Epoch 124/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7264 - val_loss: 0.6719 - val_accuracy: 0.7395\n",
      "Epoch 125/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7265 - val_loss: 0.6719 - val_accuracy: 0.7395\n",
      "Epoch 126/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7259 - val_loss: 0.6719 - val_accuracy: 0.7395\n",
      "Epoch 127/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7252 - val_loss: 0.6719 - val_accuracy: 0.7407\n",
      "Epoch 128/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7267 - val_loss: 0.6719 - val_accuracy: 0.7414\n",
      "Epoch 129/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7282 - val_loss: 0.6719 - val_accuracy: 0.7401\n",
      "Epoch 130/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7281 - val_loss: 0.6719 - val_accuracy: 0.7395\n",
      "Epoch 131/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7272 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 132/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7284 - val_loss: 0.6719 - val_accuracy: 0.7395\n",
      "Epoch 133/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7268 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 134/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7272 - val_loss: 0.6718 - val_accuracy: 0.7414\n",
      "Epoch 135/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7270 - val_loss: 0.6718 - val_accuracy: 0.7414\n",
      "Epoch 136/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7263 - val_loss: 0.6718 - val_accuracy: 0.7414\n",
      "Epoch 137/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7270 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 138/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7256 - val_loss: 0.6717 - val_accuracy: 0.7407\n",
      "Epoch 139/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7265 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 140/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7275 - val_loss: 0.6718 - val_accuracy: 0.7426\n",
      "Epoch 141/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7297 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 142/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7286 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 143/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7268 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 144/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7257 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 145/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7249 - val_loss: 0.6718 - val_accuracy: 0.7420\n",
      "Epoch 146/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7252 - val_loss: 0.6718 - val_accuracy: 0.7414\n",
      "Epoch 147/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7286 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 148/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7241 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 149/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7259 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 150/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7282 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 151/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7252 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 152/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7263 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 153/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7217 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 154/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7256 - val_loss: 0.6717 - val_accuracy: 0.7352\n",
      "Epoch 155/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7216 - val_loss: 0.6717 - val_accuracy: 0.7358\n",
      "Epoch 156/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7237 - val_loss: 0.6717 - val_accuracy: 0.7346\n",
      "Epoch 157/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7202 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 158/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7231 - val_loss: 0.6717 - val_accuracy: 0.7370\n",
      "Epoch 159/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7232 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 160/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7232 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 161/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7221 - val_loss: 0.6718 - val_accuracy: 0.7358\n",
      "Epoch 162/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7219 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 163/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7233 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 164/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7245 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 165/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7233 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 166/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7259 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 167/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7246 - val_loss: 0.6718 - val_accuracy: 0.7364\n",
      "Epoch 168/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7225 - val_loss: 0.6718 - val_accuracy: 0.7352\n",
      "Epoch 169/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7255 - val_loss: 0.6718 - val_accuracy: 0.7340\n",
      "Epoch 170/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7187 - val_loss: 0.6718 - val_accuracy: 0.7340\n",
      "Epoch 171/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7224 - val_loss: 0.6718 - val_accuracy: 0.7340\n",
      "Epoch 172/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7208 - val_loss: 0.6717 - val_accuracy: 0.7340\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 173/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7198 - val_loss: 0.6718 - val_accuracy: 0.7340\n",
      "Epoch 174/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7195 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 175/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7219 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 176/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7243 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 177/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7263 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 178/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7254 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 179/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7233 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 180/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7235 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 181/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7249 - val_loss: 0.6719 - val_accuracy: 0.7370\n",
      "Epoch 182/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7246 - val_loss: 0.6719 - val_accuracy: 0.7395\n",
      "Epoch 183/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7280 - val_loss: 0.6719 - val_accuracy: 0.7395\n",
      "Epoch 184/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7270 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 185/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7270 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 186/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7245 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 187/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7263 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 188/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7247 - val_loss: 0.6719 - val_accuracy: 0.7389\n",
      "Epoch 189/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7253 - val_loss: 0.6719 - val_accuracy: 0.7389\n",
      "Epoch 190/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7245 - val_loss: 0.6719 - val_accuracy: 0.7389\n",
      "Epoch 191/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7240 - val_loss: 0.6719 - val_accuracy: 0.7389\n",
      "Epoch 192/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7258 - val_loss: 0.6719 - val_accuracy: 0.7395\n",
      "Epoch 193/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7268 - val_loss: 0.6719 - val_accuracy: 0.7395\n",
      "Epoch 194/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7267 - val_loss: 0.6719 - val_accuracy: 0.7395\n",
      "Epoch 195/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7280 - val_loss: 0.6719 - val_accuracy: 0.7407\n",
      "Epoch 196/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7277 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 197/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7271 - val_loss: 0.6719 - val_accuracy: 0.7383\n",
      "Epoch 198/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7265 - val_loss: 0.6719 - val_accuracy: 0.7383\n",
      "Epoch 199/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7246 - val_loss: 0.6719 - val_accuracy: 0.7407\n",
      "Epoch 200/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7269 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 201/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7267 - val_loss: 0.6719 - val_accuracy: 0.7389\n",
      "Epoch 202/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7271 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 203/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7260 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 204/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7227 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 205/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7244 - val_loss: 0.6719 - val_accuracy: 0.7414\n",
      "Epoch 206/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7281 - val_loss: 0.6719 - val_accuracy: 0.7377\n",
      "Epoch 207/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7239 - val_loss: 0.6719 - val_accuracy: 0.7401\n",
      "Epoch 208/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7253 - val_loss: 0.6719 - val_accuracy: 0.7370\n",
      "Epoch 209/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7254 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 210/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7266 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 211/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7242 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 212/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7228 - val_loss: 0.6718 - val_accuracy: 0.7364\n",
      "Epoch 213/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7218 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 214/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7232 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 215/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7243 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 216/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7258 - val_loss: 0.6719 - val_accuracy: 0.7383\n",
      "Epoch 217/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7245 - val_loss: 0.6719 - val_accuracy: 0.7383\n",
      "Epoch 218/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7258 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 219/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7248 - val_loss: 0.6718 - val_accuracy: 0.7420\n",
      "Epoch 220/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7278 - val_loss: 0.6718 - val_accuracy: 0.7426\n",
      "Epoch 221/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7293 - val_loss: 0.6718 - val_accuracy: 0.7420\n",
      "Epoch 222/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7273 - val_loss: 0.6718 - val_accuracy: 0.7414\n",
      "Epoch 223/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7275 - val_loss: 0.6718 - val_accuracy: 0.7420\n",
      "Epoch 224/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7263 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 225/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7265 - val_loss: 0.6717 - val_accuracy: 0.7401\n",
      "Epoch 226/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7254 - val_loss: 0.6717 - val_accuracy: 0.7420\n",
      "Epoch 227/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7282 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 228/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7286 - val_loss: 0.6717 - val_accuracy: 0.7377\n",
      "Epoch 229/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7256 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 230/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7248 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 231/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7261 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 232/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7254 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 233/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7258 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 234/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7266 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 235/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7251 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 236/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7251 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 237/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7244 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 238/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7252 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 239/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7205 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 240/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7231 - val_loss: 0.6717 - val_accuracy: 0.7383\n",
      "Epoch 241/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7249 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 242/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7244 - val_loss: 0.6718 - val_accuracy: 0.7346\n",
      "Epoch 243/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7215 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 244/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7255 - val_loss: 0.6719 - val_accuracy: 0.7358\n",
      "Epoch 245/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7221 - val_loss: 0.6719 - val_accuracy: 0.7370\n",
      "Epoch 246/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7243 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 247/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7243 - val_loss: 0.6719 - val_accuracy: 0.7383\n",
      "Epoch 248/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7249 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 249/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7230 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 250/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7261 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 251/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7259 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 252/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7259 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 253/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7261 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 254/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7265 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 255/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7247 - val_loss: 0.6719 - val_accuracy: 0.7370\n",
      "Epoch 256/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7225 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 257/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7237 - val_loss: 0.6718 - val_accuracy: 0.7364\n",
      "Epoch 258/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7244 - val_loss: 0.6718 - val_accuracy: 0.7340\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 259/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7231 - val_loss: 0.6718 - val_accuracy: 0.7340\n",
      "Epoch 260/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7207 - val_loss: 0.6718 - val_accuracy: 0.7358\n",
      "Epoch 261/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7234 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 262/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7240 - val_loss: 0.6718 - val_accuracy: 0.7358\n",
      "Epoch 263/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7222 - val_loss: 0.6718 - val_accuracy: 0.7346\n",
      "Epoch 264/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7229 - val_loss: 0.6718 - val_accuracy: 0.7346\n",
      "Epoch 265/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7174 - val_loss: 0.6718 - val_accuracy: 0.7358\n",
      "Epoch 266/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7234 - val_loss: 0.6718 - val_accuracy: 0.7358\n",
      "Epoch 267/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7236 - val_loss: 0.6717 - val_accuracy: 0.7358\n",
      "Epoch 268/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7206 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 269/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7246 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 270/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7248 - val_loss: 0.6718 - val_accuracy: 0.7358\n",
      "Epoch 271/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7234 - val_loss: 0.6718 - val_accuracy: 0.7333\n",
      "Epoch 272/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7206 - val_loss: 0.6718 - val_accuracy: 0.7346\n",
      "Epoch 273/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7215 - val_loss: 0.6718 - val_accuracy: 0.7358\n",
      "Epoch 274/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7205 - val_loss: 0.6718 - val_accuracy: 0.7364\n",
      "Epoch 275/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7212 - val_loss: 0.6717 - val_accuracy: 0.7377\n",
      "Epoch 276/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7245 - val_loss: 0.6717 - val_accuracy: 0.7401\n",
      "Epoch 277/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7255 - val_loss: 0.6717 - val_accuracy: 0.7383\n",
      "Epoch 278/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7251 - val_loss: 0.6717 - val_accuracy: 0.7383\n",
      "Epoch 279/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7249 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 280/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7246 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 281/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7258 - val_loss: 0.6717 - val_accuracy: 0.7383\n",
      "Epoch 282/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7253 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 283/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7252 - val_loss: 0.6717 - val_accuracy: 0.7407\n",
      "Epoch 284/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7270 - val_loss: 0.6717 - val_accuracy: 0.7389\n",
      "Epoch 285/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7257 - val_loss: 0.6717 - val_accuracy: 0.7383\n",
      "Epoch 286/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7259 - val_loss: 0.6717 - val_accuracy: 0.7383\n",
      "Epoch 287/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7243 - val_loss: 0.6717 - val_accuracy: 0.7383\n",
      "Epoch 288/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7252 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 289/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7224 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 290/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7253 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 291/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7256 - val_loss: 0.6719 - val_accuracy: 0.7383\n",
      "Epoch 292/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7248 - val_loss: 0.6719 - val_accuracy: 0.7383\n",
      "Epoch 293/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7243 - val_loss: 0.6718 - val_accuracy: 0.7364\n",
      "Epoch 294/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7239 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 295/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7245 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 296/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7264 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 297/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7248 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 298/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7259 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 299/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7256 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 300/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7277 - val_loss: 0.6719 - val_accuracy: 0.7383\n",
      "Epoch 301/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7261 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 302/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7267 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 303/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7231 - val_loss: 0.6719 - val_accuracy: 0.7395\n",
      "Epoch 304/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7275 - val_loss: 0.6719 - val_accuracy: 0.7370\n",
      "Epoch 305/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7256 - val_loss: 0.6719 - val_accuracy: 0.7370\n",
      "Epoch 306/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7221 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 307/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7260 - val_loss: 0.6719 - val_accuracy: 0.7377\n",
      "Epoch 308/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7255 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 309/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7254 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 310/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7245 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 311/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7246 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 312/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7258 - val_loss: 0.6717 - val_accuracy: 0.7370\n",
      "Epoch 313/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7243 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 314/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7258 - val_loss: 0.6717 - val_accuracy: 0.7377\n",
      "Epoch 315/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7260 - val_loss: 0.6717 - val_accuracy: 0.7383\n",
      "Epoch 316/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7272 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 317/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7269 - val_loss: 0.6717 - val_accuracy: 0.7389\n",
      "Epoch 318/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7217 - val_loss: 0.6717 - val_accuracy: 0.7395\n",
      "Epoch 319/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7289 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 320/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7239 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 321/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7261 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 322/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7272 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 323/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7278 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 324/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7242 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 325/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7265 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 326/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7265 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 327/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7266 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 328/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7223 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 329/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7253 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 330/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7269 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 331/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7252 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 332/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7267 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 333/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7240 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 334/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7272 - val_loss: 0.6719 - val_accuracy: 0.7377\n",
      "Epoch 335/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7246 - val_loss: 0.6719 - val_accuracy: 0.7370\n",
      "Epoch 336/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7255 - val_loss: 0.6719 - val_accuracy: 0.7370\n",
      "Epoch 337/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7246 - val_loss: 0.6719 - val_accuracy: 0.7383\n",
      "Epoch 338/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7269 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 339/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7261 - val_loss: 0.6718 - val_accuracy: 0.7364\n",
      "Epoch 340/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7247 - val_loss: 0.6718 - val_accuracy: 0.7358\n",
      "Epoch 341/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7232 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 342/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7221 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 343/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7244 - val_loss: 0.6718 - val_accuracy: 0.7420\n",
      "Epoch 344/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7253 - val_loss: 0.6718 - val_accuracy: 0.7407\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 345/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7281 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 346/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7264 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 347/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7279 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 348/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7268 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 349/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7270 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 350/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7267 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 351/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7251 - val_loss: 0.6719 - val_accuracy: 0.7383\n",
      "Epoch 352/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7268 - val_loss: 0.6719 - val_accuracy: 0.7407\n",
      "Epoch 353/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7263 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 354/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7265 - val_loss: 0.6718 - val_accuracy: 0.7420\n",
      "Epoch 355/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7267 - val_loss: 0.6718 - val_accuracy: 0.7432\n",
      "Epoch 356/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7277 - val_loss: 0.6718 - val_accuracy: 0.7426\n",
      "Epoch 357/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7283 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 358/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7269 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 359/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7252 - val_loss: 0.6717 - val_accuracy: 0.7401\n",
      "Epoch 360/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7263 - val_loss: 0.6717 - val_accuracy: 0.7420\n",
      "Epoch 361/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7254 - val_loss: 0.6717 - val_accuracy: 0.7420\n",
      "Epoch 362/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7289 - val_loss: 0.6717 - val_accuracy: 0.7383\n",
      "Epoch 363/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7231 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 364/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7272 - val_loss: 0.6717 - val_accuracy: 0.7407\n",
      "Epoch 365/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7281 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 366/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7259 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 367/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7266 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 368/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7269 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 369/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7275 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 370/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7253 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 371/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7272 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 372/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7269 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 373/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7266 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 374/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7243 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 375/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7253 - val_loss: 0.6717 - val_accuracy: 0.7395\n",
      "Epoch 376/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7268 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 377/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7232 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 378/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7266 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 379/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7249 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 380/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7245 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 381/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7272 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 382/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7249 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 383/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7237 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 384/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7253 - val_loss: 0.6717 - val_accuracy: 0.7364\n",
      "Epoch 385/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7222 - val_loss: 0.6717 - val_accuracy: 0.7377\n",
      "Epoch 386/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7245 - val_loss: 0.6717 - val_accuracy: 0.7352\n",
      "Epoch 387/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7218 - val_loss: 0.6717 - val_accuracy: 0.7377\n",
      "Epoch 388/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7228 - val_loss: 0.6717 - val_accuracy: 0.7383\n",
      "Epoch 389/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7244 - val_loss: 0.6717 - val_accuracy: 0.7401\n",
      "Epoch 390/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7256 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 391/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7246 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 392/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7269 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 393/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7231 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 394/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7249 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 395/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7270 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 396/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7256 - val_loss: 0.6718 - val_accuracy: 0.7352\n",
      "Epoch 397/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7228 - val_loss: 0.6717 - val_accuracy: 0.7383\n",
      "Epoch 398/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7234 - val_loss: 0.6717 - val_accuracy: 0.7370\n",
      "Epoch 399/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7222 - val_loss: 0.6717 - val_accuracy: 0.7389\n",
      "Epoch 400/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7256 - val_loss: 0.6717 - val_accuracy: 0.7383\n",
      "Epoch 401/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7255 - val_loss: 0.6717 - val_accuracy: 0.7383\n",
      "Epoch 402/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7254 - val_loss: 0.6717 - val_accuracy: 0.7383\n",
      "Epoch 403/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7244 - val_loss: 0.6717 - val_accuracy: 0.7383\n",
      "Epoch 404/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7265 - val_loss: 0.6717 - val_accuracy: 0.7370\n",
      "Epoch 405/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7237 - val_loss: 0.6717 - val_accuracy: 0.7370\n",
      "Epoch 406/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7249 - val_loss: 0.6717 - val_accuracy: 0.7383\n",
      "Epoch 407/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7230 - val_loss: 0.6717 - val_accuracy: 0.7383\n",
      "Epoch 408/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7227 - val_loss: 0.6716 - val_accuracy: 0.7395\n",
      "Epoch 409/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7239 - val_loss: 0.6717 - val_accuracy: 0.7377\n",
      "Epoch 410/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7224 - val_loss: 0.6716 - val_accuracy: 0.7389\n",
      "Epoch 411/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7260 - val_loss: 0.6716 - val_accuracy: 0.7358\n",
      "Epoch 412/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7247 - val_loss: 0.6716 - val_accuracy: 0.7340\n",
      "Epoch 413/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7227 - val_loss: 0.6717 - val_accuracy: 0.7333\n",
      "Epoch 414/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7220 - val_loss: 0.6717 - val_accuracy: 0.7346\n",
      "Epoch 415/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7212 - val_loss: 0.6717 - val_accuracy: 0.7364\n",
      "Epoch 416/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7231 - val_loss: 0.6717 - val_accuracy: 0.7383\n",
      "Epoch 417/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7241 - val_loss: 0.6717 - val_accuracy: 0.7377\n",
      "Epoch 418/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7243 - val_loss: 0.6717 - val_accuracy: 0.7364\n",
      "Epoch 419/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7221 - val_loss: 0.6717 - val_accuracy: 0.7340\n",
      "Epoch 420/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7204 - val_loss: 0.6717 - val_accuracy: 0.7352\n",
      "Epoch 421/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7217 - val_loss: 0.6717 - val_accuracy: 0.7352\n",
      "Epoch 422/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7219 - val_loss: 0.6717 - val_accuracy: 0.7352\n",
      "Epoch 423/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7200 - val_loss: 0.6717 - val_accuracy: 0.7377\n",
      "Epoch 424/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7243 - val_loss: 0.6717 - val_accuracy: 0.7389\n",
      "Epoch 425/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7263 - val_loss: 0.6717 - val_accuracy: 0.7383\n",
      "Epoch 426/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7272 - val_loss: 0.6718 - val_accuracy: 0.7364\n",
      "Epoch 427/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7225 - val_loss: 0.6718 - val_accuracy: 0.7364\n",
      "Epoch 428/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7242 - val_loss: 0.6718 - val_accuracy: 0.7364\n",
      "Epoch 429/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7243 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 430/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7266 - val_loss: 0.6718 - val_accuracy: 0.7383\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 431/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7231 - val_loss: 0.6717 - val_accuracy: 0.7383\n",
      "Epoch 432/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7232 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 433/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7249 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 434/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7263 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 435/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7247 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 436/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7266 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 437/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7255 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 438/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7235 - val_loss: 0.6719 - val_accuracy: 0.7395\n",
      "Epoch 439/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7265 - val_loss: 0.6719 - val_accuracy: 0.7383\n",
      "Epoch 440/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7248 - val_loss: 0.6719 - val_accuracy: 0.7383\n",
      "Epoch 441/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7266 - val_loss: 0.6719 - val_accuracy: 0.7370\n",
      "Epoch 442/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7253 - val_loss: 0.6719 - val_accuracy: 0.7370\n",
      "Epoch 443/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7242 - val_loss: 0.6719 - val_accuracy: 0.7364\n",
      "Epoch 444/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7204 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 445/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7276 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 446/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7257 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 447/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7247 - val_loss: 0.6719 - val_accuracy: 0.7401\n",
      "Epoch 448/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7268 - val_loss: 0.6719 - val_accuracy: 0.7395\n",
      "Epoch 449/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7265 - val_loss: 0.6719 - val_accuracy: 0.7383\n",
      "Epoch 450/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7254 - val_loss: 0.6719 - val_accuracy: 0.7358\n",
      "Epoch 451/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7237 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 452/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7249 - val_loss: 0.6718 - val_accuracy: 0.7370\n",
      "Epoch 453/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7229 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 454/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7254 - val_loss: 0.6718 - val_accuracy: 0.7364\n",
      "Epoch 455/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7237 - val_loss: 0.6719 - val_accuracy: 0.7383\n",
      "Epoch 456/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7247 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 457/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7253 - val_loss: 0.6718 - val_accuracy: 0.7364\n",
      "Epoch 458/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7239 - val_loss: 0.6718 - val_accuracy: 0.7352\n",
      "Epoch 459/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7242 - val_loss: 0.6718 - val_accuracy: 0.7358\n",
      "Epoch 460/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7245 - val_loss: 0.6718 - val_accuracy: 0.7340\n",
      "Epoch 461/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7225 - val_loss: 0.6718 - val_accuracy: 0.7333\n",
      "Epoch 462/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7211 - val_loss: 0.6717 - val_accuracy: 0.7346\n",
      "Epoch 463/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7224 - val_loss: 0.6717 - val_accuracy: 0.7358\n",
      "Epoch 464/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7220 - val_loss: 0.6718 - val_accuracy: 0.7346\n",
      "Epoch 465/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7245 - val_loss: 0.6718 - val_accuracy: 0.7346\n",
      "Epoch 466/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7235 - val_loss: 0.6718 - val_accuracy: 0.7340\n",
      "Epoch 467/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7210 - val_loss: 0.6718 - val_accuracy: 0.7346\n",
      "Epoch 468/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7220 - val_loss: 0.6718 - val_accuracy: 0.7389\n",
      "Epoch 469/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7247 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 470/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7263 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 471/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7260 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 472/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7244 - val_loss: 0.6719 - val_accuracy: 0.7383\n",
      "Epoch 473/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7243 - val_loss: 0.6719 - val_accuracy: 0.7407\n",
      "Epoch 474/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7259 - val_loss: 0.6719 - val_accuracy: 0.7414\n",
      "Epoch 475/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7265 - val_loss: 0.6719 - val_accuracy: 0.7414\n",
      "Epoch 476/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7252 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 477/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7261 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 478/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7261 - val_loss: 0.6718 - val_accuracy: 0.7364\n",
      "Epoch 479/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7217 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 480/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7258 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 481/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7255 - val_loss: 0.6719 - val_accuracy: 0.7395\n",
      "Epoch 482/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7256 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 483/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7241 - val_loss: 0.6718 - val_accuracy: 0.7414\n",
      "Epoch 484/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7267 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 485/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7256 - val_loss: 0.6718 - val_accuracy: 0.7407\n",
      "Epoch 486/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7251 - val_loss: 0.6718 - val_accuracy: 0.7414\n",
      "Epoch 487/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7288 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 488/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7261 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 489/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7278 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 490/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7254 - val_loss: 0.6718 - val_accuracy: 0.7377\n",
      "Epoch 491/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7232 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 492/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7224 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 493/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7247 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 494/500\n",
      " - 0s - loss: 0.6787 - accuracy: 0.7230 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 495/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7255 - val_loss: 0.6718 - val_accuracy: 0.7395\n",
      "Epoch 496/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7256 - val_loss: 0.6718 - val_accuracy: 0.7383\n",
      "Epoch 497/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7245 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 498/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7248 - val_loss: 0.6718 - val_accuracy: 0.7401\n",
      "Epoch 499/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7239 - val_loss: 0.6718 - val_accuracy: 0.7414\n",
      "Epoch 500/500\n",
      " - 0s - loss: 0.6788 - accuracy: 0.7260 - val_loss: 0.6719 - val_accuracy: 0.7414\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.callbacks.History at 0x1cbe9975160>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(train[:, 1:], train[:, 0], epochs=500, batch_size=500, validation_split=0.15, verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_pred = model.predict(test[:, 1:]).reshape((-1,)) >= 0.5\n",
    "\n",
    "conf_mat = confusion_matrix(y_test, y_pred)\n",
    "\n",
    "lang = ['No Fault (0)', 'Fault (1)']\n",
    "ax = plt.axes()\n",
    "sns.heatmap(conf_mat, square=True, annot=True, cmap='Blues', fmt='d', cbar=True, ax=ax, xticklabels=lang, yticklabels=lang)\n",
    "ax.set_title('Confusion Matrix')\n",
    "plt.xlabel('Predicted Label')\n",
    "plt.ylabel('True Label')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The Evaluation Metrics of MLP Model-1 (on the test set) is as follows:\n",
      "\n",
      "True Positives: 234\n",
      "True Negatives: 651\n",
      "False Positives: 0\n",
      "False Negatives: 316\n",
      "\n",
      "Precision: 1.0\n",
      "Recall: 0.4254545454545455\n",
      "\n",
      "Accuracy: 0.7368859283930058\n"
     ]
    }
   ],
   "source": [
    "true_positive = conf_mat[1, 1]\n",
    "true_negative = conf_mat[0, 0]\n",
    "false_positive = conf_mat[0, 1]\n",
    "false_negative = conf_mat[1, 0]\n",
    "accuracy = accuracy_score(y_pred, y_test)\n",
    "\n",
    "print(\"The Evaluation Metrics of MLP Model-1 (on the test set) is as follows:\\n\")\n",
    "\n",
    "print(\"True Positives:\", true_positive)\n",
    "print(\"True Negatives:\", true_negative)\n",
    "print(\"False Positives:\", false_positive)\n",
    "print(\"False Negatives:\", false_negative, end='\\n\\n')\n",
    "\n",
    "print(\"Precision:\", true_positive/(true_positive+false_positive))\n",
    "print(\"Recall:\", true_positive/(true_positive+false_negative), end='\\n\\n')\n",
    "\n",
    "print(\"Accuracy:\", accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### (ii) Model 2: (1 hidden layer, relu activation, Adam optimizer)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "model.add(Dense(10, input_dim=6, activation='relu'))\n",
    "model.add(Dense(1, activation='sigmoid'))\n",
    "model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/500\n",
      " - 0s - loss: 0.7338 - accuracy: 0.4436\n",
      "Epoch 2/500\n",
      " - 0s - loss: 0.6813 - accuracy: 0.4496\n",
      "Epoch 3/500\n",
      " - 0s - loss: 0.6377 - accuracy: 0.4648\n",
      "Epoch 4/500\n",
      " - 0s - loss: 0.6003 - accuracy: 0.5514\n",
      "Epoch 5/500\n",
      " - 0s - loss: 0.5675 - accuracy: 0.6481\n",
      "Epoch 6/500\n",
      " - 0s - loss: 0.5375 - accuracy: 0.7394\n",
      "Epoch 7/500\n",
      " - 0s - loss: 0.5096 - accuracy: 0.8161\n",
      "Epoch 8/500\n",
      " - 0s - loss: 0.4833 - accuracy: 0.8994\n",
      "Epoch 9/500\n",
      " - 0s - loss: 0.4580 - accuracy: 0.9323\n",
      "Epoch 10/500\n",
      " - 0s - loss: 0.4338 - accuracy: 0.9526\n",
      "Epoch 11/500\n",
      " - 0s - loss: 0.4113 - accuracy: 0.9666\n",
      "Epoch 12/500\n",
      " - 0s - loss: 0.3900 - accuracy: 0.9678\n",
      "Epoch 13/500\n",
      " - 0s - loss: 0.3701 - accuracy: 0.9676\n",
      "Epoch 14/500\n",
      " - 0s - loss: 0.3513 - accuracy: 0.9680\n",
      "Epoch 15/500\n",
      " - 0s - loss: 0.3334 - accuracy: 0.9681\n",
      "Epoch 16/500\n",
      " - 0s - loss: 0.3166 - accuracy: 0.9675\n",
      "Epoch 17/500\n",
      " - 0s - loss: 0.3009 - accuracy: 0.9672\n",
      "Epoch 18/500\n",
      " - 0s - loss: 0.2863 - accuracy: 0.9671\n",
      "Epoch 19/500\n",
      " - 0s - loss: 0.2729 - accuracy: 0.9673\n",
      "Epoch 20/500\n",
      " - 0s - loss: 0.2605 - accuracy: 0.9671\n",
      "Epoch 21/500\n",
      " - 0s - loss: 0.2490 - accuracy: 0.9669\n",
      "Epoch 22/500\n",
      " - 0s - loss: 0.2385 - accuracy: 0.9669\n",
      "Epoch 23/500\n",
      " - 0s - loss: 0.2287 - accuracy: 0.9669\n",
      "Epoch 24/500\n",
      " - 0s - loss: 0.2198 - accuracy: 0.9667\n",
      "Epoch 25/500\n",
      " - 0s - loss: 0.2118 - accuracy: 0.9667\n",
      "Epoch 26/500\n",
      " - 0s - loss: 0.2045 - accuracy: 0.9671\n",
      "Epoch 27/500\n",
      " - 0s - loss: 0.1979 - accuracy: 0.9670\n",
      "Epoch 28/500\n",
      " - 0s - loss: 0.1918 - accuracy: 0.9671\n",
      "Epoch 29/500\n",
      " - 0s - loss: 0.1861 - accuracy: 0.9672\n",
      "Epoch 30/500\n",
      " - 0s - loss: 0.1808 - accuracy: 0.9673\n",
      "Epoch 31/500\n",
      " - 0s - loss: 0.1759 - accuracy: 0.9673\n",
      "Epoch 32/500\n",
      " - 0s - loss: 0.1712 - accuracy: 0.9676\n",
      "Epoch 33/500\n",
      " - 0s - loss: 0.1669 - accuracy: 0.9677\n",
      "Epoch 34/500\n",
      " - 0s - loss: 0.1627 - accuracy: 0.9677\n",
      "Epoch 35/500\n",
      " - 0s - loss: 0.1587 - accuracy: 0.9678\n",
      "Epoch 36/500\n",
      " - 0s - loss: 0.1550 - accuracy: 0.9680\n",
      "Epoch 37/500\n",
      " - 0s - loss: 0.1513 - accuracy: 0.9681\n",
      "Epoch 38/500\n",
      " - 0s - loss: 0.1479 - accuracy: 0.9684\n",
      "Epoch 39/500\n",
      " - 0s - loss: 0.1446 - accuracy: 0.9684\n",
      "Epoch 40/500\n",
      " - 0s - loss: 0.1413 - accuracy: 0.9686\n",
      "Epoch 41/500\n",
      " - 0s - loss: 0.1382 - accuracy: 0.9686\n",
      "Epoch 42/500\n",
      " - 0s - loss: 0.1351 - accuracy: 0.9688\n",
      "Epoch 43/500\n",
      " - 0s - loss: 0.1322 - accuracy: 0.9694\n",
      "Epoch 44/500\n",
      " - 0s - loss: 0.1293 - accuracy: 0.9695\n",
      "Epoch 45/500\n",
      " - 0s - loss: 0.1265 - accuracy: 0.9697\n",
      "Epoch 46/500\n",
      " - 0s - loss: 0.1238 - accuracy: 0.9698\n",
      "Epoch 47/500\n",
      " - 0s - loss: 0.1211 - accuracy: 0.9698\n",
      "Epoch 48/500\n",
      " - 0s - loss: 0.1185 - accuracy: 0.9702\n",
      "Epoch 49/500\n",
      " - 0s - loss: 0.1161 - accuracy: 0.9704\n",
      "Epoch 50/500\n",
      " - 0s - loss: 0.1136 - accuracy: 0.9708\n",
      "Epoch 51/500\n",
      " - 0s - loss: 0.1112 - accuracy: 0.9710\n",
      "Epoch 52/500\n",
      " - 0s - loss: 0.1089 - accuracy: 0.9714\n",
      "Epoch 53/500\n",
      " - 0s - loss: 0.1066 - accuracy: 0.9719\n",
      "Epoch 54/500\n",
      " - 0s - loss: 0.1044 - accuracy: 0.9722\n",
      "Epoch 55/500\n",
      " - 0s - loss: 0.1023 - accuracy: 0.9726\n",
      "Epoch 56/500\n",
      " - 0s - loss: 0.1003 - accuracy: 0.9729\n",
      "Epoch 57/500\n",
      " - 0s - loss: 0.0984 - accuracy: 0.9734\n",
      "Epoch 58/500\n",
      " - 0s - loss: 0.0965 - accuracy: 0.9736\n",
      "Epoch 59/500\n",
      " - 0s - loss: 0.0947 - accuracy: 0.9738\n",
      "Epoch 60/500\n",
      " - 0s - loss: 0.0930 - accuracy: 0.9741\n",
      "Epoch 61/500\n",
      " - 0s - loss: 0.0913 - accuracy: 0.9743\n",
      "Epoch 62/500\n",
      " - 0s - loss: 0.0897 - accuracy: 0.9745\n",
      "Epoch 63/500\n",
      " - 0s - loss: 0.0883 - accuracy: 0.9748\n",
      "Epoch 64/500\n",
      " - 0s - loss: 0.0869 - accuracy: 0.9752\n",
      "Epoch 65/500\n",
      " - 0s - loss: 0.0855 - accuracy: 0.9755\n",
      "Epoch 66/500\n",
      " - 0s - loss: 0.0842 - accuracy: 0.9758\n",
      "Epoch 67/500\n",
      " - 0s - loss: 0.0830 - accuracy: 0.9758\n",
      "Epoch 68/500\n",
      " - 0s - loss: 0.0819 - accuracy: 0.9764\n",
      "Epoch 69/500\n",
      " - 0s - loss: 0.0808 - accuracy: 0.9766\n",
      "Epoch 70/500\n",
      " - 0s - loss: 0.0798 - accuracy: 0.9766\n",
      "Epoch 71/500\n",
      " - 0s - loss: 0.0788 - accuracy: 0.9767\n",
      "Epoch 72/500\n",
      " - 0s - loss: 0.0779 - accuracy: 0.9768\n",
      "Epoch 73/500\n",
      " - 0s - loss: 0.0770 - accuracy: 0.9772\n",
      "Epoch 74/500\n",
      " - 0s - loss: 0.0762 - accuracy: 0.9775\n",
      "Epoch 75/500\n",
      " - 0s - loss: 0.0754 - accuracy: 0.9776\n",
      "Epoch 76/500\n",
      " - 0s - loss: 0.0746 - accuracy: 0.9776\n",
      "Epoch 77/500\n",
      " - 0s - loss: 0.0738 - accuracy: 0.9778\n",
      "Epoch 78/500\n",
      " - 0s - loss: 0.0731 - accuracy: 0.9780\n",
      "Epoch 79/500\n",
      " - 0s - loss: 0.0724 - accuracy: 0.9780\n",
      "Epoch 80/500\n",
      " - 0s - loss: 0.0717 - accuracy: 0.9782\n",
      "Epoch 81/500\n",
      " - 0s - loss: 0.0711 - accuracy: 0.9782\n",
      "Epoch 82/500\n",
      " - 0s - loss: 0.0704 - accuracy: 0.9785\n",
      "Epoch 83/500\n",
      " - 0s - loss: 0.0698 - accuracy: 0.9787\n",
      "Epoch 84/500\n",
      " - 0s - loss: 0.0692 - accuracy: 0.9790\n",
      "Epoch 85/500\n",
      " - 0s - loss: 0.0686 - accuracy: 0.9791\n",
      "Epoch 86/500\n",
      " - 0s - loss: 0.0680 - accuracy: 0.9792\n",
      "Epoch 87/500\n",
      " - 0s - loss: 0.0674 - accuracy: 0.9793\n",
      "Epoch 88/500\n",
      " - 0s - loss: 0.0669 - accuracy: 0.9795\n",
      "Epoch 89/500\n",
      " - 0s - loss: 0.0663 - accuracy: 0.9797\n",
      "Epoch 90/500\n",
      " - 0s - loss: 0.0658 - accuracy: 0.9798\n",
      "Epoch 91/500\n",
      " - 0s - loss: 0.0652 - accuracy: 0.9799\n",
      "Epoch 92/500\n",
      " - 0s - loss: 0.0647 - accuracy: 0.9801\n",
      "Epoch 93/500\n",
      " - 0s - loss: 0.0642 - accuracy: 0.9802\n",
      "Epoch 94/500\n",
      " - 0s - loss: 0.0637 - accuracy: 0.9806\n",
      "Epoch 95/500\n",
      " - 0s - loss: 0.0632 - accuracy: 0.9806\n",
      "Epoch 96/500\n",
      " - 0s - loss: 0.0628 - accuracy: 0.9807\n",
      "Epoch 97/500\n",
      " - 0s - loss: 0.0623 - accuracy: 0.9811\n",
      "Epoch 98/500\n",
      " - 0s - loss: 0.0618 - accuracy: 0.9812\n",
      "Epoch 99/500\n",
      " - 0s - loss: 0.0614 - accuracy: 0.9814\n",
      "Epoch 100/500\n",
      " - 0s - loss: 0.0609 - accuracy: 0.9814\n",
      "Epoch 101/500\n",
      " - 0s - loss: 0.0605 - accuracy: 0.9816\n",
      "Epoch 102/500\n",
      " - 0s - loss: 0.0600 - accuracy: 0.9819\n",
      "Epoch 103/500\n",
      " - 0s - loss: 0.0597 - accuracy: 0.9820\n",
      "Epoch 104/500\n",
      " - 0s - loss: 0.0592 - accuracy: 0.9820\n",
      "Epoch 105/500\n",
      " - 0s - loss: 0.0588 - accuracy: 0.9821\n",
      "Epoch 106/500\n",
      " - 0s - loss: 0.0584 - accuracy: 0.9821\n",
      "Epoch 107/500\n",
      " - 0s - loss: 0.0580 - accuracy: 0.9824\n",
      "Epoch 108/500\n",
      " - 0s - loss: 0.0576 - accuracy: 0.9825\n",
      "Epoch 109/500\n",
      " - 0s - loss: 0.0572 - accuracy: 0.9829\n",
      "Epoch 110/500\n",
      " - 0s - loss: 0.0568 - accuracy: 0.9829\n",
      "Epoch 111/500\n",
      " - 0s - loss: 0.0565 - accuracy: 0.9831\n",
      "Epoch 112/500\n",
      " - 0s - loss: 0.0561 - accuracy: 0.9832\n",
      "Epoch 113/500\n",
      " - 0s - loss: 0.0557 - accuracy: 0.9832\n",
      "Epoch 114/500\n",
      " - 0s - loss: 0.0553 - accuracy: 0.9833\n",
      "Epoch 115/500\n",
      " - 0s - loss: 0.0550 - accuracy: 0.9835\n",
      "Epoch 116/500\n",
      " - 0s - loss: 0.0547 - accuracy: 0.9835\n",
      "Epoch 117/500\n",
      " - 0s - loss: 0.0543 - accuracy: 0.9836\n",
      "Epoch 118/500\n",
      " - 0s - loss: 0.0540 - accuracy: 0.9836\n",
      "Epoch 119/500\n",
      " - 0s - loss: 0.0536 - accuracy: 0.9836\n",
      "Epoch 120/500\n",
      " - 0s - loss: 0.0533 - accuracy: 0.9839\n",
      "Epoch 121/500\n",
      " - 0s - loss: 0.0530 - accuracy: 0.9838\n",
      "Epoch 122/500\n",
      " - 0s - loss: 0.0527 - accuracy: 0.9841\n",
      "Epoch 123/500\n",
      " - 0s - loss: 0.0524 - accuracy: 0.9841\n",
      "Epoch 124/500\n",
      " - 0s - loss: 0.0521 - accuracy: 0.9841\n",
      "Epoch 125/500\n",
      " - 0s - loss: 0.0517 - accuracy: 0.9841\n",
      "Epoch 126/500\n",
      " - 0s - loss: 0.0515 - accuracy: 0.9844\n",
      "Epoch 127/500\n",
      " - 0s - loss: 0.0512 - accuracy: 0.9843\n",
      "Epoch 128/500\n",
      " - 0s - loss: 0.0509 - accuracy: 0.9844\n",
      "Epoch 129/500\n",
      " - 0s - loss: 0.0506 - accuracy: 0.9845\n",
      "Epoch 130/500\n",
      " - 0s - loss: 0.0503 - accuracy: 0.9844\n",
      "Epoch 131/500\n",
      " - 0s - loss: 0.0500 - accuracy: 0.9845\n",
      "Epoch 132/500\n",
      " - 0s - loss: 0.0497 - accuracy: 0.9846\n",
      "Epoch 133/500\n",
      " - 0s - loss: 0.0494 - accuracy: 0.9846\n",
      "Epoch 134/500\n",
      " - 0s - loss: 0.0492 - accuracy: 0.9846\n",
      "Epoch 135/500\n",
      " - 0s - loss: 0.0490 - accuracy: 0.9849\n",
      "Epoch 136/500\n",
      " - 0s - loss: 0.0487 - accuracy: 0.9848\n",
      "Epoch 137/500\n",
      " - 0s - loss: 0.0484 - accuracy: 0.9849\n",
      "Epoch 138/500\n",
      " - 0s - loss: 0.0481 - accuracy: 0.9849\n",
      "Epoch 139/500\n",
      " - 0s - loss: 0.0479 - accuracy: 0.9849\n",
      "Epoch 140/500\n",
      " - 0s - loss: 0.0476 - accuracy: 0.9849\n",
      "Epoch 141/500\n",
      " - 0s - loss: 0.0474 - accuracy: 0.9851\n",
      "Epoch 142/500\n",
      " - 0s - loss: 0.0472 - accuracy: 0.9851\n",
      "Epoch 143/500\n",
      " - 0s - loss: 0.0469 - accuracy: 0.9851\n",
      "Epoch 144/500\n",
      " - 0s - loss: 0.0467 - accuracy: 0.9852\n",
      "Epoch 145/500\n",
      " - 0s - loss: 0.0465 - accuracy: 0.9851\n",
      "Epoch 146/500\n",
      " - 0s - loss: 0.0463 - accuracy: 0.9852\n",
      "Epoch 147/500\n",
      " - 0s - loss: 0.0460 - accuracy: 0.9853\n",
      "Epoch 148/500\n",
      " - 0s - loss: 0.0458 - accuracy: 0.9854\n",
      "Epoch 149/500\n",
      " - 0s - loss: 0.0456 - accuracy: 0.9855\n",
      "Epoch 150/500\n",
      " - 0s - loss: 0.0454 - accuracy: 0.9854\n",
      "Epoch 151/500\n",
      " - 0s - loss: 0.0453 - accuracy: 0.9854\n",
      "Epoch 152/500\n",
      " - 0s - loss: 0.0451 - accuracy: 0.9854\n",
      "Epoch 153/500\n",
      " - 0s - loss: 0.0448 - accuracy: 0.9854\n",
      "Epoch 154/500\n",
      " - 0s - loss: 0.0447 - accuracy: 0.9857\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 155/500\n",
      " - 0s - loss: 0.0445 - accuracy: 0.9856\n",
      "Epoch 156/500\n",
      " - 0s - loss: 0.0443 - accuracy: 0.9857\n",
      "Epoch 157/500\n",
      " - 0s - loss: 0.0442 - accuracy: 0.9859\n",
      "Epoch 158/500\n",
      " - 0s - loss: 0.0440 - accuracy: 0.9858\n",
      "Epoch 159/500\n",
      " - 0s - loss: 0.0438 - accuracy: 0.9859\n",
      "Epoch 160/500\n",
      " - 0s - loss: 0.0436 - accuracy: 0.9861\n",
      "Epoch 161/500\n",
      " - 0s - loss: 0.0434 - accuracy: 0.9862\n",
      "Epoch 162/500\n",
      " - 0s - loss: 0.0432 - accuracy: 0.9865\n",
      "Epoch 163/500\n",
      " - 0s - loss: 0.0431 - accuracy: 0.9865\n",
      "Epoch 164/500\n",
      " - 0s - loss: 0.0429 - accuracy: 0.9866\n",
      "Epoch 165/500\n",
      " - 0s - loss: 0.0428 - accuracy: 0.9868\n",
      "Epoch 166/500\n",
      " - 0s - loss: 0.0426 - accuracy: 0.9868\n",
      "Epoch 167/500\n",
      " - 0s - loss: 0.0424 - accuracy: 0.9869\n",
      "Epoch 168/500\n",
      " - 0s - loss: 0.0423 - accuracy: 0.9869\n",
      "Epoch 169/500\n",
      " - 0s - loss: 0.0421 - accuracy: 0.9870\n",
      "Epoch 170/500\n",
      " - 0s - loss: 0.0420 - accuracy: 0.9870\n",
      "Epoch 171/500\n",
      " - 0s - loss: 0.0418 - accuracy: 0.9869\n",
      "Epoch 172/500\n",
      " - 0s - loss: 0.0417 - accuracy: 0.9870\n",
      "Epoch 173/500\n",
      " - 0s - loss: 0.0415 - accuracy: 0.9869\n",
      "Epoch 174/500\n",
      " - 0s - loss: 0.0414 - accuracy: 0.9870\n",
      "Epoch 175/500\n",
      " - 0s - loss: 0.0412 - accuracy: 0.9869\n",
      "Epoch 176/500\n",
      " - 0s - loss: 0.0412 - accuracy: 0.9869\n",
      "Epoch 177/500\n",
      " - 0s - loss: 0.0410 - accuracy: 0.9871\n",
      "Epoch 178/500\n",
      " - 0s - loss: 0.0408 - accuracy: 0.9871\n",
      "Epoch 179/500\n",
      " - 0s - loss: 0.0407 - accuracy: 0.9871\n",
      "Epoch 180/500\n",
      " - 0s - loss: 0.0406 - accuracy: 0.9871\n",
      "Epoch 181/500\n",
      " - 0s - loss: 0.0404 - accuracy: 0.9871\n",
      "Epoch 182/500\n",
      " - 0s - loss: 0.0403 - accuracy: 0.9872\n",
      "Epoch 183/500\n",
      " - 0s - loss: 0.0402 - accuracy: 0.9873\n",
      "Epoch 184/500\n",
      " - 0s - loss: 0.0401 - accuracy: 0.9874\n",
      "Epoch 185/500\n",
      " - 0s - loss: 0.0399 - accuracy: 0.9876\n",
      "Epoch 186/500\n",
      " - 0s - loss: 0.0398 - accuracy: 0.9876\n",
      "Epoch 187/500\n",
      " - 0s - loss: 0.0397 - accuracy: 0.9877\n",
      "Epoch 188/500\n",
      " - 0s - loss: 0.0395 - accuracy: 0.9878\n",
      "Epoch 189/500\n",
      " - 0s - loss: 0.0395 - accuracy: 0.9879\n",
      "Epoch 190/500\n",
      " - 0s - loss: 0.0393 - accuracy: 0.9879\n",
      "Epoch 191/500\n",
      " - 0s - loss: 0.0392 - accuracy: 0.9878\n",
      "Epoch 192/500\n",
      " - 0s - loss: 0.0391 - accuracy: 0.9878\n",
      "Epoch 193/500\n",
      " - 0s - loss: 0.0390 - accuracy: 0.9880\n",
      "Epoch 194/500\n",
      " - 0s - loss: 0.0389 - accuracy: 0.9881\n",
      "Epoch 195/500\n",
      " - 0s - loss: 0.0388 - accuracy: 0.9879\n",
      "Epoch 196/500\n",
      " - 0s - loss: 0.0387 - accuracy: 0.9880\n",
      "Epoch 197/500\n",
      " - 0s - loss: 0.0386 - accuracy: 0.9880\n",
      "Epoch 198/500\n",
      " - 0s - loss: 0.0384 - accuracy: 0.9881\n",
      "Epoch 199/500\n",
      " - 0s - loss: 0.0384 - accuracy: 0.9881\n",
      "Epoch 200/500\n",
      " - 0s - loss: 0.0382 - accuracy: 0.9881\n",
      "Epoch 201/500\n",
      " - 0s - loss: 0.0381 - accuracy: 0.9881\n",
      "Epoch 202/500\n",
      " - 0s - loss: 0.0380 - accuracy: 0.9881\n",
      "Epoch 203/500\n",
      " - 0s - loss: 0.0379 - accuracy: 0.9881\n",
      "Epoch 204/500\n",
      " - 0s - loss: 0.0378 - accuracy: 0.9881\n",
      "Epoch 205/500\n",
      " - 0s - loss: 0.0377 - accuracy: 0.9881\n",
      "Epoch 206/500\n",
      " - 0s - loss: 0.0376 - accuracy: 0.9881\n",
      "Epoch 207/500\n",
      " - 0s - loss: 0.0376 - accuracy: 0.9882\n",
      "Epoch 208/500\n",
      " - 0s - loss: 0.0375 - accuracy: 0.9885\n",
      "Epoch 209/500\n",
      " - 0s - loss: 0.0374 - accuracy: 0.9885\n",
      "Epoch 210/500\n",
      " - 0s - loss: 0.0373 - accuracy: 0.9885\n",
      "Epoch 211/500\n",
      " - 0s - loss: 0.0372 - accuracy: 0.9885\n",
      "Epoch 212/500\n",
      " - 0s - loss: 0.0371 - accuracy: 0.9885\n",
      "Epoch 213/500\n",
      " - 0s - loss: 0.0370 - accuracy: 0.9885\n",
      "Epoch 214/500\n",
      " - 0s - loss: 0.0369 - accuracy: 0.9885\n",
      "Epoch 215/500\n",
      " - 0s - loss: 0.0369 - accuracy: 0.9885\n",
      "Epoch 216/500\n",
      " - 0s - loss: 0.0368 - accuracy: 0.9886\n",
      "Epoch 217/500\n",
      " - 0s - loss: 0.0367 - accuracy: 0.9888\n",
      "Epoch 218/500\n",
      " - 0s - loss: 0.0366 - accuracy: 0.9887\n",
      "Epoch 219/500\n",
      " - 0s - loss: 0.0366 - accuracy: 0.9887\n",
      "Epoch 220/500\n",
      " - 0s - loss: 0.0364 - accuracy: 0.9887\n",
      "Epoch 221/500\n",
      " - 0s - loss: 0.0364 - accuracy: 0.9887\n",
      "Epoch 222/500\n",
      " - 0s - loss: 0.0363 - accuracy: 0.9887\n",
      "Epoch 223/500\n",
      " - 0s - loss: 0.0362 - accuracy: 0.9887\n",
      "Epoch 224/500\n",
      " - 0s - loss: 0.0362 - accuracy: 0.9887\n",
      "Epoch 225/500\n",
      " - 0s - loss: 0.0361 - accuracy: 0.9888\n",
      "Epoch 226/500\n",
      " - 0s - loss: 0.0360 - accuracy: 0.9888\n",
      "Epoch 227/500\n",
      " - 0s - loss: 0.0359 - accuracy: 0.9889\n",
      "Epoch 228/500\n",
      " - 0s - loss: 0.0358 - accuracy: 0.9889\n",
      "Epoch 229/500\n",
      " - 0s - loss: 0.0358 - accuracy: 0.9889\n",
      "Epoch 230/500\n",
      " - 0s - loss: 0.0357 - accuracy: 0.9889\n",
      "Epoch 231/500\n",
      " - 0s - loss: 0.0356 - accuracy: 0.9888\n",
      "Epoch 232/500\n",
      " - 0s - loss: 0.0356 - accuracy: 0.9888\n",
      "Epoch 233/500\n",
      " - 0s - loss: 0.0355 - accuracy: 0.9889\n",
      "Epoch 234/500\n",
      " - 0s - loss: 0.0354 - accuracy: 0.9888\n",
      "Epoch 235/500\n",
      " - 0s - loss: 0.0354 - accuracy: 0.9888\n",
      "Epoch 236/500\n",
      " - 0s - loss: 0.0353 - accuracy: 0.9888\n",
      "Epoch 237/500\n",
      " - 0s - loss: 0.0353 - accuracy: 0.9888\n",
      "Epoch 238/500\n",
      " - 0s - loss: 0.0352 - accuracy: 0.9888\n",
      "Epoch 239/500\n",
      " - 0s - loss: 0.0351 - accuracy: 0.9889\n",
      "Epoch 240/500\n",
      " - 0s - loss: 0.0351 - accuracy: 0.9889\n",
      "Epoch 241/500\n",
      " - 0s - loss: 0.0350 - accuracy: 0.9889\n",
      "Epoch 242/500\n",
      " - 0s - loss: 0.0350 - accuracy: 0.9889\n",
      "Epoch 243/500\n",
      " - 0s - loss: 0.0349 - accuracy: 0.9889\n",
      "Epoch 244/500\n",
      " - 0s - loss: 0.0348 - accuracy: 0.9890\n",
      "Epoch 245/500\n",
      " - 0s - loss: 0.0348 - accuracy: 0.9890\n",
      "Epoch 246/500\n",
      " - 0s - loss: 0.0347 - accuracy: 0.9890\n",
      "Epoch 247/500\n",
      " - 0s - loss: 0.0347 - accuracy: 0.9890\n",
      "Epoch 248/500\n",
      " - 0s - loss: 0.0346 - accuracy: 0.9890\n",
      "Epoch 249/500\n",
      " - 0s - loss: 0.0346 - accuracy: 0.9890\n",
      "Epoch 250/500\n",
      " - 0s - loss: 0.0345 - accuracy: 0.9889\n",
      "Epoch 251/500\n",
      " - 0s - loss: 0.0344 - accuracy: 0.9890\n",
      "Epoch 252/500\n",
      " - 0s - loss: 0.0344 - accuracy: 0.9890\n",
      "Epoch 253/500\n",
      " - 0s - loss: 0.0344 - accuracy: 0.9889\n",
      "Epoch 254/500\n",
      " - 0s - loss: 0.0343 - accuracy: 0.9889\n",
      "Epoch 255/500\n",
      " - 0s - loss: 0.0343 - accuracy: 0.9889\n",
      "Epoch 256/500\n",
      " - 0s - loss: 0.0343 - accuracy: 0.9890\n",
      "Epoch 257/500\n",
      " - 0s - loss: 0.0342 - accuracy: 0.9893\n",
      "Epoch 258/500\n",
      " - 0s - loss: 0.0341 - accuracy: 0.9892\n",
      "Epoch 259/500\n",
      " - 0s - loss: 0.0341 - accuracy: 0.9892\n",
      "Epoch 260/500\n",
      " - 0s - loss: 0.0340 - accuracy: 0.9893\n",
      "Epoch 261/500\n",
      " - 0s - loss: 0.0340 - accuracy: 0.9894\n",
      "Epoch 262/500\n",
      " - 0s - loss: 0.0339 - accuracy: 0.9894\n",
      "Epoch 263/500\n",
      " - 0s - loss: 0.0339 - accuracy: 0.9894\n",
      "Epoch 264/500\n",
      " - 0s - loss: 0.0338 - accuracy: 0.9895\n",
      "Epoch 265/500\n",
      " - 0s - loss: 0.0338 - accuracy: 0.9895\n",
      "Epoch 266/500\n",
      " - 0s - loss: 0.0338 - accuracy: 0.9894\n",
      "Epoch 267/500\n",
      " - 0s - loss: 0.0337 - accuracy: 0.9895\n",
      "Epoch 268/500\n",
      " - 0s - loss: 0.0337 - accuracy: 0.9895\n",
      "Epoch 269/500\n",
      " - 0s - loss: 0.0337 - accuracy: 0.9895\n",
      "Epoch 270/500\n",
      " - 0s - loss: 0.0336 - accuracy: 0.9894\n",
      "Epoch 271/500\n",
      " - 0s - loss: 0.0336 - accuracy: 0.9896\n",
      "Epoch 272/500\n",
      " - 0s - loss: 0.0335 - accuracy: 0.9895\n",
      "Epoch 273/500\n",
      " - 0s - loss: 0.0335 - accuracy: 0.9896\n",
      "Epoch 274/500\n",
      " - 0s - loss: 0.0335 - accuracy: 0.9895\n",
      "Epoch 275/500\n",
      " - 0s - loss: 0.0334 - accuracy: 0.9896\n",
      "Epoch 276/500\n",
      " - 0s - loss: 0.0334 - accuracy: 0.9898\n",
      "Epoch 277/500\n",
      " - 0s - loss: 0.0334 - accuracy: 0.9895\n",
      "Epoch 278/500\n",
      " - 0s - loss: 0.0333 - accuracy: 0.9896\n",
      "Epoch 279/500\n",
      " - 0s - loss: 0.0333 - accuracy: 0.9898\n",
      "Epoch 280/500\n",
      " - 0s - loss: 0.0332 - accuracy: 0.9896\n",
      "Epoch 281/500\n",
      " - 0s - loss: 0.0332 - accuracy: 0.9898\n",
      "Epoch 282/500\n",
      " - 0s - loss: 0.0332 - accuracy: 0.9897\n",
      "Epoch 283/500\n",
      " - 0s - loss: 0.0332 - accuracy: 0.9899\n",
      "Epoch 284/500\n",
      " - 0s - loss: 0.0331 - accuracy: 0.9899\n",
      "Epoch 285/500\n",
      " - 0s - loss: 0.0331 - accuracy: 0.9899\n",
      "Epoch 286/500\n",
      " - 0s - loss: 0.0331 - accuracy: 0.9899\n",
      "Epoch 287/500\n",
      " - 0s - loss: 0.0330 - accuracy: 0.9899\n",
      "Epoch 288/500\n",
      " - 0s - loss: 0.0330 - accuracy: 0.9899\n",
      "Epoch 289/500\n",
      " - 0s - loss: 0.0329 - accuracy: 0.9899\n",
      "Epoch 290/500\n",
      " - 0s - loss: 0.0329 - accuracy: 0.9899\n",
      "Epoch 291/500\n",
      " - 0s - loss: 0.0328 - accuracy: 0.9899\n",
      "Epoch 292/500\n",
      " - 0s - loss: 0.0329 - accuracy: 0.9899\n",
      "Epoch 293/500\n",
      " - 0s - loss: 0.0328 - accuracy: 0.9899\n",
      "Epoch 294/500\n",
      " - 0s - loss: 0.0328 - accuracy: 0.9899\n",
      "Epoch 295/500\n",
      " - 0s - loss: 0.0327 - accuracy: 0.9899\n",
      "Epoch 296/500\n",
      " - 0s - loss: 0.0327 - accuracy: 0.9899\n",
      "Epoch 297/500\n",
      " - 0s - loss: 0.0327 - accuracy: 0.9899\n",
      "Epoch 298/500\n",
      " - 0s - loss: 0.0327 - accuracy: 0.9899\n",
      "Epoch 299/500\n",
      " - 0s - loss: 0.0326 - accuracy: 0.9899\n",
      "Epoch 300/500\n",
      " - 0s - loss: 0.0326 - accuracy: 0.9899\n",
      "Epoch 301/500\n",
      " - 0s - loss: 0.0326 - accuracy: 0.9899\n",
      "Epoch 302/500\n",
      " - 0s - loss: 0.0325 - accuracy: 0.9899\n",
      "Epoch 303/500\n",
      " - 0s - loss: 0.0325 - accuracy: 0.9899\n",
      "Epoch 304/500\n",
      " - 0s - loss: 0.0324 - accuracy: 0.9899\n",
      "Epoch 305/500\n",
      " - 0s - loss: 0.0324 - accuracy: 0.9899\n",
      "Epoch 306/500\n",
      " - 0s - loss: 0.0324 - accuracy: 0.9899\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 307/500\n",
      " - 0s - loss: 0.0324 - accuracy: 0.9899\n",
      "Epoch 308/500\n",
      " - 0s - loss: 0.0324 - accuracy: 0.9898\n",
      "Epoch 309/500\n",
      " - 0s - loss: 0.0323 - accuracy: 0.9899\n",
      "Epoch 310/500\n",
      " - 0s - loss: 0.0323 - accuracy: 0.9899\n",
      "Epoch 311/500\n",
      " - 0s - loss: 0.0323 - accuracy: 0.9899\n",
      "Epoch 312/500\n",
      " - 0s - loss: 0.0322 - accuracy: 0.9899\n",
      "Epoch 313/500\n",
      " - 0s - loss: 0.0322 - accuracy: 0.9899\n",
      "Epoch 314/500\n",
      " - 0s - loss: 0.0322 - accuracy: 0.9901\n",
      "Epoch 315/500\n",
      " - 0s - loss: 0.0322 - accuracy: 0.9899\n",
      "Epoch 316/500\n",
      " - 0s - loss: 0.0321 - accuracy: 0.9901\n",
      "Epoch 317/500\n",
      " - 0s - loss: 0.0321 - accuracy: 0.9901\n",
      "Epoch 318/500\n",
      " - 0s - loss: 0.0321 - accuracy: 0.9901\n",
      "Epoch 319/500\n",
      " - 0s - loss: 0.0322 - accuracy: 0.9898\n",
      "Epoch 320/500\n",
      " - 0s - loss: 0.0320 - accuracy: 0.9900\n",
      "Epoch 321/500\n",
      " - 0s - loss: 0.0320 - accuracy: 0.9901\n",
      "Epoch 322/500\n",
      " - 0s - loss: 0.0320 - accuracy: 0.9901\n",
      "Epoch 323/500\n",
      " - 0s - loss: 0.0319 - accuracy: 0.9901\n",
      "Epoch 324/500\n",
      " - 0s - loss: 0.0319 - accuracy: 0.9900\n",
      "Epoch 325/500\n",
      " - 0s - loss: 0.0319 - accuracy: 0.9900\n",
      "Epoch 326/500\n",
      " - 0s - loss: 0.0318 - accuracy: 0.9901\n",
      "Epoch 327/500\n",
      " - 0s - loss: 0.0318 - accuracy: 0.9900\n",
      "Epoch 328/500\n",
      " - 0s - loss: 0.0318 - accuracy: 0.9900\n",
      "Epoch 329/500\n",
      " - 0s - loss: 0.0318 - accuracy: 0.9900\n",
      "Epoch 330/500\n",
      " - 0s - loss: 0.0318 - accuracy: 0.9900\n",
      "Epoch 331/500\n",
      " - 0s - loss: 0.0317 - accuracy: 0.9900\n",
      "Epoch 332/500\n",
      " - 0s - loss: 0.0317 - accuracy: 0.9900\n",
      "Epoch 333/500\n",
      " - 0s - loss: 0.0317 - accuracy: 0.9900\n",
      "Epoch 334/500\n",
      " - 0s - loss: 0.0317 - accuracy: 0.9900\n",
      "Epoch 335/500\n",
      " - 0s - loss: 0.0316 - accuracy: 0.9900\n",
      "Epoch 336/500\n",
      " - 0s - loss: 0.0316 - accuracy: 0.9900\n",
      "Epoch 337/500\n",
      " - 0s - loss: 0.0316 - accuracy: 0.9900\n",
      "Epoch 338/500\n",
      " - 0s - loss: 0.0316 - accuracy: 0.9902\n",
      "Epoch 339/500\n",
      " - 0s - loss: 0.0316 - accuracy: 0.9901\n",
      "Epoch 340/500\n",
      " - 0s - loss: 0.0315 - accuracy: 0.9900\n",
      "Epoch 341/500\n",
      " - 0s - loss: 0.0315 - accuracy: 0.9900\n",
      "Epoch 342/500\n",
      " - 0s - loss: 0.0315 - accuracy: 0.9902\n",
      "Epoch 343/500\n",
      " - 0s - loss: 0.0315 - accuracy: 0.9902\n",
      "Epoch 344/500\n",
      " - 0s - loss: 0.0315 - accuracy: 0.9902\n",
      "Epoch 345/500\n",
      " - 0s - loss: 0.0314 - accuracy: 0.9902\n",
      "Epoch 346/500\n",
      " - 0s - loss: 0.0314 - accuracy: 0.9902\n",
      "Epoch 347/500\n",
      " - 0s - loss: 0.0313 - accuracy: 0.9902\n",
      "Epoch 348/500\n",
      " - 0s - loss: 0.0314 - accuracy: 0.9900\n",
      "Epoch 349/500\n",
      " - 0s - loss: 0.0313 - accuracy: 0.9903\n",
      "Epoch 350/500\n",
      " - 0s - loss: 0.0313 - accuracy: 0.9903\n",
      "Epoch 351/500\n",
      " - 0s - loss: 0.0312 - accuracy: 0.9903\n",
      "Epoch 352/500\n",
      " - 0s - loss: 0.0312 - accuracy: 0.9904\n",
      "Epoch 353/500\n",
      " - 0s - loss: 0.0312 - accuracy: 0.9903\n",
      "Epoch 354/500\n",
      " - 0s - loss: 0.0312 - accuracy: 0.9903\n",
      "Epoch 355/500\n",
      " - 0s - loss: 0.0312 - accuracy: 0.9904\n",
      "Epoch 356/500\n",
      " - 0s - loss: 0.0311 - accuracy: 0.9902\n",
      "Epoch 357/500\n",
      " - 0s - loss: 0.0312 - accuracy: 0.9903\n",
      "Epoch 358/500\n",
      " - 0s - loss: 0.0311 - accuracy: 0.9904\n",
      "Epoch 359/500\n",
      " - 0s - loss: 0.0311 - accuracy: 0.9902\n",
      "Epoch 360/500\n",
      " - 0s - loss: 0.0310 - accuracy: 0.9903\n",
      "Epoch 361/500\n",
      " - 0s - loss: 0.0310 - accuracy: 0.9901\n",
      "Epoch 362/500\n",
      " - 0s - loss: 0.0310 - accuracy: 0.9902\n",
      "Epoch 363/500\n",
      " - 0s - loss: 0.0310 - accuracy: 0.9903\n",
      "Epoch 364/500\n",
      " - 0s - loss: 0.0309 - accuracy: 0.9902\n",
      "Epoch 365/500\n",
      " - 0s - loss: 0.0309 - accuracy: 0.9902\n",
      "Epoch 366/500\n",
      " - 0s - loss: 0.0308 - accuracy: 0.9902\n",
      "Epoch 367/500\n",
      " - 0s - loss: 0.0308 - accuracy: 0.9902\n",
      "Epoch 368/500\n",
      " - 0s - loss: 0.0308 - accuracy: 0.9902\n",
      "Epoch 369/500\n",
      " - 0s - loss: 0.0308 - accuracy: 0.9901\n",
      "Epoch 370/500\n",
      " - 0s - loss: 0.0307 - accuracy: 0.9904\n",
      "Epoch 371/500\n",
      " - 0s - loss: 0.0307 - accuracy: 0.9904\n",
      "Epoch 372/500\n",
      " - 0s - loss: 0.0307 - accuracy: 0.9904\n",
      "Epoch 373/500\n",
      " - 0s - loss: 0.0306 - accuracy: 0.9906\n",
      "Epoch 374/500\n",
      " - 0s - loss: 0.0306 - accuracy: 0.9904\n",
      "Epoch 375/500\n",
      " - 0s - loss: 0.0306 - accuracy: 0.9904\n",
      "Epoch 376/500\n",
      " - 0s - loss: 0.0305 - accuracy: 0.9905\n",
      "Epoch 377/500\n",
      " - 0s - loss: 0.0306 - accuracy: 0.9902\n",
      "Epoch 378/500\n",
      " - 0s - loss: 0.0305 - accuracy: 0.9904\n",
      "Epoch 379/500\n",
      " - 0s - loss: 0.0304 - accuracy: 0.9904\n",
      "Epoch 380/500\n",
      " - 0s - loss: 0.0304 - accuracy: 0.9904\n",
      "Epoch 381/500\n",
      " - 0s - loss: 0.0304 - accuracy: 0.9905\n",
      "Epoch 382/500\n",
      " - 0s - loss: 0.0303 - accuracy: 0.9904\n",
      "Epoch 383/500\n",
      " - 0s - loss: 0.0303 - accuracy: 0.9906\n",
      "Epoch 384/500\n",
      " - 0s - loss: 0.0302 - accuracy: 0.9905\n",
      "Epoch 385/500\n",
      " - 0s - loss: 0.0303 - accuracy: 0.9905\n",
      "Epoch 386/500\n",
      " - 0s - loss: 0.0301 - accuracy: 0.9904\n",
      "Epoch 387/500\n",
      " - 0s - loss: 0.0302 - accuracy: 0.9906\n",
      "Epoch 388/500\n",
      " - 0s - loss: 0.0302 - accuracy: 0.9905\n",
      "Epoch 389/500\n",
      " - 0s - loss: 0.0301 - accuracy: 0.9905\n",
      "Epoch 390/500\n",
      " - 0s - loss: 0.0300 - accuracy: 0.9906\n",
      "Epoch 391/500\n",
      " - 0s - loss: 0.0300 - accuracy: 0.9906\n",
      "Epoch 392/500\n",
      " - 0s - loss: 0.0299 - accuracy: 0.9906\n",
      "Epoch 393/500\n",
      " - 0s - loss: 0.0299 - accuracy: 0.9905\n",
      "Epoch 394/500\n",
      " - 0s - loss: 0.0299 - accuracy: 0.9905\n",
      "Epoch 395/500\n",
      " - 0s - loss: 0.0299 - accuracy: 0.9905\n",
      "Epoch 396/500\n",
      " - 0s - loss: 0.0298 - accuracy: 0.9906\n",
      "Epoch 397/500\n",
      " - 0s - loss: 0.0298 - accuracy: 0.9906\n",
      "Epoch 398/500\n",
      " - 0s - loss: 0.0297 - accuracy: 0.9906\n",
      "Epoch 399/500\n",
      " - 0s - loss: 0.0297 - accuracy: 0.9906\n",
      "Epoch 400/500\n",
      " - 0s - loss: 0.0297 - accuracy: 0.9907\n",
      "Epoch 401/500\n",
      " - 0s - loss: 0.0297 - accuracy: 0.9906\n",
      "Epoch 402/500\n",
      " - 0s - loss: 0.0297 - accuracy: 0.9906\n",
      "Epoch 403/500\n",
      " - 0s - loss: 0.0296 - accuracy: 0.9907\n",
      "Epoch 404/500\n",
      " - 0s - loss: 0.0296 - accuracy: 0.9906\n",
      "Epoch 405/500\n",
      " - 0s - loss: 0.0296 - accuracy: 0.9906\n",
      "Epoch 406/500\n",
      " - 0s - loss: 0.0295 - accuracy: 0.9907\n",
      "Epoch 407/500\n",
      " - 0s - loss: 0.0295 - accuracy: 0.9907\n",
      "Epoch 408/500\n",
      " - 0s - loss: 0.0295 - accuracy: 0.9907\n",
      "Epoch 409/500\n",
      " - 0s - loss: 0.0294 - accuracy: 0.9907\n",
      "Epoch 410/500\n",
      " - 0s - loss: 0.0294 - accuracy: 0.9907\n",
      "Epoch 411/500\n",
      " - 0s - loss: 0.0293 - accuracy: 0.9907\n",
      "Epoch 412/500\n",
      " - 0s - loss: 0.0293 - accuracy: 0.9906\n",
      "Epoch 413/500\n",
      " - 0s - loss: 0.0293 - accuracy: 0.9907\n",
      "Epoch 414/500\n",
      " - 0s - loss: 0.0293 - accuracy: 0.9908\n",
      "Epoch 415/500\n",
      " - 0s - loss: 0.0293 - accuracy: 0.9906\n",
      "Epoch 416/500\n",
      " - 0s - loss: 0.0293 - accuracy: 0.9908\n",
      "Epoch 417/500\n",
      " - 0s - loss: 0.0292 - accuracy: 0.9907\n",
      "Epoch 418/500\n",
      " - 0s - loss: 0.0291 - accuracy: 0.9907\n",
      "Epoch 419/500\n",
      " - 0s - loss: 0.0291 - accuracy: 0.9907\n",
      "Epoch 420/500\n",
      " - 0s - loss: 0.0291 - accuracy: 0.9909\n",
      "Epoch 421/500\n",
      " - 0s - loss: 0.0291 - accuracy: 0.9907\n",
      "Epoch 422/500\n",
      " - 0s - loss: 0.0290 - accuracy: 0.9908\n",
      "Epoch 423/500\n",
      " - 0s - loss: 0.0290 - accuracy: 0.9911\n",
      "Epoch 424/500\n",
      " - 0s - loss: 0.0290 - accuracy: 0.9910\n",
      "Epoch 425/500\n",
      " - 0s - loss: 0.0289 - accuracy: 0.9910\n",
      "Epoch 426/500\n",
      " - 0s - loss: 0.0289 - accuracy: 0.9910\n",
      "Epoch 427/500\n",
      " - 0s - loss: 0.0289 - accuracy: 0.9909\n",
      "Epoch 428/500\n",
      " - 0s - loss: 0.0289 - accuracy: 0.9909\n",
      "Epoch 429/500\n",
      " - 0s - loss: 0.0289 - accuracy: 0.9910\n",
      "Epoch 430/500\n",
      " - 0s - loss: 0.0288 - accuracy: 0.9910\n",
      "Epoch 431/500\n",
      " - 0s - loss: 0.0288 - accuracy: 0.9911\n",
      "Epoch 432/500\n",
      " - 0s - loss: 0.0287 - accuracy: 0.9910\n",
      "Epoch 433/500\n",
      " - 0s - loss: 0.0287 - accuracy: 0.9909\n",
      "Epoch 434/500\n",
      " - 0s - loss: 0.0287 - accuracy: 0.9910\n",
      "Epoch 435/500\n",
      " - 0s - loss: 0.0287 - accuracy: 0.9911\n",
      "Epoch 436/500\n",
      " - 0s - loss: 0.0287 - accuracy: 0.9911\n",
      "Epoch 437/500\n",
      " - 0s - loss: 0.0286 - accuracy: 0.9911\n",
      "Epoch 438/500\n",
      " - 0s - loss: 0.0286 - accuracy: 0.9912\n",
      "Epoch 439/500\n",
      " - 0s - loss: 0.0286 - accuracy: 0.9912\n",
      "Epoch 440/500\n",
      " - 0s - loss: 0.0285 - accuracy: 0.9911\n",
      "Epoch 441/500\n",
      " - 0s - loss: 0.0285 - accuracy: 0.9912\n",
      "Epoch 442/500\n",
      " - 0s - loss: 0.0285 - accuracy: 0.9911\n",
      "Epoch 443/500\n",
      " - 0s - loss: 0.0285 - accuracy: 0.9912\n",
      "Epoch 444/500\n",
      " - 0s - loss: 0.0285 - accuracy: 0.9912\n",
      "Epoch 445/500\n",
      " - 0s - loss: 0.0284 - accuracy: 0.9912\n",
      "Epoch 446/500\n",
      " - 0s - loss: 0.0284 - accuracy: 0.9912\n",
      "Epoch 447/500\n",
      " - 0s - loss: 0.0284 - accuracy: 0.9912\n",
      "Epoch 448/500\n",
      " - 0s - loss: 0.0284 - accuracy: 0.9911\n",
      "Epoch 449/500\n",
      " - 0s - loss: 0.0283 - accuracy: 0.9912\n",
      "Epoch 450/500\n",
      " - 0s - loss: 0.0283 - accuracy: 0.9911\n",
      "Epoch 451/500\n",
      " - 0s - loss: 0.0283 - accuracy: 0.9912\n",
      "Epoch 452/500\n",
      " - 0s - loss: 0.0283 - accuracy: 0.9912\n",
      "Epoch 453/500\n",
      " - 0s - loss: 0.0282 - accuracy: 0.9912\n",
      "Epoch 454/500\n",
      " - 0s - loss: 0.0282 - accuracy: 0.9912\n",
      "Epoch 455/500\n",
      " - 0s - loss: 0.0282 - accuracy: 0.9911\n",
      "Epoch 456/500\n",
      " - 0s - loss: 0.0282 - accuracy: 0.9912\n",
      "Epoch 457/500\n",
      " - 0s - loss: 0.0281 - accuracy: 0.9912\n",
      "Epoch 458/500\n",
      " - 0s - loss: 0.0281 - accuracy: 0.9912\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 459/500\n",
      " - 0s - loss: 0.0281 - accuracy: 0.9913\n",
      "Epoch 460/500\n",
      " - 0s - loss: 0.0282 - accuracy: 0.9911\n",
      "Epoch 461/500\n",
      " - 0s - loss: 0.0281 - accuracy: 0.9914\n",
      "Epoch 462/500\n",
      " - 0s - loss: 0.0281 - accuracy: 0.9912\n",
      "Epoch 463/500\n",
      " - 0s - loss: 0.0281 - accuracy: 0.9913\n",
      "Epoch 464/500\n",
      " - 0s - loss: 0.0280 - accuracy: 0.9911\n",
      "Epoch 465/500\n",
      " - 0s - loss: 0.0280 - accuracy: 0.9912\n",
      "Epoch 466/500\n",
      " - 0s - loss: 0.0280 - accuracy: 0.9914\n",
      "Epoch 467/500\n",
      " - 0s - loss: 0.0279 - accuracy: 0.9915\n",
      "Epoch 468/500\n",
      " - 0s - loss: 0.0279 - accuracy: 0.9912\n",
      "Epoch 469/500\n",
      " - 0s - loss: 0.0279 - accuracy: 0.9911\n",
      "Epoch 470/500\n",
      " - 0s - loss: 0.0279 - accuracy: 0.9913\n",
      "Epoch 471/500\n",
      " - 0s - loss: 0.0278 - accuracy: 0.9913\n",
      "Epoch 472/500\n",
      " - 0s - loss: 0.0279 - accuracy: 0.9915\n",
      "Epoch 473/500\n",
      " - 0s - loss: 0.0278 - accuracy: 0.9913\n",
      "Epoch 474/500\n",
      " - 0s - loss: 0.0278 - accuracy: 0.9911\n",
      "Epoch 475/500\n",
      " - 0s - loss: 0.0278 - accuracy: 0.9916\n",
      "Epoch 476/500\n",
      " - 0s - loss: 0.0278 - accuracy: 0.9915\n",
      "Epoch 477/500\n",
      " - 0s - loss: 0.0278 - accuracy: 0.9915\n",
      "Epoch 478/500\n",
      " - 0s - loss: 0.0277 - accuracy: 0.9915\n",
      "Epoch 479/500\n",
      " - 0s - loss: 0.0278 - accuracy: 0.9916\n",
      "Epoch 480/500\n",
      " - 0s - loss: 0.0277 - accuracy: 0.9915\n",
      "Epoch 481/500\n",
      " - 0s - loss: 0.0277 - accuracy: 0.9915\n",
      "Epoch 482/500\n",
      " - 0s - loss: 0.0276 - accuracy: 0.9916\n",
      "Epoch 483/500\n",
      " - 0s - loss: 0.0276 - accuracy: 0.9914\n",
      "Epoch 484/500\n",
      " - 0s - loss: 0.0276 - accuracy: 0.9915\n",
      "Epoch 485/500\n",
      " - 0s - loss: 0.0276 - accuracy: 0.9915\n",
      "Epoch 486/500\n",
      " - 0s - loss: 0.0276 - accuracy: 0.9916\n",
      "Epoch 487/500\n",
      " - 0s - loss: 0.0275 - accuracy: 0.9915\n",
      "Epoch 488/500\n",
      " - 0s - loss: 0.0275 - accuracy: 0.9916\n",
      "Epoch 489/500\n",
      " - 0s - loss: 0.0274 - accuracy: 0.9915\n",
      "Epoch 490/500\n",
      " - 0s - loss: 0.0275 - accuracy: 0.9915\n",
      "Epoch 491/500\n",
      " - 0s - loss: 0.0275 - accuracy: 0.9916\n",
      "Epoch 492/500\n",
      " - 0s - loss: 0.0275 - accuracy: 0.9915\n",
      "Epoch 493/500\n",
      " - 0s - loss: 0.0274 - accuracy: 0.9915\n",
      "Epoch 494/500\n",
      " - 0s - loss: 0.0274 - accuracy: 0.9916\n",
      "Epoch 495/500\n",
      " - 0s - loss: 0.0273 - accuracy: 0.9916\n",
      "Epoch 496/500\n",
      " - 0s - loss: 0.0274 - accuracy: 0.9915\n",
      "Epoch 497/500\n",
      " - 0s - loss: 0.0273 - accuracy: 0.9916\n",
      "Epoch 498/500\n",
      " - 0s - loss: 0.0273 - accuracy: 0.9916\n",
      "Epoch 499/500\n",
      " - 0s - loss: 0.0273 - accuracy: 0.9916\n",
      "Epoch 500/500\n",
      " - 0s - loss: 0.0273 - accuracy: 0.9914\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.callbacks.History at 0x1cbebc8ad60>"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(train[:, 1:], train[:, 0], epochs=500, batch_size=500, verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_pred = model.predict(test[:, 1:]).reshape((-1,)) >= 0.5\n",
    "\n",
    "conf_mat = confusion_matrix(y_test, y_pred)\n",
    "\n",
    "lang = ['No Fault (0)', 'Fault (1)']\n",
    "ax = plt.axes()\n",
    "sns.heatmap(conf_mat, square=True, annot=True, cmap='Blues', fmt='d', cbar=True, ax=ax, xticklabels=lang, yticklabels=lang)\n",
    "ax.set_title('Confusion Matrix')\n",
    "plt.xlabel('Predicted Label')\n",
    "plt.ylabel('True Label')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The Evaluation Metrics of the MLP Model-2 (on the test set) is as follows:\n",
      "\n",
      "True Positives: 537\n",
      "True Negatives: 650\n",
      "False Positives: 1\n",
      "False Negatives: 13\n",
      "\n",
      "Precision: 0.9981412639405205\n",
      "Recall: 0.9763636363636363\n",
      "\n",
      "Accuracy: 0.9883430474604497\n"
     ]
    }
   ],
   "source": [
    "true_positive = conf_mat[1, 1]\n",
    "true_negative = conf_mat[0, 0]\n",
    "false_positive = conf_mat[0, 1]\n",
    "false_negative = conf_mat[1, 0]\n",
    "accuracy = accuracy_score(y_pred, y_test)\n",
    "\n",
    "print(\"The Evaluation Metrics of the MLP Model-2 (on the test set) is as follows:\\n\")\n",
    "\n",
    "print(\"True Positives:\", true_positive)\n",
    "print(\"True Negatives:\", true_negative)\n",
    "print(\"False Positives:\", false_positive)\n",
    "print(\"False Negatives:\", false_negative, end='\\n\\n')\n",
    "\n",
    "print(\"Precision:\", true_positive/(true_positive+false_positive))\n",
    "print(\"Recall:\", true_positive/(true_positive+false_negative), end='\\n\\n')\n",
    "\n",
    "print(\"Accuracy:\", accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### (iii) Model 3: (1 hidden layer, LeakyReLU activation, RMSprop optimizer)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "model.add(Dense(10, input_dim=6))\n",
    "model.add(LeakyReLU(alpha=0.05))\n",
    "model.add(Dense(1, activation='sigmoid'))\n",
    "model.compile(loss='binary_crossentropy', optimizer='RMSprop', metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/500\n",
      " - 0s - loss: 0.7502 - accuracy: 0.5019\n",
      "Epoch 2/500\n",
      " - 0s - loss: 0.6917 - accuracy: 0.5532\n",
      "Epoch 3/500\n",
      " - 0s - loss: 0.6435 - accuracy: 0.6111\n",
      "Epoch 4/500\n",
      " - 0s - loss: 0.5993 - accuracy: 0.6676\n",
      "Epoch 5/500\n",
      " - 0s - loss: 0.5575 - accuracy: 0.7006\n",
      "Epoch 6/500\n",
      " - 0s - loss: 0.5185 - accuracy: 0.7342\n",
      "Epoch 7/500\n",
      " - 0s - loss: 0.4828 - accuracy: 0.8281\n",
      "Epoch 8/500\n",
      " - 0s - loss: 0.4498 - accuracy: 0.8690\n",
      "Epoch 9/500\n",
      " - 0s - loss: 0.4200 - accuracy: 0.8811\n",
      "Epoch 10/500\n",
      " - 0s - loss: 0.3923 - accuracy: 0.8836\n",
      "Epoch 11/500\n",
      " - 0s - loss: 0.3663 - accuracy: 0.8912\n",
      "Epoch 12/500\n",
      " - 0s - loss: 0.3420 - accuracy: 0.9053\n",
      "Epoch 13/500\n",
      " - 0s - loss: 0.3195 - accuracy: 0.9156\n",
      "Epoch 14/500\n",
      " - 0s - loss: 0.2994 - accuracy: 0.9238\n",
      "Epoch 15/500\n",
      " - 0s - loss: 0.2808 - accuracy: 0.9286\n",
      "Epoch 16/500\n",
      " - 0s - loss: 0.2634 - accuracy: 0.9320\n",
      "Epoch 17/500\n",
      " - 0s - loss: 0.2469 - accuracy: 0.9366\n",
      "Epoch 18/500\n",
      " - 0s - loss: 0.2316 - accuracy: 0.9391\n",
      "Epoch 19/500\n",
      " - 0s - loss: 0.2176 - accuracy: 0.9432\n",
      "Epoch 20/500\n",
      " - 0s - loss: 0.2048 - accuracy: 0.9462\n",
      "Epoch 21/500\n",
      " - 0s - loss: 0.1933 - accuracy: 0.9493\n",
      "Epoch 22/500\n",
      " - 0s - loss: 0.1830 - accuracy: 0.9519\n",
      "Epoch 23/500\n",
      " - 0s - loss: 0.1738 - accuracy: 0.9532\n",
      "Epoch 24/500\n",
      " - 0s - loss: 0.1653 - accuracy: 0.9546\n",
      "Epoch 25/500\n",
      " - 0s - loss: 0.1576 - accuracy: 0.9552\n",
      "Epoch 26/500\n",
      " - 0s - loss: 0.1504 - accuracy: 0.9566\n",
      "Epoch 27/500\n",
      " - 0s - loss: 0.1439 - accuracy: 0.9576\n",
      "Epoch 28/500\n",
      " - 0s - loss: 0.1378 - accuracy: 0.9589\n",
      "Epoch 29/500\n",
      " - 0s - loss: 0.1322 - accuracy: 0.9595\n",
      "Epoch 30/500\n",
      " - 0s - loss: 0.1270 - accuracy: 0.9608\n",
      "Epoch 31/500\n",
      " - 0s - loss: 0.1221 - accuracy: 0.9619\n",
      "Epoch 32/500\n",
      " - 0s - loss: 0.1176 - accuracy: 0.9629\n",
      "Epoch 33/500\n",
      " - 0s - loss: 0.1134 - accuracy: 0.9644\n",
      "Epoch 34/500\n",
      " - 0s - loss: 0.1094 - accuracy: 0.9651\n",
      "Epoch 35/500\n",
      " - 0s - loss: 0.1057 - accuracy: 0.9657\n",
      "Epoch 36/500\n",
      " - 0s - loss: 0.1022 - accuracy: 0.9669\n",
      "Epoch 37/500\n",
      " - 0s - loss: 0.0989 - accuracy: 0.9680\n",
      "Epoch 38/500\n",
      " - 0s - loss: 0.0958 - accuracy: 0.9688\n",
      "Epoch 39/500\n",
      " - 0s - loss: 0.0928 - accuracy: 0.9692\n",
      "Epoch 40/500\n",
      " - 0s - loss: 0.0900 - accuracy: 0.9705\n",
      "Epoch 41/500\n",
      " - 0s - loss: 0.0874 - accuracy: 0.9709\n",
      "Epoch 42/500\n",
      " - 0s - loss: 0.0850 - accuracy: 0.9719\n",
      "Epoch 43/500\n",
      " - 0s - loss: 0.0826 - accuracy: 0.9727\n",
      "Epoch 44/500\n",
      " - 0s - loss: 0.0803 - accuracy: 0.9732\n",
      "Epoch 45/500\n",
      " - 0s - loss: 0.0782 - accuracy: 0.9737\n",
      "Epoch 46/500\n",
      " - 0s - loss: 0.0762 - accuracy: 0.9747\n",
      "Epoch 47/500\n",
      " - 0s - loss: 0.0742 - accuracy: 0.9749\n",
      "Epoch 48/500\n",
      " - 0s - loss: 0.0724 - accuracy: 0.9759\n",
      "Epoch 49/500\n",
      " - 0s - loss: 0.0707 - accuracy: 0.9764\n",
      "Epoch 50/500\n",
      " - 0s - loss: 0.0690 - accuracy: 0.9765\n",
      "Epoch 51/500\n",
      " - 0s - loss: 0.0675 - accuracy: 0.9776\n",
      "Epoch 52/500\n",
      " - 0s - loss: 0.0660 - accuracy: 0.9777\n",
      "Epoch 53/500\n",
      " - 0s - loss: 0.0646 - accuracy: 0.9780\n",
      "Epoch 54/500\n",
      " - 0s - loss: 0.0633 - accuracy: 0.9787\n",
      "Epoch 55/500\n",
      " - 0s - loss: 0.0619 - accuracy: 0.9794\n",
      "Epoch 56/500\n",
      " - 0s - loss: 0.0607 - accuracy: 0.9794\n",
      "Epoch 57/500\n",
      " - 0s - loss: 0.0595 - accuracy: 0.9799\n",
      "Epoch 58/500\n",
      " - 0s - loss: 0.0583 - accuracy: 0.9802\n",
      "Epoch 59/500\n",
      " - 0s - loss: 0.0572 - accuracy: 0.9810\n",
      "Epoch 60/500\n",
      " - 0s - loss: 0.0562 - accuracy: 0.9814\n",
      "Epoch 61/500\n",
      " - 0s - loss: 0.0552 - accuracy: 0.9813\n",
      "Epoch 62/500\n",
      " - 0s - loss: 0.0541 - accuracy: 0.9816\n",
      "Epoch 63/500\n",
      " - 0s - loss: 0.0532 - accuracy: 0.9819\n",
      "Epoch 64/500\n",
      " - 0s - loss: 0.0523 - accuracy: 0.9821\n",
      "Epoch 65/500\n",
      " - 0s - loss: 0.0514 - accuracy: 0.9831\n",
      "Epoch 66/500\n",
      " - 0s - loss: 0.0506 - accuracy: 0.9828\n",
      "Epoch 67/500\n",
      " - 0s - loss: 0.0498 - accuracy: 0.9832\n",
      "Epoch 68/500\n",
      " - 0s - loss: 0.0490 - accuracy: 0.9832\n",
      "Epoch 69/500\n",
      " - 0s - loss: 0.0483 - accuracy: 0.9833\n",
      "Epoch 70/500\n",
      " - 0s - loss: 0.0476 - accuracy: 0.9834\n",
      "Epoch 71/500\n",
      " - 0s - loss: 0.0468 - accuracy: 0.9838\n",
      "Epoch 72/500\n",
      " - 0s - loss: 0.0462 - accuracy: 0.9840\n",
      "Epoch 73/500\n",
      " - 0s - loss: 0.0455 - accuracy: 0.9841\n",
      "Epoch 74/500\n",
      " - 0s - loss: 0.0449 - accuracy: 0.9842\n",
      "Epoch 75/500\n",
      " - 0s - loss: 0.0443 - accuracy: 0.9846\n",
      "Epoch 76/500\n",
      " - 0s - loss: 0.0437 - accuracy: 0.9847\n",
      "Epoch 77/500\n",
      " - 0s - loss: 0.0431 - accuracy: 0.9850\n",
      "Epoch 78/500\n",
      " - 0s - loss: 0.0426 - accuracy: 0.9851\n",
      "Epoch 79/500\n",
      " - 0s - loss: 0.0420 - accuracy: 0.9852\n",
      "Epoch 80/500\n",
      " - 0s - loss: 0.0415 - accuracy: 0.9852\n",
      "Epoch 81/500\n",
      " - 0s - loss: 0.0410 - accuracy: 0.9852\n",
      "Epoch 82/500\n",
      " - 0s - loss: 0.0405 - accuracy: 0.9855\n",
      "Epoch 83/500\n",
      " - 0s - loss: 0.0401 - accuracy: 0.9856\n",
      "Epoch 84/500\n",
      " - 0s - loss: 0.0397 - accuracy: 0.9856\n",
      "Epoch 85/500\n",
      " - 0s - loss: 0.0393 - accuracy: 0.9856\n",
      "Epoch 86/500\n",
      " - 0s - loss: 0.0389 - accuracy: 0.9862\n",
      "Epoch 87/500\n",
      " - 0s - loss: 0.0385 - accuracy: 0.9861\n",
      "Epoch 88/500\n",
      " - 0s - loss: 0.0381 - accuracy: 0.9866\n",
      "Epoch 89/500\n",
      " - 0s - loss: 0.0377 - accuracy: 0.9868\n",
      "Epoch 90/500\n",
      " - 0s - loss: 0.0374 - accuracy: 0.9868\n",
      "Epoch 91/500\n",
      " - 0s - loss: 0.0371 - accuracy: 0.9869\n",
      "Epoch 92/500\n",
      " - 0s - loss: 0.0367 - accuracy: 0.9869\n",
      "Epoch 93/500\n",
      " - 0s - loss: 0.0364 - accuracy: 0.9870\n",
      "Epoch 94/500\n",
      " - 0s - loss: 0.0361 - accuracy: 0.9871\n",
      "Epoch 95/500\n",
      " - 0s - loss: 0.0358 - accuracy: 0.9874\n",
      "Epoch 96/500\n",
      " - 0s - loss: 0.0355 - accuracy: 0.9873\n",
      "Epoch 97/500\n",
      " - 0s - loss: 0.0352 - accuracy: 0.9874\n",
      "Epoch 98/500\n",
      " - 0s - loss: 0.0349 - accuracy: 0.9873\n",
      "Epoch 99/500\n",
      " - 0s - loss: 0.0346 - accuracy: 0.9875\n",
      "Epoch 100/500\n",
      " - 0s - loss: 0.0344 - accuracy: 0.9875\n",
      "Epoch 101/500\n",
      " - 0s - loss: 0.0341 - accuracy: 0.9877\n",
      "Epoch 102/500\n",
      " - 0s - loss: 0.0339 - accuracy: 0.9876\n",
      "Epoch 103/500\n",
      " - 0s - loss: 0.0336 - accuracy: 0.9880\n",
      "Epoch 104/500\n",
      " - 0s - loss: 0.0334 - accuracy: 0.9881\n",
      "Epoch 105/500\n",
      " - 0s - loss: 0.0332 - accuracy: 0.9882\n",
      "Epoch 106/500\n",
      " - 0s - loss: 0.0330 - accuracy: 0.9881\n",
      "Epoch 107/500\n",
      " - 0s - loss: 0.0328 - accuracy: 0.9882\n",
      "Epoch 108/500\n",
      " - 0s - loss: 0.0326 - accuracy: 0.9882\n",
      "Epoch 109/500\n",
      " - 0s - loss: 0.0324 - accuracy: 0.9883\n",
      "Epoch 110/500\n",
      " - 0s - loss: 0.0323 - accuracy: 0.9884\n",
      "Epoch 111/500\n",
      " - 0s - loss: 0.0321 - accuracy: 0.9884\n",
      "Epoch 112/500\n",
      " - 0s - loss: 0.0319 - accuracy: 0.9888\n",
      "Epoch 113/500\n",
      " - 0s - loss: 0.0317 - accuracy: 0.9885\n",
      "Epoch 114/500\n",
      " - 0s - loss: 0.0315 - accuracy: 0.9889\n",
      "Epoch 115/500\n",
      " - 0s - loss: 0.0313 - accuracy: 0.9890\n",
      "Epoch 116/500\n",
      " - 0s - loss: 0.0312 - accuracy: 0.9888\n",
      "Epoch 117/500\n",
      " - 0s - loss: 0.0310 - accuracy: 0.9890\n",
      "Epoch 118/500\n",
      " - 0s - loss: 0.0309 - accuracy: 0.9890\n",
      "Epoch 119/500\n",
      " - 0s - loss: 0.0307 - accuracy: 0.9889\n",
      "Epoch 120/500\n",
      " - 0s - loss: 0.0306 - accuracy: 0.9890\n",
      "Epoch 121/500\n",
      " - 0s - loss: 0.0304 - accuracy: 0.9890\n",
      "Epoch 122/500\n",
      " - 0s - loss: 0.0303 - accuracy: 0.9890\n",
      "Epoch 123/500\n",
      " - 0s - loss: 0.0301 - accuracy: 0.9890\n",
      "Epoch 124/500\n",
      " - 0s - loss: 0.0300 - accuracy: 0.9890\n",
      "Epoch 125/500\n",
      " - 0s - loss: 0.0299 - accuracy: 0.9891\n",
      "Epoch 126/500\n",
      " - 0s - loss: 0.0297 - accuracy: 0.9892\n",
      "Epoch 127/500\n",
      " - 0s - loss: 0.0296 - accuracy: 0.9893\n",
      "Epoch 128/500\n",
      " - 0s - loss: 0.0294 - accuracy: 0.9893\n",
      "Epoch 129/500\n",
      " - 0s - loss: 0.0293 - accuracy: 0.9894\n",
      "Epoch 130/500\n",
      " - 0s - loss: 0.0292 - accuracy: 0.9892\n",
      "Epoch 131/500\n",
      " - 0s - loss: 0.0291 - accuracy: 0.9893\n",
      "Epoch 132/500\n",
      " - 0s - loss: 0.0290 - accuracy: 0.9893\n",
      "Epoch 133/500\n",
      " - 0s - loss: 0.0288 - accuracy: 0.9894\n",
      "Epoch 134/500\n",
      " - 0s - loss: 0.0288 - accuracy: 0.9896\n",
      "Epoch 135/500\n",
      " - 0s - loss: 0.0286 - accuracy: 0.9894\n",
      "Epoch 136/500\n",
      " - 0s - loss: 0.0285 - accuracy: 0.9896\n",
      "Epoch 137/500\n",
      " - 0s - loss: 0.0285 - accuracy: 0.9898\n",
      "Epoch 138/500\n",
      " - 0s - loss: 0.0283 - accuracy: 0.9897\n",
      "Epoch 139/500\n",
      " - 0s - loss: 0.0283 - accuracy: 0.9898\n",
      "Epoch 140/500\n",
      " - 0s - loss: 0.0281 - accuracy: 0.9899\n",
      "Epoch 141/500\n",
      " - 0s - loss: 0.0280 - accuracy: 0.9898\n",
      "Epoch 142/500\n",
      " - 0s - loss: 0.0279 - accuracy: 0.9898\n",
      "Epoch 143/500\n",
      " - 0s - loss: 0.0278 - accuracy: 0.9898\n",
      "Epoch 144/500\n",
      " - 0s - loss: 0.0278 - accuracy: 0.9899\n",
      "Epoch 145/500\n",
      " - 0s - loss: 0.0277 - accuracy: 0.9900\n",
      "Epoch 146/500\n",
      " - 0s - loss: 0.0276 - accuracy: 0.9900\n",
      "Epoch 147/500\n",
      " - 0s - loss: 0.0275 - accuracy: 0.9900\n",
      "Epoch 148/500\n",
      " - 0s - loss: 0.0274 - accuracy: 0.9901\n",
      "Epoch 149/500\n",
      " - 0s - loss: 0.0273 - accuracy: 0.9900\n",
      "Epoch 150/500\n",
      " - 0s - loss: 0.0273 - accuracy: 0.9901\n",
      "Epoch 151/500\n",
      " - 0s - loss: 0.0272 - accuracy: 0.9901\n",
      "Epoch 152/500\n",
      " - 0s - loss: 0.0271 - accuracy: 0.9903\n",
      "Epoch 153/500\n",
      " - 0s - loss: 0.0270 - accuracy: 0.9902\n",
      "Epoch 154/500\n",
      " - 0s - loss: 0.0270 - accuracy: 0.9903\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 155/500\n",
      " - 0s - loss: 0.0269 - accuracy: 0.9904\n",
      "Epoch 156/500\n",
      " - 0s - loss: 0.0268 - accuracy: 0.9905\n",
      "Epoch 157/500\n",
      " - 0s - loss: 0.0267 - accuracy: 0.9903\n",
      "Epoch 158/500\n",
      " - 0s - loss: 0.0266 - accuracy: 0.9906\n",
      "Epoch 159/500\n",
      " - 0s - loss: 0.0265 - accuracy: 0.9905\n",
      "Epoch 160/500\n",
      " - 0s - loss: 0.0265 - accuracy: 0.9906\n",
      "Epoch 161/500\n",
      " - 0s - loss: 0.0264 - accuracy: 0.9906\n",
      "Epoch 162/500\n",
      " - 0s - loss: 0.0263 - accuracy: 0.9906\n",
      "Epoch 163/500\n",
      " - 0s - loss: 0.0263 - accuracy: 0.9905\n",
      "Epoch 164/500\n",
      " - 0s - loss: 0.0262 - accuracy: 0.9906\n",
      "Epoch 165/500\n",
      " - 0s - loss: 0.0261 - accuracy: 0.9907\n",
      "Epoch 166/500\n",
      " - 0s - loss: 0.0261 - accuracy: 0.9907\n",
      "Epoch 167/500\n",
      " - 0s - loss: 0.0260 - accuracy: 0.9906\n",
      "Epoch 168/500\n",
      " - 0s - loss: 0.0260 - accuracy: 0.9906\n",
      "Epoch 169/500\n",
      " - 0s - loss: 0.0259 - accuracy: 0.9909\n",
      "Epoch 170/500\n",
      " - 0s - loss: 0.0258 - accuracy: 0.9906\n",
      "Epoch 171/500\n",
      " - 0s - loss: 0.0258 - accuracy: 0.9911\n",
      "Epoch 172/500\n",
      " - 0s - loss: 0.0257 - accuracy: 0.9909\n",
      "Epoch 173/500\n",
      " - 0s - loss: 0.0256 - accuracy: 0.9910\n",
      "Epoch 174/500\n",
      " - 0s - loss: 0.0256 - accuracy: 0.9908\n",
      "Epoch 175/500\n",
      " - 0s - loss: 0.0255 - accuracy: 0.9911\n",
      "Epoch 176/500\n",
      " - 0s - loss: 0.0255 - accuracy: 0.9909\n",
      "Epoch 177/500\n",
      " - 0s - loss: 0.0253 - accuracy: 0.9910\n",
      "Epoch 178/500\n",
      " - 0s - loss: 0.0253 - accuracy: 0.9909\n",
      "Epoch 179/500\n",
      " - 0s - loss: 0.0252 - accuracy: 0.9911\n",
      "Epoch 180/500\n",
      " - 0s - loss: 0.0252 - accuracy: 0.9908\n",
      "Epoch 181/500\n",
      " - 0s - loss: 0.0251 - accuracy: 0.9910\n",
      "Epoch 182/500\n",
      " - 0s - loss: 0.0251 - accuracy: 0.9911\n",
      "Epoch 183/500\n",
      " - 0s - loss: 0.0250 - accuracy: 0.9910\n",
      "Epoch 184/500\n",
      " - 0s - loss: 0.0250 - accuracy: 0.9910\n",
      "Epoch 185/500\n",
      " - 0s - loss: 0.0249 - accuracy: 0.9910\n",
      "Epoch 186/500\n",
      " - 0s - loss: 0.0248 - accuracy: 0.9910\n",
      "Epoch 187/500\n",
      " - 0s - loss: 0.0248 - accuracy: 0.9910\n",
      "Epoch 188/500\n",
      " - 0s - loss: 0.0247 - accuracy: 0.9910\n",
      "Epoch 189/500\n",
      " - 0s - loss: 0.0247 - accuracy: 0.9912\n",
      "Epoch 190/500\n",
      " - 0s - loss: 0.0246 - accuracy: 0.9909\n",
      "Epoch 191/500\n",
      " - 0s - loss: 0.0245 - accuracy: 0.9912\n",
      "Epoch 192/500\n",
      " - 0s - loss: 0.0245 - accuracy: 0.9914\n",
      "Epoch 193/500\n",
      " - 0s - loss: 0.0245 - accuracy: 0.9913\n",
      "Epoch 194/500\n",
      " - 0s - loss: 0.0243 - accuracy: 0.9915\n",
      "Epoch 195/500\n",
      " - 0s - loss: 0.0243 - accuracy: 0.9914\n",
      "Epoch 196/500\n",
      " - 0s - loss: 0.0243 - accuracy: 0.9913\n",
      "Epoch 197/500\n",
      " - 0s - loss: 0.0242 - accuracy: 0.9916\n",
      "Epoch 198/500\n",
      " - 0s - loss: 0.0242 - accuracy: 0.9915\n",
      "Epoch 199/500\n",
      " - 0s - loss: 0.0242 - accuracy: 0.9916\n",
      "Epoch 200/500\n",
      " - 0s - loss: 0.0241 - accuracy: 0.9914\n",
      "Epoch 201/500\n",
      " - 0s - loss: 0.0240 - accuracy: 0.9915\n",
      "Epoch 202/500\n",
      " - 0s - loss: 0.0240 - accuracy: 0.9915\n",
      "Epoch 203/500\n",
      " - 0s - loss: 0.0240 - accuracy: 0.9915\n",
      "Epoch 204/500\n",
      " - 0s - loss: 0.0239 - accuracy: 0.9916\n",
      "Epoch 205/500\n",
      " - 0s - loss: 0.0239 - accuracy: 0.9915\n",
      "Epoch 206/500\n",
      " - 0s - loss: 0.0239 - accuracy: 0.9915\n",
      "Epoch 207/500\n",
      " - 0s - loss: 0.0238 - accuracy: 0.9915\n",
      "Epoch 208/500\n",
      " - 0s - loss: 0.0237 - accuracy: 0.9916\n",
      "Epoch 209/500\n",
      " - 0s - loss: 0.0237 - accuracy: 0.9915\n",
      "Epoch 210/500\n",
      " - 0s - loss: 0.0237 - accuracy: 0.9916\n",
      "Epoch 211/500\n",
      " - 0s - loss: 0.0236 - accuracy: 0.9914\n",
      "Epoch 212/500\n",
      " - 0s - loss: 0.0236 - accuracy: 0.9916\n",
      "Epoch 213/500\n",
      " - 0s - loss: 0.0235 - accuracy: 0.9915\n",
      "Epoch 214/500\n",
      " - 0s - loss: 0.0235 - accuracy: 0.9916\n",
      "Epoch 215/500\n",
      " - 0s - loss: 0.0235 - accuracy: 0.9916\n",
      "Epoch 216/500\n",
      " - 0s - loss: 0.0234 - accuracy: 0.9917\n",
      "Epoch 217/500\n",
      " - 0s - loss: 0.0234 - accuracy: 0.9917\n",
      "Epoch 218/500\n",
      " - 0s - loss: 0.0234 - accuracy: 0.9917\n",
      "Epoch 219/500\n",
      " - 0s - loss: 0.0233 - accuracy: 0.9916\n",
      "Epoch 220/500\n",
      " - 0s - loss: 0.0233 - accuracy: 0.9915\n",
      "Epoch 221/500\n",
      " - 0s - loss: 0.0233 - accuracy: 0.9917\n",
      "Epoch 222/500\n",
      " - 0s - loss: 0.0232 - accuracy: 0.9917\n",
      "Epoch 223/500\n",
      " - 0s - loss: 0.0232 - accuracy: 0.9917\n",
      "Epoch 224/500\n",
      " - 0s - loss: 0.0232 - accuracy: 0.9917\n",
      "Epoch 225/500\n",
      " - 0s - loss: 0.0231 - accuracy: 0.9917\n",
      "Epoch 226/500\n",
      " - 0s - loss: 0.0231 - accuracy: 0.9917\n",
      "Epoch 227/500\n",
      " - 0s - loss: 0.0230 - accuracy: 0.9916\n",
      "Epoch 228/500\n",
      " - 0s - loss: 0.0230 - accuracy: 0.9917\n",
      "Epoch 229/500\n",
      " - 0s - loss: 0.0230 - accuracy: 0.9917\n",
      "Epoch 230/500\n",
      " - 0s - loss: 0.0229 - accuracy: 0.9917\n",
      "Epoch 231/500\n",
      " - 0s - loss: 0.0229 - accuracy: 0.9917\n",
      "Epoch 232/500\n",
      " - 0s - loss: 0.0229 - accuracy: 0.9917\n",
      "Epoch 233/500\n",
      " - 0s - loss: 0.0228 - accuracy: 0.9916\n",
      "Epoch 234/500\n",
      " - 0s - loss: 0.0229 - accuracy: 0.9917\n",
      "Epoch 235/500\n",
      " - 0s - loss: 0.0228 - accuracy: 0.9917\n",
      "Epoch 236/500\n",
      " - 0s - loss: 0.0227 - accuracy: 0.9916\n",
      "Epoch 237/500\n",
      " - 0s - loss: 0.0227 - accuracy: 0.9917\n",
      "Epoch 238/500\n",
      " - 0s - loss: 0.0227 - accuracy: 0.9917\n",
      "Epoch 239/500\n",
      " - 0s - loss: 0.0227 - accuracy: 0.9918\n",
      "Epoch 240/500\n",
      " - 0s - loss: 0.0226 - accuracy: 0.9916\n",
      "Epoch 241/500\n",
      " - 0s - loss: 0.0226 - accuracy: 0.9916\n",
      "Epoch 242/500\n",
      " - 0s - loss: 0.0226 - accuracy: 0.9916\n",
      "Epoch 243/500\n",
      " - 0s - loss: 0.0226 - accuracy: 0.9917\n",
      "Epoch 244/500\n",
      " - 0s - loss: 0.0225 - accuracy: 0.9917\n",
      "Epoch 245/500\n",
      " - 0s - loss: 0.0225 - accuracy: 0.9917\n",
      "Epoch 246/500\n",
      " - 0s - loss: 0.0224 - accuracy: 0.9916\n",
      "Epoch 247/500\n",
      " - 0s - loss: 0.0224 - accuracy: 0.9917\n",
      "Epoch 248/500\n",
      " - 0s - loss: 0.0224 - accuracy: 0.9917\n",
      "Epoch 249/500\n",
      " - 0s - loss: 0.0224 - accuracy: 0.9917\n",
      "Epoch 250/500\n",
      " - 0s - loss: 0.0224 - accuracy: 0.9918\n",
      "Epoch 251/500\n",
      " - 0s - loss: 0.0223 - accuracy: 0.9917\n",
      "Epoch 252/500\n",
      " - 0s - loss: 0.0223 - accuracy: 0.9917\n",
      "Epoch 253/500\n",
      " - 0s - loss: 0.0223 - accuracy: 0.9917\n",
      "Epoch 254/500\n",
      " - 0s - loss: 0.0223 - accuracy: 0.9918\n",
      "Epoch 255/500\n",
      " - 0s - loss: 0.0222 - accuracy: 0.9918\n",
      "Epoch 256/500\n",
      " - 0s - loss: 0.0222 - accuracy: 0.9918\n",
      "Epoch 257/500\n",
      " - 0s - loss: 0.0221 - accuracy: 0.9917\n",
      "Epoch 258/500\n",
      " - 0s - loss: 0.0221 - accuracy: 0.9918\n",
      "Epoch 259/500\n",
      " - 0s - loss: 0.0221 - accuracy: 0.9918\n",
      "Epoch 260/500\n",
      " - 0s - loss: 0.0221 - accuracy: 0.9919\n",
      "Epoch 261/500\n",
      " - 0s - loss: 0.0220 - accuracy: 0.9918\n",
      "Epoch 262/500\n",
      " - 0s - loss: 0.0220 - accuracy: 0.9919\n",
      "Epoch 263/500\n",
      " - 0s - loss: 0.0220 - accuracy: 0.9919\n",
      "Epoch 264/500\n",
      " - 0s - loss: 0.0220 - accuracy: 0.9919\n",
      "Epoch 265/500\n",
      " - 0s - loss: 0.0219 - accuracy: 0.9919\n",
      "Epoch 266/500\n",
      " - 0s - loss: 0.0219 - accuracy: 0.9919\n",
      "Epoch 267/500\n",
      " - 0s - loss: 0.0219 - accuracy: 0.9918\n",
      "Epoch 268/500\n",
      " - 0s - loss: 0.0219 - accuracy: 0.9918\n",
      "Epoch 269/500\n",
      " - 0s - loss: 0.0218 - accuracy: 0.9919\n",
      "Epoch 270/500\n",
      " - 0s - loss: 0.0218 - accuracy: 0.9919\n",
      "Epoch 271/500\n",
      " - 0s - loss: 0.0218 - accuracy: 0.9917\n",
      "Epoch 272/500\n",
      " - 0s - loss: 0.0218 - accuracy: 0.9919\n",
      "Epoch 273/500\n",
      " - 0s - loss: 0.0218 - accuracy: 0.9919\n",
      "Epoch 274/500\n",
      " - 0s - loss: 0.0217 - accuracy: 0.9919\n",
      "Epoch 275/500\n",
      " - 0s - loss: 0.0217 - accuracy: 0.9919\n",
      "Epoch 276/500\n",
      " - 0s - loss: 0.0217 - accuracy: 0.9919\n",
      "Epoch 277/500\n",
      " - 0s - loss: 0.0217 - accuracy: 0.9918\n",
      "Epoch 278/500\n",
      " - 0s - loss: 0.0217 - accuracy: 0.9919\n",
      "Epoch 279/500\n",
      " - 0s - loss: 0.0216 - accuracy: 0.9919\n",
      "Epoch 280/500\n",
      " - 0s - loss: 0.0216 - accuracy: 0.9919\n",
      "Epoch 281/500\n",
      " - 0s - loss: 0.0215 - accuracy: 0.9922\n",
      "Epoch 282/500\n",
      " - 0s - loss: 0.0215 - accuracy: 0.9919\n",
      "Epoch 283/500\n",
      " - 0s - loss: 0.0216 - accuracy: 0.9919\n",
      "Epoch 284/500\n",
      " - 0s - loss: 0.0215 - accuracy: 0.9920\n",
      "Epoch 285/500\n",
      " - 0s - loss: 0.0215 - accuracy: 0.9919\n",
      "Epoch 286/500\n",
      " - 0s - loss: 0.0214 - accuracy: 0.9922\n",
      "Epoch 287/500\n",
      " - 0s - loss: 0.0214 - accuracy: 0.9919\n",
      "Epoch 288/500\n",
      " - 0s - loss: 0.0214 - accuracy: 0.9922\n",
      "Epoch 289/500\n",
      " - 0s - loss: 0.0214 - accuracy: 0.9922\n",
      "Epoch 290/500\n",
      " - 0s - loss: 0.0214 - accuracy: 0.9924\n",
      "Epoch 291/500\n",
      " - 0s - loss: 0.0213 - accuracy: 0.9921\n",
      "Epoch 292/500\n",
      " - 0s - loss: 0.0213 - accuracy: 0.9920\n",
      "Epoch 293/500\n",
      " - 0s - loss: 0.0213 - accuracy: 0.9922\n",
      "Epoch 294/500\n",
      " - 0s - loss: 0.0212 - accuracy: 0.9922\n",
      "Epoch 295/500\n",
      " - 0s - loss: 0.0212 - accuracy: 0.9922\n",
      "Epoch 296/500\n",
      " - 0s - loss: 0.0212 - accuracy: 0.9921\n",
      "Epoch 297/500\n",
      " - 0s - loss: 0.0212 - accuracy: 0.9921\n",
      "Epoch 298/500\n",
      " - 0s - loss: 0.0211 - accuracy: 0.9922\n",
      "Epoch 299/500\n",
      " - 0s - loss: 0.0211 - accuracy: 0.9923\n",
      "Epoch 300/500\n",
      " - 0s - loss: 0.0211 - accuracy: 0.9922\n",
      "Epoch 301/500\n",
      " - 0s - loss: 0.0211 - accuracy: 0.9920\n",
      "Epoch 302/500\n",
      " - 0s - loss: 0.0211 - accuracy: 0.9920\n",
      "Epoch 303/500\n",
      " - 0s - loss: 0.0211 - accuracy: 0.9922\n",
      "Epoch 304/500\n",
      " - 0s - loss: 0.0210 - accuracy: 0.9923\n",
      "Epoch 305/500\n",
      " - 0s - loss: 0.0210 - accuracy: 0.9919\n",
      "Epoch 306/500\n",
      " - 0s - loss: 0.0210 - accuracy: 0.9921\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 307/500\n",
      " - 0s - loss: 0.0210 - accuracy: 0.9922\n",
      "Epoch 308/500\n",
      " - 0s - loss: 0.0209 - accuracy: 0.9921\n",
      "Epoch 309/500\n",
      " - 0s - loss: 0.0209 - accuracy: 0.9920\n",
      "Epoch 310/500\n",
      " - 0s - loss: 0.0209 - accuracy: 0.9923\n",
      "Epoch 311/500\n",
      " - 0s - loss: 0.0209 - accuracy: 0.9923\n",
      "Epoch 312/500\n",
      " - 0s - loss: 0.0208 - accuracy: 0.9922\n",
      "Epoch 313/500\n",
      " - 0s - loss: 0.0208 - accuracy: 0.9920\n",
      "Epoch 314/500\n",
      " - 0s - loss: 0.0208 - accuracy: 0.9924\n",
      "Epoch 315/500\n",
      " - 0s - loss: 0.0208 - accuracy: 0.9923\n",
      "Epoch 316/500\n",
      " - 0s - loss: 0.0207 - accuracy: 0.9923\n",
      "Epoch 317/500\n",
      " - 0s - loss: 0.0207 - accuracy: 0.9921\n",
      "Epoch 318/500\n",
      " - 0s - loss: 0.0207 - accuracy: 0.9923\n",
      "Epoch 319/500\n",
      " - 0s - loss: 0.0207 - accuracy: 0.9923\n",
      "Epoch 320/500\n",
      " - 0s - loss: 0.0207 - accuracy: 0.9921\n",
      "Epoch 321/500\n",
      " - 0s - loss: 0.0207 - accuracy: 0.9922\n",
      "Epoch 322/500\n",
      " - 0s - loss: 0.0206 - accuracy: 0.9922\n",
      "Epoch 323/500\n",
      " - 0s - loss: 0.0206 - accuracy: 0.9921\n",
      "Epoch 324/500\n",
      " - 0s - loss: 0.0206 - accuracy: 0.9922\n",
      "Epoch 325/500\n",
      " - 0s - loss: 0.0206 - accuracy: 0.9921\n",
      "Epoch 326/500\n",
      " - 0s - loss: 0.0205 - accuracy: 0.9922\n",
      "Epoch 327/500\n",
      " - 0s - loss: 0.0205 - accuracy: 0.9923\n",
      "Epoch 328/500\n",
      " - 0s - loss: 0.0205 - accuracy: 0.9925\n",
      "Epoch 329/500\n",
      " - 0s - loss: 0.0204 - accuracy: 0.9924\n",
      "Epoch 330/500\n",
      " - 0s - loss: 0.0205 - accuracy: 0.9922\n",
      "Epoch 331/500\n",
      " - 0s - loss: 0.0204 - accuracy: 0.9923\n",
      "Epoch 332/500\n",
      " - 0s - loss: 0.0204 - accuracy: 0.9924\n",
      "Epoch 333/500\n",
      " - 0s - loss: 0.0204 - accuracy: 0.9922\n",
      "Epoch 334/500\n",
      " - 0s - loss: 0.0203 - accuracy: 0.9921\n",
      "Epoch 335/500\n",
      " - 0s - loss: 0.0203 - accuracy: 0.9922\n",
      "Epoch 336/500\n",
      " - 0s - loss: 0.0203 - accuracy: 0.9922\n",
      "Epoch 337/500\n",
      " - 0s - loss: 0.0203 - accuracy: 0.9922\n",
      "Epoch 338/500\n",
      " - 0s - loss: 0.0202 - accuracy: 0.9922\n",
      "Epoch 339/500\n",
      " - 0s - loss: 0.0202 - accuracy: 0.9922\n",
      "Epoch 340/500\n",
      " - 0s - loss: 0.0201 - accuracy: 0.9922\n",
      "Epoch 341/500\n",
      " - 0s - loss: 0.0201 - accuracy: 0.9923\n",
      "Epoch 342/500\n",
      " - 0s - loss: 0.0201 - accuracy: 0.9923\n",
      "Epoch 343/500\n",
      " - 0s - loss: 0.0201 - accuracy: 0.9925\n",
      "Epoch 344/500\n",
      " - 0s - loss: 0.0200 - accuracy: 0.9922\n",
      "Epoch 345/500\n",
      " - 0s - loss: 0.0200 - accuracy: 0.9924\n",
      "Epoch 346/500\n",
      " - 0s - loss: 0.0200 - accuracy: 0.9921\n",
      "Epoch 347/500\n",
      " - 0s - loss: 0.0200 - accuracy: 0.9922\n",
      "Epoch 348/500\n",
      " - 0s - loss: 0.0200 - accuracy: 0.9923\n",
      "Epoch 349/500\n",
      " - 0s - loss: 0.0199 - accuracy: 0.9922\n",
      "Epoch 350/500\n",
      " - 0s - loss: 0.0199 - accuracy: 0.9922\n",
      "Epoch 351/500\n",
      " - 0s - loss: 0.0198 - accuracy: 0.9923\n",
      "Epoch 352/500\n",
      " - 0s - loss: 0.0198 - accuracy: 0.9921\n",
      "Epoch 353/500\n",
      " - 0s - loss: 0.0198 - accuracy: 0.9922\n",
      "Epoch 354/500\n",
      " - 0s - loss: 0.0198 - accuracy: 0.9923\n",
      "Epoch 355/500\n",
      " - 0s - loss: 0.0197 - accuracy: 0.9921\n",
      "Epoch 356/500\n",
      " - 0s - loss: 0.0197 - accuracy: 0.9924\n",
      "Epoch 357/500\n",
      " - 0s - loss: 0.0197 - accuracy: 0.9921\n",
      "Epoch 358/500\n",
      " - 0s - loss: 0.0197 - accuracy: 0.9923\n",
      "Epoch 359/500\n",
      " - 0s - loss: 0.0196 - accuracy: 0.9922\n",
      "Epoch 360/500\n",
      " - 0s - loss: 0.0196 - accuracy: 0.9923\n",
      "Epoch 361/500\n",
      " - 0s - loss: 0.0196 - accuracy: 0.9924\n",
      "Epoch 362/500\n",
      " - 0s - loss: 0.0195 - accuracy: 0.9922\n",
      "Epoch 363/500\n",
      " - 0s - loss: 0.0195 - accuracy: 0.9922\n",
      "Epoch 364/500\n",
      " - 0s - loss: 0.0195 - accuracy: 0.9921\n",
      "Epoch 365/500\n",
      " - 0s - loss: 0.0195 - accuracy: 0.9923\n",
      "Epoch 366/500\n",
      " - 0s - loss: 0.0194 - accuracy: 0.9920\n",
      "Epoch 367/500\n",
      " - 0s - loss: 0.0193 - accuracy: 0.9921\n",
      "Epoch 368/500\n",
      " - 0s - loss: 0.0193 - accuracy: 0.9921\n",
      "Epoch 369/500\n",
      " - 0s - loss: 0.0193 - accuracy: 0.9922\n",
      "Epoch 370/500\n",
      " - 0s - loss: 0.0193 - accuracy: 0.9920\n",
      "Epoch 371/500\n",
      " - 0s - loss: 0.0192 - accuracy: 0.9921\n",
      "Epoch 372/500\n",
      " - 0s - loss: 0.0192 - accuracy: 0.9920\n",
      "Epoch 373/500\n",
      " - 0s - loss: 0.0191 - accuracy: 0.9923\n",
      "Epoch 374/500\n",
      " - 0s - loss: 0.0191 - accuracy: 0.9923\n",
      "Epoch 375/500\n",
      " - 0s - loss: 0.0190 - accuracy: 0.9922\n",
      "Epoch 376/500\n",
      " - 0s - loss: 0.0190 - accuracy: 0.9921\n",
      "Epoch 377/500\n",
      " - 0s - loss: 0.0190 - accuracy: 0.9923\n",
      "Epoch 378/500\n",
      " - 0s - loss: 0.0189 - accuracy: 0.9922\n",
      "Epoch 379/500\n",
      " - 0s - loss: 0.0189 - accuracy: 0.9923\n",
      "Epoch 380/500\n",
      " - 0s - loss: 0.0189 - accuracy: 0.9922\n",
      "Epoch 381/500\n",
      " - 0s - loss: 0.0188 - accuracy: 0.9923\n",
      "Epoch 382/500\n",
      " - 0s - loss: 0.0188 - accuracy: 0.9923\n",
      "Epoch 383/500\n",
      " - 0s - loss: 0.0187 - accuracy: 0.9921\n",
      "Epoch 384/500\n",
      " - 0s - loss: 0.0188 - accuracy: 0.9922\n",
      "Epoch 385/500\n",
      " - 0s - loss: 0.0187 - accuracy: 0.9923\n",
      "Epoch 386/500\n",
      " - 0s - loss: 0.0187 - accuracy: 0.9923\n",
      "Epoch 387/500\n",
      " - 0s - loss: 0.0186 - accuracy: 0.9923\n",
      "Epoch 388/500\n",
      " - 0s - loss: 0.0186 - accuracy: 0.9923\n",
      "Epoch 389/500\n",
      " - 0s - loss: 0.0185 - accuracy: 0.9924\n",
      "Epoch 390/500\n",
      " - 0s - loss: 0.0185 - accuracy: 0.9922\n",
      "Epoch 391/500\n",
      " - 0s - loss: 0.0185 - accuracy: 0.9922\n",
      "Epoch 392/500\n",
      " - 0s - loss: 0.0184 - accuracy: 0.9923\n",
      "Epoch 393/500\n",
      " - 0s - loss: 0.0184 - accuracy: 0.9923\n",
      "Epoch 394/500\n",
      " - 0s - loss: 0.0183 - accuracy: 0.9924\n",
      "Epoch 395/500\n",
      " - 0s - loss: 0.0184 - accuracy: 0.9924\n",
      "Epoch 396/500\n",
      " - 0s - loss: 0.0183 - accuracy: 0.9923\n",
      "Epoch 397/500\n",
      " - 0s - loss: 0.0183 - accuracy: 0.9924\n",
      "Epoch 398/500\n",
      " - 0s - loss: 0.0183 - accuracy: 0.9924\n",
      "Epoch 399/500\n",
      " - 0s - loss: 0.0182 - accuracy: 0.9923\n",
      "Epoch 400/500\n",
      " - 0s - loss: 0.0182 - accuracy: 0.9923\n",
      "Epoch 401/500\n",
      " - 0s - loss: 0.0181 - accuracy: 0.9925\n",
      "Epoch 402/500\n",
      " - 0s - loss: 0.0182 - accuracy: 0.9925\n",
      "Epoch 403/500\n",
      " - 0s - loss: 0.0181 - accuracy: 0.9926\n",
      "Epoch 404/500\n",
      " - 0s - loss: 0.0181 - accuracy: 0.9925\n",
      "Epoch 405/500\n",
      " - 0s - loss: 0.0180 - accuracy: 0.9926\n",
      "Epoch 406/500\n",
      " - 0s - loss: 0.0180 - accuracy: 0.9927\n",
      "Epoch 407/500\n",
      " - 0s - loss: 0.0179 - accuracy: 0.9926\n",
      "Epoch 408/500\n",
      " - 0s - loss: 0.0179 - accuracy: 0.9924\n",
      "Epoch 409/500\n",
      " - 0s - loss: 0.0179 - accuracy: 0.9924\n",
      "Epoch 410/500\n",
      " - 0s - loss: 0.0179 - accuracy: 0.9928\n",
      "Epoch 411/500\n",
      " - 0s - loss: 0.0178 - accuracy: 0.9924\n",
      "Epoch 412/500\n",
      " - 0s - loss: 0.0178 - accuracy: 0.9928\n",
      "Epoch 413/500\n",
      " - 0s - loss: 0.0178 - accuracy: 0.9928\n",
      "Epoch 414/500\n",
      " - 0s - loss: 0.0178 - accuracy: 0.9925\n",
      "Epoch 415/500\n",
      " - 0s - loss: 0.0177 - accuracy: 0.9927\n",
      "Epoch 416/500\n",
      " - 0s - loss: 0.0177 - accuracy: 0.9929\n",
      "Epoch 417/500\n",
      " - 0s - loss: 0.0177 - accuracy: 0.9929\n",
      "Epoch 418/500\n",
      " - 0s - loss: 0.0176 - accuracy: 0.9927\n",
      "Epoch 419/500\n",
      " - 0s - loss: 0.0176 - accuracy: 0.9927\n",
      "Epoch 420/500\n",
      " - 0s - loss: 0.0176 - accuracy: 0.9926\n",
      "Epoch 421/500\n",
      " - 0s - loss: 0.0176 - accuracy: 0.9928\n",
      "Epoch 422/500\n",
      " - 0s - loss: 0.0175 - accuracy: 0.9926\n",
      "Epoch 423/500\n",
      " - 0s - loss: 0.0175 - accuracy: 0.9927\n",
      "Epoch 424/500\n",
      " - 0s - loss: 0.0174 - accuracy: 0.9927\n",
      "Epoch 425/500\n",
      " - 0s - loss: 0.0175 - accuracy: 0.9927\n",
      "Epoch 426/500\n",
      " - 0s - loss: 0.0174 - accuracy: 0.9928\n",
      "Epoch 427/500\n",
      " - 0s - loss: 0.0174 - accuracy: 0.9928\n",
      "Epoch 428/500\n",
      " - 0s - loss: 0.0174 - accuracy: 0.9930\n",
      "Epoch 429/500\n",
      " - 0s - loss: 0.0173 - accuracy: 0.9927\n",
      "Epoch 430/500\n",
      " - 0s - loss: 0.0173 - accuracy: 0.9927\n",
      "Epoch 431/500\n",
      " - 0s - loss: 0.0173 - accuracy: 0.9931\n",
      "Epoch 432/500\n",
      " - 0s - loss: 0.0173 - accuracy: 0.9930\n",
      "Epoch 433/500\n",
      " - 0s - loss: 0.0172 - accuracy: 0.9927\n",
      "Epoch 434/500\n",
      " - 0s - loss: 0.0172 - accuracy: 0.9930\n",
      "Epoch 435/500\n",
      " - 0s - loss: 0.0172 - accuracy: 0.9930\n",
      "Epoch 436/500\n",
      " - 0s - loss: 0.0172 - accuracy: 0.9930\n",
      "Epoch 437/500\n",
      " - 0s - loss: 0.0172 - accuracy: 0.9928\n",
      "Epoch 438/500\n",
      " - 0s - loss: 0.0172 - accuracy: 0.9930\n",
      "Epoch 439/500\n",
      " - 0s - loss: 0.0171 - accuracy: 0.9929\n",
      "Epoch 440/500\n",
      " - 0s - loss: 0.0171 - accuracy: 0.9930\n",
      "Epoch 441/500\n",
      " - 0s - loss: 0.0171 - accuracy: 0.9929\n",
      "Epoch 442/500\n",
      " - 0s - loss: 0.0170 - accuracy: 0.9929\n",
      "Epoch 443/500\n",
      " - 0s - loss: 0.0170 - accuracy: 0.9929\n",
      "Epoch 444/500\n",
      " - 0s - loss: 0.0170 - accuracy: 0.9930\n",
      "Epoch 445/500\n",
      " - 0s - loss: 0.0170 - accuracy: 0.9930\n",
      "Epoch 446/500\n",
      " - 0s - loss: 0.0170 - accuracy: 0.9930\n",
      "Epoch 447/500\n",
      " - 0s - loss: 0.0169 - accuracy: 0.9931\n",
      "Epoch 448/500\n",
      " - 0s - loss: 0.0169 - accuracy: 0.9931\n",
      "Epoch 449/500\n",
      " - 0s - loss: 0.0169 - accuracy: 0.9930\n",
      "Epoch 450/500\n",
      " - 0s - loss: 0.0169 - accuracy: 0.9932\n",
      "Epoch 451/500\n",
      " - 0s - loss: 0.0169 - accuracy: 0.9931\n",
      "Epoch 452/500\n",
      " - 0s - loss: 0.0168 - accuracy: 0.9931\n",
      "Epoch 453/500\n",
      " - 0s - loss: 0.0168 - accuracy: 0.9929\n",
      "Epoch 454/500\n",
      " - 0s - loss: 0.0168 - accuracy: 0.9929\n",
      "Epoch 455/500\n",
      " - 0s - loss: 0.0168 - accuracy: 0.9934\n",
      "Epoch 456/500\n",
      " - 0s - loss: 0.0167 - accuracy: 0.9929\n",
      "Epoch 457/500\n",
      " - 0s - loss: 0.0167 - accuracy: 0.9930\n",
      "Epoch 458/500\n",
      " - 0s - loss: 0.0167 - accuracy: 0.9931\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 459/500\n",
      " - 0s - loss: 0.0167 - accuracy: 0.9931\n",
      "Epoch 460/500\n",
      " - 0s - loss: 0.0167 - accuracy: 0.9928\n",
      "Epoch 461/500\n",
      " - 0s - loss: 0.0166 - accuracy: 0.9931\n",
      "Epoch 462/500\n",
      " - 0s - loss: 0.0166 - accuracy: 0.9927\n",
      "Epoch 463/500\n",
      " - 0s - loss: 0.0166 - accuracy: 0.9930\n",
      "Epoch 464/500\n",
      " - 0s - loss: 0.0166 - accuracy: 0.9931\n",
      "Epoch 465/500\n",
      " - 0s - loss: 0.0166 - accuracy: 0.9931\n",
      "Epoch 466/500\n",
      " - 0s - loss: 0.0165 - accuracy: 0.9931\n",
      "Epoch 467/500\n",
      " - 0s - loss: 0.0165 - accuracy: 0.9932\n",
      "Epoch 468/500\n",
      " - 0s - loss: 0.0165 - accuracy: 0.9931\n",
      "Epoch 469/500\n",
      " - 0s - loss: 0.0165 - accuracy: 0.9930\n",
      "Epoch 470/500\n",
      " - 0s - loss: 0.0164 - accuracy: 0.9931\n",
      "Epoch 471/500\n",
      " - 0s - loss: 0.0164 - accuracy: 0.9930\n",
      "Epoch 472/500\n",
      " - 0s - loss: 0.0164 - accuracy: 0.9933\n",
      "Epoch 473/500\n",
      " - 0s - loss: 0.0164 - accuracy: 0.9931\n",
      "Epoch 474/500\n",
      " - 0s - loss: 0.0164 - accuracy: 0.9933\n",
      "Epoch 475/500\n",
      " - 0s - loss: 0.0164 - accuracy: 0.9933\n",
      "Epoch 476/500\n",
      " - 0s - loss: 0.0164 - accuracy: 0.9933\n",
      "Epoch 477/500\n",
      " - 0s - loss: 0.0163 - accuracy: 0.9931\n",
      "Epoch 478/500\n",
      " - 0s - loss: 0.0163 - accuracy: 0.9932\n",
      "Epoch 479/500\n",
      " - 0s - loss: 0.0163 - accuracy: 0.9931\n",
      "Epoch 480/500\n",
      " - 0s - loss: 0.0162 - accuracy: 0.9933\n",
      "Epoch 481/500\n",
      " - 0s - loss: 0.0162 - accuracy: 0.9932\n",
      "Epoch 482/500\n",
      " - 0s - loss: 0.0162 - accuracy: 0.9931\n",
      "Epoch 483/500\n",
      " - 0s - loss: 0.0162 - accuracy: 0.9933\n",
      "Epoch 484/500\n",
      " - 0s - loss: 0.0162 - accuracy: 0.9931\n",
      "Epoch 485/500\n",
      " - 0s - loss: 0.0161 - accuracy: 0.9931\n",
      "Epoch 486/500\n",
      " - 0s - loss: 0.0161 - accuracy: 0.9932\n",
      "Epoch 487/500\n",
      " - 0s - loss: 0.0161 - accuracy: 0.9931\n",
      "Epoch 488/500\n",
      " - 0s - loss: 0.0161 - accuracy: 0.9935\n",
      "Epoch 489/500\n",
      " - 0s - loss: 0.0161 - accuracy: 0.9932\n",
      "Epoch 490/500\n",
      " - 0s - loss: 0.0161 - accuracy: 0.9933\n",
      "Epoch 491/500\n",
      " - 0s - loss: 0.0160 - accuracy: 0.9933\n",
      "Epoch 492/500\n",
      " - 0s - loss: 0.0161 - accuracy: 0.9931\n",
      "Epoch 493/500\n",
      " - 0s - loss: 0.0160 - accuracy: 0.9933\n",
      "Epoch 494/500\n",
      " - 0s - loss: 0.0160 - accuracy: 0.9933\n",
      "Epoch 495/500\n",
      " - 0s - loss: 0.0160 - accuracy: 0.9931\n",
      "Epoch 496/500\n",
      " - 0s - loss: 0.0160 - accuracy: 0.9933\n",
      "Epoch 497/500\n",
      " - 0s - loss: 0.0159 - accuracy: 0.9934\n",
      "Epoch 498/500\n",
      " - 0s - loss: 0.0159 - accuracy: 0.9931\n",
      "Epoch 499/500\n",
      " - 0s - loss: 0.0159 - accuracy: 0.9931\n",
      "Epoch 500/500\n",
      " - 0s - loss: 0.0159 - accuracy: 0.9931\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.callbacks.History at 0x1cb91d45430>"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(train[:, 1:], train[:, 0], epochs=500, batch_size=500, verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_pred = model.predict(test[:, 1:]).reshape((-1,)) >= 0.5\n",
    "\n",
    "conf_mat = confusion_matrix(y_test, y_pred)\n",
    "\n",
    "lang = ['No Fault (0)', 'Fault (1)']\n",
    "ax = plt.axes()\n",
    "sns.heatmap(conf_mat, square=True, annot=True, cmap='Blues', fmt='d', cbar=True, ax=ax, xticklabels=lang, yticklabels=lang)\n",
    "ax.set_title('Confusion Matrix')\n",
    "plt.xlabel('Predicted Label')\n",
    "plt.ylabel('True Label')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The Evaluation Metrics of the MLP Model-3 (on the test set) is as follows:\n",
      "\n",
      "True Positives: 542\n",
      "True Negatives: 650\n",
      "False Positives: 1\n",
      "False Negatives: 8\n",
      "\n",
      "Precision: 0.998158379373849\n",
      "Recall: 0.9854545454545455\n",
      "\n",
      "Accuracy: 0.9925062447960034\n"
     ]
    }
   ],
   "source": [
    "true_positive = conf_mat[1, 1]\n",
    "true_negative = conf_mat[0, 0]\n",
    "false_positive = conf_mat[0, 1]\n",
    "false_negative = conf_mat[1, 0]\n",
    "accuracy = accuracy_score(y_pred, y_test)\n",
    "\n",
    "print(\"The Evaluation Metrics of the MLP Model-3 (on the test set) is as follows:\\n\")\n",
    "\n",
    "print(\"True Positives:\", true_positive)\n",
    "print(\"True Negatives:\", true_negative)\n",
    "print(\"False Positives:\", false_positive)\n",
    "print(\"False Negatives:\", false_negative, end='\\n\\n')\n",
    "\n",
    "print(\"Precision:\", true_positive/(true_positive+false_positive))\n",
    "print(\"Recall:\", true_positive/(true_positive+false_negative), end='\\n\\n')\n",
    "\n",
    "print(\"Accuracy:\", accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### (iv) Model 4: (2 hidden layers, relu activation, RMSprop optimizer)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "model.add(Dense(10, input_dim=6, activation='relu'))\n",
    "model.add(Dense(10, activation='relu'))\n",
    "model.add(Dense(1, activation='sigmoid'))\n",
    "model.compile(loss='binary_crossentropy', optimizer='RMSprop', metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/500\n",
      " - 0s - loss: 0.6790 - accuracy: 0.6061\n",
      "Epoch 2/500\n",
      " - 0s - loss: 0.6136 - accuracy: 0.7532\n",
      "Epoch 3/500\n",
      " - 0s - loss: 0.5536 - accuracy: 0.8138\n",
      "Epoch 4/500\n",
      " - 0s - loss: 0.4948 - accuracy: 0.8456\n",
      "Epoch 5/500\n",
      " - 0s - loss: 0.4371 - accuracy: 0.8845\n",
      "Epoch 6/500\n",
      " - 0s - loss: 0.3814 - accuracy: 0.9106\n",
      "Epoch 7/500\n",
      " - 0s - loss: 0.3290 - accuracy: 0.9306\n",
      "Epoch 8/500\n",
      " - 0s - loss: 0.2818 - accuracy: 0.9419\n",
      "Epoch 9/500\n",
      " - 0s - loss: 0.2403 - accuracy: 0.9476\n",
      "Epoch 10/500\n",
      " - 0s - loss: 0.2059 - accuracy: 0.9519\n",
      "Epoch 11/500\n",
      " - 0s - loss: 0.1773 - accuracy: 0.9557\n",
      "Epoch 12/500\n",
      " - 0s - loss: 0.1537 - accuracy: 0.9598\n",
      "Epoch 13/500\n",
      " - 0s - loss: 0.1343 - accuracy: 0.9636\n",
      "Epoch 14/500\n",
      " - 0s - loss: 0.1187 - accuracy: 0.9656\n",
      "Epoch 15/500\n",
      " - 0s - loss: 0.1056 - accuracy: 0.9691\n",
      "Epoch 16/500\n",
      " - 0s - loss: 0.0948 - accuracy: 0.9710\n",
      "Epoch 17/500\n",
      " - 0s - loss: 0.0856 - accuracy: 0.9737\n",
      "Epoch 18/500\n",
      " - 0s - loss: 0.0781 - accuracy: 0.9759\n",
      "Epoch 19/500\n",
      " - 0s - loss: 0.0718 - accuracy: 0.9771\n",
      "Epoch 20/500\n",
      " - 0s - loss: 0.0663 - accuracy: 0.9786\n",
      "Epoch 21/500\n",
      " - 0s - loss: 0.0615 - accuracy: 0.9800\n",
      "Epoch 22/500\n",
      " - 0s - loss: 0.0574 - accuracy: 0.9808\n",
      "Epoch 23/500\n",
      " - 0s - loss: 0.0539 - accuracy: 0.9819\n",
      "Epoch 24/500\n",
      " - 0s - loss: 0.0510 - accuracy: 0.9830\n",
      "Epoch 25/500\n",
      " - 0s - loss: 0.0481 - accuracy: 0.9838\n",
      "Epoch 26/500\n",
      " - 0s - loss: 0.0457 - accuracy: 0.9847\n",
      "Epoch 27/500\n",
      " - 0s - loss: 0.0435 - accuracy: 0.9846\n",
      "Epoch 28/500\n",
      " - 0s - loss: 0.0416 - accuracy: 0.9856\n",
      "Epoch 29/500\n",
      " - 0s - loss: 0.0399 - accuracy: 0.9860\n",
      "Epoch 30/500\n",
      " - 0s - loss: 0.0384 - accuracy: 0.9866\n",
      "Epoch 31/500\n",
      " - 0s - loss: 0.0370 - accuracy: 0.9869\n",
      "Epoch 32/500\n",
      " - 0s - loss: 0.0359 - accuracy: 0.9869\n",
      "Epoch 33/500\n",
      " - 0s - loss: 0.0349 - accuracy: 0.9875\n",
      "Epoch 34/500\n",
      " - 0s - loss: 0.0339 - accuracy: 0.9878\n",
      "Epoch 35/500\n",
      " - 0s - loss: 0.0330 - accuracy: 0.9881\n",
      "Epoch 36/500\n",
      " - 0s - loss: 0.0321 - accuracy: 0.9882\n",
      "Epoch 37/500\n",
      " - 0s - loss: 0.0314 - accuracy: 0.9889\n",
      "Epoch 38/500\n",
      " - 0s - loss: 0.0306 - accuracy: 0.9886\n",
      "Epoch 39/500\n",
      " - 0s - loss: 0.0300 - accuracy: 0.9892\n",
      "Epoch 40/500\n",
      " - 0s - loss: 0.0294 - accuracy: 0.9892\n",
      "Epoch 41/500\n",
      " - 0s - loss: 0.0288 - accuracy: 0.9893\n",
      "Epoch 42/500\n",
      " - 0s - loss: 0.0284 - accuracy: 0.9898\n",
      "Epoch 43/500\n",
      " - 0s - loss: 0.0280 - accuracy: 0.9902\n",
      "Epoch 44/500\n",
      " - 0s - loss: 0.0275 - accuracy: 0.9901\n",
      "Epoch 45/500\n",
      " - 0s - loss: 0.0271 - accuracy: 0.9905\n",
      "Epoch 46/500\n",
      " - 0s - loss: 0.0268 - accuracy: 0.9905\n",
      "Epoch 47/500\n",
      " - 0s - loss: 0.0264 - accuracy: 0.9907\n",
      "Epoch 48/500\n",
      " - 0s - loss: 0.0261 - accuracy: 0.9906\n",
      "Epoch 49/500\n",
      " - 0s - loss: 0.0257 - accuracy: 0.9906\n",
      "Epoch 50/500\n",
      " - 0s - loss: 0.0253 - accuracy: 0.9909\n",
      "Epoch 51/500\n",
      " - 0s - loss: 0.0251 - accuracy: 0.9911\n",
      "Epoch 52/500\n",
      " - 0s - loss: 0.0247 - accuracy: 0.9910\n",
      "Epoch 53/500\n",
      " - 0s - loss: 0.0245 - accuracy: 0.9910\n",
      "Epoch 54/500\n",
      " - 0s - loss: 0.0241 - accuracy: 0.9913\n",
      "Epoch 55/500\n",
      " - 0s - loss: 0.0240 - accuracy: 0.9909\n",
      "Epoch 56/500\n",
      " - 0s - loss: 0.0237 - accuracy: 0.9911\n",
      "Epoch 57/500\n",
      " - 0s - loss: 0.0235 - accuracy: 0.9912\n",
      "Epoch 58/500\n",
      " - 0s - loss: 0.0231 - accuracy: 0.9912\n",
      "Epoch 59/500\n",
      " - 0s - loss: 0.0229 - accuracy: 0.9912\n",
      "Epoch 60/500\n",
      " - 0s - loss: 0.0227 - accuracy: 0.9919\n",
      "Epoch 61/500\n",
      " - 0s - loss: 0.0224 - accuracy: 0.9918\n",
      "Epoch 62/500\n",
      " - 0s - loss: 0.0223 - accuracy: 0.9915\n",
      "Epoch 63/500\n",
      " - 0s - loss: 0.0221 - accuracy: 0.9916\n",
      "Epoch 64/500\n",
      " - 0s - loss: 0.0218 - accuracy: 0.9915\n",
      "Epoch 65/500\n",
      " - 0s - loss: 0.0216 - accuracy: 0.9918\n",
      "Epoch 66/500\n",
      " - 0s - loss: 0.0215 - accuracy: 0.9919\n",
      "Epoch 67/500\n",
      " - 0s - loss: 0.0213 - accuracy: 0.9925\n",
      "Epoch 68/500\n",
      " - 0s - loss: 0.0210 - accuracy: 0.9919\n",
      "Epoch 69/500\n",
      " - 0s - loss: 0.0207 - accuracy: 0.9922\n",
      "Epoch 70/500\n",
      " - 0s - loss: 0.0205 - accuracy: 0.9922\n",
      "Epoch 71/500\n",
      " - 0s - loss: 0.0205 - accuracy: 0.9921\n",
      "Epoch 72/500\n",
      " - 0s - loss: 0.0201 - accuracy: 0.9925\n",
      "Epoch 73/500\n",
      " - 0s - loss: 0.0200 - accuracy: 0.9922\n",
      "Epoch 74/500\n",
      " - 0s - loss: 0.0198 - accuracy: 0.9925\n",
      "Epoch 75/500\n",
      " - 0s - loss: 0.0197 - accuracy: 0.9922\n",
      "Epoch 76/500\n",
      " - 0s - loss: 0.0193 - accuracy: 0.9926\n",
      "Epoch 77/500\n",
      " - 0s - loss: 0.0193 - accuracy: 0.9925\n",
      "Epoch 78/500\n",
      " - 0s - loss: 0.0191 - accuracy: 0.9927\n",
      "Epoch 79/500\n",
      " - 0s - loss: 0.0190 - accuracy: 0.9933\n",
      "Epoch 80/500\n",
      " - 0s - loss: 0.0188 - accuracy: 0.9924\n",
      "Epoch 81/500\n",
      " - 0s - loss: 0.0188 - accuracy: 0.9930\n",
      "Epoch 82/500\n",
      " - 0s - loss: 0.0185 - accuracy: 0.9929\n",
      "Epoch 83/500\n",
      " - 0s - loss: 0.0185 - accuracy: 0.9928\n",
      "Epoch 84/500\n",
      " - 0s - loss: 0.0183 - accuracy: 0.9928\n",
      "Epoch 85/500\n",
      " - 0s - loss: 0.0181 - accuracy: 0.9926\n",
      "Epoch 86/500\n",
      " - 0s - loss: 0.0180 - accuracy: 0.9928\n",
      "Epoch 87/500\n",
      " - 0s - loss: 0.0180 - accuracy: 0.9931\n",
      "Epoch 88/500\n",
      " - 0s - loss: 0.0178 - accuracy: 0.9931\n",
      "Epoch 89/500\n",
      " - 0s - loss: 0.0177 - accuracy: 0.9931\n",
      "Epoch 90/500\n",
      " - 0s - loss: 0.0176 - accuracy: 0.9930\n",
      "Epoch 91/500\n",
      " - 0s - loss: 0.0175 - accuracy: 0.9932\n",
      "Epoch 92/500\n",
      " - 0s - loss: 0.0173 - accuracy: 0.9933\n",
      "Epoch 93/500\n",
      " - 0s - loss: 0.0173 - accuracy: 0.9933\n",
      "Epoch 94/500\n",
      " - 0s - loss: 0.0172 - accuracy: 0.9931\n",
      "Epoch 95/500\n",
      " - 0s - loss: 0.0171 - accuracy: 0.9934\n",
      "Epoch 96/500\n",
      " - 0s - loss: 0.0171 - accuracy: 0.9930\n",
      "Epoch 97/500\n",
      " - 0s - loss: 0.0169 - accuracy: 0.9932\n",
      "Epoch 98/500\n",
      " - 0s - loss: 0.0168 - accuracy: 0.9931\n",
      "Epoch 99/500\n",
      " - 0s - loss: 0.0165 - accuracy: 0.9937\n",
      "Epoch 100/500\n",
      " - 0s - loss: 0.0167 - accuracy: 0.9932\n",
      "Epoch 101/500\n",
      " - 0s - loss: 0.0166 - accuracy: 0.9935\n",
      "Epoch 102/500\n",
      " - 0s - loss: 0.0163 - accuracy: 0.9937\n",
      "Epoch 103/500\n",
      " - 0s - loss: 0.0166 - accuracy: 0.9935\n",
      "Epoch 104/500\n",
      " - 0s - loss: 0.0164 - accuracy: 0.9932\n",
      "Epoch 105/500\n",
      " - 0s - loss: 0.0162 - accuracy: 0.9936\n",
      "Epoch 106/500\n",
      " - 0s - loss: 0.0164 - accuracy: 0.9940\n",
      "Epoch 107/500\n",
      " - 0s - loss: 0.0162 - accuracy: 0.9936\n",
      "Epoch 108/500\n",
      " - 0s - loss: 0.0160 - accuracy: 0.9935\n",
      "Epoch 109/500\n",
      " - 0s - loss: 0.0159 - accuracy: 0.9938\n",
      "Epoch 110/500\n",
      " - 0s - loss: 0.0159 - accuracy: 0.9942\n",
      "Epoch 111/500\n",
      " - 0s - loss: 0.0159 - accuracy: 0.9935\n",
      "Epoch 112/500\n",
      " - 0s - loss: 0.0158 - accuracy: 0.9942\n",
      "Epoch 113/500\n",
      " - 0s - loss: 0.0157 - accuracy: 0.9934\n",
      "Epoch 114/500\n",
      " - 0s - loss: 0.0156 - accuracy: 0.9935\n",
      "Epoch 115/500\n",
      " - 0s - loss: 0.0157 - accuracy: 0.9938\n",
      "Epoch 116/500\n",
      " - 0s - loss: 0.0157 - accuracy: 0.9938\n",
      "Epoch 117/500\n",
      " - 0s - loss: 0.0154 - accuracy: 0.9938\n",
      "Epoch 118/500\n",
      " - 0s - loss: 0.0154 - accuracy: 0.9935\n",
      "Epoch 119/500\n",
      " - 0s - loss: 0.0154 - accuracy: 0.9940\n",
      "Epoch 120/500\n",
      " - 0s - loss: 0.0153 - accuracy: 0.9936\n",
      "Epoch 121/500\n",
      " - 0s - loss: 0.0152 - accuracy: 0.9939\n",
      "Epoch 122/500\n",
      " - 0s - loss: 0.0152 - accuracy: 0.9940\n",
      "Epoch 123/500\n",
      " - 0s - loss: 0.0150 - accuracy: 0.9938\n",
      "Epoch 124/500\n",
      " - 0s - loss: 0.0152 - accuracy: 0.9936\n",
      "Epoch 125/500\n",
      " - 0s - loss: 0.0149 - accuracy: 0.9939\n",
      "Epoch 126/500\n",
      " - 0s - loss: 0.0151 - accuracy: 0.9941\n",
      "Epoch 127/500\n",
      " - 0s - loss: 0.0150 - accuracy: 0.9937\n",
      "Epoch 128/500\n",
      " - 0s - loss: 0.0149 - accuracy: 0.9943\n",
      "Epoch 129/500\n",
      " - 0s - loss: 0.0148 - accuracy: 0.9940\n",
      "Epoch 130/500\n",
      " - 0s - loss: 0.0148 - accuracy: 0.9942\n",
      "Epoch 131/500\n",
      " - 0s - loss: 0.0147 - accuracy: 0.9938\n",
      "Epoch 132/500\n",
      " - 0s - loss: 0.0147 - accuracy: 0.9940\n",
      "Epoch 133/500\n",
      " - 0s - loss: 0.0145 - accuracy: 0.9937\n",
      "Epoch 134/500\n",
      " - 0s - loss: 0.0145 - accuracy: 0.9940\n",
      "Epoch 135/500\n",
      " - 0s - loss: 0.0143 - accuracy: 0.9938\n",
      "Epoch 136/500\n",
      " - 0s - loss: 0.0145 - accuracy: 0.9942\n",
      "Epoch 137/500\n",
      " - 0s - loss: 0.0144 - accuracy: 0.9941\n",
      "Epoch 138/500\n",
      " - 0s - loss: 0.0141 - accuracy: 0.9941\n",
      "Epoch 139/500\n",
      " - 0s - loss: 0.0143 - accuracy: 0.9940\n",
      "Epoch 140/500\n",
      " - 0s - loss: 0.0142 - accuracy: 0.9942\n",
      "Epoch 141/500\n",
      " - 0s - loss: 0.0141 - accuracy: 0.9943\n",
      "Epoch 142/500\n",
      " - 0s - loss: 0.0141 - accuracy: 0.9942\n",
      "Epoch 143/500\n",
      " - 0s - loss: 0.0140 - accuracy: 0.9940\n",
      "Epoch 144/500\n",
      " - 0s - loss: 0.0140 - accuracy: 0.9939\n",
      "Epoch 145/500\n",
      " - 0s - loss: 0.0139 - accuracy: 0.9940\n",
      "Epoch 146/500\n",
      " - 0s - loss: 0.0139 - accuracy: 0.9945\n",
      "Epoch 147/500\n",
      " - 0s - loss: 0.0138 - accuracy: 0.9943\n",
      "Epoch 148/500\n",
      " - 0s - loss: 0.0136 - accuracy: 0.9944\n",
      "Epoch 149/500\n",
      " - 0s - loss: 0.0136 - accuracy: 0.9944\n",
      "Epoch 150/500\n",
      " - 0s - loss: 0.0137 - accuracy: 0.9941\n",
      "Epoch 151/500\n",
      " - 0s - loss: 0.0136 - accuracy: 0.9945\n",
      "Epoch 152/500\n",
      " - 0s - loss: 0.0135 - accuracy: 0.9943\n",
      "Epoch 153/500\n",
      " - 0s - loss: 0.0135 - accuracy: 0.9944\n",
      "Epoch 154/500\n",
      " - 0s - loss: 0.0136 - accuracy: 0.9942\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 155/500\n",
      " - 0s - loss: 0.0135 - accuracy: 0.9942\n",
      "Epoch 156/500\n",
      " - 0s - loss: 0.0135 - accuracy: 0.9944\n",
      "Epoch 157/500\n",
      " - 0s - loss: 0.0134 - accuracy: 0.9943\n",
      "Epoch 158/500\n",
      " - 0s - loss: 0.0132 - accuracy: 0.9945\n",
      "Epoch 159/500\n",
      " - 0s - loss: 0.0133 - accuracy: 0.9945\n",
      "Epoch 160/500\n",
      " - 0s - loss: 0.0131 - accuracy: 0.9945\n",
      "Epoch 161/500\n",
      " - 0s - loss: 0.0131 - accuracy: 0.9943\n",
      "Epoch 162/500\n",
      " - 0s - loss: 0.0132 - accuracy: 0.9942\n",
      "Epoch 163/500\n",
      " - 0s - loss: 0.0130 - accuracy: 0.9945\n",
      "Epoch 164/500\n",
      " - 0s - loss: 0.0131 - accuracy: 0.9943\n",
      "Epoch 165/500\n",
      " - 0s - loss: 0.0130 - accuracy: 0.9945\n",
      "Epoch 166/500\n",
      " - 0s - loss: 0.0130 - accuracy: 0.9944\n",
      "Epoch 167/500\n",
      " - 0s - loss: 0.0129 - accuracy: 0.9946\n",
      "Epoch 168/500\n",
      " - 0s - loss: 0.0129 - accuracy: 0.9946\n",
      "Epoch 169/500\n",
      " - 0s - loss: 0.0127 - accuracy: 0.9944\n",
      "Epoch 170/500\n",
      " - 0s - loss: 0.0128 - accuracy: 0.9948\n",
      "Epoch 171/500\n",
      " - 0s - loss: 0.0126 - accuracy: 0.9944\n",
      "Epoch 172/500\n",
      " - 0s - loss: 0.0126 - accuracy: 0.9944\n",
      "Epoch 173/500\n",
      " - 0s - loss: 0.0126 - accuracy: 0.9949\n",
      "Epoch 174/500\n",
      " - 0s - loss: 0.0126 - accuracy: 0.9943\n",
      "Epoch 175/500\n",
      " - 0s - loss: 0.0125 - accuracy: 0.9941\n",
      "Epoch 176/500\n",
      " - 0s - loss: 0.0125 - accuracy: 0.9946\n",
      "Epoch 177/500\n",
      " - 0s - loss: 0.0124 - accuracy: 0.9946\n",
      "Epoch 178/500\n",
      " - 0s - loss: 0.0123 - accuracy: 0.9949\n",
      "Epoch 179/500\n",
      " - 0s - loss: 0.0123 - accuracy: 0.9946\n",
      "Epoch 180/500\n",
      " - 0s - loss: 0.0123 - accuracy: 0.9944\n",
      "Epoch 181/500\n",
      " - 0s - loss: 0.0122 - accuracy: 0.9943\n",
      "Epoch 182/500\n",
      " - 0s - loss: 0.0121 - accuracy: 0.9951\n",
      "Epoch 183/500\n",
      " - 0s - loss: 0.0122 - accuracy: 0.9943\n",
      "Epoch 184/500\n",
      " - 0s - loss: 0.0120 - accuracy: 0.9949\n",
      "Epoch 185/500\n",
      " - 0s - loss: 0.0121 - accuracy: 0.9946\n",
      "Epoch 186/500\n",
      " - 0s - loss: 0.0120 - accuracy: 0.9949\n",
      "Epoch 187/500\n",
      " - 0s - loss: 0.0118 - accuracy: 0.9947\n",
      "Epoch 188/500\n",
      " - 0s - loss: 0.0119 - accuracy: 0.9952\n",
      "Epoch 189/500\n",
      " - 0s - loss: 0.0119 - accuracy: 0.9945\n",
      "Epoch 190/500\n",
      " - 0s - loss: 0.0117 - accuracy: 0.9950\n",
      "Epoch 191/500\n",
      " - 0s - loss: 0.0118 - accuracy: 0.9948\n",
      "Epoch 192/500\n",
      " - 0s - loss: 0.0116 - accuracy: 0.9948\n",
      "Epoch 193/500\n",
      " - 0s - loss: 0.0116 - accuracy: 0.9949\n",
      "Epoch 194/500\n",
      " - 0s - loss: 0.0115 - accuracy: 0.9949\n",
      "Epoch 195/500\n",
      " - 0s - loss: 0.0116 - accuracy: 0.9948\n",
      "Epoch 196/500\n",
      " - 0s - loss: 0.0117 - accuracy: 0.9947\n",
      "Epoch 197/500\n",
      " - 0s - loss: 0.0115 - accuracy: 0.9951\n",
      "Epoch 198/500\n",
      " - 0s - loss: 0.0114 - accuracy: 0.9953\n",
      "Epoch 199/500\n",
      " - 0s - loss: 0.0114 - accuracy: 0.9950\n",
      "Epoch 200/500\n",
      " - 0s - loss: 0.0115 - accuracy: 0.9949\n",
      "Epoch 201/500\n",
      " - 0s - loss: 0.0113 - accuracy: 0.9952\n",
      "Epoch 202/500\n",
      " - 0s - loss: 0.0113 - accuracy: 0.9950\n",
      "Epoch 203/500\n",
      " - 0s - loss: 0.0112 - accuracy: 0.9949\n",
      "Epoch 204/500\n",
      " - 0s - loss: 0.0113 - accuracy: 0.9950\n",
      "Epoch 205/500\n",
      " - 0s - loss: 0.0114 - accuracy: 0.9951\n",
      "Epoch 206/500\n",
      " - 0s - loss: 0.0112 - accuracy: 0.9952\n",
      "Epoch 207/500\n",
      " - 0s - loss: 0.0111 - accuracy: 0.9952\n",
      "Epoch 208/500\n",
      " - 0s - loss: 0.0112 - accuracy: 0.9952\n",
      "Epoch 209/500\n",
      " - 0s - loss: 0.0110 - accuracy: 0.9953\n",
      "Epoch 210/500\n",
      " - 0s - loss: 0.0112 - accuracy: 0.9948\n",
      "Epoch 211/500\n",
      " - 0s - loss: 0.0111 - accuracy: 0.9953\n",
      "Epoch 212/500\n",
      " - 0s - loss: 0.0110 - accuracy: 0.9950\n",
      "Epoch 213/500\n",
      " - 0s - loss: 0.0110 - accuracy: 0.9953\n",
      "Epoch 214/500\n",
      " - 0s - loss: 0.0110 - accuracy: 0.9955\n",
      "Epoch 215/500\n",
      " - 0s - loss: 0.0110 - accuracy: 0.9953\n",
      "Epoch 216/500\n",
      " - 0s - loss: 0.0108 - accuracy: 0.9952\n",
      "Epoch 217/500\n",
      " - 0s - loss: 0.0109 - accuracy: 0.9951\n",
      "Epoch 218/500\n",
      " - 0s - loss: 0.0108 - accuracy: 0.9951\n",
      "Epoch 219/500\n",
      " - 0s - loss: 0.0108 - accuracy: 0.9954\n",
      "Epoch 220/500\n",
      " - 0s - loss: 0.0108 - accuracy: 0.9954\n",
      "Epoch 221/500\n",
      " - 0s - loss: 0.0108 - accuracy: 0.9954\n",
      "Epoch 222/500\n",
      " - 0s - loss: 0.0108 - accuracy: 0.9951\n",
      "Epoch 223/500\n",
      " - 0s - loss: 0.0107 - accuracy: 0.9954\n",
      "Epoch 224/500\n",
      " - 0s - loss: 0.0107 - accuracy: 0.9951\n",
      "Epoch 225/500\n",
      " - 0s - loss: 0.0107 - accuracy: 0.9953\n",
      "Epoch 226/500\n",
      " - 0s - loss: 0.0106 - accuracy: 0.9955\n",
      "Epoch 227/500\n",
      " - 0s - loss: 0.0106 - accuracy: 0.9954\n",
      "Epoch 228/500\n",
      " - 0s - loss: 0.0106 - accuracy: 0.9952\n",
      "Epoch 229/500\n",
      " - 0s - loss: 0.0106 - accuracy: 0.9952\n",
      "Epoch 230/500\n",
      " - 0s - loss: 0.0105 - accuracy: 0.9956\n",
      "Epoch 231/500\n",
      " - 0s - loss: 0.0106 - accuracy: 0.9951\n",
      "Epoch 232/500\n",
      " - 0s - loss: 0.0105 - accuracy: 0.9956\n",
      "Epoch 233/500\n",
      " - 0s - loss: 0.0104 - accuracy: 0.9957\n",
      "Epoch 234/500\n",
      " - 0s - loss: 0.0103 - accuracy: 0.9954\n",
      "Epoch 235/500\n",
      " - 0s - loss: 0.0102 - accuracy: 0.9956\n",
      "Epoch 236/500\n",
      " - 0s - loss: 0.0102 - accuracy: 0.9954\n",
      "Epoch 237/500\n",
      " - 0s - loss: 0.0104 - accuracy: 0.9949\n",
      "Epoch 238/500\n",
      " - 0s - loss: 0.0101 - accuracy: 0.9956\n",
      "Epoch 239/500\n",
      " - 0s - loss: 0.0102 - accuracy: 0.9952\n",
      "Epoch 240/500\n",
      " - 0s - loss: 0.0101 - accuracy: 0.9951\n",
      "Epoch 241/500\n",
      " - 0s - loss: 0.0102 - accuracy: 0.9955\n",
      "Epoch 242/500\n",
      " - 0s - loss: 0.0101 - accuracy: 0.9954\n",
      "Epoch 243/500\n",
      " - 0s - loss: 0.0102 - accuracy: 0.9955\n",
      "Epoch 244/500\n",
      " - 0s - loss: 0.0100 - accuracy: 0.9955\n",
      "Epoch 245/500\n",
      " - 0s - loss: 0.0101 - accuracy: 0.9955\n",
      "Epoch 246/500\n",
      " - 0s - loss: 0.0099 - accuracy: 0.9957\n",
      "Epoch 247/500\n",
      " - 0s - loss: 0.0098 - accuracy: 0.9956\n",
      "Epoch 248/500\n",
      " - 0s - loss: 0.0100 - accuracy: 0.9955\n",
      "Epoch 249/500\n",
      " - 0s - loss: 0.0099 - accuracy: 0.9955\n",
      "Epoch 250/500\n",
      " - 0s - loss: 0.0100 - accuracy: 0.9956\n",
      "Epoch 251/500\n",
      " - 0s - loss: 0.0099 - accuracy: 0.9957\n",
      "Epoch 252/500\n",
      " - 0s - loss: 0.0099 - accuracy: 0.9954\n",
      "Epoch 253/500\n",
      " - 0s - loss: 0.0098 - accuracy: 0.9953\n",
      "Epoch 254/500\n",
      " - 0s - loss: 0.0097 - accuracy: 0.9956\n",
      "Epoch 255/500\n",
      " - 0s - loss: 0.0098 - accuracy: 0.9958\n",
      "Epoch 256/500\n",
      " - 0s - loss: 0.0095 - accuracy: 0.9958\n",
      "Epoch 257/500\n",
      " - 0s - loss: 0.0099 - accuracy: 0.9958\n",
      "Epoch 258/500\n",
      " - 0s - loss: 0.0096 - accuracy: 0.9958\n",
      "Epoch 259/500\n",
      " - 0s - loss: 0.0098 - accuracy: 0.9957\n",
      "Epoch 260/500\n",
      " - 0s - loss: 0.0095 - accuracy: 0.9958\n",
      "Epoch 261/500\n",
      " - 0s - loss: 0.0097 - accuracy: 0.9956\n",
      "Epoch 262/500\n",
      " - 0s - loss: 0.0096 - accuracy: 0.9956\n",
      "Epoch 263/500\n",
      " - 0s - loss: 0.0095 - accuracy: 0.9961\n",
      "Epoch 264/500\n",
      " - 0s - loss: 0.0096 - accuracy: 0.9957\n",
      "Epoch 265/500\n",
      " - 0s - loss: 0.0095 - accuracy: 0.9961\n",
      "Epoch 266/500\n",
      " - 0s - loss: 0.0096 - accuracy: 0.9955\n",
      "Epoch 267/500\n",
      " - 0s - loss: 0.0096 - accuracy: 0.9961\n",
      "Epoch 268/500\n",
      " - 0s - loss: 0.0095 - accuracy: 0.9956\n",
      "Epoch 269/500\n",
      " - 0s - loss: 0.0095 - accuracy: 0.9956\n",
      "Epoch 270/500\n",
      " - 0s - loss: 0.0094 - accuracy: 0.9959\n",
      "Epoch 271/500\n",
      " - 0s - loss: 0.0094 - accuracy: 0.9958\n",
      "Epoch 272/500\n",
      " - 0s - loss: 0.0094 - accuracy: 0.9956\n",
      "Epoch 273/500\n",
      " - 0s - loss: 0.0093 - accuracy: 0.9956\n",
      "Epoch 274/500\n",
      " - 0s - loss: 0.0094 - accuracy: 0.9957\n",
      "Epoch 275/500\n",
      " - 0s - loss: 0.0095 - accuracy: 0.9957\n",
      "Epoch 276/500\n",
      " - 0s - loss: 0.0093 - accuracy: 0.9956\n",
      "Epoch 277/500\n",
      " - 0s - loss: 0.0094 - accuracy: 0.9956\n",
      "Epoch 278/500\n",
      " - 0s - loss: 0.0092 - accuracy: 0.9959\n",
      "Epoch 279/500\n",
      " - 0s - loss: 0.0093 - accuracy: 0.9962\n",
      "Epoch 280/500\n",
      " - 0s - loss: 0.0093 - accuracy: 0.9962\n",
      "Epoch 281/500\n",
      " - 0s - loss: 0.0092 - accuracy: 0.9963\n",
      "Epoch 282/500\n",
      " - 0s - loss: 0.0090 - accuracy: 0.9961\n",
      "Epoch 283/500\n",
      " - 0s - loss: 0.0093 - accuracy: 0.9962\n",
      "Epoch 284/500\n",
      " - 0s - loss: 0.0091 - accuracy: 0.9961\n",
      "Epoch 285/500\n",
      " - 0s - loss: 0.0092 - accuracy: 0.9961\n",
      "Epoch 286/500\n",
      " - 0s - loss: 0.0091 - accuracy: 0.9959\n",
      "Epoch 287/500\n",
      " - 0s - loss: 0.0091 - accuracy: 0.9962\n",
      "Epoch 288/500\n",
      " - 0s - loss: 0.0090 - accuracy: 0.9961\n",
      "Epoch 289/500\n",
      " - 0s - loss: 0.0090 - accuracy: 0.9965\n",
      "Epoch 290/500\n",
      " - 0s - loss: 0.0088 - accuracy: 0.9964\n",
      "Epoch 291/500\n",
      " - 0s - loss: 0.0092 - accuracy: 0.9961\n",
      "Epoch 292/500\n",
      " - 0s - loss: 0.0092 - accuracy: 0.9958\n",
      "Epoch 293/500\n",
      " - 0s - loss: 0.0090 - accuracy: 0.9964\n",
      "Epoch 294/500\n",
      " - 0s - loss: 0.0088 - accuracy: 0.9961\n",
      "Epoch 295/500\n",
      " - 0s - loss: 0.0092 - accuracy: 0.9960\n",
      "Epoch 296/500\n",
      " - 0s - loss: 0.0088 - accuracy: 0.9959\n",
      "Epoch 297/500\n",
      " - 0s - loss: 0.0090 - accuracy: 0.9959\n",
      "Epoch 298/500\n",
      " - 0s - loss: 0.0089 - accuracy: 0.9963\n",
      "Epoch 299/500\n",
      " - 0s - loss: 0.0089 - accuracy: 0.9961\n",
      "Epoch 300/500\n",
      " - 0s - loss: 0.0088 - accuracy: 0.9961\n",
      "Epoch 301/500\n",
      " - 0s - loss: 0.0089 - accuracy: 0.9963\n",
      "Epoch 302/500\n",
      " - 0s - loss: 0.0090 - accuracy: 0.9964\n",
      "Epoch 303/500\n",
      " - 0s - loss: 0.0088 - accuracy: 0.9962\n",
      "Epoch 304/500\n",
      " - 0s - loss: 0.0088 - accuracy: 0.9963\n",
      "Epoch 305/500\n",
      " - 0s - loss: 0.0087 - accuracy: 0.9965\n",
      "Epoch 306/500\n",
      " - 0s - loss: 0.0088 - accuracy: 0.9962\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 307/500\n",
      " - 0s - loss: 0.0087 - accuracy: 0.9963\n",
      "Epoch 308/500\n",
      " - 0s - loss: 0.0089 - accuracy: 0.9959\n",
      "Epoch 309/500\n",
      " - 0s - loss: 0.0088 - accuracy: 0.9962\n",
      "Epoch 310/500\n",
      " - 0s - loss: 0.0087 - accuracy: 0.9961\n",
      "Epoch 311/500\n",
      " - 0s - loss: 0.0086 - accuracy: 0.9963\n",
      "Epoch 312/500\n",
      " - 0s - loss: 0.0089 - accuracy: 0.9961\n",
      "Epoch 313/500\n",
      " - 0s - loss: 0.0087 - accuracy: 0.9963\n",
      "Epoch 314/500\n",
      " - 0s - loss: 0.0084 - accuracy: 0.9964\n",
      "Epoch 315/500\n",
      " - 0s - loss: 0.0087 - accuracy: 0.9960\n",
      "Epoch 316/500\n",
      " - 0s - loss: 0.0086 - accuracy: 0.9962\n",
      "Epoch 317/500\n",
      " - 0s - loss: 0.0083 - accuracy: 0.9967\n",
      "Epoch 318/500\n",
      " - 0s - loss: 0.0085 - accuracy: 0.9962\n",
      "Epoch 319/500\n",
      " - 0s - loss: 0.0084 - accuracy: 0.9965\n",
      "Epoch 320/500\n",
      " - 0s - loss: 0.0086 - accuracy: 0.9962\n",
      "Epoch 321/500\n",
      " - 0s - loss: 0.0084 - accuracy: 0.9961\n",
      "Epoch 322/500\n",
      " - 0s - loss: 0.0086 - accuracy: 0.9969\n",
      "Epoch 323/500\n",
      " - 0s - loss: 0.0084 - accuracy: 0.9965\n",
      "Epoch 324/500\n",
      " - 0s - loss: 0.0084 - accuracy: 0.9966\n",
      "Epoch 325/500\n",
      " - 0s - loss: 0.0084 - accuracy: 0.9965\n",
      "Epoch 326/500\n",
      " - 0s - loss: 0.0085 - accuracy: 0.9966\n",
      "Epoch 327/500\n",
      " - 0s - loss: 0.0083 - accuracy: 0.9967\n",
      "Epoch 328/500\n",
      " - 0s - loss: 0.0082 - accuracy: 0.9968\n",
      "Epoch 329/500\n",
      " - 0s - loss: 0.0085 - accuracy: 0.9967\n",
      "Epoch 330/500\n",
      " - 0s - loss: 0.0083 - accuracy: 0.9966\n",
      "Epoch 331/500\n",
      " - 0s - loss: 0.0084 - accuracy: 0.9965\n",
      "Epoch 332/500\n",
      " - 0s - loss: 0.0083 - accuracy: 0.9966\n",
      "Epoch 333/500\n",
      " - 0s - loss: 0.0085 - accuracy: 0.9965\n",
      "Epoch 334/500\n",
      " - 0s - loss: 0.0083 - accuracy: 0.9969\n",
      "Epoch 335/500\n",
      " - 0s - loss: 0.0083 - accuracy: 0.9965\n",
      "Epoch 336/500\n",
      " - 0s - loss: 0.0084 - accuracy: 0.9966\n",
      "Epoch 337/500\n",
      " - 0s - loss: 0.0083 - accuracy: 0.9967\n",
      "Epoch 338/500\n",
      " - 0s - loss: 0.0083 - accuracy: 0.9969\n",
      "Epoch 339/500\n",
      " - 0s - loss: 0.0083 - accuracy: 0.9968\n",
      "Epoch 340/500\n",
      " - 0s - loss: 0.0080 - accuracy: 0.9967\n",
      "Epoch 341/500\n",
      " - 0s - loss: 0.0082 - accuracy: 0.9969\n",
      "Epoch 342/500\n",
      " - 0s - loss: 0.0081 - accuracy: 0.9967\n",
      "Epoch 343/500\n",
      " - 0s - loss: 0.0081 - accuracy: 0.9962\n",
      "Epoch 344/500\n",
      " - 0s - loss: 0.0081 - accuracy: 0.9969\n",
      "Epoch 345/500\n",
      " - 0s - loss: 0.0082 - accuracy: 0.9970\n",
      "Epoch 346/500\n",
      " - 0s - loss: 0.0081 - accuracy: 0.9970\n",
      "Epoch 347/500\n",
      " - 0s - loss: 0.0080 - accuracy: 0.9970\n",
      "Epoch 348/500\n",
      " - 0s - loss: 0.0082 - accuracy: 0.9968\n",
      "Epoch 349/500\n",
      " - 0s - loss: 0.0080 - accuracy: 0.9967\n",
      "Epoch 350/500\n",
      " - 0s - loss: 0.0081 - accuracy: 0.9968\n",
      "Epoch 351/500\n",
      " - 0s - loss: 0.0081 - accuracy: 0.9963\n",
      "Epoch 352/500\n",
      " - 0s - loss: 0.0080 - accuracy: 0.9965\n",
      "Epoch 353/500\n",
      " - 0s - loss: 0.0080 - accuracy: 0.9969\n",
      "Epoch 354/500\n",
      " - 0s - loss: 0.0079 - accuracy: 0.9967\n",
      "Epoch 355/500\n",
      " - 0s - loss: 0.0081 - accuracy: 0.9969\n",
      "Epoch 356/500\n",
      " - 0s - loss: 0.0079 - accuracy: 0.9968\n",
      "Epoch 357/500\n",
      " - 0s - loss: 0.0080 - accuracy: 0.9971\n",
      "Epoch 358/500\n",
      " - 0s - loss: 0.0080 - accuracy: 0.9971\n",
      "Epoch 359/500\n",
      " - 0s - loss: 0.0080 - accuracy: 0.9969\n",
      "Epoch 360/500\n",
      " - 0s - loss: 0.0079 - accuracy: 0.9969\n",
      "Epoch 361/500\n",
      " - 0s - loss: 0.0079 - accuracy: 0.9973\n",
      "Epoch 362/500\n",
      " - 0s - loss: 0.0079 - accuracy: 0.9972\n",
      "Epoch 363/500\n",
      " - 0s - loss: 0.0078 - accuracy: 0.9969\n",
      "Epoch 364/500\n",
      " - 0s - loss: 0.0077 - accuracy: 0.9966\n",
      "Epoch 365/500\n",
      " - 0s - loss: 0.0081 - accuracy: 0.9967\n",
      "Epoch 366/500\n",
      " - 0s - loss: 0.0079 - accuracy: 0.9968\n",
      "Epoch 367/500\n",
      " - 0s - loss: 0.0078 - accuracy: 0.9969\n",
      "Epoch 368/500\n",
      " - 0s - loss: 0.0077 - accuracy: 0.9970\n",
      "Epoch 369/500\n",
      " - 0s - loss: 0.0077 - accuracy: 0.9969\n",
      "Epoch 370/500\n",
      " - 0s - loss: 0.0077 - accuracy: 0.9968\n",
      "Epoch 371/500\n",
      " - 0s - loss: 0.0077 - accuracy: 0.9971\n",
      "Epoch 372/500\n",
      " - 0s - loss: 0.0078 - accuracy: 0.9969\n",
      "Epoch 373/500\n",
      " - 0s - loss: 0.0078 - accuracy: 0.9971\n",
      "Epoch 374/500\n",
      " - 0s - loss: 0.0076 - accuracy: 0.9973\n",
      "Epoch 375/500\n",
      " - 0s - loss: 0.0078 - accuracy: 0.9969\n",
      "Epoch 376/500\n",
      " - 0s - loss: 0.0078 - accuracy: 0.9967\n",
      "Epoch 377/500\n",
      " - 0s - loss: 0.0076 - accuracy: 0.9970\n",
      "Epoch 378/500\n",
      " - 0s - loss: 0.0077 - accuracy: 0.9969\n",
      "Epoch 379/500\n",
      " - 0s - loss: 0.0075 - accuracy: 0.9967\n",
      "Epoch 380/500\n",
      " - 0s - loss: 0.0077 - accuracy: 0.9971\n",
      "Epoch 381/500\n",
      " - 0s - loss: 0.0076 - accuracy: 0.9974\n",
      "Epoch 382/500\n",
      " - 0s - loss: 0.0077 - accuracy: 0.9971\n",
      "Epoch 383/500\n",
      " - 0s - loss: 0.0078 - accuracy: 0.9969\n",
      "Epoch 384/500\n",
      " - 0s - loss: 0.0076 - accuracy: 0.9971\n",
      "Epoch 385/500\n",
      " - 0s - loss: 0.0077 - accuracy: 0.9969\n",
      "Epoch 386/500\n",
      " - 0s - loss: 0.0074 - accuracy: 0.9972\n",
      "Epoch 387/500\n",
      " - 0s - loss: 0.0078 - accuracy: 0.9968\n",
      "Epoch 388/500\n",
      " - 0s - loss: 0.0076 - accuracy: 0.9972\n",
      "Epoch 389/500\n",
      " - 0s - loss: 0.0076 - accuracy: 0.9972\n",
      "Epoch 390/500\n",
      " - 0s - loss: 0.0073 - accuracy: 0.9972\n",
      "Epoch 391/500\n",
      " - 0s - loss: 0.0076 - accuracy: 0.9971\n",
      "Epoch 392/500\n",
      " - 0s - loss: 0.0076 - accuracy: 0.9970\n",
      "Epoch 393/500\n",
      " - 0s - loss: 0.0076 - accuracy: 0.9968\n",
      "Epoch 394/500\n",
      " - 0s - loss: 0.0075 - accuracy: 0.9969\n",
      "Epoch 395/500\n",
      " - 0s - loss: 0.0074 - accuracy: 0.9970\n",
      "Epoch 396/500\n",
      " - 0s - loss: 0.0076 - accuracy: 0.9970\n",
      "Epoch 397/500\n",
      " - 0s - loss: 0.0076 - accuracy: 0.9971\n",
      "Epoch 398/500\n",
      " - 0s - loss: 0.0075 - accuracy: 0.9969\n",
      "Epoch 399/500\n",
      " - 0s - loss: 0.0075 - accuracy: 0.9967\n",
      "Epoch 400/500\n",
      " - 0s - loss: 0.0074 - accuracy: 0.9971\n",
      "Epoch 401/500\n",
      " - 0s - loss: 0.0074 - accuracy: 0.9970\n",
      "Epoch 402/500\n",
      " - 0s - loss: 0.0074 - accuracy: 0.9972\n",
      "Epoch 403/500\n",
      " - 0s - loss: 0.0074 - accuracy: 0.9969\n",
      "Epoch 404/500\n",
      " - 0s - loss: 0.0075 - accuracy: 0.9973\n",
      "Epoch 405/500\n",
      " - 0s - loss: 0.0073 - accuracy: 0.9975\n",
      "Epoch 406/500\n",
      " - 0s - loss: 0.0075 - accuracy: 0.9973\n",
      "Epoch 407/500\n",
      " - 0s - loss: 0.0073 - accuracy: 0.9972\n",
      "Epoch 408/500\n",
      " - 0s - loss: 0.0073 - accuracy: 0.9969\n",
      "Epoch 409/500\n",
      " - 0s - loss: 0.0074 - accuracy: 0.9974\n",
      "Epoch 410/500\n",
      " - 0s - loss: 0.0073 - accuracy: 0.9972\n",
      "Epoch 411/500\n",
      " - 0s - loss: 0.0074 - accuracy: 0.9974\n",
      "Epoch 412/500\n",
      " - 0s - loss: 0.0073 - accuracy: 0.9970\n",
      "Epoch 413/500\n",
      " - 0s - loss: 0.0071 - accuracy: 0.9972\n",
      "Epoch 414/500\n",
      " - 0s - loss: 0.0074 - accuracy: 0.9972\n",
      "Epoch 415/500\n",
      " - 0s - loss: 0.0074 - accuracy: 0.9969\n",
      "Epoch 416/500\n",
      " - 0s - loss: 0.0073 - accuracy: 0.9969\n",
      "Epoch 417/500\n",
      " - 0s - loss: 0.0073 - accuracy: 0.9970\n",
      "Epoch 418/500\n",
      " - 0s - loss: 0.0072 - accuracy: 0.9973\n",
      "Epoch 419/500\n",
      " - 0s - loss: 0.0073 - accuracy: 0.9974\n",
      "Epoch 420/500\n",
      " - 0s - loss: 0.0073 - accuracy: 0.9972\n",
      "Epoch 421/500\n",
      " - 0s - loss: 0.0072 - accuracy: 0.9969\n",
      "Epoch 422/500\n",
      " - 0s - loss: 0.0072 - accuracy: 0.9969\n",
      "Epoch 423/500\n",
      " - 0s - loss: 0.0072 - accuracy: 0.9974\n",
      "Epoch 424/500\n",
      " - 0s - loss: 0.0074 - accuracy: 0.9970\n",
      "Epoch 425/500\n",
      " - 0s - loss: 0.0072 - accuracy: 0.9970\n",
      "Epoch 426/500\n",
      " - 0s - loss: 0.0070 - accuracy: 0.9977\n",
      "Epoch 427/500\n",
      " - 0s - loss: 0.0072 - accuracy: 0.9971\n",
      "Epoch 428/500\n",
      " - 0s - loss: 0.0070 - accuracy: 0.9973\n",
      "Epoch 429/500\n",
      " - 0s - loss: 0.0073 - accuracy: 0.9971\n",
      "Epoch 430/500\n",
      " - 0s - loss: 0.0073 - accuracy: 0.9973\n",
      "Epoch 431/500\n",
      " - 0s - loss: 0.0069 - accuracy: 0.9977\n",
      "Epoch 432/500\n",
      " - 0s - loss: 0.0072 - accuracy: 0.9973\n",
      "Epoch 433/500\n",
      " - 0s - loss: 0.0068 - accuracy: 0.9976\n",
      "Epoch 434/500\n",
      " - 0s - loss: 0.0072 - accuracy: 0.9971\n",
      "Epoch 435/500\n",
      " - 0s - loss: 0.0071 - accuracy: 0.9969\n",
      "Epoch 436/500\n",
      " - 0s - loss: 0.0068 - accuracy: 0.9971\n",
      "Epoch 437/500\n",
      " - 0s - loss: 0.0073 - accuracy: 0.9970\n",
      "Epoch 438/500\n",
      " - 0s - loss: 0.0071 - accuracy: 0.9975\n",
      "Epoch 439/500\n",
      " - 0s - loss: 0.0071 - accuracy: 0.9974\n",
      "Epoch 440/500\n",
      " - 0s - loss: 0.0070 - accuracy: 0.9972\n",
      "Epoch 441/500\n",
      " - 0s - loss: 0.0070 - accuracy: 0.9975\n",
      "Epoch 442/500\n",
      " - 0s - loss: 0.0072 - accuracy: 0.9969\n",
      "Epoch 443/500\n",
      " - 0s - loss: 0.0071 - accuracy: 0.9973\n",
      "Epoch 444/500\n",
      " - 0s - loss: 0.0071 - accuracy: 0.9967\n",
      "Epoch 445/500\n",
      " - 0s - loss: 0.0070 - accuracy: 0.9973\n",
      "Epoch 446/500\n",
      " - 0s - loss: 0.0071 - accuracy: 0.9975\n",
      "Epoch 447/500\n",
      " - 0s - loss: 0.0070 - accuracy: 0.9974\n",
      "Epoch 448/500\n",
      " - 0s - loss: 0.0070 - accuracy: 0.9974\n",
      "Epoch 449/500\n",
      " - 0s - loss: 0.0069 - accuracy: 0.9972\n",
      "Epoch 450/500\n",
      " - 0s - loss: 0.0069 - accuracy: 0.9970\n",
      "Epoch 451/500\n",
      " - 0s - loss: 0.0072 - accuracy: 0.9973\n",
      "Epoch 452/500\n",
      " - 0s - loss: 0.0068 - accuracy: 0.9976\n",
      "Epoch 453/500\n",
      " - 0s - loss: 0.0071 - accuracy: 0.9970\n",
      "Epoch 454/500\n",
      " - 0s - loss: 0.0069 - accuracy: 0.9973\n",
      "Epoch 455/500\n",
      " - 0s - loss: 0.0069 - accuracy: 0.9977\n",
      "Epoch 456/500\n",
      " - 0s - loss: 0.0069 - accuracy: 0.9972\n",
      "Epoch 457/500\n",
      " - 0s - loss: 0.0071 - accuracy: 0.9975\n",
      "Epoch 458/500\n",
      " - 0s - loss: 0.0070 - accuracy: 0.9970\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 459/500\n",
      " - 0s - loss: 0.0069 - accuracy: 0.9975\n",
      "Epoch 460/500\n",
      " - 0s - loss: 0.0068 - accuracy: 0.9974\n",
      "Epoch 461/500\n",
      " - 0s - loss: 0.0070 - accuracy: 0.9972\n",
      "Epoch 462/500\n",
      " - 0s - loss: 0.0069 - accuracy: 0.9969\n",
      "Epoch 463/500\n",
      " - 0s - loss: 0.0069 - accuracy: 0.9974\n",
      "Epoch 464/500\n",
      " - 0s - loss: 0.0071 - accuracy: 0.9972\n",
      "Epoch 465/500\n",
      " - 0s - loss: 0.0069 - accuracy: 0.9975\n",
      "Epoch 466/500\n",
      " - 0s - loss: 0.0068 - accuracy: 0.9975\n",
      "Epoch 467/500\n",
      " - 0s - loss: 0.0069 - accuracy: 0.9975\n",
      "Epoch 468/500\n",
      " - 0s - loss: 0.0070 - accuracy: 0.9973\n",
      "Epoch 469/500\n",
      " - 0s - loss: 0.0067 - accuracy: 0.9974\n",
      "Epoch 470/500\n",
      " - 0s - loss: 0.0071 - accuracy: 0.9971\n",
      "Epoch 471/500\n",
      " - 0s - loss: 0.0068 - accuracy: 0.9975\n",
      "Epoch 472/500\n",
      " - 0s - loss: 0.0070 - accuracy: 0.9969\n",
      "Epoch 473/500\n",
      " - 0s - loss: 0.0069 - accuracy: 0.9973\n",
      "Epoch 474/500\n",
      " - 0s - loss: 0.0068 - accuracy: 0.9973\n",
      "Epoch 475/500\n",
      " - 0s - loss: 0.0067 - accuracy: 0.9975\n",
      "Epoch 476/500\n",
      " - 0s - loss: 0.0067 - accuracy: 0.9973\n",
      "Epoch 477/500\n",
      " - 0s - loss: 0.0069 - accuracy: 0.9973\n",
      "Epoch 478/500\n",
      " - 0s - loss: 0.0067 - accuracy: 0.9972\n",
      "Epoch 479/500\n",
      " - 0s - loss: 0.0070 - accuracy: 0.9971\n",
      "Epoch 480/500\n",
      " - 0s - loss: 0.0069 - accuracy: 0.9974\n",
      "Epoch 481/500\n",
      " - 0s - loss: 0.0068 - accuracy: 0.9973\n",
      "Epoch 482/500\n",
      " - 0s - loss: 0.0067 - accuracy: 0.9972\n",
      "Epoch 483/500\n",
      " - 0s - loss: 0.0069 - accuracy: 0.9973\n",
      "Epoch 484/500\n",
      " - 0s - loss: 0.0068 - accuracy: 0.9973\n",
      "Epoch 485/500\n",
      " - 0s - loss: 0.0069 - accuracy: 0.9973\n",
      "Epoch 486/500\n",
      " - 0s - loss: 0.0068 - accuracy: 0.9976\n",
      "Epoch 487/500\n",
      " - 0s - loss: 0.0069 - accuracy: 0.9972\n",
      "Epoch 488/500\n",
      " - 0s - loss: 0.0069 - accuracy: 0.9972\n",
      "Epoch 489/500\n",
      " - 0s - loss: 0.0066 - accuracy: 0.9974\n",
      "Epoch 490/500\n",
      " - 0s - loss: 0.0068 - accuracy: 0.9973\n",
      "Epoch 491/500\n",
      " - 0s - loss: 0.0065 - accuracy: 0.9975\n",
      "Epoch 492/500\n",
      " - 0s - loss: 0.0067 - accuracy: 0.9973\n",
      "Epoch 493/500\n",
      " - 0s - loss: 0.0068 - accuracy: 0.9970\n",
      "Epoch 494/500\n",
      " - 0s - loss: 0.0068 - accuracy: 0.9973\n",
      "Epoch 495/500\n",
      " - 0s - loss: 0.0066 - accuracy: 0.9975\n",
      "Epoch 496/500\n",
      " - 0s - loss: 0.0067 - accuracy: 0.9974\n",
      "Epoch 497/500\n",
      " - 0s - loss: 0.0067 - accuracy: 0.9978\n",
      "Epoch 498/500\n",
      " - 0s - loss: 0.0066 - accuracy: 0.9973\n",
      "Epoch 499/500\n",
      " - 0s - loss: 0.0066 - accuracy: 0.9975\n",
      "Epoch 500/500\n",
      " - 0s - loss: 0.0066 - accuracy: 0.9974\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.callbacks.History at 0x1cd8ba96ac0>"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(train[:, 1:], train[:, 0], epochs=500, batch_size=500, verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_pred = model.predict(test[:, 1:]).reshape((-1,)) >= 0.5\n",
    "\n",
    "conf_mat = confusion_matrix(y_test, y_pred)\n",
    "\n",
    "lang = ['No Fault (0)', 'Fault (1)']\n",
    "ax = plt.axes()\n",
    "sns.heatmap(conf_mat, square=True, annot=True, cmap='Blues', fmt='d', cbar=True, ax=ax, xticklabels=lang, yticklabels=lang)\n",
    "ax.set_title('Confusion Matrix')\n",
    "plt.xlabel('Predicted Label')\n",
    "plt.ylabel('True Label')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The Evaluation Metrics of the MLP Model-4 (on the test set) is as follows:\n",
      "\n",
      "True Positives: 548\n",
      "True Negatives: 650\n",
      "False Positives: 1\n",
      "False Negatives: 2\n",
      "\n",
      "Precision: 0.9981785063752276\n",
      "Recall: 0.9963636363636363\n",
      "\n",
      "Accuracy: 0.9975020815986678\n"
     ]
    }
   ],
   "source": [
    "true_positive = conf_mat[1, 1]\n",
    "true_negative = conf_mat[0, 0]\n",
    "false_positive = conf_mat[0, 1]\n",
    "false_negative = conf_mat[1, 0]\n",
    "accuracy = accuracy_score(y_pred, y_test)\n",
    "\n",
    "print(\"The Evaluation Metrics of the MLP Model-4 (on the test set) is as follows:\\n\")\n",
    "\n",
    "print(\"True Positives:\", true_positive)\n",
    "print(\"True Negatives:\", true_negative)\n",
    "print(\"False Positives:\", false_positive)\n",
    "print(\"False Negatives:\", false_negative, end='\\n\\n')\n",
    "\n",
    "print(\"Precision:\", true_positive/(true_positive+false_positive))\n",
    "print(\"Recall:\", true_positive/(true_positive+false_negative), end='\\n\\n')\n",
    "\n",
    "print(\"Accuracy:\", accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## D) Results & Inferences:\n",
    "\n",
    "The results of all the 7 models is as follows:<br><br>\n",
    "\n",
    "Well tuned SVM, Decision Tree & Random Forest: Tie with **99.83% accuracy**<br>\n",
    "MLP Model 1: **73.68% accuracy**<br>\n",
    "MLP Model 2: **98.83% accuracy**<br>\n",
    "MLP Model 3: **99.25% accuracy**<br>\n",
    "MLP Model 4: **99.75% accuracy**<br><br>\n",
    "\n",
    "Performance of models & Reasons for performance:<br><br>\n",
    "\n",
    "(i) MLP vs non-MLP:<br>\n",
    "- It can be seen that, that highest accuracy is obtained from the non-MLP models, i.e. the SVM, Decision Tree & Random Forest. But that is expected to be the case only with this dataset and less likely to be applicable to conventional datasets.<br>\n",
    "- The reason for the non-NLP methods to outperform NLP methods, is since this is a simple dataset where the data has very less noise and is well balanced. Since neural networks tend to more exhibit the nature of function approximators, it is less likely to reach 100% accuracy in any case, and the 99.75% accuracy accomplished is good already.<br>\n",
    "- In case of datasets, where the accuracy accomplished by non-MLP methods is less than 80%, MLPs are very likely to exhibit a better performance than them.<br><br>\n",
    "\n",
    "(ii) Comparision b/w MLP models:<br>\n",
    "- For all models, I have set only 10 perceptrons per layer since, the performance did not improve when the count is increased beyond that. And for the sake of fair comparision of the activation functions & optimizers I kept the perceptrons count the same. Also the final layer was always 'sigmoid' and 'binary_crossentropy' loss since this is a binary classification problem.<br>\n",
    "- From MLP Model-1, it was observed that, 'linear' activativation was **insufficient** to predict this dataset since it gave only 73.68% accuracy, and SGD optimizer took relatively more number of episodes for the loss to converge to the saturation value compared to Adam & RMSProp.<br>\n",
    "- From MLP Model-2 it was observed that, 'relu' activation function performed far better than 'linear' with an accuracy of 98.83%, and 'Adam' optimiser converged faster than 'SGD' optimizer.<br>\n",
    "- From MLP Model-3, it was observed that, 'LeakyReLU' provided a slightly better accuracy than 'relu' activation function with an accuracy of 99.25%, and 'RMSProp' converged faster than both 'Adam' & 'SGD' optimizers.<br>\n",
    "- From MLP Model-4, it was observed that, an additional layer in MLP, increased both the accuracy of the model and also caused the model to converge at an earlier number of episodes, thereby accomplishing the best accuracy among MLP models of 99.75%."
   ]
  }
 ],
 "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}