GRU_Experiment1_Padding_and_masking.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "python version:  3.7.3 (default, Mar 27 2019, 22:11:17) \n",
      "[GCC 7.3.0]\n",
      "pandas version:  0.24.2\n",
      "numpy version:  1.16.4\n",
      "scipy version:  1.3.0\n",
      "sklearn version:  0.21.2\n",
      "keras version:  2.2.4\n",
      "keras backend:  tensorflow\n",
      "tensorflow version:  1.14.0\n"
     ]
    }
   ],
   "source": [
    "import sys\n",
    "print(\"python version: \", sys.version)\n",
    "\n",
    "import pandas as pd\n",
    "print(\"pandas version: \", pd.__version__)\n",
    "\n",
    "import numpy as np\n",
    "print('numpy version: ', np.__version__)\n",
    "\n",
    "import scipy as sp\n",
    "print('scipy version: ', sp.__version__)\n",
    "\n",
    "import sklearn as skl\n",
    "print('sklearn version: ', skl.__version__)\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import os\n",
    "os.environ[\"CUDA_DEVICE_ORDER\"] = \"PCI_BUS_ID\"\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"\"\n",
    "\n",
    "import keras as ks\n",
    "print('keras version: ', ks.__version__)\n",
    "print('keras backend: ', ks.backend.backend())\n",
    "\n",
    "import tensorflow as tf\n",
    "print('tensorflow version: ', tf.__version__)\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "from keras import backend as K\n",
    "from sklearn.metrics import confusion_matrix, precision_score, recall_score\n",
    "from sklearn.metrics import roc_curve, auc, f1_score, accuracy_score, precision_recall_curve\n",
    "\n",
    "import random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(7584, 208)\n"
     ]
    }
   ],
   "source": [
    "excel_file = r'geriomop-data-from-egems-paper-UPDATED-COL-FIX-ATC_2-and-behavior-change-022819.tsv'\n",
    "df = pd.DataFrame(pd.read_csv(excel_file, sep='\\t'))\n",
    "print(df.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Drop features\n",
    "f_drop = [\"Blank-1\", \"ws-sudden-stop\", \"ws-sudden-stop-drugs\", \"ws-prolonged-psychotropics\", \"mds-toilet-prgm-atmptd\"] + \\\n",
    "         ['Deleted-%d' % i for i in range(1, 13)] + ['Psychotropic-%d' % i for i in range(1, 6)] + \\\n",
    "         ['Psychotropic-%d-average-daily-dose' % i for i in range(1, 6)] + [\"Entry-discharge-type\"]\n",
    "\n",
    "# Missing value means not relevant\n",
    "f_one_hot = [\"Mds-fall-2-6-months-to-admission\", \"Mds-fall-last-month-to-admission\", \"mds-pain-last-five-days\", \\\n",
    "             \"mds-delirium-scale\", \"mds-long-term-memory-ok\", \"mds-short-term-memory-ok\", \"mds-staff-assess-pain\", \\\n",
    "             \"recent-start-other-fall-risk-rx\"] + [\"mds-pressure-ulcer-stage-%d\" % i for i in range(1, 5)]\n",
    "cate_one_hot = [['Yes', 'No', 'Unable to answer']] * 3 + [['Yes', 'No']] * 9\n",
    "\n",
    "f_period = [(\"Psychotropic-%d-Start-date\" % i, 'Psychotropic-%d-End-date' % i, 'Psychotropic-%d-Period' % i) for i in range(1, 6)]\n",
    "\n",
    "f_label = ['mds-cognitive-scale', 'mds-pain-freq-last-five-days', 'mds-pain-intensity']\n",
    "dic_label = [{'Independent': 0, 'Modified Independence': 1, 'Moderately Impaired': 2, 'Severely Impaired': 3}, \\\n",
    "             {'Continuous': 4, 'Frequent': 3, 'Occasional': 2, 'Rare': 1, 'Unable to respond': 0}, \\\n",
    "             {'None': 0, 'Mild': 1, 'Moderate': 2, 'Severe': 3, 'Very severe, horrible': 4}]\n",
    "\n",
    "f_transform = [('mds-cognitive-scale', 3, 0), ('mds-pain-freq-last-five-days', 4, 0), ('mds-pain-intensity', 4, 0)] + \\\n",
    "              [('Psychotropic-%d-Period' % i, 0, 0) for i in range(1, 6)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Encode features (originally) with no missing values\n",
    "f_drop += [\"predictor-date\", \"projected-date\", \"PPS-assessment\", \"Federal-assessment\"]\n",
    "\n",
    "f_period += [('episode-start-date', 'episode-end-date', 'episode-period'),\\\n",
    "             ('stay-start-date', 'stay-end-date', 'stay-period')]\n",
    "\n",
    "f_label += [\"cms-long-stay\", \"mds-stay-trans-short-to-long\", \"mds-antianxiety-medication\", \\\n",
    "            \"mds-antidepressant-medication\", \"mds-antipsychotic-medication\", \"mds-antibiotic-medication\",\n",
    "            \"mds-anticoagulant-medication\", \"mds-diuretic-medication\", \\\n",
    "            \"mds-hypnotic-medication\", \"mds-behavioral-symptoms\", \"mds-behavioral-symptoms-to-others\", \\\n",
    "            \"mds-dehydrated\", \"mds-depression\", \"mds-fever\", \"mds-impaired-mobility\", \"mds-impaired-transfer\", \\\n",
    "            \"mds-internal-bleeding\", \"mds-malnutrition\", \"mds-no-problem-conditions\", \"mds-vomiting\", \\\n",
    "            \"mds-impaired-walk-in-room\", \"mds-impaired-walk-in-corridor\", \"mds-impaired-locomot-unit\", \\\n",
    "            \"mds-impaired-locomot-other\", \"ws-antibiotic-anticoag-coexposure\", \\\n",
    "            \"ws-psychotropic-with-not-ordered-weight-loss\", \\\n",
    "            \"ws-diuretic-adl\", \"ws-tramadol-antidepressant-coexposure\"] + [\"mds-pain-non-verbal\"]\n",
    "dic_label += [{'Yes': 1, 'No': 0}] * 28 + [{'None/Mild':0,'Moderate/Severe':1}]\n",
    "\n",
    "f_one_hot += [\"race\", \"facility\", \"gender\", \"mds-bims-summary-ranking\", \"ws-meclizine-psych-coexposure\"]\n",
    "cate_one_hot += [['White', 'Black', 'Other'], \\\n",
    "                 ['Sugar Creek', 'Heritage Place', 'Canterbury Place', 'Senaca Place', 'Cranberry Place'], \\\n",
    "                 ['Female', 'Male'], ['Intact or Moderately Intact', 'Moderate Impairment'], \\\n",
    "                 ['No', 'meclAndPsychStartSameR', 'meclPrecedesPsychR', 'psychPrecedesMeclR', 'meclAndPsychOlderStart']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [],
   "source": [
    "f_drop += [\"min_effective_dose_%d\" % i for i in range(1, 6)] + \\\n",
    "          [\"Psychotropic-%d-average-daily-dose\" % i for i in range(1, 6)] + \\\n",
    "          [\"psych_sdd_%d\" % i for i in range(1, 6)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [],
   "source": [
    "# generate target\n",
    "target = df['outcome-Mds-fall-since-prior-assessment'].map({'None': 0, 'Yes': 1})\n",
    "dff = df.drop(['outcome-mds-fall-no-injury', 'outcome-mds-fall-minor-injury', 'outcome-mds-fall-major-injury', \\\n",
    "               'outcome-Mds-fall-since-prior-assessment', 'outcome-riskmaster-fall-incident'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [],
   "source": [
    "# in sdd 0 means NA\n",
    "for i in range(1, 6):\n",
    "    df.loc[df['psych_sdd_%d' % i] == 0, 'psych_sdd_%d' % i] = np.nan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [],
   "source": [
    "# train test split\n",
    "from sklearn.model_selection import GroupShuffleSplit\n",
    "def train_test_split(df, target, groupby, radio, random_state):\n",
    "    train_idx, test_idx = next(GroupShuffleSplit(train_size=radio, test_size=1-radio, random_state=random_state).split(df, target, df[groupby]))\n",
    "    return df.iloc[train_idx], df.iloc[test_idx], target.iloc[train_idx], target.iloc[test_idx]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total number of patients: 3133\n",
      "Number of patients in training set: 2193\n"
     ]
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(dff, target, 'patient-id', 0.7, 0)\n",
    "print('Total number of patients: %d' % len(df.groupby('patient-id')))\n",
    "print('Number of patients in training set: %d' % len(X_train.groupby('patient-id')))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [],
   "source": [
    "# drop feature\n",
    "def drop_features(df, f):\n",
    "    df.drop(f, axis=1, inplace=True)\n",
    "    \n",
    "# one hot encoding\n",
    "def one_hot_encoding(df, features, categories):\n",
    "    for f, c in zip(features, categories):\n",
    "        df[f] = df[f].astype('category', categories=c)\n",
    "    return pd.get_dummies(df, columns=features, prefix=features)\n",
    "\n",
    "# compute date difference\n",
    "def date_diff(start, end):\n",
    "    return (pd.to_datetime(end) - pd.to_datetime(start)) / np.timedelta64(1, 'D')\n",
    "\n",
    "# generate date diff features\n",
    "def generate_period(df, f):\n",
    "    for start, end, period in f:\n",
    "        df[period] = date_diff(df[start], df[end])\n",
    "        df.drop([start, end], axis=1, inplace=True)\n",
    "        \n",
    "# label encoding\n",
    "def label_encoding(df, features, dictionaries):\n",
    "    for f, dic in zip(features, dictionaries):\n",
    "        df[f] = df[f].map(dic)\n",
    "        \n",
    "# sin&cos transform\n",
    "def col_transform(df, col, mmax, mmin, df_ref=None):\n",
    "    if(mmax <= mmin):\n",
    "        # for test set, if max&min not set, use training set value\n",
    "        if(df_ref is not None): \n",
    "            mmax = df_ref[col].max()\n",
    "            mmin = df_ref[col].max()\n",
    "        else:\n",
    "            mmax = df[col].max()\n",
    "            mmin = df[col].min()\n",
    "    angle = 0.25 * np.pi * (df[col] - mmin) / (mmax - mmin) + 0.125 * np.pi\n",
    "    return (np.cos(angle), np.sin(angle))\n",
    "\n",
    "def df_transform(df, f, df_ref=None):\n",
    "    for col, mmax, mmin in f:\n",
    "        df[col + '_x'], df[col + '_y'] = col_transform(df, col, mmax, mmin, df_ref)\n",
    "        df[col + '_x'].fillna(0, inplace=True)\n",
    "        df[col + '_y'].fillna(0, inplace=True)\n",
    "        df.drop(col, axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 46.1 s, sys: 176 ms, total: 46.3 s\n",
      "Wall time: 40.8 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "generate_period(X_train, f_period)\n",
    "label_encoding(X_train, f_label, dic_label)\n",
    "X_train_original = X_train.copy() # store value range\n",
    "df_transform(X_train, f_transform)\n",
    "X_train = one_hot_encoding(X_train, f_one_hot, cate_one_hot)\n",
    "drop_features(X_train, f_drop)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_use = X_train.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 39 s, sys: 0 ns, total: 39 s\n",
      "Wall time: 38.9 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "generate_period(X_test, f_period)\n",
    "label_encoding(X_test, f_label, dic_label)\n",
    "df_transform(X_test, f_transform, X_train_original)\n",
    "X_test = one_hot_encoding(X_test, f_one_hot, cate_one_hot)\n",
    "drop_features(X_test, f_drop)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test_use = X_test.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5301, 189)"
      ]
     },
     "execution_count": 109,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_use.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_use.to_csv (r'/tmp/export_exp1_train_dataframe.csv', index = False, header=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2283, 189)"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test_use.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**RNN - GRU**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from keras.models import Sequential\n",
    "from keras.layers import LSTM, GRU, Dense, Dropout, Masking, TimeDistributed\n",
    "from numpy.random import seed\n",
    "from tensorflow import set_random_seed\n",
    "from copy import deepcopy\n",
    "from keras.preprocessing.sequence import pad_sequences"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [],
   "source": [
    "# missing data describe\n",
    "def missing_summary(df):\n",
    "    missing_cnt = df.isna().sum()\n",
    "    missing_data = pd.concat([missing_cnt, missing_cnt/df.shape[0]], axis=1, keys=['count', 'percentage'])\n",
    "    return missing_data[missing_data['percentage'] != 0].sort_values(by='percentage', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [],
   "source": [
    "idx = missing_summary(X_train_use).index.tolist()\n",
    "X_train_use.drop(columns=idx, axis=1, inplace=True)\n",
    "X_test_use.drop(columns=idx, axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [],
   "source": [
    "def scaler_transform(X_train_use, X_test_use):    \n",
    "    scaler = MinMaxScaler(feature_range=(0, 1)).fit(X_train_use.drop('patient-id', axis=1))\n",
    "    X_train_transformed = pd.DataFrame(scaler.transform(X_train_use.drop('patient-id', axis=1)), \\\n",
    "                                       columns=X_train_use.columns[1:], index=X_train_use.index)                                   \n",
    "    X_train_transformed['patient-id'] = X_train_use['patient-id']\n",
    "    X_test_transformed = pd.DataFrame(scaler.transform(X_test_use.drop('patient-id', axis=1)), \\\n",
    "                                       columns=X_test_use.columns[1:], index=X_test_use.index) \n",
    "    X_test_transformed['patient-id'] = X_test_use['patient-id']\n",
    "    return X_train_transformed, X_test_transformed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [],
   "source": [
    "def add_previous_outcome(X, Y):\n",
    "    X = X.copy()\n",
    "    x1, x2 = [], []\n",
    "    pids = X['patient-id'].unique()\n",
    "    for pid in pids:\n",
    "        df = Y[X[X['patient-id']==pid].index]\n",
    "        if(df.shape[0] == 1):\n",
    "            x1.append(pd.Series([0]))\n",
    "            x2.append(pd.Series([0]))\n",
    "        else:\n",
    "            # input 1=(1, 0), 0=(0, 1), NA=(0, 0)\n",
    "            x1.append(pd.Series([0]))\n",
    "            x1.append(df[:-1])\n",
    "            x2.append(pd.Series([0]))\n",
    "            x2.append(df[:-1].map({0: 1, 1: 0}))\n",
    "    X['input_outcome_x'] = pd.Series(pd.concat(x1, ignore_index=True).tolist(), index=X.index)\n",
    "    X['input_outcome_y'] = pd.Series(pd.concat(x2, ignore_index=True).tolist(), index=X.index)\n",
    "    return X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [],
   "source": [
    "## This step converts the outcome to a point in x,y coordinates where fall (1) = (1,0); no fall (0) = (0,1); and NA = (0,0) \n",
    "X_train_use = add_previous_outcome(X_train_use, y_train)\n",
    "X_test_use = add_previous_outcome(X_test_use, y_test)\n",
    "\n",
    "## This step transforms the data to range between 0 and 1\n",
    "X_train_transformed, X_test_transformed = scaler_transform(X_train_use, X_test_use)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_transformed.to_csv (r'/tmp/exp1_train_transformed_df.csv', index = False, header=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5301, 102)"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_use.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2283, 102)"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test_use.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>patient-id</th>\n",
       "      <th>current-stay-days</th>\n",
       "      <th>cumulative-days-in-facility</th>\n",
       "      <th>cms-long-stay</th>\n",
       "      <th>mds-stay-trans-short-to-long</th>\n",
       "      <th>age</th>\n",
       "      <th>mds-antianxiety-medication</th>\n",
       "      <th>mds-antidepressant-medication</th>\n",
       "      <th>mds-antipsychotic-medication</th>\n",
       "      <th>mds-antibiotic-medication</th>\n",
       "      <th>...</th>\n",
       "      <th>gender_Male</th>\n",
       "      <th>mds-bims-summary-ranking_Intact or Moderately Intact</th>\n",
       "      <th>mds-bims-summary-ranking_Moderate Impairment</th>\n",
       "      <th>ws-meclizine-psych-coexposure_No</th>\n",
       "      <th>ws-meclizine-psych-coexposure_meclAndPsychStartSameR</th>\n",
       "      <th>ws-meclizine-psych-coexposure_meclPrecedesPsychR</th>\n",
       "      <th>ws-meclizine-psych-coexposure_psychPrecedesMeclR</th>\n",
       "      <th>ws-meclizine-psych-coexposure_meclAndPsychOlderStart</th>\n",
       "      <th>input_outcome_x</th>\n",
       "      <th>input_outcome_y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10353</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>67</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>10364</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>84</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10386</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>77</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5000</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>90</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>5003</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>90</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 102 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   patient-id  current-stay-days  cumulative-days-in-facility  cms-long-stay  \\\n",
       "0       10353                  7                            7              0   \n",
       "1       10364                  7                            7              0   \n",
       "4       10386                  7                            7              0   \n",
       "5        5000                  7                            7              0   \n",
       "6        5003                  7                            7              0   \n",
       "\n",
       "   mds-stay-trans-short-to-long  age  mds-antianxiety-medication  \\\n",
       "0                             0   67                           1   \n",
       "1                             1   84                           0   \n",
       "4                             1   77                           1   \n",
       "5                             0   90                           0   \n",
       "6                             1   90                           0   \n",
       "\n",
       "   mds-antidepressant-medication  mds-antipsychotic-medication  \\\n",
       "0                              1                             0   \n",
       "1                              0                             1   \n",
       "4                              1                             1   \n",
       "5                              1                             1   \n",
       "6                              1                             1   \n",
       "\n",
       "   mds-antibiotic-medication  ...  gender_Male  \\\n",
       "0                          0  ...            0   \n",
       "1                          1  ...            0   \n",
       "4                          0  ...            0   \n",
       "5                          0  ...            0   \n",
       "6                          1  ...            1   \n",
       "\n",
       "   mds-bims-summary-ranking_Intact or Moderately Intact  \\\n",
       "0                                                  1      \n",
       "1                                                  1      \n",
       "4                                                  1      \n",
       "5                                                  1      \n",
       "6                                                  1      \n",
       "\n",
       "   mds-bims-summary-ranking_Moderate Impairment  \\\n",
       "0                                             0   \n",
       "1                                             0   \n",
       "4                                             0   \n",
       "5                                             0   \n",
       "6                                             0   \n",
       "\n",
       "   ws-meclizine-psych-coexposure_No  \\\n",
       "0                                 1   \n",
       "1                                 1   \n",
       "4                                 1   \n",
       "5                                 1   \n",
       "6                                 1   \n",
       "\n",
       "   ws-meclizine-psych-coexposure_meclAndPsychStartSameR  \\\n",
       "0                                                  0      \n",
       "1                                                  0      \n",
       "4                                                  0      \n",
       "5                                                  0      \n",
       "6                                                  0      \n",
       "\n",
       "   ws-meclizine-psych-coexposure_meclPrecedesPsychR  \\\n",
       "0                                                 0   \n",
       "1                                                 0   \n",
       "4                                                 0   \n",
       "5                                                 0   \n",
       "6                                                 0   \n",
       "\n",
       "   ws-meclizine-psych-coexposure_psychPrecedesMeclR  \\\n",
       "0                                                 0   \n",
       "1                                                 0   \n",
       "4                                                 0   \n",
       "5                                                 0   \n",
       "6                                                 0   \n",
       "\n",
       "   ws-meclizine-psych-coexposure_meclAndPsychOlderStart  input_outcome_x  \\\n",
       "0                                                  0                   0   \n",
       "1                                                  0                   0   \n",
       "4                                                  0                   0   \n",
       "5                                                  0                   0   \n",
       "6                                                  0                   0   \n",
       "\n",
       "   input_outcome_y  \n",
       "0                0  \n",
       "1                0  \n",
       "4                0  \n",
       "5                0  \n",
       "6                0  \n",
       "\n",
       "[5 rows x 102 columns]"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_use.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>current-stay-days</th>\n",
       "      <th>cumulative-days-in-facility</th>\n",
       "      <th>cms-long-stay</th>\n",
       "      <th>mds-stay-trans-short-to-long</th>\n",
       "      <th>age</th>\n",
       "      <th>mds-antianxiety-medication</th>\n",
       "      <th>mds-antidepressant-medication</th>\n",
       "      <th>mds-antipsychotic-medication</th>\n",
       "      <th>mds-antibiotic-medication</th>\n",
       "      <th>mds-anticoagulant-medication</th>\n",
       "      <th>...</th>\n",
       "      <th>mds-bims-summary-ranking_Intact or Moderately Intact</th>\n",
       "      <th>mds-bims-summary-ranking_Moderate Impairment</th>\n",
       "      <th>ws-meclizine-psych-coexposure_No</th>\n",
       "      <th>ws-meclizine-psych-coexposure_meclAndPsychStartSameR</th>\n",
       "      <th>ws-meclizine-psych-coexposure_meclPrecedesPsychR</th>\n",
       "      <th>ws-meclizine-psych-coexposure_psychPrecedesMeclR</th>\n",
       "      <th>ws-meclizine-psych-coexposure_meclAndPsychOlderStart</th>\n",
       "      <th>input_outcome_x</th>\n",
       "      <th>input_outcome_y</th>\n",
       "      <th>patient-id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.006369</td>\n",
       "      <td>0.006342</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.558140</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>10353</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.006369</td>\n",
       "      <td>0.006342</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.755814</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>10364</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.006369</td>\n",
       "      <td>0.006342</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.674419</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>10386</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.006369</td>\n",
       "      <td>0.006342</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.825581</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.006369</td>\n",
       "      <td>0.006342</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.825581</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5003</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 102 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   current-stay-days  cumulative-days-in-facility  cms-long-stay  \\\n",
       "0           0.006369                     0.006342            0.0   \n",
       "1           0.006369                     0.006342            0.0   \n",
       "4           0.006369                     0.006342            0.0   \n",
       "5           0.006369                     0.006342            0.0   \n",
       "6           0.006369                     0.006342            0.0   \n",
       "\n",
       "   mds-stay-trans-short-to-long       age  mds-antianxiety-medication  \\\n",
       "0                           0.0  0.558140                         1.0   \n",
       "1                           1.0  0.755814                         0.0   \n",
       "4                           1.0  0.674419                         1.0   \n",
       "5                           0.0  0.825581                         0.0   \n",
       "6                           1.0  0.825581                         0.0   \n",
       "\n",
       "   mds-antidepressant-medication  mds-antipsychotic-medication  \\\n",
       "0                            1.0                           0.0   \n",
       "1                            0.0                           1.0   \n",
       "4                            1.0                           1.0   \n",
       "5                            1.0                           1.0   \n",
       "6                            1.0                           1.0   \n",
       "\n",
       "   mds-antibiotic-medication  mds-anticoagulant-medication  ...  \\\n",
       "0                        0.0                           0.0  ...   \n",
       "1                        1.0                           0.0  ...   \n",
       "4                        0.0                           0.0  ...   \n",
       "5                        0.0                           0.0  ...   \n",
       "6                        1.0                           0.0  ...   \n",
       "\n",
       "   mds-bims-summary-ranking_Intact or Moderately Intact  \\\n",
       "0                                                1.0      \n",
       "1                                                1.0      \n",
       "4                                                1.0      \n",
       "5                                                1.0      \n",
       "6                                                1.0      \n",
       "\n",
       "   mds-bims-summary-ranking_Moderate Impairment  \\\n",
       "0                                           0.0   \n",
       "1                                           0.0   \n",
       "4                                           0.0   \n",
       "5                                           0.0   \n",
       "6                                           0.0   \n",
       "\n",
       "   ws-meclizine-psych-coexposure_No  \\\n",
       "0                               1.0   \n",
       "1                               1.0   \n",
       "4                               1.0   \n",
       "5                               1.0   \n",
       "6                               1.0   \n",
       "\n",
       "   ws-meclizine-psych-coexposure_meclAndPsychStartSameR  \\\n",
       "0                                                0.0      \n",
       "1                                                0.0      \n",
       "4                                                0.0      \n",
       "5                                                0.0      \n",
       "6                                                0.0      \n",
       "\n",
       "   ws-meclizine-psych-coexposure_meclPrecedesPsychR  \\\n",
       "0                                               0.0   \n",
       "1                                               0.0   \n",
       "4                                               0.0   \n",
       "5                                               0.0   \n",
       "6                                               0.0   \n",
       "\n",
       "   ws-meclizine-psych-coexposure_psychPrecedesMeclR  \\\n",
       "0                                               0.0   \n",
       "1                                               0.0   \n",
       "4                                               0.0   \n",
       "5                                               0.0   \n",
       "6                                               0.0   \n",
       "\n",
       "   ws-meclizine-psych-coexposure_meclAndPsychOlderStart  input_outcome_x  \\\n",
       "0                                                0.0                 0.0   \n",
       "1                                                0.0                 0.0   \n",
       "4                                                0.0                 0.0   \n",
       "5                                                0.0                 0.0   \n",
       "6                                                0.0                 0.0   \n",
       "\n",
       "   input_outcome_y  patient-id  \n",
       "0              0.0       10353  \n",
       "1              0.0       10364  \n",
       "4              0.0       10386  \n",
       "5              0.0        5000  \n",
       "6              0.0        5003  \n",
       "\n",
       "[5 rows x 102 columns]"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_transformed.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "def generate_nested_list(X, Y):\n",
    "    x_list, y_list = [], [] \n",
    "    pids = X['patient-id'].unique()\n",
    "    for pid in pids:\n",
    "        df = X[X['patient-id']==pid]\n",
    "        y_list.append(Y[df.index].values)\n",
    "        x_list.append([record for record in df.drop('patient-id', axis=1).values])\n",
    "    return x_list, y_list\n",
    "\n",
    "def mygenerator(x_list, y_list=None):\n",
    "    if(y_list is not None):\n",
    "        while True:\n",
    "            for x, y in zip(x_list, y_list):\n",
    "                yield np.array(x).reshape((len(x), 1, x[0].shape[0])), y\n",
    "    else:\n",
    "        while True:\n",
    "            for x in x_list:\n",
    "                yield np.array(x).reshape((len(x), 1, x[0].shape[0])) \n",
    "                \n",
    "def fit_model(model, epoch, X_train, y_train, X_valid=None, y_valid=None, verbose=0):\n",
    "    history = model.fit(X_train, y_train, epochs=epoch, verbose=verbose)\n",
    "    if(verbose != 0):\n",
    "        plot_history(history)\n",
    "    return history\n",
    "\n",
    "def plot_roc_curve(fpr, tpr):\n",
    "    plt.figure()\n",
    "    plt.plot(fpr, tpr, color='darkorange', label='ROC curve (area = %0.2f)' % auc(fpr, tpr))\n",
    "    plt.plot([0, 1], [0, 1], color='navy', linestyle='--')\n",
    "    plt.xlim([0.0, 1.0])\n",
    "    plt.ylim([0.0, 1.05])\n",
    "    plt.xlabel('False Positive Rate')\n",
    "    plt.ylabel('True Positive Rate')\n",
    "    plt.legend(loc=\"lower right\")\n",
    "    plt.show()\n",
    "    \n",
    "def plot_history(history, loss=None, val_loss=None):\n",
    "    if(history is not None):\n",
    "        loss = history.history['loss']\n",
    "    plt.figure()\n",
    "    plt.plot(loss)\n",
    "    plt.title('model loss')\n",
    "    plt.ylabel('loss')\n",
    "    plt.xlabel('epoch')\n",
    "    if('val_loss' in history.history):\n",
    "        val_loss = history.history['val_loss']\n",
    "        plt.plot(val_loss)\n",
    "    plt.legend(['train', 'test'], loc='upper right')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "def pad_data(X, maxlen):\n",
    "    return pad_sequences(X, maxlen=maxlen, dtype='float64', padding='post')\n",
    "\n",
    "def pad_all(X_train, y_train, X_test, y_test, maxlen):\n",
    "    return pad_data(X_train, maxlen), pad_data(y_train, maxlen), pad_data(X_test, maxlen), pad_data(y_test, maxlen)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "def auroc(y_predict, y_true, plot=False):\n",
    "    fpr, tpr, thresholds = roc_curve(y_true, y_predict, pos_label=1)    \n",
    "    if(plot):\n",
    "        plot_roc_curve(fpr, tpr)\n",
    "    return auc(fpr, tpr)\n",
    "\n",
    "\n",
    "def find_best_f1(y_predict, y_true):\n",
    "    fpr, tpr, thresholds = roc_curve(y_true, y_predict, pos_label=1) \n",
    "    max_f1, th = 0, 0\n",
    "    for threshold in thresholds:\n",
    "        f1 = f1_score(y_true, y_predict > threshold)\n",
    "        if(f1 > max_f1):\n",
    "            max_f1 = f1\n",
    "            th = threshold\n",
    "    return max_f1, th\n",
    "\n",
    "def other_metrics(y_predict, y_true, threshold):\n",
    "    y_threshold = y_predict > threshold\n",
    "    tn, fp, fn, tp = confusion_matrix(y_true, y_threshold).ravel()\n",
    "    return precision_score(y_true, y_threshold), recall_score(y_true, y_threshold),\\\n",
    "            accuracy_score(y_true, y_threshold), float(tn) / (tn + fp)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import GroupKFold\n",
    "def kfold_validation(unit, ratio, epoch, X, y, groupby, k=5, verbose=0, max_timestamps=20, n_features=101):\n",
    "    auroc_list = []\n",
    "    other_metric_list = []\n",
    "    for train_idx, valid_idx in GroupKFold(n_splits=5).split(X, y, X[groupby]):\n",
    "        X_train, X_valid, y_train, y_valid = X.iloc[train_idx], X.iloc[valid_idx], y.iloc[train_idx], y.iloc[valid_idx]\n",
    "        X_train, y_train = generate_nested_list(X_train, y_train)\n",
    "        X_valid, y_valid = generate_nested_list(X_valid, y_valid)\n",
    "        X_train_pad, y_train_pad, X_valid_pad, y_valid_pad = pad_all(X_train, y_train, X_valid, y_valid, max_timestamps)\n",
    "        y_train_pad = y_train_pad.reshape(y_train_pad.shape[0], y_train_pad.shape[1], 1)\n",
    "        seed(1)\n",
    "        set_random_seed(2)\n",
    "        model = build_model(unit, ratio, max_timestamps, n_features)\n",
    "        fit_model(model, epoch, X_train_pad, y_train_pad, verbose=verbose)\n",
    "        y_valid_len = [len(y) for y in y_valid]\n",
    "        y_predict = model.predict(X_valid_pad)\n",
    "        y_predict = truncate_prediction(y_predict, y_valid_len)\n",
    "        y_true = np.concatenate(y_valid)\n",
    "        auroc_list.append(auroc(y_predict, y_true))\n",
    "        f1, th = find_best_f1(y_predict, y_true)\n",
    "        precision, recall, accuracy, specificity = other_metrics(y_predict, y_true, th)\n",
    "        other_metric_list.append((precision, recall, accuracy, specificity,f1,th))\n",
    "    return (auroc_list, other_metric_list) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "# %%time\n",
    "# shape: (num_of_patients, num_of_records, num_of_features)\n",
    "X_train_list, y_train_list = generate_nested_list(X_train_transformed, y_train)\n",
    "X_test_list, y_test_list = generate_nested_list(X_test_transformed, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 2,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 3,\n",
       " 7,\n",
       " 7,\n",
       " 5,\n",
       " 7,\n",
       " 8,\n",
       " 1,\n",
       " 11,\n",
       " 10,\n",
       " 7,\n",
       " 4,\n",
       " 1,\n",
       " 2,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 7,\n",
       " 2,\n",
       " 1,\n",
       " 1,\n",
       " 2,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 2,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 2,\n",
       " 4,\n",
       " 2,\n",
       " 14,\n",
       " 4,\n",
       " 11,\n",
       " 4,\n",
       " 5,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 2,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 9,\n",
       " 1,\n",
       " 5,\n",
       " 2,\n",
       " 1,\n",
       " 4,\n",
       " 15,\n",
       " 2,\n",
       " 2,\n",
       " 5,\n",
       " 11,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 6,\n",
       " 7,\n",
       " 3,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 2,\n",
       " 1,\n",
       " 2,\n",
       " 3,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 12,\n",
       " 13,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 8,\n",
       " 1,\n",
       " 1,\n",
       " 3,\n",
       " 1,\n",
       " 2,\n",
       " 2,\n",
       " 1,\n",
       " 1,\n",
       " 2,\n",
       " 1,\n",
       " 10,\n",
       " 2,\n",
       " 5,\n",
       " 11,\n",
       " 11,\n",
       " 1,\n",
       " 2,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 3,\n",
       " 2,\n",
       " 2,\n",
       " 1,\n",
       " 1,\n",
       " 5,\n",
       " 8,\n",
       " 12,\n",
       " 1,\n",
       " 2,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 13,\n",
       " 2,\n",
       " 1,\n",
       " 1,\n",
       " 7,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 9,\n",
       " 1,\n",
       " 3,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 3,\n",
       " 2,\n",
       " 1,\n",
       " 2,\n",
       " 1,\n",
       " 1,\n",
       " 2,\n",
       " 1,\n",
       " 1,\n",
       " 2,\n",
       " 2,\n",
       " 1,\n",
       " 1,\n",
       " 3,\n",
       " 8,\n",
       " 8,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 2,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 2,\n",
       " 1,\n",
       " 2,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 2,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 1,\n",
       " 2,\n",
       " 1,\n",
       " 1,\n",
       " 3,\n",
       " 1]"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[len(x) for x in X_train_list[0:200]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[array([0.00636943, 0.00634249, 0.        , 1.        , 0.8255814 ,\n",
       "        0.        , 1.        , 1.        , 1.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 1.        , 1.        , 0.        ,\n",
       "        0.        , 0.        , 1.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.25      , 0.25      ,\n",
       "        1.        , 0.        , 1.        , 0.        , 0.67857143,\n",
       "        0.03676471, 0.        , 0.01563858, 0.01937984, 0.        ,\n",
       "        0.        , 0.41421356, 1.        , 0.60134489, 0.89997622,\n",
       "        0.99672198, 0.42204048, 0.9984043 , 0.41804512, 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 1.        , 0.        , 0.        , 1.        ,\n",
       "        0.        , 1.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 1.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 1.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 1.        , 0.        , 0.        ,\n",
       "        0.        , 1.        , 1.        , 0.        , 1.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        ]),\n",
       " array([0.00636943, 0.0422833 , 0.        , 1.        , 0.8255814 ,\n",
       "        0.        , 1.        , 1.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 1.        , 1.        , 1.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.375     , 0.25      ,\n",
       "        1.        , 1.        , 1.        , 1.        , 0.71428571,\n",
       "        0.06617647, 0.        , 0.04691573, 0.02131783, 0.        ,\n",
       "        0.        , 0.41421356, 1.        , 0.41421356, 1.        ,\n",
       "        0.9971726 , 0.42097469, 0.99767413, 0.41978471, 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 1.        , 0.        , 0.        , 1.        ,\n",
       "        0.        , 1.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 1.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 1.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 1.        , 0.        , 0.        ,\n",
       "        0.        , 1.        , 1.        , 0.        , 1.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        1.        ])]"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_list[4]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19\n"
     ]
    }
   ],
   "source": [
    "max_timestamps = df.groupby(['patient-id']).size().max()\n",
    "print(max_timestamps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "101\n"
     ]
    }
   ],
   "source": [
    "n_features = X_train_list[0][0].shape[0]\n",
    "print(n_features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Use a GRU to compare with the LSTM runs\n",
    "from keras import regularizers\n",
    "from keras.optimizers import Adagrad, Adadelta\n",
    "def build_model(units, dropout_ratio, n_timesteps, n_dimensions):\n",
    "    model = Sequential()\n",
    "    model.add(Masking(input_shape=(n_timesteps, n_dimensions)))\n",
    "    model.add(GRU(units, return_sequences=True, recurrent_dropout=dropout_ratio, dropout=0.5))\n",
    "    model.add(TimeDistributed(Dense(1, activation='sigmoid')))\n",
    "    model.compile(loss='binary_crossentropy', optimizer='adagrad', metrics=['binary_accuracy'])\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "def truncate_prediction(y_, y_len):\n",
    "    y_concat = []\n",
    "    for i, y in enumerate(y_):\n",
    "        y_concat.append(np.concatenate(y[:y_len[i], :]))\n",
    "    y_concat = np.concatenate(y_concat)\n",
    "    return y_concat"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_pad, y_train_pad, X_test_pad, y_test_pad = pad_all(X_train_list, y_train_list, X_test_list, y_test_list, max_timestamps)\n",
    "y_train_pad = y_train_pad.reshape((2193, 19, 1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(X_train_list[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "19"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(X_train_pad[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/175\n",
      "2193/2193 [==============================] - 3s 1ms/step - loss: 0.4703 - binary_accuracy: 0.8220\n",
      "Epoch 2/175\n",
      "2193/2193 [==============================] - 2s 812us/step - loss: 0.4474 - binary_accuracy: 0.8205\n",
      "Epoch 3/175\n",
      "2193/2193 [==============================] - 2s 811us/step - loss: 0.4460 - binary_accuracy: 0.8202\n",
      "Epoch 4/175\n",
      "2193/2193 [==============================] - 2s 817us/step - loss: 0.4408 - binary_accuracy: 0.8230\n",
      "Epoch 5/175\n",
      "2193/2193 [==============================] - 2s 821us/step - loss: 0.4323 - binary_accuracy: 0.8258\n",
      "Epoch 6/175\n",
      "2193/2193 [==============================] - 2s 829us/step - loss: 0.4359 - binary_accuracy: 0.8245\n",
      "Epoch 7/175\n",
      "2193/2193 [==============================] - 2s 827us/step - loss: 0.4373 - binary_accuracy: 0.8207\n",
      "Epoch 8/175\n",
      "2193/2193 [==============================] - 2s 814us/step - loss: 0.4327 - binary_accuracy: 0.8271\n",
      "Epoch 9/175\n",
      "2193/2193 [==============================] - 2s 822us/step - loss: 0.4275 - binary_accuracy: 0.8233\n",
      "Epoch 10/175\n",
      "2193/2193 [==============================] - 2s 835us/step - loss: 0.4304 - binary_accuracy: 0.8213\n",
      "Epoch 11/175\n",
      "2193/2193 [==============================] - 2s 832us/step - loss: 0.4332 - binary_accuracy: 0.8203\n",
      "Epoch 12/175\n",
      "2193/2193 [==============================] - 2s 806us/step - loss: 0.4345 - binary_accuracy: 0.8177\n",
      "Epoch 13/175\n",
      "2193/2193 [==============================] - 2s 820us/step - loss: 0.4328 - binary_accuracy: 0.8219\n",
      "Epoch 14/175\n",
      "2193/2193 [==============================] - 2s 828us/step - loss: 0.4280 - binary_accuracy: 0.8212\n",
      "Epoch 15/175\n",
      "2193/2193 [==============================] - 2s 808us/step - loss: 0.4335 - binary_accuracy: 0.8222\n",
      "Epoch 16/175\n",
      "2193/2193 [==============================] - 2s 819us/step - loss: 0.4279 - binary_accuracy: 0.8248\n",
      "Epoch 17/175\n",
      "2193/2193 [==============================] - 2s 837us/step - loss: 0.4242 - binary_accuracy: 0.8261\n",
      "Epoch 18/175\n",
      "2193/2193 [==============================] - 2s 815us/step - loss: 0.4274 - binary_accuracy: 0.8234\n",
      "Epoch 19/175\n",
      "2193/2193 [==============================] - 2s 814us/step - loss: 0.4299 - binary_accuracy: 0.8231\n",
      "Epoch 20/175\n",
      "2193/2193 [==============================] - 2s 816us/step - loss: 0.4217 - binary_accuracy: 0.8266\n",
      "Epoch 21/175\n",
      "2193/2193 [==============================] - 2s 811us/step - loss: 0.4261 - binary_accuracy: 0.8276\n",
      "Epoch 22/175\n",
      "2193/2193 [==============================] - 2s 816us/step - loss: 0.4256 - binary_accuracy: 0.8262\n",
      "Epoch 23/175\n",
      "2193/2193 [==============================] - 2s 819us/step - loss: 0.4299 - binary_accuracy: 0.8203\n",
      "Epoch 24/175\n",
      "2193/2193 [==============================] - 2s 803us/step - loss: 0.4248 - binary_accuracy: 0.8273\n",
      "Epoch 25/175\n",
      "2193/2193 [==============================] - 2s 812us/step - loss: 0.4220 - binary_accuracy: 0.8242\n",
      "Epoch 26/175\n",
      "2193/2193 [==============================] - 2s 816us/step - loss: 0.4240 - binary_accuracy: 0.8236\n",
      "Epoch 27/175\n",
      "2193/2193 [==============================] - 2s 829us/step - loss: 0.4230 - binary_accuracy: 0.8271\n",
      "Epoch 28/175\n",
      "2193/2193 [==============================] - 2s 813us/step - loss: 0.4210 - binary_accuracy: 0.8222\n",
      "Epoch 29/175\n",
      "2193/2193 [==============================] - 2s 833us/step - loss: 0.4236 - binary_accuracy: 0.8222\n",
      "Epoch 30/175\n",
      "2193/2193 [==============================] - 2s 818us/step - loss: 0.4216 - binary_accuracy: 0.8251\n",
      "Epoch 31/175\n",
      "2193/2193 [==============================] - 2s 811us/step - loss: 0.4280 - binary_accuracy: 0.8238\n",
      "Epoch 32/175\n",
      "2193/2193 [==============================] - 2s 802us/step - loss: 0.4259 - binary_accuracy: 0.8234\n",
      "Epoch 33/175\n",
      "2193/2193 [==============================] - 2s 809us/step - loss: 0.4198 - binary_accuracy: 0.8284\n",
      "Epoch 34/175\n",
      "2193/2193 [==============================] - 2s 833us/step - loss: 0.4257 - binary_accuracy: 0.8243\n",
      "Epoch 35/175\n",
      "2193/2193 [==============================] - 2s 810us/step - loss: 0.4210 - binary_accuracy: 0.8251\n",
      "Epoch 36/175\n",
      "2193/2193 [==============================] - 2s 816us/step - loss: 0.4253 - binary_accuracy: 0.8211\n",
      "Epoch 37/175\n",
      "2193/2193 [==============================] - 2s 808us/step - loss: 0.4182 - binary_accuracy: 0.8272\n",
      "Epoch 38/175\n",
      "2193/2193 [==============================] - 2s 824us/step - loss: 0.4190 - binary_accuracy: 0.8275\n",
      "Epoch 39/175\n",
      "2193/2193 [==============================] - 2s 801us/step - loss: 0.4198 - binary_accuracy: 0.8261\n",
      "Epoch 40/175\n",
      "2193/2193 [==============================] - 2s 839us/step - loss: 0.4212 - binary_accuracy: 0.8272\n",
      "Epoch 41/175\n",
      "2193/2193 [==============================] - 2s 812us/step - loss: 0.4236 - binary_accuracy: 0.8235\n",
      "Epoch 42/175\n",
      "2193/2193 [==============================] - 2s 809us/step - loss: 0.4158 - binary_accuracy: 0.8251\n",
      "Epoch 43/175\n",
      "2193/2193 [==============================] - 2s 819us/step - loss: 0.4206 - binary_accuracy: 0.8247\n",
      "Epoch 44/175\n",
      "2193/2193 [==============================] - 2s 819us/step - loss: 0.4229 - binary_accuracy: 0.8233\n",
      "Epoch 45/175\n",
      "2193/2193 [==============================] - 2s 826us/step - loss: 0.4173 - binary_accuracy: 0.8284\n",
      "Epoch 46/175\n",
      "2193/2193 [==============================] - 2s 804us/step - loss: 0.4161 - binary_accuracy: 0.8261\n",
      "Epoch 47/175\n",
      "2193/2193 [==============================] - 2s 823us/step - loss: 0.4133 - binary_accuracy: 0.8269\n",
      "Epoch 48/175\n",
      "2193/2193 [==============================] - 2s 787us/step - loss: 0.4215 - binary_accuracy: 0.8255\n",
      "Epoch 49/175\n",
      "2193/2193 [==============================] - 2s 788us/step - loss: 0.4147 - binary_accuracy: 0.8288\n",
      "Epoch 50/175\n",
      "2193/2193 [==============================] - 2s 796us/step - loss: 0.4148 - binary_accuracy: 0.8321\n",
      "Epoch 51/175\n",
      "2193/2193 [==============================] - 2s 795us/step - loss: 0.4214 - binary_accuracy: 0.8259\n",
      "Epoch 52/175\n",
      "2193/2193 [==============================] - 2s 801us/step - loss: 0.4199 - binary_accuracy: 0.8263\n",
      "Epoch 53/175\n",
      "2193/2193 [==============================] - 2s 805us/step - loss: 0.4229 - binary_accuracy: 0.8221\n",
      "Epoch 54/175\n",
      "2193/2193 [==============================] - 2s 817us/step - loss: 0.4208 - binary_accuracy: 0.8274\n",
      "Epoch 55/175\n",
      "2193/2193 [==============================] - 2s 806us/step - loss: 0.4240 - binary_accuracy: 0.8227\n",
      "Epoch 56/175\n",
      "2193/2193 [==============================] - 2s 835us/step - loss: 0.4223 - binary_accuracy: 0.8265\n",
      "Epoch 57/175\n",
      "2193/2193 [==============================] - 2s 804us/step - loss: 0.4179 - binary_accuracy: 0.8261\n",
      "Epoch 58/175\n",
      "2193/2193 [==============================] - 2s 807us/step - loss: 0.4161 - binary_accuracy: 0.8265\n",
      "Epoch 59/175\n",
      "2193/2193 [==============================] - 2s 820us/step - loss: 0.4159 - binary_accuracy: 0.8267\n",
      "Epoch 60/175\n",
      "2193/2193 [==============================] - 2s 783us/step - loss: 0.4192 - binary_accuracy: 0.8252\n",
      "Epoch 61/175\n",
      "2193/2193 [==============================] - 2s 789us/step - loss: 0.4193 - binary_accuracy: 0.8296\n",
      "Epoch 62/175\n",
      "2193/2193 [==============================] - 2s 793us/step - loss: 0.4114 - binary_accuracy: 0.8333\n",
      "Epoch 63/175\n",
      "2193/2193 [==============================] - 2s 789us/step - loss: 0.4200 - binary_accuracy: 0.8253\n",
      "Epoch 64/175\n",
      "2193/2193 [==============================] - 2s 789us/step - loss: 0.4168 - binary_accuracy: 0.8300\n",
      "Epoch 65/175\n",
      "2193/2193 [==============================] - 2s 787us/step - loss: 0.4186 - binary_accuracy: 0.8254\n",
      "Epoch 66/175\n",
      "2193/2193 [==============================] - 2s 792us/step - loss: 0.4122 - binary_accuracy: 0.8309\n",
      "Epoch 67/175\n",
      "2193/2193 [==============================] - 2s 795us/step - loss: 0.4117 - binary_accuracy: 0.8317\n",
      "Epoch 68/175\n",
      "2193/2193 [==============================] - 2s 811us/step - loss: 0.4209 - binary_accuracy: 0.8258\n",
      "Epoch 69/175\n",
      "2193/2193 [==============================] - 2s 816us/step - loss: 0.4143 - binary_accuracy: 0.8304\n",
      "Epoch 70/175\n",
      "2193/2193 [==============================] - 2s 797us/step - loss: 0.4241 - binary_accuracy: 0.8251\n",
      "Epoch 71/175\n",
      "2193/2193 [==============================] - 2s 810us/step - loss: 0.4172 - binary_accuracy: 0.8248\n",
      "Epoch 72/175\n",
      "2193/2193 [==============================] - 2s 812us/step - loss: 0.4153 - binary_accuracy: 0.8269\n",
      "Epoch 73/175\n",
      "2193/2193 [==============================] - 2s 798us/step - loss: 0.4117 - binary_accuracy: 0.8279\n",
      "Epoch 74/175\n",
      "2193/2193 [==============================] - 2s 805us/step - loss: 0.4145 - binary_accuracy: 0.8255\n",
      "Epoch 75/175\n",
      "2193/2193 [==============================] - 2s 794us/step - loss: 0.4131 - binary_accuracy: 0.8292\n",
      "Epoch 76/175\n",
      "2193/2193 [==============================] - 2s 793us/step - loss: 0.4159 - binary_accuracy: 0.8265\n",
      "Epoch 77/175\n",
      "2193/2193 [==============================] - 2s 802us/step - loss: 0.4135 - binary_accuracy: 0.8246\n",
      "Epoch 78/175\n",
      "2193/2193 [==============================] - 2s 807us/step - loss: 0.4169 - binary_accuracy: 0.8268\n",
      "Epoch 79/175\n",
      "2193/2193 [==============================] - 2s 810us/step - loss: 0.4125 - binary_accuracy: 0.8289\n",
      "Epoch 80/175\n",
      "2193/2193 [==============================] - 2s 800us/step - loss: 0.4182 - binary_accuracy: 0.8277\n",
      "Epoch 81/175\n",
      "2193/2193 [==============================] - 2s 800us/step - loss: 0.4164 - binary_accuracy: 0.8273\n",
      "Epoch 82/175\n",
      "2193/2193 [==============================] - 2s 800us/step - loss: 0.4192 - binary_accuracy: 0.8214\n",
      "Epoch 83/175\n",
      "2193/2193 [==============================] - 2s 809us/step - loss: 0.4141 - binary_accuracy: 0.8289\n",
      "Epoch 84/175\n",
      "2193/2193 [==============================] - 2s 813us/step - loss: 0.4220 - binary_accuracy: 0.8260\n",
      "Epoch 85/175\n",
      "2193/2193 [==============================] - 2s 828us/step - loss: 0.4155 - binary_accuracy: 0.8290\n",
      "Epoch 86/175\n",
      "2193/2193 [==============================] - 2s 811us/step - loss: 0.4165 - binary_accuracy: 0.8257\n",
      "Epoch 87/175\n",
      "2193/2193 [==============================] - 2s 806us/step - loss: 0.4177 - binary_accuracy: 0.8270\n",
      "Epoch 88/175\n",
      "2193/2193 [==============================] - 2s 805us/step - loss: 0.4189 - binary_accuracy: 0.8260\n",
      "Epoch 89/175\n",
      "2193/2193 [==============================] - 2s 799us/step - loss: 0.4230 - binary_accuracy: 0.8214\n",
      "Epoch 90/175\n",
      "2193/2193 [==============================] - 2s 796us/step - loss: 0.4123 - binary_accuracy: 0.8287\n",
      "Epoch 91/175\n",
      "2193/2193 [==============================] - 2s 796us/step - loss: 0.4132 - binary_accuracy: 0.8239\n",
      "Epoch 92/175\n",
      "2193/2193 [==============================] - 2s 814us/step - loss: 0.4112 - binary_accuracy: 0.8274\n",
      "Epoch 93/175\n",
      "2193/2193 [==============================] - 2s 808us/step - loss: 0.4178 - binary_accuracy: 0.8235\n",
      "Epoch 94/175\n",
      "2193/2193 [==============================] - 2s 812us/step - loss: 0.4132 - binary_accuracy: 0.8261\n",
      "Epoch 95/175\n",
      "2193/2193 [==============================] - 2s 800us/step - loss: 0.4160 - binary_accuracy: 0.8252\n",
      "Epoch 96/175\n",
      "2193/2193 [==============================] - 2s 786us/step - loss: 0.4187 - binary_accuracy: 0.8260\n",
      "Epoch 97/175\n",
      "2193/2193 [==============================] - 2s 790us/step - loss: 0.4164 - binary_accuracy: 0.8234\n",
      "Epoch 98/175\n",
      "2193/2193 [==============================] - 2s 785us/step - loss: 0.4196 - binary_accuracy: 0.8228\n",
      "Epoch 99/175\n",
      "2193/2193 [==============================] - 2s 788us/step - loss: 0.4124 - binary_accuracy: 0.8256\n",
      "Epoch 100/175\n",
      "2193/2193 [==============================] - 2s 820us/step - loss: 0.4182 - binary_accuracy: 0.8260\n",
      "Epoch 101/175\n",
      "2193/2193 [==============================] - 2s 829us/step - loss: 0.4155 - binary_accuracy: 0.8276\n",
      "Epoch 102/175\n",
      "2193/2193 [==============================] - 2s 807us/step - loss: 0.4165 - binary_accuracy: 0.8295\n",
      "Epoch 103/175\n",
      "2193/2193 [==============================] - 2s 781us/step - loss: 0.4143 - binary_accuracy: 0.8278\n",
      "Epoch 104/175\n",
      "2193/2193 [==============================] - 2s 785us/step - loss: 0.4124 - binary_accuracy: 0.8275\n",
      "Epoch 105/175\n",
      "2193/2193 [==============================] - 2s 793us/step - loss: 0.4115 - binary_accuracy: 0.8281\n",
      "Epoch 106/175\n",
      "2193/2193 [==============================] - 2s 784us/step - loss: 0.4087 - binary_accuracy: 0.8312\n",
      "Epoch 107/175\n",
      "2193/2193 [==============================] - 2s 804us/step - loss: 0.4089 - binary_accuracy: 0.8249\n",
      "Epoch 108/175\n",
      "2193/2193 [==============================] - 2s 836us/step - loss: 0.4177 - binary_accuracy: 0.8244\n",
      "Epoch 109/175\n",
      "2193/2193 [==============================] - 2s 801us/step - loss: 0.4190 - binary_accuracy: 0.8262\n",
      "Epoch 110/175\n",
      "2193/2193 [==============================] - 2s 825us/step - loss: 0.4153 - binary_accuracy: 0.8312\n",
      "Epoch 111/175\n",
      "2193/2193 [==============================] - 2s 808us/step - loss: 0.4133 - binary_accuracy: 0.8293\n",
      "Epoch 112/175\n",
      "2193/2193 [==============================] - 2s 787us/step - loss: 0.4088 - binary_accuracy: 0.8295\n",
      "Epoch 113/175\n",
      "2193/2193 [==============================] - 2s 788us/step - loss: 0.4158 - binary_accuracy: 0.8272\n",
      "Epoch 114/175\n",
      "2193/2193 [==============================] - 2s 795us/step - loss: 0.4126 - binary_accuracy: 0.8281\n",
      "Epoch 115/175\n",
      "2193/2193 [==============================] - 2s 800us/step - loss: 0.4147 - binary_accuracy: 0.8273\n",
      "Epoch 116/175\n",
      "2193/2193 [==============================] - 2s 807us/step - loss: 0.4143 - binary_accuracy: 0.8238\n",
      "Epoch 117/175\n",
      "2193/2193 [==============================] - 2s 810us/step - loss: 0.4103 - binary_accuracy: 0.8273\n",
      "Epoch 118/175\n",
      "2193/2193 [==============================] - 2s 822us/step - loss: 0.4165 - binary_accuracy: 0.8264\n",
      "Epoch 119/175\n",
      "2193/2193 [==============================] - 2s 801us/step - loss: 0.4158 - binary_accuracy: 0.8295\n",
      "Epoch 120/175\n",
      "2193/2193 [==============================] - 2s 794us/step - loss: 0.4143 - binary_accuracy: 0.8270\n",
      "Epoch 121/175\n",
      "2193/2193 [==============================] - 2s 795us/step - loss: 0.4141 - binary_accuracy: 0.8239\n",
      "Epoch 122/175\n",
      "2193/2193 [==============================] - 2s 810us/step - loss: 0.4140 - binary_accuracy: 0.8275\n",
      "Epoch 123/175\n",
      "2193/2193 [==============================] - 2s 813us/step - loss: 0.4141 - binary_accuracy: 0.8295\n",
      "Epoch 124/175\n",
      "2193/2193 [==============================] - 2s 818us/step - loss: 0.4114 - binary_accuracy: 0.8304\n",
      "Epoch 125/175\n",
      "2193/2193 [==============================] - 2s 813us/step - loss: 0.4117 - binary_accuracy: 0.8298\n",
      "Epoch 126/175\n",
      "2193/2193 [==============================] - 2s 799us/step - loss: 0.4141 - binary_accuracy: 0.8290\n",
      "Epoch 127/175\n",
      "2193/2193 [==============================] - 2s 800us/step - loss: 0.4160 - binary_accuracy: 0.8246\n",
      "Epoch 128/175\n",
      "2193/2193 [==============================] - 2s 812us/step - loss: 0.4161 - binary_accuracy: 0.8266\n",
      "Epoch 129/175\n",
      "2193/2193 [==============================] - 2s 794us/step - loss: 0.4070 - binary_accuracy: 0.8331\n",
      "Epoch 130/175\n",
      "2193/2193 [==============================] - 2s 800us/step - loss: 0.4092 - binary_accuracy: 0.8319\n",
      "Epoch 131/175\n",
      "2193/2193 [==============================] - 2s 813us/step - loss: 0.4082 - binary_accuracy: 0.8281\n",
      "Epoch 132/175\n",
      "2193/2193 [==============================] - 2s 810us/step - loss: 0.4178 - binary_accuracy: 0.8270\n",
      "Epoch 133/175\n",
      "2193/2193 [==============================] - 2s 807us/step - loss: 0.4140 - binary_accuracy: 0.8264\n",
      "Epoch 134/175\n",
      "2193/2193 [==============================] - 2s 813us/step - loss: 0.4096 - binary_accuracy: 0.8296\n",
      "Epoch 135/175\n",
      "2193/2193 [==============================] - 2s 802us/step - loss: 0.4172 - binary_accuracy: 0.8283\n",
      "Epoch 136/175\n",
      "2193/2193 [==============================] - 2s 800us/step - loss: 0.4221 - binary_accuracy: 0.8267\n",
      "Epoch 137/175\n",
      "2193/2193 [==============================] - 2s 795us/step - loss: 0.4139 - binary_accuracy: 0.8287\n",
      "Epoch 138/175\n",
      "2193/2193 [==============================] - 2s 798us/step - loss: 0.4129 - binary_accuracy: 0.8293\n",
      "Epoch 139/175\n",
      "2193/2193 [==============================] - 2s 800us/step - loss: 0.4119 - binary_accuracy: 0.8284\n",
      "Epoch 140/175\n",
      "2193/2193 [==============================] - 2s 823us/step - loss: 0.4167 - binary_accuracy: 0.8249\n",
      "Epoch 141/175\n",
      "2193/2193 [==============================] - 2s 800us/step - loss: 0.4092 - binary_accuracy: 0.8292\n",
      "Epoch 142/175\n",
      "2193/2193 [==============================] - 2s 796us/step - loss: 0.4138 - binary_accuracy: 0.8293\n",
      "Epoch 143/175\n",
      "2193/2193 [==============================] - 2s 816us/step - loss: 0.4083 - binary_accuracy: 0.8320\n",
      "Epoch 144/175\n",
      "2193/2193 [==============================] - 2s 806us/step - loss: 0.4130 - binary_accuracy: 0.8295\n",
      "Epoch 145/175\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2193/2193 [==============================] - 2s 799us/step - loss: 0.4094 - binary_accuracy: 0.8273\n",
      "Epoch 146/175\n",
      "2193/2193 [==============================] - 2s 793us/step - loss: 0.4121 - binary_accuracy: 0.8262\n",
      "Epoch 147/175\n",
      "2193/2193 [==============================] - 2s 795us/step - loss: 0.4084 - binary_accuracy: 0.8296\n",
      "Epoch 148/175\n",
      "2193/2193 [==============================] - 2s 797us/step - loss: 0.4056 - binary_accuracy: 0.8327\n",
      "Epoch 149/175\n",
      "2193/2193 [==============================] - 2s 817us/step - loss: 0.4084 - binary_accuracy: 0.8260\n",
      "Epoch 150/175\n",
      "2193/2193 [==============================] - 2s 803us/step - loss: 0.4089 - binary_accuracy: 0.8312\n",
      "Epoch 151/175\n",
      "2193/2193 [==============================] - 2s 820us/step - loss: 0.4085 - binary_accuracy: 0.8298\n",
      "Epoch 152/175\n",
      "2193/2193 [==============================] - 2s 807us/step - loss: 0.4138 - binary_accuracy: 0.8268\n",
      "Epoch 153/175\n",
      "2193/2193 [==============================] - 2s 789us/step - loss: 0.4124 - binary_accuracy: 0.8296\n",
      "Epoch 154/175\n",
      "2193/2193 [==============================] - 2s 801us/step - loss: 0.4124 - binary_accuracy: 0.8260\n",
      "Epoch 155/175\n",
      "2193/2193 [==============================] - 2s 803us/step - loss: 0.4141 - binary_accuracy: 0.8276\n",
      "Epoch 156/175\n",
      "2193/2193 [==============================] - 2s 794us/step - loss: 0.4102 - binary_accuracy: 0.8314\n",
      "Epoch 157/175\n",
      "2193/2193 [==============================] - 2s 796us/step - loss: 0.4109 - binary_accuracy: 0.8297\n",
      "Epoch 158/175\n",
      "2193/2193 [==============================] - 2s 798us/step - loss: 0.4080 - binary_accuracy: 0.8294\n",
      "Epoch 159/175\n",
      "2193/2193 [==============================] - 2s 793us/step - loss: 0.4124 - binary_accuracy: 0.8297\n",
      "Epoch 160/175\n",
      "2193/2193 [==============================] - 2s 809us/step - loss: 0.4111 - binary_accuracy: 0.8316\n",
      "Epoch 161/175\n",
      "2193/2193 [==============================] - 2s 793us/step - loss: 0.4119 - binary_accuracy: 0.8272\n",
      "Epoch 162/175\n",
      "2193/2193 [==============================] - 2s 790us/step - loss: 0.4183 - binary_accuracy: 0.8216\n",
      "Epoch 163/175\n",
      "2193/2193 [==============================] - 2s 805us/step - loss: 0.4106 - binary_accuracy: 0.8318\n",
      "Epoch 164/175\n",
      "2193/2193 [==============================] - 2s 811us/step - loss: 0.4076 - binary_accuracy: 0.8298\n",
      "Epoch 165/175\n",
      "2193/2193 [==============================] - 2s 796us/step - loss: 0.4083 - binary_accuracy: 0.8274\n",
      "Epoch 166/175\n",
      "2193/2193 [==============================] - 2s 802us/step - loss: 0.4125 - binary_accuracy: 0.8265\n",
      "Epoch 167/175\n",
      "2193/2193 [==============================] - 2s 808us/step - loss: 0.4088 - binary_accuracy: 0.8331\n",
      "Epoch 168/175\n",
      "2193/2193 [==============================] - 2s 824us/step - loss: 0.4134 - binary_accuracy: 0.8263\n",
      "Epoch 169/175\n",
      "2193/2193 [==============================] - 2s 816us/step - loss: 0.4057 - binary_accuracy: 0.8344\n",
      "Epoch 170/175\n",
      "2193/2193 [==============================] - 2s 808us/step - loss: 0.4097 - binary_accuracy: 0.8294\n",
      "Epoch 171/175\n",
      "2193/2193 [==============================] - 2s 823us/step - loss: 0.4108 - binary_accuracy: 0.8296\n",
      "Epoch 172/175\n",
      "2193/2193 [==============================] - 2s 807us/step - loss: 0.4045 - binary_accuracy: 0.8329\n",
      "Epoch 173/175\n",
      "2193/2193 [==============================] - 2s 794us/step - loss: 0.4114 - binary_accuracy: 0.8258\n",
      "Epoch 174/175\n",
      "2193/2193 [==============================] - 2s 803us/step - loss: 0.4055 - binary_accuracy: 0.8311\n",
      "Epoch 175/175\n",
      "2193/2193 [==============================] - 2s 803us/step - loss: 0.4085 - binary_accuracy: 0.8287\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "auroc:  0.7261681542974845\n",
      "f1:  0.4394546912590217\n",
      "threshold:  0.16626877\n",
      "precision:  0.3366093366093366\n",
      "recall:  0.6327944572748267\n",
      "accuracy:  0.6938239159001314\n",
      "specificity:  0.7081081081081081\n",
      "CPU times: user 17min 28s, sys: 3min 58s, total: 21min 26s\n",
      "Wall time: 5min 14s\n"
     ]
    }
   ],
   "source": [
    "### A single run w/out cross-validation\n",
    "%%time\n",
    "# with previous outcome added to the features\n",
    "# outcome is added before scaling\n",
    "seed(1)\n",
    "set_random_seed(2)\n",
    "model = build_model(32, 0.5, max_timestamps, n_features)\n",
    "fit_model(model, 175, X_train_pad, y_train_pad, verbose=1)\n",
    "y_predict = model.predict(X_test_pad)\n",
    "y_test_len = [len(y) for y in y_test_list]\n",
    "y_predict = truncate_prediction(y_predict, y_test_len)\n",
    "y_true = np.concatenate(y_test_list)\n",
    "print('auroc: ', auroc(y_predict, y_true))\n",
    "f1, th = find_best_f1(y_predict, y_true)\n",
    "print('f1: ', f1);\n",
    "print('threshold: ', th)\n",
    "precision, recall, accuracy, specificity = other_metrics(y_predict, y_true, th)\n",
    "print('precision: ', precision)\n",
    "print('recall: ', recall)\n",
    "print('accuracy: ', accuracy)\n",
    "print('specificity: ', specificity)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/225\n",
      "1755/1755 [==============================] - 5s 3ms/step - loss: 0.4752 - binary_accuracy: 0.8179\n",
      "Epoch 2/225\n",
      "1755/1755 [==============================] - 2s 855us/step - loss: 0.4523 - binary_accuracy: 0.8196\n",
      "Epoch 3/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4490 - binary_accuracy: 0.8179\n",
      "Epoch 4/225\n",
      "1755/1755 [==============================] - 1s 852us/step - loss: 0.4320 - binary_accuracy: 0.8240\n",
      "Epoch 5/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4404 - binary_accuracy: 0.8219\n",
      "Epoch 6/225\n",
      "1755/1755 [==============================] - 2s 858us/step - loss: 0.4394 - binary_accuracy: 0.8197\n",
      "Epoch 7/225\n",
      "1755/1755 [==============================] - 1s 850us/step - loss: 0.4410 - binary_accuracy: 0.8184\n",
      "Epoch 8/225\n",
      "1755/1755 [==============================] - 1s 846us/step - loss: 0.4391 - binary_accuracy: 0.8180\n",
      "Epoch 9/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4351 - binary_accuracy: 0.8222\n",
      "Epoch 10/225\n",
      "1755/1755 [==============================] - 1s 851us/step - loss: 0.4336 - binary_accuracy: 0.8263\n",
      "Epoch 11/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4291 - binary_accuracy: 0.8247\n",
      "Epoch 12/225\n",
      "1755/1755 [==============================] - 1s 854us/step - loss: 0.4283 - binary_accuracy: 0.8227\n",
      "Epoch 13/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4261 - binary_accuracy: 0.8260\n",
      "Epoch 14/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4286 - binary_accuracy: 0.8209\n",
      "Epoch 15/225\n",
      "1755/1755 [==============================] - 1s 849us/step - loss: 0.4264 - binary_accuracy: 0.8255\n",
      "Epoch 16/225\n",
      "1755/1755 [==============================] - 1s 849us/step - loss: 0.4330 - binary_accuracy: 0.8208\n",
      "Epoch 17/225\n",
      "1755/1755 [==============================] - 1s 853us/step - loss: 0.4316 - binary_accuracy: 0.8179\n",
      "Epoch 18/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4266 - binary_accuracy: 0.8253\n",
      "Epoch 19/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4228 - binary_accuracy: 0.8253\n",
      "Epoch 20/225\n",
      "1755/1755 [==============================] - 1s 847us/step - loss: 0.4224 - binary_accuracy: 0.8246\n",
      "Epoch 21/225\n",
      "1755/1755 [==============================] - 1s 846us/step - loss: 0.4260 - binary_accuracy: 0.8247\n",
      "Epoch 22/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4234 - binary_accuracy: 0.8266\n",
      "Epoch 23/225\n",
      "1755/1755 [==============================] - 1s 846us/step - loss: 0.4225 - binary_accuracy: 0.8266\n",
      "Epoch 24/225\n",
      "1755/1755 [==============================] - 1s 844us/step - loss: 0.4232 - binary_accuracy: 0.8248\n",
      "Epoch 25/225\n",
      "1755/1755 [==============================] - 1s 849us/step - loss: 0.4255 - binary_accuracy: 0.8242\n",
      "Epoch 26/225\n",
      "1755/1755 [==============================] - 1s 847us/step - loss: 0.4200 - binary_accuracy: 0.8272\n",
      "Epoch 27/225\n",
      "1755/1755 [==============================] - 1s 846us/step - loss: 0.4251 - binary_accuracy: 0.8232\n",
      "Epoch 28/225\n",
      "1755/1755 [==============================] - 1s 847us/step - loss: 0.4226 - binary_accuracy: 0.8258\n",
      "Epoch 29/225\n",
      "1755/1755 [==============================] - 1s 849us/step - loss: 0.4199 - binary_accuracy: 0.8287\n",
      "Epoch 30/225\n",
      "1755/1755 [==============================] - 1s 849us/step - loss: 0.4233 - binary_accuracy: 0.8273\n",
      "Epoch 31/225\n",
      "1755/1755 [==============================] - 1s 847us/step - loss: 0.4173 - binary_accuracy: 0.8252\n",
      "Epoch 32/225\n",
      "1755/1755 [==============================] - 1s 852us/step - loss: 0.4249 - binary_accuracy: 0.8231\n",
      "Epoch 33/225\n",
      "1755/1755 [==============================] - 1s 852us/step - loss: 0.4214 - binary_accuracy: 0.8228\n",
      "Epoch 34/225\n",
      "1755/1755 [==============================] - 1s 851us/step - loss: 0.4244 - binary_accuracy: 0.8259\n",
      "Epoch 35/225\n",
      "1755/1755 [==============================] - 1s 851us/step - loss: 0.4244 - binary_accuracy: 0.8257\n",
      "Epoch 36/225\n",
      "1755/1755 [==============================] - 2s 856us/step - loss: 0.4241 - binary_accuracy: 0.8237\n",
      "Epoch 37/225\n",
      "1755/1755 [==============================] - 1s 854us/step - loss: 0.4207 - binary_accuracy: 0.8271\n",
      "Epoch 38/225\n",
      "1755/1755 [==============================] - 2s 876us/step - loss: 0.4253 - binary_accuracy: 0.8232\n",
      "Epoch 39/225\n",
      "1755/1755 [==============================] - 2s 858us/step - loss: 0.4241 - binary_accuracy: 0.8187\n",
      "Epoch 40/225\n",
      "1755/1755 [==============================] - 2s 859us/step - loss: 0.4240 - binary_accuracy: 0.8242\n",
      "Epoch 41/225\n",
      "1755/1755 [==============================] - 1s 853us/step - loss: 0.4168 - binary_accuracy: 0.8277\n",
      "Epoch 42/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4208 - binary_accuracy: 0.8279\n",
      "Epoch 43/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4173 - binary_accuracy: 0.8253\n",
      "Epoch 44/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4176 - binary_accuracy: 0.8266\n",
      "Epoch 45/225\n",
      "1755/1755 [==============================] - 1s 846us/step - loss: 0.4183 - binary_accuracy: 0.8224\n",
      "Epoch 46/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4209 - binary_accuracy: 0.8262\n",
      "Epoch 47/225\n",
      "1755/1755 [==============================] - 1s 849us/step - loss: 0.4197 - binary_accuracy: 0.8273\n",
      "Epoch 48/225\n",
      "1755/1755 [==============================] - 1s 841us/step - loss: 0.4194 - binary_accuracy: 0.8245\n",
      "Epoch 49/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4184 - binary_accuracy: 0.8267\n",
      "Epoch 50/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4216 - binary_accuracy: 0.8217\n",
      "Epoch 51/225\n",
      "1755/1755 [==============================] - 1s 832us/step - loss: 0.4122 - binary_accuracy: 0.8284\n",
      "Epoch 52/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4109 - binary_accuracy: 0.8304\n",
      "Epoch 53/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4206 - binary_accuracy: 0.8297\n",
      "Epoch 54/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4142 - binary_accuracy: 0.8299\n",
      "Epoch 55/225\n",
      "1755/1755 [==============================] - 2s 860us/step - loss: 0.4147 - binary_accuracy: 0.8255\n",
      "Epoch 56/225\n",
      "1755/1755 [==============================] - 1s 850us/step - loss: 0.4120 - binary_accuracy: 0.8292\n",
      "Epoch 57/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4153 - binary_accuracy: 0.8314\n",
      "Epoch 58/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4126 - binary_accuracy: 0.8258\n",
      "Epoch 59/225\n",
      "1755/1755 [==============================] - 2s 876us/step - loss: 0.4225 - binary_accuracy: 0.8243\n",
      "Epoch 60/225\n",
      "1755/1755 [==============================] - 1s 846us/step - loss: 0.4142 - binary_accuracy: 0.8327\n",
      "Epoch 61/225\n",
      "1755/1755 [==============================] - 1s 851us/step - loss: 0.4121 - binary_accuracy: 0.8272\n",
      "Epoch 62/225\n",
      "1755/1755 [==============================] - 1s 852us/step - loss: 0.4115 - binary_accuracy: 0.8300\n",
      "Epoch 63/225\n",
      "1755/1755 [==============================] - 1s 850us/step - loss: 0.4185 - binary_accuracy: 0.8228\n",
      "Epoch 64/225\n",
      "1755/1755 [==============================] - 2s 855us/step - loss: 0.4118 - binary_accuracy: 0.8252\n",
      "Epoch 65/225\n",
      "1755/1755 [==============================] - 1s 853us/step - loss: 0.4192 - binary_accuracy: 0.8275\n",
      "Epoch 66/225\n",
      "1755/1755 [==============================] - 1s 846us/step - loss: 0.4186 - binary_accuracy: 0.8256\n",
      "Epoch 67/225\n",
      "1755/1755 [==============================] - 1s 850us/step - loss: 0.4116 - binary_accuracy: 0.8264\n",
      "Epoch 68/225\n",
      "1755/1755 [==============================] - 1s 849us/step - loss: 0.4153 - binary_accuracy: 0.8257\n",
      "Epoch 69/225\n",
      "1755/1755 [==============================] - 2s 856us/step - loss: 0.4119 - binary_accuracy: 0.8281\n",
      "Epoch 70/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4199 - binary_accuracy: 0.8239\n",
      "Epoch 71/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4186 - binary_accuracy: 0.8283\n",
      "Epoch 72/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4144 - binary_accuracy: 0.8293\n",
      "Epoch 73/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4106 - binary_accuracy: 0.8276\n",
      "Epoch 74/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4124 - binary_accuracy: 0.8306\n",
      "Epoch 75/225\n",
      "1755/1755 [==============================] - 1s 846us/step - loss: 0.4168 - binary_accuracy: 0.8285\n",
      "Epoch 76/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4129 - binary_accuracy: 0.8259\n",
      "Epoch 77/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4144 - binary_accuracy: 0.8302\n",
      "Epoch 78/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4035 - binary_accuracy: 0.8298\n",
      "Epoch 79/225\n",
      "1755/1755 [==============================] - 1s 840us/step - loss: 0.4077 - binary_accuracy: 0.8329\n",
      "Epoch 80/225\n",
      "1755/1755 [==============================] - 1s 846us/step - loss: 0.4116 - binary_accuracy: 0.8285\n",
      "Epoch 81/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4079 - binary_accuracy: 0.8304\n",
      "Epoch 82/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4171 - binary_accuracy: 0.8214\n",
      "Epoch 83/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4084 - binary_accuracy: 0.8307\n",
      "Epoch 84/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4204 - binary_accuracy: 0.8263\n",
      "Epoch 85/225\n",
      "1755/1755 [==============================] - 1s 835us/step - loss: 0.4174 - binary_accuracy: 0.8236\n",
      "Epoch 86/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4082 - binary_accuracy: 0.8297\n",
      "Epoch 87/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4138 - binary_accuracy: 0.8294\n",
      "Epoch 88/225\n",
      "1755/1755 [==============================] - 1s 844us/step - loss: 0.4140 - binary_accuracy: 0.8240\n",
      "Epoch 89/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4050 - binary_accuracy: 0.8342\n",
      "Epoch 90/225\n",
      "1755/1755 [==============================] - 2s 856us/step - loss: 0.4120 - binary_accuracy: 0.8312\n",
      "Epoch 91/225\n",
      "1755/1755 [==============================] - 1s 851us/step - loss: 0.4115 - binary_accuracy: 0.8284\n",
      "Epoch 92/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4118 - binary_accuracy: 0.8318\n",
      "Epoch 93/225\n",
      "1755/1755 [==============================] - 1s 844us/step - loss: 0.4126 - binary_accuracy: 0.8325\n",
      "Epoch 94/225\n",
      "1755/1755 [==============================] - 1s 854us/step - loss: 0.4168 - binary_accuracy: 0.8274\n",
      "Epoch 95/225\n",
      "1755/1755 [==============================] - 1s 846us/step - loss: 0.4105 - binary_accuracy: 0.8298\n",
      "Epoch 96/225\n",
      "1755/1755 [==============================] - 1s 846us/step - loss: 0.4167 - binary_accuracy: 0.8263\n",
      "Epoch 97/225\n",
      "1755/1755 [==============================] - 1s 850us/step - loss: 0.4095 - binary_accuracy: 0.8253\n",
      "Epoch 98/225\n",
      "1755/1755 [==============================] - 1s 851us/step - loss: 0.4118 - binary_accuracy: 0.8286\n",
      "Epoch 99/225\n",
      "1755/1755 [==============================] - 1s 852us/step - loss: 0.4095 - binary_accuracy: 0.8280\n",
      "Epoch 100/225\n",
      "1755/1755 [==============================] - 2s 858us/step - loss: 0.4110 - binary_accuracy: 0.8290\n",
      "Epoch 101/225\n",
      "1755/1755 [==============================] - 1s 844us/step - loss: 0.4137 - binary_accuracy: 0.8305\n",
      "Epoch 102/225\n",
      "1755/1755 [==============================] - 1s 847us/step - loss: 0.4108 - binary_accuracy: 0.8260\n",
      "Epoch 103/225\n",
      "1755/1755 [==============================] - 1s 854us/step - loss: 0.4150 - binary_accuracy: 0.8220\n",
      "Epoch 104/225\n",
      "1755/1755 [==============================] - 1s 850us/step - loss: 0.4175 - binary_accuracy: 0.8247\n",
      "Epoch 105/225\n",
      "1755/1755 [==============================] - 1s 846us/step - loss: 0.4058 - binary_accuracy: 0.8278\n",
      "Epoch 106/225\n",
      "1755/1755 [==============================] - 1s 852us/step - loss: 0.4135 - binary_accuracy: 0.8246\n",
      "Epoch 107/225\n",
      "1755/1755 [==============================] - 1s 847us/step - loss: 0.4111 - binary_accuracy: 0.8269\n",
      "Epoch 108/225\n",
      "1755/1755 [==============================] - 2s 856us/step - loss: 0.4161 - binary_accuracy: 0.8234\n",
      "Epoch 109/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4080 - binary_accuracy: 0.8300\n",
      "Epoch 110/225\n",
      "1755/1755 [==============================] - 1s 853us/step - loss: 0.4053 - binary_accuracy: 0.8277\n",
      "Epoch 111/225\n",
      "1755/1755 [==============================] - 2s 855us/step - loss: 0.4111 - binary_accuracy: 0.8263\n",
      "Epoch 112/225\n",
      "1755/1755 [==============================] - 2s 884us/step - loss: 0.4126 - binary_accuracy: 0.8321\n",
      "Epoch 113/225\n",
      "1755/1755 [==============================] - 2s 882us/step - loss: 0.4117 - binary_accuracy: 0.8278\n",
      "Epoch 114/225\n",
      "1755/1755 [==============================] - 2s 877us/step - loss: 0.4110 - binary_accuracy: 0.8329\n",
      "Epoch 115/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4078 - binary_accuracy: 0.8265\n",
      "Epoch 116/225\n",
      "1755/1755 [==============================] - 1s 851us/step - loss: 0.4115 - binary_accuracy: 0.8290\n",
      "Epoch 117/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4123 - binary_accuracy: 0.8275\n",
      "Epoch 118/225\n",
      "1755/1755 [==============================] - 2s 878us/step - loss: 0.4146 - binary_accuracy: 0.8278\n",
      "Epoch 119/225\n",
      "1755/1755 [==============================] - 2s 880us/step - loss: 0.4138 - binary_accuracy: 0.8263\n",
      "Epoch 120/225\n",
      "1755/1755 [==============================] - 2s 879us/step - loss: 0.4129 - binary_accuracy: 0.8276\n",
      "Epoch 121/225\n",
      "1755/1755 [==============================] - 2s 884us/step - loss: 0.4093 - binary_accuracy: 0.8282\n",
      "Epoch 122/225\n",
      "1755/1755 [==============================] - 2s 881us/step - loss: 0.4128 - binary_accuracy: 0.8276\n",
      "Epoch 123/225\n",
      "1755/1755 [==============================] - 1s 855us/step - loss: 0.4051 - binary_accuracy: 0.8328\n",
      "Epoch 124/225\n",
      "1755/1755 [==============================] - 2s 874us/step - loss: 0.4145 - binary_accuracy: 0.8254\n",
      "Epoch 125/225\n",
      "1755/1755 [==============================] - 2s 871us/step - loss: 0.4071 - binary_accuracy: 0.8275\n",
      "Epoch 126/225\n",
      "1755/1755 [==============================] - 1s 852us/step - loss: 0.4174 - binary_accuracy: 0.8277\n",
      "Epoch 127/225\n",
      "1755/1755 [==============================] - 2s 862us/step - loss: 0.4099 - binary_accuracy: 0.8300\n",
      "Epoch 128/225\n",
      "1755/1755 [==============================] - 2s 867us/step - loss: 0.4064 - binary_accuracy: 0.8305\n",
      "Epoch 129/225\n",
      "1755/1755 [==============================] - 2s 862us/step - loss: 0.4169 - binary_accuracy: 0.8248\n",
      "Epoch 130/225\n",
      "1755/1755 [==============================] - 2s 865us/step - loss: 0.4133 - binary_accuracy: 0.8246\n",
      "Epoch 131/225\n",
      "1755/1755 [==============================] - 2s 859us/step - loss: 0.4149 - binary_accuracy: 0.8248\n",
      "Epoch 132/225\n",
      "1755/1755 [==============================] - 2s 867us/step - loss: 0.4066 - binary_accuracy: 0.8338\n",
      "Epoch 133/225\n",
      "1755/1755 [==============================] - 2s 877us/step - loss: 0.4011 - binary_accuracy: 0.8322\n",
      "Epoch 134/225\n",
      "1755/1755 [==============================] - 2s 910us/step - loss: 0.4105 - binary_accuracy: 0.8282\n",
      "Epoch 135/225\n",
      "1755/1755 [==============================] - 1s 850us/step - loss: 0.4130 - binary_accuracy: 0.8277\n",
      "Epoch 136/225\n",
      "1755/1755 [==============================] - 1s 852us/step - loss: 0.4044 - binary_accuracy: 0.8322\n",
      "Epoch 137/225\n",
      "1755/1755 [==============================] - 2s 939us/step - loss: 0.4079 - binary_accuracy: 0.8358\n",
      "Epoch 138/225\n",
      "1755/1755 [==============================] - 2s 919us/step - loss: 0.4059 - binary_accuracy: 0.8303\n",
      "Epoch 139/225\n",
      "1755/1755 [==============================] - 1s 850us/step - loss: 0.4085 - binary_accuracy: 0.8261\n",
      "Epoch 140/225\n",
      "1755/1755 [==============================] - 2s 862us/step - loss: 0.4115 - binary_accuracy: 0.8271\n",
      "Epoch 141/225\n",
      "1755/1755 [==============================] - 2s 862us/step - loss: 0.4138 - binary_accuracy: 0.8264\n",
      "Epoch 142/225\n",
      "1755/1755 [==============================] - 2s 861us/step - loss: 0.4110 - binary_accuracy: 0.8275\n",
      "Epoch 143/225\n",
      "1755/1755 [==============================] - 2s 863us/step - loss: 0.4031 - binary_accuracy: 0.8290\n",
      "Epoch 144/225\n",
      "1755/1755 [==============================] - 2s 863us/step - loss: 0.4046 - binary_accuracy: 0.8327\n",
      "Epoch 145/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1755/1755 [==============================] - 2s 857us/step - loss: 0.4118 - binary_accuracy: 0.8286\n",
      "Epoch 146/225\n",
      "1755/1755 [==============================] - 1s 850us/step - loss: 0.3977 - binary_accuracy: 0.8343\n",
      "Epoch 147/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4163 - binary_accuracy: 0.8287\n",
      "Epoch 148/225\n",
      "1755/1755 [==============================] - 1s 850us/step - loss: 0.4041 - binary_accuracy: 0.8325\n",
      "Epoch 149/225\n",
      "1755/1755 [==============================] - 1s 850us/step - loss: 0.4159 - binary_accuracy: 0.8307\n",
      "Epoch 150/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4074 - binary_accuracy: 0.8326\n",
      "Epoch 151/225\n",
      "1755/1755 [==============================] - 1s 850us/step - loss: 0.4110 - binary_accuracy: 0.8256\n",
      "Epoch 152/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4092 - binary_accuracy: 0.8307\n",
      "Epoch 153/225\n",
      "1755/1755 [==============================] - 1s 847us/step - loss: 0.4099 - binary_accuracy: 0.8284\n",
      "Epoch 154/225\n",
      "1755/1755 [==============================] - 1s 849us/step - loss: 0.4112 - binary_accuracy: 0.8259\n",
      "Epoch 155/225\n",
      "1755/1755 [==============================] - 1s 850us/step - loss: 0.4138 - binary_accuracy: 0.8283\n",
      "Epoch 156/225\n",
      "1755/1755 [==============================] - 1s 851us/step - loss: 0.4121 - binary_accuracy: 0.8270\n",
      "Epoch 157/225\n",
      "1755/1755 [==============================] - 1s 851us/step - loss: 0.4053 - binary_accuracy: 0.8311\n",
      "Epoch 158/225\n",
      "1755/1755 [==============================] - 1s 853us/step - loss: 0.4094 - binary_accuracy: 0.8274\n",
      "Epoch 159/225\n",
      "1755/1755 [==============================] - 1s 840us/step - loss: 0.4044 - binary_accuracy: 0.8289\n",
      "Epoch 160/225\n",
      "1755/1755 [==============================] - 1s 852us/step - loss: 0.4069 - binary_accuracy: 0.8283\n",
      "Epoch 161/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4077 - binary_accuracy: 0.8269\n",
      "Epoch 162/225\n",
      "1755/1755 [==============================] - 1s 846us/step - loss: 0.4009 - binary_accuracy: 0.8332\n",
      "Epoch 163/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4082 - binary_accuracy: 0.8281\n",
      "Epoch 164/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4080 - binary_accuracy: 0.8319\n",
      "Epoch 165/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4055 - binary_accuracy: 0.8327\n",
      "Epoch 166/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4080 - binary_accuracy: 0.8328\n",
      "Epoch 167/225\n",
      "1755/1755 [==============================] - 1s 840us/step - loss: 0.4029 - binary_accuracy: 0.8311\n",
      "Epoch 168/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4057 - binary_accuracy: 0.8282\n",
      "Epoch 169/225\n",
      "1755/1755 [==============================] - 1s 853us/step - loss: 0.4111 - binary_accuracy: 0.8264\n",
      "Epoch 170/225\n",
      "1755/1755 [==============================] - 1s 847us/step - loss: 0.4084 - binary_accuracy: 0.8311\n",
      "Epoch 171/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4069 - binary_accuracy: 0.8324\n",
      "Epoch 172/225\n",
      "1755/1755 [==============================] - 1s 849us/step - loss: 0.4093 - binary_accuracy: 0.8290\n",
      "Epoch 173/225\n",
      "1755/1755 [==============================] - 2s 857us/step - loss: 0.4047 - binary_accuracy: 0.8312\n",
      "Epoch 174/225\n",
      "1755/1755 [==============================] - 1s 847us/step - loss: 0.4072 - binary_accuracy: 0.8337\n",
      "Epoch 175/225\n",
      "1755/1755 [==============================] - 1s 836us/step - loss: 0.4013 - binary_accuracy: 0.8344\n",
      "Epoch 176/225\n",
      "1755/1755 [==============================] - 1s 852us/step - loss: 0.4063 - binary_accuracy: 0.8273\n",
      "Epoch 177/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4014 - binary_accuracy: 0.8314\n",
      "Epoch 178/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4059 - binary_accuracy: 0.8296\n",
      "Epoch 179/225\n",
      "1755/1755 [==============================] - 1s 846us/step - loss: 0.4048 - binary_accuracy: 0.8312\n",
      "Epoch 180/225\n",
      "1755/1755 [==============================] - 1s 851us/step - loss: 0.3982 - binary_accuracy: 0.8349\n",
      "Epoch 181/225\n",
      "1755/1755 [==============================] - 2s 856us/step - loss: 0.4030 - binary_accuracy: 0.8327\n",
      "Epoch 182/225\n",
      "1755/1755 [==============================] - 1s 841us/step - loss: 0.4106 - binary_accuracy: 0.8241\n",
      "Epoch 183/225\n",
      "1755/1755 [==============================] - 2s 857us/step - loss: 0.4097 - binary_accuracy: 0.8293\n",
      "Epoch 184/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4034 - binary_accuracy: 0.8330\n",
      "Epoch 185/225\n",
      "1755/1755 [==============================] - 1s 851us/step - loss: 0.4017 - binary_accuracy: 0.8316\n",
      "Epoch 186/225\n",
      "1755/1755 [==============================] - 1s 852us/step - loss: 0.4091 - binary_accuracy: 0.8243\n",
      "Epoch 187/225\n",
      "1755/1755 [==============================] - 2s 861us/step - loss: 0.4079 - binary_accuracy: 0.8306\n",
      "Epoch 188/225\n",
      "1755/1755 [==============================] - 1s 840us/step - loss: 0.4054 - binary_accuracy: 0.8296\n",
      "Epoch 189/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4030 - binary_accuracy: 0.8327\n",
      "Epoch 190/225\n",
      "1755/1755 [==============================] - 1s 844us/step - loss: 0.4073 - binary_accuracy: 0.8273\n",
      "Epoch 191/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4020 - binary_accuracy: 0.8310\n",
      "Epoch 192/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4040 - binary_accuracy: 0.8356\n",
      "Epoch 193/225\n",
      "1755/1755 [==============================] - 1s 841us/step - loss: 0.4030 - binary_accuracy: 0.8341\n",
      "Epoch 194/225\n",
      "1755/1755 [==============================] - 1s 833us/step - loss: 0.4119 - binary_accuracy: 0.8274\n",
      "Epoch 195/225\n",
      "1755/1755 [==============================] - 1s 835us/step - loss: 0.4090 - binary_accuracy: 0.8305\n",
      "Epoch 196/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4041 - binary_accuracy: 0.8333\n",
      "Epoch 197/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4030 - binary_accuracy: 0.8317\n",
      "Epoch 198/225\n",
      "1755/1755 [==============================] - 1s 847us/step - loss: 0.4014 - binary_accuracy: 0.8294\n",
      "Epoch 199/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4026 - binary_accuracy: 0.8337\n",
      "Epoch 200/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4013 - binary_accuracy: 0.8354\n",
      "Epoch 201/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4016 - binary_accuracy: 0.8311\n",
      "Epoch 202/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4085 - binary_accuracy: 0.8303\n",
      "Epoch 203/225\n",
      "1755/1755 [==============================] - 2s 855us/step - loss: 0.4067 - binary_accuracy: 0.8298\n",
      "Epoch 204/225\n",
      "1755/1755 [==============================] - 2s 858us/step - loss: 0.4083 - binary_accuracy: 0.8319\n",
      "Epoch 205/225\n",
      "1755/1755 [==============================] - 1s 846us/step - loss: 0.4095 - binary_accuracy: 0.8251\n",
      "Epoch 206/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4061 - binary_accuracy: 0.8334\n",
      "Epoch 207/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4034 - binary_accuracy: 0.8296\n",
      "Epoch 208/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4074 - binary_accuracy: 0.8276\n",
      "Epoch 209/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4088 - binary_accuracy: 0.8245\n",
      "Epoch 210/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4012 - binary_accuracy: 0.8296\n",
      "Epoch 211/225\n",
      "1755/1755 [==============================] - 1s 844us/step - loss: 0.4082 - binary_accuracy: 0.8333\n",
      "Epoch 212/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4086 - binary_accuracy: 0.8253\n",
      "Epoch 213/225\n",
      "1755/1755 [==============================] - 1s 840us/step - loss: 0.4082 - binary_accuracy: 0.8300\n",
      "Epoch 214/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4044 - binary_accuracy: 0.8282\n",
      "Epoch 215/225\n",
      "1755/1755 [==============================] - 1s 844us/step - loss: 0.4099 - binary_accuracy: 0.8309\n",
      "Epoch 216/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4012 - binary_accuracy: 0.8323\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 217/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4067 - binary_accuracy: 0.8312\n",
      "Epoch 218/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4061 - binary_accuracy: 0.8346\n",
      "Epoch 219/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4045 - binary_accuracy: 0.8311\n",
      "Epoch 220/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4079 - binary_accuracy: 0.8286\n",
      "Epoch 221/225\n",
      "1755/1755 [==============================] - 1s 852us/step - loss: 0.3984 - binary_accuracy: 0.8373\n",
      "Epoch 222/225\n",
      "1755/1755 [==============================] - 1s 841us/step - loss: 0.4055 - binary_accuracy: 0.8277\n",
      "Epoch 223/225\n",
      "1755/1755 [==============================] - 1s 850us/step - loss: 0.4015 - binary_accuracy: 0.8289\n",
      "Epoch 224/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.3993 - binary_accuracy: 0.8353\n",
      "Epoch 225/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4035 - binary_accuracy: 0.8287\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/225\n",
      "1755/1755 [==============================] - 5s 3ms/step - loss: 0.4745 - binary_accuracy: 0.8230\n",
      "Epoch 2/225\n",
      "1755/1755 [==============================] - 1s 833us/step - loss: 0.4380 - binary_accuracy: 0.8261\n",
      "Epoch 3/225\n",
      "1755/1755 [==============================] - 1s 835us/step - loss: 0.4413 - binary_accuracy: 0.8197\n",
      "Epoch 4/225\n",
      "1755/1755 [==============================] - 1s 832us/step - loss: 0.4382 - binary_accuracy: 0.8271\n",
      "Epoch 5/225\n",
      "1755/1755 [==============================] - 1s 834us/step - loss: 0.4343 - binary_accuracy: 0.8240\n",
      "Epoch 6/225\n",
      "1755/1755 [==============================] - 1s 832us/step - loss: 0.4336 - binary_accuracy: 0.8266\n",
      "Epoch 7/225\n",
      "1755/1755 [==============================] - 1s 823us/step - loss: 0.4423 - binary_accuracy: 0.8225\n",
      "Epoch 8/225\n",
      "1755/1755 [==============================] - 1s 833us/step - loss: 0.4409 - binary_accuracy: 0.8212\n",
      "Epoch 9/225\n",
      "1755/1755 [==============================] - 1s 834us/step - loss: 0.4227 - binary_accuracy: 0.8288\n",
      "Epoch 10/225\n",
      "1755/1755 [==============================] - 1s 831us/step - loss: 0.4277 - binary_accuracy: 0.8302\n",
      "Epoch 11/225\n",
      "1755/1755 [==============================] - 1s 844us/step - loss: 0.4318 - binary_accuracy: 0.8254\n",
      "Epoch 12/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4306 - binary_accuracy: 0.8257\n",
      "Epoch 13/225\n",
      "1755/1755 [==============================] - 1s 828us/step - loss: 0.4291 - binary_accuracy: 0.8254\n",
      "Epoch 14/225\n",
      "1755/1755 [==============================] - 1s 829us/step - loss: 0.4319 - binary_accuracy: 0.8223\n",
      "Epoch 15/225\n",
      "1755/1755 [==============================] - 1s 835us/step - loss: 0.4286 - binary_accuracy: 0.8263\n",
      "Epoch 16/225\n",
      "1755/1755 [==============================] - 1s 836us/step - loss: 0.4277 - binary_accuracy: 0.8255\n",
      "Epoch 17/225\n",
      "1755/1755 [==============================] - 1s 832us/step - loss: 0.4261 - binary_accuracy: 0.8266\n",
      "Epoch 18/225\n",
      "1755/1755 [==============================] - 1s 835us/step - loss: 0.4301 - binary_accuracy: 0.8226\n",
      "Epoch 19/225\n",
      "1755/1755 [==============================] - 1s 854us/step - loss: 0.4205 - binary_accuracy: 0.8341\n",
      "Epoch 20/225\n",
      "1755/1755 [==============================] - 1s 834us/step - loss: 0.4229 - binary_accuracy: 0.8283\n",
      "Epoch 21/225\n",
      "1755/1755 [==============================] - 1s 832us/step - loss: 0.4247 - binary_accuracy: 0.8237\n",
      "Epoch 22/225\n",
      "1755/1755 [==============================] - 1s 832us/step - loss: 0.4240 - binary_accuracy: 0.8265\n",
      "Epoch 23/225\n",
      "1755/1755 [==============================] - 1s 832us/step - loss: 0.4267 - binary_accuracy: 0.8237\n",
      "Epoch 24/225\n",
      "1755/1755 [==============================] - 1s 833us/step - loss: 0.4175 - binary_accuracy: 0.8325\n",
      "Epoch 25/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4273 - binary_accuracy: 0.8228\n",
      "Epoch 26/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4237 - binary_accuracy: 0.8274\n",
      "Epoch 27/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4249 - binary_accuracy: 0.8266\n",
      "Epoch 28/225\n",
      "1755/1755 [==============================] - 1s 841us/step - loss: 0.4214 - binary_accuracy: 0.8302\n",
      "Epoch 29/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4265 - binary_accuracy: 0.8259\n",
      "Epoch 30/225\n",
      "1755/1755 [==============================] - 1s 844us/step - loss: 0.4291 - binary_accuracy: 0.8252\n",
      "Epoch 31/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4200 - binary_accuracy: 0.8300\n",
      "Epoch 32/225\n",
      "1755/1755 [==============================] - 1s 847us/step - loss: 0.4197 - binary_accuracy: 0.8241\n",
      "Epoch 33/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4280 - binary_accuracy: 0.8253\n",
      "Epoch 34/225\n",
      "1755/1755 [==============================] - 1s 840us/step - loss: 0.4195 - binary_accuracy: 0.8242\n",
      "Epoch 35/225\n",
      "1755/1755 [==============================] - 1s 834us/step - loss: 0.4222 - binary_accuracy: 0.8240\n",
      "Epoch 36/225\n",
      "1755/1755 [==============================] - 1s 840us/step - loss: 0.4204 - binary_accuracy: 0.8250\n",
      "Epoch 37/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4204 - binary_accuracy: 0.8303\n",
      "Epoch 38/225\n",
      "1755/1755 [==============================] - 1s 835us/step - loss: 0.4171 - binary_accuracy: 0.8331\n",
      "Epoch 39/225\n",
      "1755/1755 [==============================] - 1s 840us/step - loss: 0.4159 - binary_accuracy: 0.8307\n",
      "Epoch 40/225\n",
      "1755/1755 [==============================] - 1s 836us/step - loss: 0.4161 - binary_accuracy: 0.8312\n",
      "Epoch 41/225\n",
      "1755/1755 [==============================] - 1s 832us/step - loss: 0.4180 - binary_accuracy: 0.8260\n",
      "Epoch 42/225\n",
      "1755/1755 [==============================] - 1s 840us/step - loss: 0.4229 - binary_accuracy: 0.8291\n",
      "Epoch 43/225\n",
      "1755/1755 [==============================] - 1s 840us/step - loss: 0.4220 - binary_accuracy: 0.8253\n",
      "Epoch 44/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4167 - binary_accuracy: 0.8332\n",
      "Epoch 45/225\n",
      "1755/1755 [==============================] - 1s 836us/step - loss: 0.4202 - binary_accuracy: 0.8293\n",
      "Epoch 46/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4184 - binary_accuracy: 0.8277\n",
      "Epoch 47/225\n",
      "1755/1755 [==============================] - 1s 833us/step - loss: 0.4206 - binary_accuracy: 0.8284\n",
      "Epoch 48/225\n",
      "1755/1755 [==============================] - 2s 856us/step - loss: 0.4148 - binary_accuracy: 0.8352\n",
      "Epoch 49/225\n",
      "1755/1755 [==============================] - 1s 835us/step - loss: 0.4199 - binary_accuracy: 0.8247\n",
      "Epoch 50/225\n",
      "1755/1755 [==============================] - 1s 836us/step - loss: 0.4155 - binary_accuracy: 0.8293\n",
      "Epoch 51/225\n",
      "1755/1755 [==============================] - 1s 841us/step - loss: 0.4138 - binary_accuracy: 0.8313\n",
      "Epoch 52/225\n",
      "1755/1755 [==============================] - 1s 835us/step - loss: 0.4186 - binary_accuracy: 0.8269\n",
      "Epoch 53/225\n",
      "1755/1755 [==============================] - 1s 841us/step - loss: 0.4212 - binary_accuracy: 0.8253\n",
      "Epoch 54/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4178 - binary_accuracy: 0.8288\n",
      "Epoch 55/225\n",
      "1755/1755 [==============================] - 1s 834us/step - loss: 0.4233 - binary_accuracy: 0.8277\n",
      "Epoch 56/225\n",
      "1755/1755 [==============================] - 1s 836us/step - loss: 0.4178 - binary_accuracy: 0.8317\n",
      "Epoch 57/225\n",
      "1755/1755 [==============================] - 1s 841us/step - loss: 0.4177 - binary_accuracy: 0.8249\n",
      "Epoch 58/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4157 - binary_accuracy: 0.8286\n",
      "Epoch 59/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4159 - binary_accuracy: 0.8307\n",
      "Epoch 60/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4139 - binary_accuracy: 0.8278\n",
      "Epoch 61/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4144 - binary_accuracy: 0.8344\n",
      "Epoch 62/225\n",
      "1755/1755 [==============================] - 1s 830us/step - loss: 0.4184 - binary_accuracy: 0.8325\n",
      "Epoch 63/225\n",
      "1755/1755 [==============================] - 1s 833us/step - loss: 0.4239 - binary_accuracy: 0.8254\n",
      "Epoch 64/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4125 - binary_accuracy: 0.8331\n",
      "Epoch 65/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4161 - binary_accuracy: 0.8286\n",
      "Epoch 66/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4198 - binary_accuracy: 0.8269\n",
      "Epoch 67/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4185 - binary_accuracy: 0.8253\n",
      "Epoch 68/225\n",
      "1755/1755 [==============================] - 1s 834us/step - loss: 0.4129 - binary_accuracy: 0.8324\n",
      "Epoch 69/225\n",
      "1755/1755 [==============================] - 1s 835us/step - loss: 0.4139 - binary_accuracy: 0.8272\n",
      "Epoch 70/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4172 - binary_accuracy: 0.8285\n",
      "Epoch 71/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4151 - binary_accuracy: 0.8324\n",
      "Epoch 72/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4128 - binary_accuracy: 0.8321\n",
      "Epoch 73/225\n",
      "1755/1755 [==============================] - 1s 840us/step - loss: 0.4152 - binary_accuracy: 0.8326\n",
      "Epoch 74/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4186 - binary_accuracy: 0.8295\n",
      "Epoch 75/225\n",
      "1755/1755 [==============================] - 1s 844us/step - loss: 0.4181 - binary_accuracy: 0.8322\n",
      "Epoch 76/225\n",
      "1755/1755 [==============================] - 1s 836us/step - loss: 0.4165 - binary_accuracy: 0.8264\n",
      "Epoch 77/225\n",
      "1755/1755 [==============================] - 1s 836us/step - loss: 0.4113 - binary_accuracy: 0.8316\n",
      "Epoch 78/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4168 - binary_accuracy: 0.8265\n",
      "Epoch 79/225\n",
      "1755/1755 [==============================] - 1s 846us/step - loss: 0.4199 - binary_accuracy: 0.8248\n",
      "Epoch 80/225\n",
      "1755/1755 [==============================] - 1s 835us/step - loss: 0.4119 - binary_accuracy: 0.8325\n",
      "Epoch 81/225\n",
      "1755/1755 [==============================] - 1s 844us/step - loss: 0.4138 - binary_accuracy: 0.8300\n",
      "Epoch 82/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4185 - binary_accuracy: 0.8260\n",
      "Epoch 83/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4148 - binary_accuracy: 0.8289\n",
      "Epoch 84/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4150 - binary_accuracy: 0.8284\n",
      "Epoch 85/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4151 - binary_accuracy: 0.8304\n",
      "Epoch 86/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4151 - binary_accuracy: 0.8287\n",
      "Epoch 87/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4176 - binary_accuracy: 0.8313\n",
      "Epoch 88/225\n",
      "1755/1755 [==============================] - 1s 840us/step - loss: 0.4125 - binary_accuracy: 0.8303\n",
      "Epoch 89/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4114 - binary_accuracy: 0.8318\n",
      "Epoch 90/225\n",
      "1755/1755 [==============================] - 1s 847us/step - loss: 0.4124 - binary_accuracy: 0.8285\n",
      "Epoch 91/225\n",
      "1755/1755 [==============================] - 1s 841us/step - loss: 0.4126 - binary_accuracy: 0.8313\n",
      "Epoch 92/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4114 - binary_accuracy: 0.8313\n",
      "Epoch 93/225\n",
      "1755/1755 [==============================] - 1s 846us/step - loss: 0.4124 - binary_accuracy: 0.8311\n",
      "Epoch 94/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4220 - binary_accuracy: 0.8257\n",
      "Epoch 95/225\n",
      "1755/1755 [==============================] - 1s 841us/step - loss: 0.4089 - binary_accuracy: 0.8324\n",
      "Epoch 96/225\n",
      "1755/1755 [==============================] - 1s 841us/step - loss: 0.4092 - binary_accuracy: 0.8331\n",
      "Epoch 97/225\n",
      "1755/1755 [==============================] - 1s 841us/step - loss: 0.4161 - binary_accuracy: 0.8325\n",
      "Epoch 98/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4171 - binary_accuracy: 0.8272\n",
      "Epoch 99/225\n",
      "1755/1755 [==============================] - 1s 831us/step - loss: 0.4099 - binary_accuracy: 0.8331\n",
      "Epoch 100/225\n",
      "1755/1755 [==============================] - 1s 849us/step - loss: 0.4155 - binary_accuracy: 0.8254\n",
      "Epoch 101/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4194 - binary_accuracy: 0.8287\n",
      "Epoch 102/225\n",
      "1755/1755 [==============================] - 1s 835us/step - loss: 0.4169 - binary_accuracy: 0.8240\n",
      "Epoch 103/225\n",
      "1755/1755 [==============================] - 1s 840us/step - loss: 0.4060 - binary_accuracy: 0.8331\n",
      "Epoch 104/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4134 - binary_accuracy: 0.8277\n",
      "Epoch 105/225\n",
      "1755/1755 [==============================] - 1s 849us/step - loss: 0.4126 - binary_accuracy: 0.8284\n",
      "Epoch 106/225\n",
      "1755/1755 [==============================] - 2s 859us/step - loss: 0.4119 - binary_accuracy: 0.8311\n",
      "Epoch 107/225\n",
      "1755/1755 [==============================] - 1s 850us/step - loss: 0.4068 - binary_accuracy: 0.8318\n",
      "Epoch 108/225\n",
      "1755/1755 [==============================] - 1s 851us/step - loss: 0.4129 - binary_accuracy: 0.8301\n",
      "Epoch 109/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4156 - binary_accuracy: 0.8270\n",
      "Epoch 110/225\n",
      "1755/1755 [==============================] - 1s 844us/step - loss: 0.4095 - binary_accuracy: 0.8304\n",
      "Epoch 111/225\n",
      "1755/1755 [==============================] - 1s 849us/step - loss: 0.4142 - binary_accuracy: 0.8294\n",
      "Epoch 112/225\n",
      "1755/1755 [==============================] - 1s 851us/step - loss: 0.4093 - binary_accuracy: 0.8331\n",
      "Epoch 113/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4093 - binary_accuracy: 0.8290\n",
      "Epoch 114/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4073 - binary_accuracy: 0.8358\n",
      "Epoch 115/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4098 - binary_accuracy: 0.8317\n",
      "Epoch 116/225\n",
      "1755/1755 [==============================] - 2s 858us/step - loss: 0.4114 - binary_accuracy: 0.8328\n",
      "Epoch 117/225\n",
      "1755/1755 [==============================] - 2s 855us/step - loss: 0.4086 - binary_accuracy: 0.8309\n",
      "Epoch 118/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4175 - binary_accuracy: 0.8302\n",
      "Epoch 119/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4161 - binary_accuracy: 0.8288\n",
      "Epoch 120/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4144 - binary_accuracy: 0.8258\n",
      "Epoch 121/225\n",
      "1755/1755 [==============================] - 1s 841us/step - loss: 0.4117 - binary_accuracy: 0.8327\n",
      "Epoch 122/225\n",
      "1755/1755 [==============================] - 1s 836us/step - loss: 0.4138 - binary_accuracy: 0.8266\n",
      "Epoch 123/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4105 - binary_accuracy: 0.8313\n",
      "Epoch 124/225\n",
      "1755/1755 [==============================] - 1s 847us/step - loss: 0.4061 - binary_accuracy: 0.8334\n",
      "Epoch 125/225\n",
      "1755/1755 [==============================] - 2s 857us/step - loss: 0.4092 - binary_accuracy: 0.8320\n",
      "Epoch 126/225\n",
      "1755/1755 [==============================] - 1s 844us/step - loss: 0.4095 - binary_accuracy: 0.8299\n",
      "Epoch 127/225\n",
      "1755/1755 [==============================] - 1s 851us/step - loss: 0.4114 - binary_accuracy: 0.8291\n",
      "Epoch 128/225\n",
      "1755/1755 [==============================] - 1s 836us/step - loss: 0.4086 - binary_accuracy: 0.8287\n",
      "Epoch 129/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4084 - binary_accuracy: 0.8361\n",
      "Epoch 130/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4079 - binary_accuracy: 0.8320\n",
      "Epoch 131/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4063 - binary_accuracy: 0.8347\n",
      "Epoch 132/225\n",
      "1755/1755 [==============================] - 2s 861us/step - loss: 0.4065 - binary_accuracy: 0.8308\n",
      "Epoch 133/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4159 - binary_accuracy: 0.8254\n",
      "Epoch 134/225\n",
      "1755/1755 [==============================] - 1s 849us/step - loss: 0.4129 - binary_accuracy: 0.8288\n",
      "Epoch 135/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4104 - binary_accuracy: 0.8275\n",
      "Epoch 136/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4076 - binary_accuracy: 0.8343\n",
      "Epoch 137/225\n",
      "1755/1755 [==============================] - 1s 849us/step - loss: 0.4145 - binary_accuracy: 0.8279\n",
      "Epoch 138/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4104 - binary_accuracy: 0.8252\n",
      "Epoch 139/225\n",
      "1755/1755 [==============================] - 1s 847us/step - loss: 0.4089 - binary_accuracy: 0.8311\n",
      "Epoch 140/225\n",
      "1755/1755 [==============================] - 1s 844us/step - loss: 0.4056 - binary_accuracy: 0.8364\n",
      "Epoch 141/225\n",
      "1755/1755 [==============================] - 1s 847us/step - loss: 0.4167 - binary_accuracy: 0.8262\n",
      "Epoch 142/225\n",
      "1755/1755 [==============================] - 1s 844us/step - loss: 0.4108 - binary_accuracy: 0.8301\n",
      "Epoch 143/225\n",
      "1755/1755 [==============================] - 1s 835us/step - loss: 0.4054 - binary_accuracy: 0.8305\n",
      "Epoch 144/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4087 - binary_accuracy: 0.8320\n",
      "Epoch 145/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1755/1755 [==============================] - 1s 835us/step - loss: 0.4069 - binary_accuracy: 0.8358\n",
      "Epoch 146/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4129 - binary_accuracy: 0.8291\n",
      "Epoch 147/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4165 - binary_accuracy: 0.8295\n",
      "Epoch 148/225\n",
      "1755/1755 [==============================] - 1s 836us/step - loss: 0.4091 - binary_accuracy: 0.8337\n",
      "Epoch 149/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4129 - binary_accuracy: 0.8301\n",
      "Epoch 150/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.3996 - binary_accuracy: 0.8379\n",
      "Epoch 151/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4076 - binary_accuracy: 0.8344\n",
      "Epoch 152/225\n",
      "1755/1755 [==============================] - 1s 846us/step - loss: 0.4058 - binary_accuracy: 0.8300\n",
      "Epoch 153/225\n",
      "1755/1755 [==============================] - 1s 840us/step - loss: 0.4064 - binary_accuracy: 0.8342\n",
      "Epoch 154/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4112 - binary_accuracy: 0.8332\n",
      "Epoch 155/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4045 - binary_accuracy: 0.8340\n",
      "Epoch 156/225\n",
      "1755/1755 [==============================] - 1s 854us/step - loss: 0.4059 - binary_accuracy: 0.8314\n",
      "Epoch 157/225\n",
      "1755/1755 [==============================] - 1s 846us/step - loss: 0.4127 - binary_accuracy: 0.8330\n",
      "Epoch 158/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4121 - binary_accuracy: 0.8322\n",
      "Epoch 159/225\n",
      "1755/1755 [==============================] - 1s 847us/step - loss: 0.4078 - binary_accuracy: 0.8301\n",
      "Epoch 160/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4046 - binary_accuracy: 0.8328\n",
      "Epoch 161/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4038 - binary_accuracy: 0.8360\n",
      "Epoch 162/225\n",
      "1755/1755 [==============================] - 1s 835us/step - loss: 0.4045 - binary_accuracy: 0.8344\n",
      "Epoch 163/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4075 - binary_accuracy: 0.8320\n",
      "Epoch 164/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4042 - binary_accuracy: 0.8291\n",
      "Epoch 165/225\n",
      "1755/1755 [==============================] - 1s 849us/step - loss: 0.4106 - binary_accuracy: 0.8318\n",
      "Epoch 166/225\n",
      "1755/1755 [==============================] - 1s 851us/step - loss: 0.4062 - binary_accuracy: 0.8338\n",
      "Epoch 167/225\n",
      "1755/1755 [==============================] - 1s 844us/step - loss: 0.4075 - binary_accuracy: 0.8302\n",
      "Epoch 168/225\n",
      "1755/1755 [==============================] - 1s 853us/step - loss: 0.4123 - binary_accuracy: 0.8269\n",
      "Epoch 169/225\n",
      "1755/1755 [==============================] - 1s 847us/step - loss: 0.4096 - binary_accuracy: 0.8328\n",
      "Epoch 170/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4144 - binary_accuracy: 0.8299\n",
      "Epoch 171/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4033 - binary_accuracy: 0.8321\n",
      "Epoch 172/225\n",
      "1755/1755 [==============================] - 1s 836us/step - loss: 0.4145 - binary_accuracy: 0.8277\n",
      "Epoch 173/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4079 - binary_accuracy: 0.8336\n",
      "Epoch 174/225\n",
      "1755/1755 [==============================] - 1s 840us/step - loss: 0.4098 - binary_accuracy: 0.8297\n",
      "Epoch 175/225\n",
      "1755/1755 [==============================] - 1s 833us/step - loss: 0.4090 - binary_accuracy: 0.8278\n",
      "Epoch 176/225\n",
      "1755/1755 [==============================] - 1s 835us/step - loss: 0.4007 - binary_accuracy: 0.8341\n",
      "Epoch 177/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4083 - binary_accuracy: 0.8308\n",
      "Epoch 178/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4093 - binary_accuracy: 0.8318\n",
      "Epoch 179/225\n",
      "1755/1755 [==============================] - 1s 849us/step - loss: 0.4020 - binary_accuracy: 0.8372\n",
      "Epoch 180/225\n",
      "1755/1755 [==============================] - 1s 840us/step - loss: 0.4020 - binary_accuracy: 0.8346\n",
      "Epoch 181/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4046 - binary_accuracy: 0.8357\n",
      "Epoch 182/225\n",
      "1755/1755 [==============================] - 1s 835us/step - loss: 0.4045 - binary_accuracy: 0.8347\n",
      "Epoch 183/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4088 - binary_accuracy: 0.8312\n",
      "Epoch 184/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4102 - binary_accuracy: 0.8291\n",
      "Epoch 185/225\n",
      "1755/1755 [==============================] - 1s 841us/step - loss: 0.4103 - binary_accuracy: 0.8334\n",
      "Epoch 186/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4094 - binary_accuracy: 0.8280\n",
      "Epoch 187/225\n",
      "1755/1755 [==============================] - 1s 835us/step - loss: 0.4068 - binary_accuracy: 0.8303\n",
      "Epoch 188/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4094 - binary_accuracy: 0.8288\n",
      "Epoch 189/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4048 - binary_accuracy: 0.8329\n",
      "Epoch 190/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4066 - binary_accuracy: 0.8335\n",
      "Epoch 191/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4036 - binary_accuracy: 0.8322\n",
      "Epoch 192/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4107 - binary_accuracy: 0.8280\n",
      "Epoch 193/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4088 - binary_accuracy: 0.8315\n",
      "Epoch 194/225\n",
      "1755/1755 [==============================] - 1s 834us/step - loss: 0.4058 - binary_accuracy: 0.8314\n",
      "Epoch 195/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4087 - binary_accuracy: 0.8318\n",
      "Epoch 196/225\n",
      "1755/1755 [==============================] - 1s 841us/step - loss: 0.4072 - binary_accuracy: 0.8325\n",
      "Epoch 197/225\n",
      "1755/1755 [==============================] - 1s 853us/step - loss: 0.4005 - binary_accuracy: 0.8360\n",
      "Epoch 198/225\n",
      "1755/1755 [==============================] - 1s 849us/step - loss: 0.4110 - binary_accuracy: 0.8291\n",
      "Epoch 199/225\n",
      "1755/1755 [==============================] - 1s 850us/step - loss: 0.4027 - binary_accuracy: 0.8352\n",
      "Epoch 200/225\n",
      "1755/1755 [==============================] - 1s 853us/step - loss: 0.4105 - binary_accuracy: 0.8333\n",
      "Epoch 201/225\n",
      "1755/1755 [==============================] - 1s 841us/step - loss: 0.4070 - binary_accuracy: 0.8307\n",
      "Epoch 202/225\n",
      "1755/1755 [==============================] - 1s 850us/step - loss: 0.4063 - binary_accuracy: 0.8320\n",
      "Epoch 203/225\n",
      "1755/1755 [==============================] - 1s 844us/step - loss: 0.4061 - binary_accuracy: 0.8337\n",
      "Epoch 204/225\n",
      "1755/1755 [==============================] - 1s 832us/step - loss: 0.4023 - binary_accuracy: 0.8353\n",
      "Epoch 205/225\n",
      "1755/1755 [==============================] - 1s 841us/step - loss: 0.4091 - binary_accuracy: 0.8312\n",
      "Epoch 206/225\n",
      "1755/1755 [==============================] - 1s 842us/step - loss: 0.4039 - binary_accuracy: 0.8316\n",
      "Epoch 207/225\n",
      "1755/1755 [==============================] - 2s 855us/step - loss: 0.4064 - binary_accuracy: 0.8298\n",
      "Epoch 208/225\n",
      "1755/1755 [==============================] - 1s 848us/step - loss: 0.4057 - binary_accuracy: 0.8328\n",
      "Epoch 209/225\n",
      "1755/1755 [==============================] - 1s 833us/step - loss: 0.4087 - binary_accuracy: 0.8313\n",
      "Epoch 210/225\n",
      "1755/1755 [==============================] - 1s 834us/step - loss: 0.4029 - binary_accuracy: 0.8332\n",
      "Epoch 211/225\n",
      "1755/1755 [==============================] - 1s 836us/step - loss: 0.4081 - binary_accuracy: 0.8318\n",
      "Epoch 212/225\n",
      "1755/1755 [==============================] - 1s 851us/step - loss: 0.4055 - binary_accuracy: 0.8345\n",
      "Epoch 213/225\n",
      "1755/1755 [==============================] - 1s 841us/step - loss: 0.4070 - binary_accuracy: 0.8305\n",
      "Epoch 214/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4105 - binary_accuracy: 0.8259\n",
      "Epoch 215/225\n",
      "1755/1755 [==============================] - 1s 849us/step - loss: 0.4064 - binary_accuracy: 0.8303\n",
      "Epoch 216/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4078 - binary_accuracy: 0.8284\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 217/225\n",
      "1755/1755 [==============================] - 1s 838us/step - loss: 0.4030 - binary_accuracy: 0.8319\n",
      "Epoch 218/225\n",
      "1755/1755 [==============================] - 1s 839us/step - loss: 0.4063 - binary_accuracy: 0.8303\n",
      "Epoch 219/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4052 - binary_accuracy: 0.8354\n",
      "Epoch 220/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4052 - binary_accuracy: 0.8360\n",
      "Epoch 221/225\n",
      "1755/1755 [==============================] - 1s 843us/step - loss: 0.4039 - binary_accuracy: 0.8324\n",
      "Epoch 222/225\n",
      "1755/1755 [==============================] - 1s 845us/step - loss: 0.4016 - binary_accuracy: 0.8314\n",
      "Epoch 223/225\n",
      "1755/1755 [==============================] - 1s 837us/step - loss: 0.4049 - binary_accuracy: 0.8324\n",
      "Epoch 224/225\n",
      "1755/1755 [==============================] - 1s 846us/step - loss: 0.4101 - binary_accuracy: 0.8326\n",
      "Epoch 225/225\n",
      "1755/1755 [==============================] - 1s 844us/step - loss: 0.4043 - binary_accuracy: 0.8334\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEWCAYAAAB8LwAVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAAgAElEQVR4nOy9eZxddX3//3zffe7MvbPPZJJJMtkhCSFACCA7irKjlSouba2t1F9L0a+oRfutrdpavvhtba20it+qdUUUrSmrIgTECCSBkD1kZyaTzJLZl7t/fn+cZc6dubMluQkk7+fjkUfmnnvOuZ87gc/rvHcxxqAoiqIoo/Gd6gUoiqIob0xUIBRFUZSCqEAoiqIoBVGBUBRFUQqiAqEoiqIURAVCURRFKYgKhKIcJyLyHRH5+ymee0BE3na891GUk4EKhKIoilIQFQhFURSlICoQyhmB7dr5lIhsFpFBEflPEakXkcdFpF9EnhKRSs/5t4jINhHpEZG1InK2573zRORl+7ofA5FRn3WTiGyyr10nIiuOcc0fEZE9ItIlImtEZKZ9XETkKyLSLiK99ndabr93g4hst9d2SEQ+eUy/MEVBBUI5s3g3cC2wGLgZeBz4LFCD9f/CXQAishj4EfBxoBZ4DPgfEQmJSAj4b+B7QBXwE/u+2NeeD3wL+DOgGvgGsEZEwtNZqIhcA/wj8B6gATgIPGi//XbgCvt7VADvBY7a7/0n8GfGmBiwHHh6Op+rKF5UIJQziX8zxrQZYw4BvwFeNMa8YoxJAj8HzrPPey/wqDHmV8aYNPB/gRLgLcDFQBD4F2NM2hjzU2C95zM+AnzDGPOiMSZrjPkvIGlfNx0+AHzLGPOyvb7PAJeISBOQBmLAWYAYY3YYYw7b16WBpSISN8Z0G2NenubnKoqLCoRyJtHm+Xm4wOsy++eZWE/sABhjckAzMMt+75DJ73J50PPzXOBu273UIyI9wGz7uukweg0DWFbCLGPM08DXgPuBNhF5QETi9qnvBm4ADorIsyJyyTQ/V1FcVCAUZSytWBs9YPn8sTb5Q8BhYJZ9zGGO5+dm4B+MMRWeP1FjzI+Ocw2lWC6rQwDGmK8aYy4AlmG5mj5lH19vjLkVqMNyhT00zc9VFBcVCEUZy0PAjSLyVhEJAndjuYnWAb8DMsBdIhIQkd8DVnuu/SbwURG5yA4ml4rIjSISm+Yafgj8sYistOMXX8JyiR0QkQvt+weBQSABZO0YyQdEpNx2jfUB2eP4PShnOCoQijIKY8wu4IPAvwGdWAHtm40xKWNMCvg94ENAN1a84meeazdgxSG+Zr+/xz53umv4NfA3wMNYVssC4Hb77TiWEHVjuaGOYsVJAP4AOCAifcBH7e+hKMeE6MAgRVEUpRBqQSiKoigFUYFQFEVRCqICoSiKohREBUJRFEUpSOBUL+BEUVNTY5qamk71MhRFUd5UbNy4sdMYU1vovdNGIJqamtiwYcOpXoaiKMqbChE5ON576mJSFEVRCqICoSiKohREBUJRFEUpyGkTg1AURTkW0uk0LS0tJBKJU72UohKJRGhsbCQYDE75GhUIRVHOaFpaWojFYjQ1NZHfpPf0wRjD0aNHaWlpYd68eVO+Tl1MiqKc0SQSCaqrq09bcQAQEaqrq6dtJalAKIpyxnM6i4PDsXzHM14gDvcO88+/3MW+joFTvRRFUZQ3FGe8QLT3Jfnq03s4cHTwVC9FUZQzkJ6eHv793/992tfdcMMN9PT0FGFFI5zxAuH3WWZXJqtzMRRFOfmMJxDZ7MTDAB977DEqKiqKtSxAs5hcgcjmVCAURTn53HPPPezdu5eVK1cSDAYpKyujoaGBTZs2sX37dt75znfS3NxMIpHgYx/7GHfccQcw0l5oYGCA66+/nssuu4x169Yxa9YsfvGLX1BSUnLcazvjBSLgCIRO1lOUM57P/882trf2ndB7Lp0Z529vXjbu+/feey9bt25l06ZNrF27lhtvvJGtW7e66ajf+ta3qKqqYnh4mAsvvJB3v/vdVFdX591j9+7d/OhHP+Kb3/wm73nPe3j44Yf54AePf9rsGS8QPrUgFEV5A7F69eq8WoWvfvWr/PznPwegubmZ3bt3jxGIefPmsXLlSgAuuOACDhw4cELWcsYLREAFQlEUm4me9E8WpaWl7s9r167lqaee4ne/+x3RaJSrrrqqYC1DOBx2f/b7/QwPD5+QtWiQ2glSq0AoinIKiMVi9Pf3F3yvt7eXyspKotEoO3fu5IUXXjipazvjLQgNUiuKciqprq7m0ksvZfny5ZSUlFBfX+++d9111/H1r3+dFStWsGTJEi6++OKTujYVCBUIRVFOMT/84Q8LHg+Hwzz++OMF33PiDDU1NWzdutU9/slPfvKErauoLiYRuU5EdonIHhG5Z4LzbhMRIyKr7NcfEJFNnj85EVlZjDX6RQVCURSlEEUTCBHxA/cD1wNLgfeJyNIC58WAu4AXnWPGmB8YY1YaY1YCfwAcMMZsKsY6Az7rV6AxCEVRlHyKaUGsBvYYY/YZY1LAg8CtBc77InAfMF6bwfcBPyrOEsHvtyyInAqEopyxmDOgDupYvmMxBWIW0Ox53WIfcxGR84DZxphHJrjPexlHIETkDhHZICIbOjo6jmmRjotJLQhFOTOJRCIcPXr0tBYJZx5EJBKZ1nXFDFIX6i3r/guIiA/4CvChcW8gchEwZIzZWuh9Y8wDwAMAq1atOqZ/XSdInTuN/+NQFGV8GhsbaWlp4VgfMt8sOBPlpkMxBaIFmO153Qi0el7HgOXAWrtP+QxgjYjcYozZYJ9zO0V0L8FIoZw261OUM5NgMDitKWtnEsUUiPXAIhGZBxzC2uzf77xpjOkFapzXIrIW+KQjDraF8fvAFUVco6fVRq6YH6MoivKmo2gxCGNMBrgTeBLYATxkjNkmIl8QkVumcIsrgBZjzL5irdEh4BNt1qcoijKKohbKGWMeAx4bdexz45x71ajXa4GTUjbo84kGqRVFUUZxxvdiAtuC0BiEoihKHioQWJlM6mJSFEXJRwUCWyDUxaQoipKHCgS2i0kFQlEUJQ8VCNSCUBRFKYQKBFa7Dc1iUhRFyUcFAqthnzbrUxRFyUcFArUgFEVRCqECgaa5KoqiFEIFAmtokBbKKYqi5KMCgbbaUBRFKYQKBFYdhM6DUBRFyUcFAisGoRaEoihKPioQOIVyOg9CURTFiwoEWkmtKIpSCBUIrDoIFQhFUZR8VCCAgF8FQlEUZTRFFQgRuU5EdonIHhG5Z4LzbhMRIyKrPMdWiMjvRGSbiGwRkUix1qkuJkVRlLEUbeSoiPiB+4FrgRZgvYisMcZsH3VeDLgLeNFzLAB8H/gDY8yrIlINpIu1Vm21oSiKMpZiWhCrgT3GmH3GmBTwIHBrgfO+CNwHJDzH3g5sNsa8CmCMOWqMyRZroWpBKIqijKWYAjELaPa8brGPuYjIecBsY8wjo65dDBgReVJEXhaRTxf6ABG5Q0Q2iMiGjo6OY16oCoSiKMpYiikQUuCYuwuLiA/4CnB3gfMCwGXAB+y/3yUibx1zM2MeMMasMsasqq2tPeaFqkAoiqKMpZgC0QLM9rxuBFo9r2PAcmCtiBwALgbW2IHqFuBZY0ynMWYIeAw4v1gLDWg3V0VRlDEUUyDWA4tEZJ6IhIDbgTXOm8aYXmNMjTGmyRjTBLwA3GKM2QA8CawQkagdsL4S2D72I04MPp+Q0W6uiqIoeRRNIIwxGeBOrM1+B/CQMWabiHxBRG6Z5Npu4J+xRGYT8LIx5tFirVWb9SmKooylaGmuAMaYx7DcQ95jnxvn3KtGvf4+Vqpr0fH7fJrmqiiKMgqtpAb8PjRIrSiKMgoVCOyJcioQiqIoeahAAD5t1qcoijIGFQisZn0ZnQehKIqShwoEVqGc6oOiKEo+KhA4zfpUIRRFUbyoQGBbEAaM1kIoiqK4qEBgFcqBproqiqJ4UYHAarUBaLGcoiiKBxUIRiwIbbehKIoyggoEVgwC1IJQFEXxogLBiEDkVCAURVFcVCAYcTGpBaEoijKCCgQjQWrNYlIURRlBBQJNc1UURSmECgTWPAhQgVAURfGiAoE1DwI0BqEoiuJFBQK1IBRFUQpRVIEQketEZJeI7BGReyY47zYRMSKyyn7dJCLDIrLJ/vP1Yq7TLxqDUBRFGU3RZlKLiB+4H7gWaAHWi8gaY8z2UefFgLuAF0fdYq8xZmWx1ufFr0FqRVGUMRTTglgN7DHG7DPGpIAHgVsLnPdF4D4gUcS1TIhmMSmKooylmAIxC2j2vG6xj7mIyHnAbGPMIwWunycir4jIsyJyeaEPEJE7RGSDiGzo6Og45oWOtNrQmRCKoigOxRQIKXDMfUQXER/wFeDuAucdBuYYY84DPgH8UETiY25mzAPGmFXGmFW1tbXHvFC/NutTFEUZQzEFogWY7XndCLR6XseA5cBaETkAXAysEZFVxpikMeYogDFmI7AXWFyshboWRFYFQlEUxaGYArEeWCQi80QkBNwOrHHeNMb0GmNqjDFNxpgm4AXgFmPMBhGptYPciMh8YBGwr1gL1SC1oijKWIqWxWSMyYjIncCTgB/4ljFmm4h8AdhgjFkzweVXAF8QkQyQBT5qjOkq1lrdILW6mBRFUVyKJhAAxpjHgMdGHfvcOOde5fn5YeDhYq7Ni06UUxRFGYtWUuOZKKcCoSiK4qICgU6UUxRFKYQKBBqkVhRFKYQKBFpJrSiKUggVCMCnzfoURVHGoAIBBOx23xqDUBRFGUEFAvD7NYtJURRlNCoQjMyDUAtCURRlBBUIPFlMWkmtKIriogKBJ4spq+2+FUVRHFQg0FYbiqIohVCBwNNqQ11MiqIoLioQaKsNRVGUQqhA4JkoN45APLH1MN2DqZO5JEVRlFOOCgQTp7n2DqX56Pdf5p9+tetkL0tRFOWUogKBFaQWKdxqo3c4DcCaTa0k0tmTvTRFUZRThgqETcAnPLrlMO/+j3V5QtGXSNt/Z3hy25FTtTxFUZSTTlEFQkSuE5FdIrJHRO6Z4LzbRMSIyKpRx+eIyICIfLKY6wQrDrGvY5CNB7tp7hpyj/cnMvZa4OevHCr2MhRFUd4wFE0gRMQP3A9cDywF3iciSwucFwPuAl4scJuvAI8Xa41enDgEwGtt/e7PjgUxq6KEwz2Jk7EURVGUNwTFtCBWA3uMMfuMMSngQeDWAud9EbgPyNt9ReSdwD5gWxHX6OIUywHsbh9wf3YsiOqyMCmttFYU5QyimAIxC2j2vG6xj7mIyHnAbGPMI6OOlwJ/BXx+og8QkTtEZIOIbOjo6DiuxTpCAPkWRL9tQdSUhkhlVCAURTlzKKZASIFjbvRXRHxYLqS7C5z3eeArxpiBAu+N3MyYB4wxq4wxq2pra49rsQ4lQT+vtRWyIEIkM5rFpCjKmcOUBEJEPiYicbH4TxF5WUTePsllLcBsz+tGoNXzOgYsB9aKyAHgYmCNHai+CLjPPv5x4LMicueUvtFxcvVZteztGHAzmfoTaSJBH9FQgKRaEIqinEFM1YL4sDGmD3g7UAv8MXDvJNesBxaJyDwRCQG3A2ucN40xvcaYGmNMkzGmCXgBuMUYs8EYc7nn+L8AXzLGfG1a3+wYuXJxLalMjoNHBwHLgohHgoQDPnUxKYpyRjFVgXDcRTcA3zbGvEphF5KLMSYD3Ak8CewAHjLGbBORL4jILce64GKzZEYcwHUz9ScyxCIBQgEfqWwOow39FEU5QwhM8byNIvJLYB7wGTs1ddLHaWPMY8Bjo459bpxzrxrn+N9NcY3HzZyqKPNqSgF4vcuyIPoSaWKRICG/D2OsdhxB/4TaqCiKclowVYH4E2AlsM8YMyQiVVhuptOGrZ9/BwGfEAn6KS8J8rpdLNeXyBCPBAgHLWMrmckR9GsBuqIopz9T3ekuAXYZY3pE5IPA/wZ6i7esk09ZOEAk6AdgdlUJzV3DgBWkjtsWBKBxCEVRzhimKhD/AQyJyLnAp4GDwHeLtqpTzOzKKM3dlgXRn8gQLwkQCljioQKhKMqZwlQFImOs6OytwL8aY/4VK031tGR2VZSW7mFyOUO/E4MIqAWhKMqZxVRjEP0i8hngD4DL7T5LweIt69Qyu7KEVCZHa+8wiXSOWDhAOODEILRYTlGUM4OpWhDvBZJY9RBHsFpmfLloqzrFzK6KArCttQ/ATXMFtFhOUZQzhikJhC0KPwDKReQmIGGMOX1jELZAbHcFwuNi0oZ9iqKcIUy11cZ7gJeA3wfeA7woIrcVc2GnklkVJQBsP2wJRLwkSFizmBRFOcOYagzir4ELjTHtACJSCzwF/LRYCzuVRIJ+6uNhjwURIGC3A1eBUBTlTGGqMQifIw42R6dx7ZuSs2bEOdRj1ULEIgHCdpqrxiAURTlTmOom/4SIPCkiHxKRDwGPMqqFxunG39y0lGjIEoW4prkqinIGMtUg9aeAB4AVwLnAA8aYvyrmwk41C+vKuO+2FZwzq5zaWNgTpNY0V0VRzgymGoPAGPMw8HAR1/KG46YVM7lpxUwAtSAURTnjmFAgRKQfzxQ471uAMcbEi7KqNyBhrYNQFOUMY0KBMMactu00pstkFkQmmyORyVEWnrJRpiiK8obmtM5EOpE43VwdC+IHLx7kzh++7A4Q+o+1e3nbPz2rA4UURTltUIGYIqPbfT+1vY1HNh9m48FuAF460MWRvgRH+hKnbI2KoignkqIKhIhcJyK7RGSPiNwzwXm3iYgRkVX269Uissn+86qIvKuY65wKPp8Q8vvcVhst3VaNxLfXHQBgx+F+APZ1DJ6S9SmKopxoiiYQdsfX+4HrgaXA+0RkaYHzYsBdwIuew1uBVcaYlcB1wDdE5JQ790MBH8m0NZe6pXuYkN/HE1uPsPVQL50DSQD2dgyc4lUqiqKcGIppQawG9hhj9hljUsCDWPMkRvNF4D7A9c0YY4aMMRn7ZYTCmVQnnVDARyqbpWswxXA6y/svmkM2Z/jyk7vcc/a25wvE7/YeJZHW2glFUd58FFMgZgHNntct9jEXETkPmG2MeWT0xSJykYhsA7YAH/UIxikj5PeRyuRc99KlC2tYNjPOs691ANao0r0eF1NL9xDv++YL/M+rradkvYqiKMdDMQVCChxzLQER8QFfAe4udLEx5kVjzDLgQuAzIhIZ8wEid4jIBhHZ0NHRcYKWPT6hQL5ANFaWcMM5DQDUxcJcOLcqz8V08Kg1trR7KFX0tSmKopxoiikQLcBsz+tGwPsoHQOWA2tF5ABwMbDGCVQ7GGN2AIP2uYx67wFjzCpjzKra2toTvPyxhAM+kpkcLfa86lmVJdxoC8TZDXEW1JVxuDfBQNIydpq7rPMGEsdu/PzgxYM8vuXwca5cURRl+hRTINYDi0RknoiEgNuBNc6bxpheY0yNMabJGNMEvADcYozZYF8TABCRucAS4EAR1zolvBZEeUmQeCRIU00p779oDr93/iwW1JYCsN92MzmWRn/y2ATCGMM///I1fvjS6yfmCyiKokyDomUGGWMyInIn8CTgB75ljNkmIl8ANhhj1kxw+WXAPSKSBnLAnxtjOou11qliBalzHOoZprGyxD3+pXedA1gWgwg8sqWVcxrLabYtjcFjFIgjfQmODqboPw4LRFEU5VgpauqoMeYxRrUFN8Z8bpxzr/L8/D3ge8Vc27EQ8lsupra+BPNqSse8P7sqyrtWzuLbvz3AH17S5FoQA8coEFsP9Y25/v5n9jCcyvLJdyw5pnsqiqJMFa2kngbhoN+OQQwzqyJa8JxPvH0xAF97eo8bgzhWC2DroV77+rR77Nc72nhqR9sx3U9RFGU6qEBMg5Dfx9GBJEOpLDMrxiRVAdBYGeXGcxp4dHMr7f1W8dxkFsQ3n9vHe77xuzHHHYHwBrn7Ehl1OSmKclJQgZgG4YCPVnsMaV28sEAAvGPZDPrsTVykcAwilzO88rrVx2lbay/rD3SNKajb2moJxGAqSzZnZQj3J9J5FoWiKEqxUIGYBqGAD3ufpj4WHve8KxfXEglav9qm6tKCaa5PbDvCu/59HXvaBxhIZjEGXrddUgAd/Una+pLMqrCC4Y4V0jecoT+ZIZd7QxSXK4pyGqMCMQ2cjq4AM8rHtyBKQn6uXGzVZZw1I1YwzXVTcw8AnQNJBpKWRbC/c6QK2xGLFY3lgGU5pDI5htOWmAym1M2kKEpxOeUN8N5MhIMjAlEXG18gAD5y+XxKwwEayiM8se0IxhhERorLt7WOxBcGk5Zr6eDREYFot9uGL6wrs85LZigJjriW+hMZYpHgcX4jRVGU8VELYho4FkQ8EqAk5J/w3FVNVfzze1YSjwQxBoZSI/EFYwzbWq0U1v5k2nUf7e8ccTG1jRKI/lHB6T6NQyiKUmTUgpgGztjR+gkC1KMpi1i/4sFkhtJwwM2C6hmyNviBRMYViAOdg/zPq600VZfS1p8k6BdmV0Xd85y52HDsqbOKoihTRQViGjgCMVH8YTTOjOr+ZIbwcJrL73uG2ZUjNRT9yYwbxN7a2svHf7yJa8+uJxr2UxeLELfdSH2JNEG/VyDUglAUpbioQEyDcMByK00Wf/DiCMRAIsPOw/0MpbLsauvHJyAi9A6nGU5nCQV8rlWwv3OQ2liYuniYmG2BDCQzeQLRN/zGtSBSmRwBn+DzFWroqyjKmwWNQUyDERfT+Cmuo3EFIpnhmV3tlJcEWdFYzvJZ5ZSXBGnrtWINZ8+IAVbdxP6jgxzuHaY+FnEFoj+RoW/YG6R+Y1oQxhiu+ae1fMcexaooypsXFYhpcDwxiP5Emmdf6+CKxbU8eMfFfO/DF1EWDnDYFogrl9SxpD7Gh97SRCqTY1/nIPXxMCVBP36fMDAmSH1yLYhDPcN0DU4+12IwlaWlezgvI0tRlDcnKhDTIOyfvkDEwlYM4cX9XXT0J7lqcS3RUIDyaJCycIAjdrbSoroynvxfV/COZTMAMMaq1hYRysIB+hNp+hJpRKxsqqlmMQ0kM+w43Dedr1mQj/zXBr74yPZJz+uw24sMpnTMqqK82VGBmAZOHcS0XEy2BfHLbVaDvcsX17jvxSIjFoRznrdLrCNEZeEA/UnLxRSPBImXBKecxfTt5/dz41d/w/bW4xOJ5q6hvErv8XAEYlgFQlHe9KhATINL5lfz/ovmsHRmfMrXlIatwPahnmFmV5XkBbhjkQCpTA4YiVXUxcJE7RoLR4hikYAVg0hkiEUCxCMB+obTfOPZvbzW1j/h57/eNUTOwBce2YYxx9aeYziVpT+ZcTf/iRixIN64QXRFUaaGCsQ0qItH+NK7znGzmaZCOOB3C+xWzq7Me88RBe/PIuJaEY4FEYsEGEiMWBCxSIC9HYP84+M7+fqzeyf8/CN9Cfw+4YV9XTy/59hmLnUOWJt+R39yUpFxzh16E1gQf/a9DXzlV6+d6mUoyhsWFYiTgOM+Wjm7Iu+4t1WGVyxcgYhF3PMGklaQOl4SIF4SZOcRy2X0m92dE27abX0JLphrCdNrbQPHtH6nbflwOjtpbMGxIIbeBBbE5pZetp+A+IyinK6oQJwEnM1/tEA4wgFQ6hGIC5uqmFVRQrwk4F7vBKkdC8LRhI7+JDsOj+9mOtKbYEl9jKBfaO9PTGvdj205zMcefCXPtTSZm2lEIN74FsRAIjOmxbqiKCMUVSBE5DoR2SUie0TkngnOu01EjIissl9fKyIbRWSL/fc1xVxnsSkLBwj6hWWjYhdeq8GJVQD84SVzef6vrnab+8UiAQacIHVJ0M2MKi+x/n72tY6CnzucytKXyDCjPEJtWZiOvvzNvXMgyZHe8UXjZy8f4hebWjngSVmdVCAcF1Pyjb3xGmMYSGU0mK4oE1C0SmoR8QP3A9cCLcB6EVljjNk+6rwYcBfwoudwJ3CzMaZVRJYDTwKzirXWYlNdFiIcLCcSzI9dxG0LIuT35cU1vF1fwbI0+hIZQn4f8UgQp6B69bwqmruG+PZv97O3Y4DP37IszxJxUmhnxCPUxiN0DCTJZHMc7k0wuyrKJx56laFkhp/+f28puG5not1L+7vcY1O3IN7YLqahlNU2fVgtCEUZl2JaEKuBPcaYfcaYFPAgcGuB874I3Ae4j7LGmFeMMa32y21ARESmnlv6BuNL7zqHr95+3pjjjovJ62oqRDwSJJXJMZC0YhBO7OLshjh/cfVC6uJhfrqxhfUHuvKuc6yDGeUR6mJh2vuS/GRjC9f801qau4ZYv7+Llm5rQt4zu9rdFuNgbfSOwKzf3+UG2jsmcVMVcjF1DaaO25WTzRl++OLrbtbX8eI0SFSBUJTxKaZAzAKaPa9bGGUFiMh5wGxjzCMT3OfdwCvGmDGPriJyh4hsEJENHR2F3SxvBGZXRd2urF7KbFeR171UiCsX17otN5wYBMDShjg3nzuT7374IgD2tOcHoZ2W4fVxWyD6E2xv7SOdNdz/zB6G01mODiZJZXL86X9t4D9/u9+91rEewGooOK+mlIBPXBdSIXI5Q+dAEr9PyOSMu5nf9vV1fPZnWyb8jpPx/J5OPvvzLfz2GDOxRuPUkSSm4GLK6vQ+5QylmAJRqFOb+3+aiPiArwB3j3sDkWXA/wH+rND7xpgHjDGrjDGramtrj3O5Jx9noy8NTWxBLJ9VzjOfvIp7rj+Lm8+dydzqKKGAj3NnW9PmqkpDVEaD7O3Ib29xxBWIMHWxCN1Dabdu4icbWwBIZw172gfI5oxrTQBssQWipswy3OriYarLQhO6mHqH02RyhpkVVvbVcCrLcCrLvo5BHtl8eEqtOsZjt73unuFjv4eXqVoQ6w90sexvn+BQz/CE5ynK6UgxBaIFmO153Qi0el7HgOXAWhE5AFwMrPEEqhuBnwN/aIyZONn/TYoTpI5N4mICa6P+6JULqI2FuXpJHes/+zYaykvc9xfWlbF3lAVxpDdBachPLBKkzi66c0adep+Knel2rT35AjG/ptQNrNfGwtTGwhMKhGNdNFVbabqDqYxbfZ3K5vjZyy2Tfs9MtrALyRG23qET06RwcIoCsen1HhLpHOv3d014nqKcjhRTINYDi0RknhgWuzwAACAASURBVIiEgNuBNc6bxpheY0yNMabJGNMEvADcYozZICIVwKPAZ4wxvy3iGk8prgURnl6ugIhQHs0fN7qgtoy9HfkC0d6foN6eXVEXswQimcmxtMHa9GfYhXhOLYBXILa39rF8Vjnza63NvjYWtjKhJnAxOeIxx3anDaWybtO+imiQhzY0j3stQDqb45J7n+avf76F3Ci3zm5b/JwmhclMlk//9FVauidv/1EI18WUzo35LC+OwG1u6R33HEU5XSmaQBhjMsCdWBlIO4CHjDHbROQLInLLJJffCSwE/kZENtl/6oq11lOFE2wum6ZAFGJhXRlHB1N0e9w4R3oTrgh4W3x88OK5zK8t5YMXzwFw+zS191vxCCvTaZi51VHm20V7tWWTWxBOUNwp9BtKZTh41Npg331+I6+1DZDMjP/E/nrXEB39SX7w4uv8nyd2useNMeyxi/x67ZbnOw/389CGFn61vW0qv54xOC4mgMQEa2q2BWjLoZ5j+hxFeTNT1DoIY8xjxpjFxpgFxph/sI99zhizpsC5VxljNtg//70xptQYs9Lzp72Yaz0VOMHpEyEQC2qt2dWOFZHLGfZ3DtJYabmhamMjSWDLZ8V5+u6ruH21LRC2BWGMFdjuGEiSM1b203z7vl4X0yd/8iqHe8f65Ju7hxAZmaM9lMpy4OggFdGge+zowPgxhP12DGXZzDjfXnfA3cTb+pL02z87MzGcJoeOAHnJ5kxekL0QA55uuBPVQjgWxNZDfeO6vxTldEUrqU8h4YCfsnCAimjouO/lbMCOQGw/3Ef3UJpLFlQDUFMWwimvaLKf8CujIXySP9/6UM/wSHpsPMKqpkr+4uoFXLWkjpvPnclVS+p4+OUWfrx+rLuopdsacuR8H8eCmFtd6ga7nV5NqUyOA535QfX99uu/vGYRqUyO5+wCQG9DQqfNuSNQhTrM/mLTIW76t+d5vYB4OHgtiNFxCEcwcnbgvj4eZjidHZMEcLw0dw1x3hd+yeYWtU6UNyYqEKeY7/3Jav708nnHfZ+ZFSVEgj43++i53dbmeulCq714wO+jujRETVnInXPt9wlVpdZm7hTtHe4dzkuPDQf8fOodZ1FeEuSsGXG+9aELWT6znHV7jo5ZQ0v3ELOrStxutI4F0VQddS0Yx0X1gxcP8vavPJeX2bSvc4Cq0hBvO7uOymiQJ7cdAUbiDwtqS10XkyNiBwoMJtpwsNtdz3gMeCq9vTUa33vhIOd/8Vf0J9Kuy+365Q0AJ3wj39sxQPdQmv9ad/CE3ldRThQqEKeY8+ZUuk/Xx4PfJ1y7dAZrNrUylMrw/O5OzpoRy4s91McjefMmYCSN9Vy7T1RrTyKvwK4Qb1lYzSvN3WOqpZu7hmmsjLoC0TOUprVn2LYgLCFyLIith/pIZXNs8BT37esYZH5NKQG/j7edXc/TO9tJZXLs7xygvCTIvJpSdxZ3q73Glq7hMXUKzkbeNkFR30DS62KyXEeJdJav/no3w+ksbX1JN/5w5ZJaQn4fezqOrdnheDiW22NbDp+SEbIfe/AVntk5vud248Fu9rRP3E5eOb1RgTiN+KNL5tKXyPDt3x5gw4FuLl9Uk/f+F9+5nL+9eVneMUcg5lRFqS4NWS6mviRBv1A1juvr0gU1pLMmrwVHJpvjSF+CxsoSonZdx+62fnIGmqqj7uc4FoSz2TpP+2C5mBwBu+asOvoTGbYf7qO9L0lDeYR4SdBjQVguppT9uQ6JdJaddvPCtr7xA+oDibEuph+++Lq7vt7hlOuiaqoupSwSOOH9pRyBGE5nWfNq6yRnn1gy2Ry/2NTK2l3jC8T/+vEmvvzkrpO4KuWNhgrEacQFcys5uyHOl5/cRTqX47rlM/LeP39OJctnlecdq7af7OvjEWZWlNDaM8yR3mHqYhF8vkK1jla32aBfuO+JXfz5DzaSSGc53JsgmzO2QFgWhBP8nlsdJRL0E4sE6BxIYYxhn+02ckTGcenMs9NqncrzI70J2vuT1MbCxCNBTwwi4bqtvPOvtx/uI2NbFG19E1kQYwXikc2tbsJA92Ca17usoPvMigjRkP+ED0FyrIZZFSU8uvnwCbnnxx58ha89vXvS84bs79w5TtJAMpOlpXuI7sGTb9kobxxUIE4jRITP37KMO66Yz6N/eTkXzK2a9Brnyb4+HmZmRYSW7mGO9CVoGMe9BFAS8nPFolp2tfXz2JYjbDzY7bpjZldGCQd8+AT3SX5u9UiqbEd/kvZ+KyupMhpk66FehlNZDnRa1ztptU5hX3t/go7+JLVlYeIl1lyMTDZHW1+Ci+dbAXhvMHqzXQgYiwRon8iCSGaI2WLgBKW7BlNusL9nOE1z9xANdhymNBRwi+vGY82rrVx679NT7hc1kMzgE7hpRQMv7e86IW6mF/d18asdkyf8OdZQ5zh1Lc1dw+TMiatcV96cqECcZqyeV8Vnbzh7ymNRR1ppRFg+s5w97QO81jbgFtiNxzf+4AI2/PXbCPiEdXs73TYdjZVRRIRoyJqjXRYOUG0HwmtiVqGdU/H9e+c3kskZNjX3uNlX82qsDbq6NIzfJ27abW0sTHlJEGOswHQ6azh/TgVBv/DktiN84sebuO0/1vG1Z/ZSGwuztCE+qQXhWCBOkLp7KO26uHqGUhzpTdBQYaUJR8P+SWdcrN3VzqGeYY4OTj6aFSwXU1k4wDVn1ZHJmRPSZ2ogmbFce5P0j3KsofEEwskw6zlBlevKmxMViDMcZ5Osj0W4+iyrFrFrMOUW2I1HwO+jsjTEisZy1u09Skv3MD6BBrsPk+NmmlsddduX15aF6exPuvGH9144GxGr39Hmll4iQZ9bue33CbVlYfa0D5DK5GwXk/XEv+uIdX1jZZTGyijP7Org2dc68PuE5bPi3HXNQmaURyYOUicy1NjffTidJZsz9CXSzK4swe8TeobSdPQn3Qr0qVgQTsHhRLUeXvoSaWKRIOfPrSQWCfD0BAHjqZDLGQaSGYZS2bzeUcaYMQkFjgVxdJz+WE52mBPzOZW09gyfsBYryvQo2jwI5c3B9ctnkM7mOLshBliupra+5KQC4fCWBTX8x7N7KQn6aSgvIWi3BXcEwunLBJYYPbfbsiDKwgEW1ZWxpD7G+gNdDCQznDOr3L0eLDfT1kN97rUl9jyNXfa41YbyCH913RLa+5O8Z9XsvHkbzd3DtPVZM7RHz9cA60l7mSMQqSy9w2mMgcrSEOUlQXqGU3QMJF03VjTkH/dpGywrxOmm2zmQpNmuzyjUxdehP5EhFgkQ9Pu4YnEtz+zqGHe9U8EbI3mtrd/97Me3HuHTP93Mus9c46Y4O+f2DKVJZ3N5v3cYEYhkJkcinR0zy6TYfPd3B3hy2xF+8KcX8yf/tYFzG8u5990rTuoaFLUgznhKwwHet3oOIoKIcI1tRUzmYnJ4y4JqsjnDur1Hedd5I93cnUymudUjG2RNWYj+RIatrX0sqC1FRLiwqYqXD3az7VAf582pzLt3XSziPgk7MQiAnUes2EZDeYTrljfwh5c0jdnA6mJhUpncuE/AA8kMtWUjFkT3kPUkXRkNUVESpL0vSc9Q2rWwSsOBCYPUu9sG3OD40YEUf/XwZj7101cn+tXRn0i7/bgunldFR3+Sw70JeofSeUH0qeK9xjt/fE/7AAPJTF5hoteiKNRl11uhfircTC/u62KjneHW0Z/UbrqnCBUIJY+3L7Myn+aPqpcYj/PnVnLWjBh/cfUC7n77Yve418Xk4Gy2Gw92s3qeFUBf1VTJYCpLKpsbM7O7Ph7Ou9YZsfry6z2UhQNukV8h6m0LqL1A76h0NkcinaOq1KokH05l6bEFoiIapCIadN1gzpqjIb/rlinUj8rpiAvWhnvw6BDNXRNvagPJjNuPa+nMcvs+fXzoOy/xVz/dPOG1Be+XyLcgHBzx87ZzH/Sk7Bb6Pvs7B4kEre3hVLiZjvQlSKRzZHOWe+x4WsUrx44KhJLH1Uvq+PXdV45Jhx2PSNDPEx+/gk+946w810g07FgQI0LjBMRDAR9/evl8AFcoAM6bky8Q3iK/2tiIBdE5kOTi+VUTumIcgSgUqHZiCbFIgGgoYFkQdjpnZTRERTTkPkE7VoZjQWxp6eXCf3hqTN3CttY+ysIBQn4fHQNJ2voStPcnJgwWOy4mgLNmxBCB3+09yqbmHrcifjo4/aqCfskXCHtzbfa0JfH2nxodh0hmsrT2DLPcFi1HPE8mTrHmYMqKqWiw/NSgAqGMwWn8dzxEg2NjEM6G//7Vc9wNvKG8hFkVJcyIR/LmW8CIBRHy+ygvCboWBMBlC/OLAEfjXOstlutPpMlkc26BWmk4QCToL+hicqqzvRZEIp3jYJflpvm7Ndvynmq3H+5jaUOc6rIQu470k8kZ0llD1wSbq1cgSsMB5tWU8pONzRhjNT6c7phWx4JYPquc3e0DbnPBbntzzbMgPC6mzlEWREu3leLqVtf3DnPdvzzHxoOTz8RIpLN8/dm9x2V15HLGFfYuO+CvFsSpQQVCKQrRkJ9I0OdmAYHVpfVzNy3lf71tcd65d711IXdes3DMPRwRqY2FERFKQ36c2r3LFk0sEI4Yvd41RM9Qij//wUbO+8Kv+PtHd7jFdrFwgJKQj4TnCbWiNJjXPNERCKeA7nCPvXENpvjsz0bmVhzuGWZ2VZTqslBeJ1nnSXg0xhj6E2l37CxYI2Qd8TKmcKfaiXBiEBfNqyaVybluMscCaPb0phrKsyDyBcKpHznbnhuy/kA3O4/0s83O0pqIDQe6uffxndz90KuTptqOJpszPPdaB52DSTee48wfGU5nJ+y6qxQHFQilKLz7gkY+ce3ivGpsn0/48GXzxgw7eu+Fc/jgxXPH3MMplnPSUUWEeEmQGfHIpFZOScjP6qYqHlrfzL2P7+TJbW0srCvjx+ub+dFLryMCS2fGKbEtiK6hFAGfEAsHqPCsz6k0d4LuTrD07msX88S2I/zb03swxnB0MEV1WYjq0nCey6Z9nFTbZCZHOmvypgk6tStObGX0AKjJcCyISxdamVebmy2hKmhBJDMEfEI44BtTTe0IhpNy7PS2msoG7VhMT+1o4/svTq8J4bOvtfOH33qJR14dqSr3xke6T4Gr60xHBUIpCpcurOGOKxYc1z0cK6DW08xwYW0ZN5zTMKVU0DuvWciRvgQPrm/m/avn8H9//1yG01m+/8Lr3LC8gbnVpa5A9AylqIgGERFXIMpLgoQDlqvMmd3R2jNMyO/jzmsWcv3yGdy/dg/9yQzJTI7q0pArKA5HegunxjqWQtwjEMtsn/+7z7eywUaPkJ0MJwaxfGY5sXCAzfaQo5Eg9RDGWE/mQ6kspeEANWXhMem7jstpblUUv0/YZWeNTTaeFUbiHdWlIX6ze3qFf05Q/5fbj4ysxbM2dTOdfFQglDcs1aUhAj6hNjay6T70Z5fw1zeePaXrL19Uw8rZFZSG/Nz11kUsn1XuZkp99EpLvCJBP8MpK0jtuJacWId3yJJjQbT2DlNuC8nqeVWWK8feyKvLwm7VeMjvQ4S8RoJguZbu+O4Gfvji68DIVEGwemm99aw6bl89h1kVJWMsiGzO8KFvv8SvdxSeoucNvp/TWM7mll7SdsylqjREIp1zrYXBZIZoyE9NWaiABZHCJyPxmHTWEpVCApHJ5nh6Z5srPI4YLZtVzqHu6aWmOkOg1h8YaeDYqRbEKaWoAiEi14nILhHZIyL3THDebSJiRGSV/bpaRJ4RkQER+Vox16i8cfH5hM/dvJT3r56bd8w/ThPB0YgI//HB83noo5e4m/3f3bKMv715Kec0Wk/rJSE/CTtIXWlbDpW2UHgtl9KQY0EkqLAFxImR7LCbEloWhHXNjPII1aVh2kcJxJZDvfxyexs/eskRiBELoiwc4D8/dCELasuYX1s6ZkDR5pYe1u7q4NEthRv7DSQzlAT9BPw+zmksZ8fhPtdFc46dlebMyBhKZW2BCHN0tAUxkKKqNIzPlz/7PJke22PqV9vb+PB3Nri1KT1DaeKRAHOqSqZdu+AEpr3t2zveYBbELzYd4oP/78VTvYyTRtEEQkT8wP3A9cBS4H0isrTAeTHgLsD7W08AfwN8sljrU94c/OElTe5mfiw0lJe4rhuAlbMr+ONLRwY0jbiYRiwIx8WUZ0HYQequwZT7vpMp5QpEWci1IBrKI8woD4+xIB7bYrlPnOPjjZtdUFvGvo4B98kc4LnXOu3PKzyjoT+RocwWnBWzKkhnDS/sswY7rbB/h832U/1gKuO6mEbXQRwdSLrzO7yZY4ViEE5cw6k36RpMUVkaYlZFlN7hiQv+drf18+HvrHfv6x1j6zwDdPSPiEL3cQqE93c5Hvc+vpOPfHfDuO+/8noPz+/pJJ3NsW5PJ1tapp+OXIjHthyesPX6qaKYFsRqYI8xZp8xJgU8CNxa4LwvAvdhiQIAxphBY8zz3mOKUgxKPGmuzvyLihLbgoiNtSAAyu33nRiJ04Opuizs1no0lEeYEY/kpdkaY3h8a/7Tv9fF5GV+bSmDqWze5u1MCdzT3l+wY6y3Q60jCM/aY1tXNFYQ8ImbYeVYELUxK6jufWp3Au7W78IjEAVcTI5byNm8u4dSVERDzLJnoU/kZlq39yhP72xnd/vI/A6nwLKx0iqwzLMgjrEWYu2udq768jMs/t+PT7oJv9rcw693tI1b++G48foTGf52zTbufWLHMa1pNPc9sZOvPb1nWtdksjm++dy+omZ3FVMgZgHewcUt9jEXETkPmG2MeeRYPkBE7hCRDSKyoaOj49hXqpyxREJ+u5I6TUWptRlWlYUI+oVZFSN1GVHPk/7oGIXjXvEGqWeUl1Afj+QV6m0/3MfBo0O8Y1m9e8zrYvLifLYzOa93OM2m5h6aqqOks6ZghtNAIu1aEI2VJcQiAdbttSyIhvIIly6s4bEth93mfaWhAHXxMNmcyXPfHB1IUl1qfTdvym8hgXC+nxMf6BlKUxUNuus/1DN+qq6T7dXaM4wxhsO9w259i5NB5cQgRKzz7vrRK64gTwVjDPc+vpN01jCjPMLfrtlGMjP+hto9lCJn4LcFRurCSHpwfyJNz3Ca/SdgTnk6m6O5ezgvy2wqvPx6D//w2A6eKaLlUUyBKOQodh9TRMQHfAW4+1g/wBjzgDFmlTFmVW1t7bHeRjmDKQn6OTqYIpXNubGHsnCA//6LS3n/RXPc87wWhONiigT9VESDVkZQyE8k6HfjEnOqotTHI3QNpvjFpkP0JdK8uM8qNPvzq0ZqPuLjWBAzK0aewAeTGb706A6yOeNWoDtuLS8DSWvTByv+cnZD3LVAKktD3LiigZbuYTa39DKUzBINB9w4i9dSOTowYkE4YhjwScHCPcct5FgQXYMpKqMhGqdgQTjXtHQP0zecIZHOcWFTFQ3lEVY0ViAyksVUH4vw5NYjrHm1lbsefCVvLRsPdrFunFbpWw71svNIPx+9agFfetc5HDw6xHd+e4BUJsdnfrY5b9gUjPSd+s3uwg+cjsusbzhD73Ca1t7EcT/BN3cNkc0Z2voTE4rXaBxxnqiJ5PFSTIFoAWZ7XjcC3v4EMWA5sFZEDgAXA2ucQLWinAyiIT+Oa9rbwXbZzPK8BoBOFhPku12cQsBqd/BShB9+5CJ+7/xZ7vChjz24ia+v3cvmlh7q42FWNJa7tQ5O+uxoHIFo7RnmMz/bwkMbm/nwpfN474WzCQV8BQXCG4MAq/DOoSoa4u1L6wn4hMe2HLZiECF/3mAmsCqh+5MZ11XmCMT82tKCG6HjQutyLQjLxVRbFibk99EyQaDauaa1J8HhPuu8hooIT3z8Cu68eiGloQBJ25U2syJCfzJjzQZvH+Df7Kl5iXSWj37/Zb74aGFXz0MbmgkHfNxy7kwuX1TLytkV/HJ7G5tbevjRS838ant+RpgzIOm51zoKxiycJoft/QnXzedU1zvkcob3f/MFnt5ZONtsNPvtJorGWL+LqeIKRIFeWieKYgrEemCRiMwTkRBwO7DGedMY02uMqTHGNBljmoAXgFuMMeNHiBTlBHN2Q5z6eJi/vGYhN65oGPe8UMBH0G8Zxd5COsdi8DYOfMuCGiJBP9cvn8Gzn7qK8+ZU8PyeTja39NpPxsLZDTGiISvjqBDxSICycIBDPcNsPNjNTStm8rmblxL0+1hcX8b6A91jJtB5YxAwUngXDvgoCfmpiIa4ZEE1a3d1WBZEKODGUTr6k/zpf63n/9ozqJ1g+60rZ/Lp65bQWBkd42LytsToHkyTzGQZTGWpKg3i8wkNFZEJLQinjYY15ta6z4x4hPKSIKGAz41HlAT9rgBfsbiWVXMr3VTYn2xsoaM/OW7M4ImtbVy7tN4VupWzK9je2sdmO7jsrXRPpLMk0jlmV5XQ2pvgwNEhHtrQzN0PjXTlHbCbHHr7Wo12M/UOp1m39yjrRrmpfrKhmY8/+MqYNe73dNlt6Z569bxj9XUWMburaAJhjMkAdwJPAjuAh4wx20TkCyJyy2TX21bFPwMfEpGWQhlQinK83HBOAy9+9m3c/fYlY2YijMaxIuJ5FoS1wdaUje0sKyLMrS7lysW1bDnUy77OQbcO48ZzZroDmgohIsysiLC7vZ9DPcOcNSPmvvf2pTPY1NzDJf/4NL/YdMg9PpAsbEFUeuIISxvi7O8ctLOY/G4cpblriF/vbOe7L1jVz86GPL+2jD+/aiElIf8YgfC2xOgeSo20K7E/b1bFxKmuTtyitdcjEJ42845AlIb9bgLBlUtqqYhabeOzOcM3nt0LFG5JPpzK0jmQdFuGgNWnajid5ZHNljPjsEcgnPVcschyV+843MeaTa08/HIL++yYj2NBNHuEb19nvkA492kb9WT/6JbDPGrHgPoTaddNtr9z0E3dnqwDsBcnc+zNakFgjHnMGLPYGLPAGPMP9rHPGWPWFDj3Kq/1YFsWVcaYMmNMozFmezHXqiiT4aSkegO3jotmotbjly2scd1YTnbR+y+aw/3vP3/Cz5tVUcL6/daTsuOuAvjLaxby339xKUsb4nzswU384MWDGGMYsEeYOiysKyPgEyo9a1tQW0YqmyNnLMGLBP3EIgFe2N+FMbhuk9EV4SVBP4lRLiZnUw/6ha7BVF7DQ2f9E1kQ3iC1k/br7eDrCHI0FHC/w5WLaomXBOgbTtM5kKSle5j6eJjhdHZMjKTVjo/MrBi55/JZlli8/LpVZe5NrXVEZlVTJT6BXUf62WkPp3p0s5V95mQxeS2IA+MIxOgamNeO9JPOGvqGM9z+wAv8/aPWlnbg6CDLZsYJ+ITm7iEe3tgypjalEI71Nt5UwBOBVlIryhRxnmi9MYj6UTGIQpxrV3ODVZ8wVWZWlJCyO7Iurh+xIESElbMr+MFHLuLcxnK+/8LrJDM5MjmTZ0FEgn4W1cfy0nUXeITGiX/UxsJssjdMh5rS/O9TEvSTyOTY3trHn3xnPUOpjCsQC+tidA+l3EyoSjsbbHZVlPb+ZN5wImMMX392L4d7h+ketPpfdQ6kONA5SE1ZmFBgZEty1hcN+bntgkb+5qalzKmOEo8E6Uuk3Y6xzhzxvlEdZFtt62Wmp0vwgtqyvM84UsCCqI9HaKouZd3eTrfK/BFXICwRcjKOSkP+PBcR4LaO984i6Uuk3Yy0joEku9sHeMFOWtjfMciC2jJmVpTw6ObD3P2TV3n45RYmw7Ug3qRBakU5rYi6FsTYGET1BBZE0O/j0oU1LK4vG9OocCKcQHUo4GNOgdGlQb+Pm1bMZMfhPrbbQevYqMK7f719JX9384h3dkHtSPt15wm9LhYmlc0h4g26j7Ig7HTgF/Yd5dc723nutU73qX9pQ5zuwbT7BO5YEI7Vs8/jo9/TPsC9j+/kO789QCZnXOF7YtsRzpkV936kx4Lws7CujD+5zCpwjJcEGUhm3Cwop6V8z3gC4UlXDvp9nG276+bXlNLWn3RrQLzrX1wfc+Mc1y6tZ1dbP7vb+l03mxMrOKex3B3P6lDIgnjtyEhx4/7OQVKZHHs7BugcSNLam2BeTSmzq0p43bZMphKsdu4/1Rnox4IKhKJMkVLXghjrYhq9oY7mvttW8N0PXzStz3NqCRbUlo3bXuQd9gTAn9lPnGWj6ioW18eY7+l8WxENufESxyJy3DqNlSVct3wG5SVB9z0HZ26Gswk/taONI70JAj5hYZ3ltnI2zdECscfTdHCXPcho/QHr6dlpAZJI57h99UhaMYxYEKWjRC8eCWDMSGfdOfbUwtFxiEM9CUTy4xpg9YkCeOvZdWRzhkPdw7y0v8sTQwmyxBPzed9qKxnT2+68z262uLqpis6BFM97GhM69xlMZd202F2eAU7O9EFj4Ed2T65FdWU0Vow8BIzXJt4hkc7SZ7sUB5KZac8OmSoqEIoyRaKhACL5xW0rGiu48+qFXLOkfoIrrY159EY1Gc6T7+L68Vubz6mOctaMGA9vtILV3vkS4+EIhiMCjgtqQW0Zn77uLH7+528Z0y23xE75dTJnntnZzmttA9THI644OpaCY2E1VZfi94lbKQ24nWGdiXnL7ZhMXSzszkMf+cxA3mc7OLUjTkB3bpVtQYzKZGrtGaY+FhmTfHDbBY28/6I5XDTPaot+7xM7eM83fsdWe+OujIZcgagpC7Oozvq5UIbRR66Yz8K6Mj7+403uxu8dEvXbPZ1c/6+/4YmtR1yR9xb6ff3ZvUSCPq5YXOvWjlREgxzuS9A1mOL+Z/YUrJofmdlhra1YbiYVCEWZIqVhP/FIMG/GRdDv45PvWDIt19FUcdpVeOMPhfjjS5torCzhDy6ey8XzqyY8F0YmBjpP5nUegSgLB/IsDocSez61NzD61I42rls+w80w2tcxaA+Ksjb0UMDH3OpovgVhC4TTIXZpQ5x4JMAHLpo7ZiMfxzhh6QAADuhJREFU14IosV47A5CcueeFXEzeALXD+XMq+dK7znEF+JfbrHqFF/YdJRL0EQn63d/5WTNiruCNrnSOhQPEIkH+/QPnk0xnufGrz/OPj+/IE6ofr29mx+E+frO7k2Uz44jADjvwHfQLg6ks1y6dQWk4wG2rGvncTUt561n1tPUmeGRzK19+chf/M2q8LYzUrTiZaqM78p4oVCAUZYpcu7Se2y+cPfmJJ4hZFSXcd9sK3jfK9TKa9144h1994kq++M7l4/Z28uK4fgpZEONRYp97pDdBY2UJFdEgt66cyWdvONvNMNrU0pM3gxws18loF5PXOKmLhXn2U1fzlwUmCnpjEF5GLIghRKxgOEDvUCGByB9j66XBtuicVN19HYOu+7CpOkosbLVNLwsHCPrFFYiA/YDgpDsvro/x3Kev5uL5VTyx9Qjdg2lCttj9dk+nu/5lM+NURUOu5XPxfMuCeefKmfZ6SvjwZfOYWRGhvT/hiun/e37/mKI9p0DRqXUpVqpr4UYwiqKM4aYVM7lpxcyT+pnvWXXiBemtZ9Xxu72dbvaPYzE48YBCOFZBe3+ChXVlfOtDF7obuJPim8rkuHVl/u9nYV0ZT+1oJ5XJkcnleL1riAvnVvGSHYOoLgvlVal7KQ0VtiBiHoGIR4LEIwECPnGroMEq4mvtTbgxmkJURINEgj4SnjbmjrUQ8Pt45K7LqCmzxt1WRkOui6k+HuFQz3Bep9vK0hDnz6nkgef2UVMWZn5tKTuP9JPM5LjhnBncdkEji+tjvHywh6ODKUpDfm5dOYuO/iSXL8pvEzSjPELOWBaNT6x6jM/+fCtXLal1v8+IBWH9m40eG3uiUAtCUc4wmmpK+X9/NLLBr5xdwbp7rpmwrbojEJ0DKeKRYN6m7riYROCdK/P6cbKoLkY2Zzh4dJDdbQMYAzecY21y4YBvTHzBi5M1NiYGYbuYjvQlKC8ZmQLoDVIfHUyRyuQmtCBEhIbyEkqCftc1581Qm1td6opTZTTkBsUdy8MrENb5UTI5w87DfcytjhKx3XIrGiu45qx6GiujbrymLh7htgsaeeLjV+Sl3Xrvv7djkBvOaeDshjg/2dDMXT96xXVftfcn3QQBUBeToihFZKKNFPI36dEbYywSwO8TLl1QMyYQv8gOsL+w7yib7cD0FYtriYb8VJeGJhwdW+qppPbiuJhyhrzxsE4M4l+eeo0rv/zMlL7XRfOquPncBtd68lade6ksHZms53xHR6gc5tjB8sFUlspoyM0OW+GxzJweV3Wx8etmZsRH1rx0ZpzHP3Y5//OXl5HM5PjpRitb7eDRQWZXRSkJ+SkLB4oWpFYXk6Iok1LiiQPERwmEzyd87qalXDC3csx1SxvirJpbyb88tdvtMNtUXcqC2jJykwzwcT5ztAvKm0XmiFVFNOTGIJ7e2U5pOMCymXHOmzNxYeK9714BWE39nPsUwiscjuiMFkon3RYsl1N9PMzrXUNuphaQZ0GMh1dknbjQ2Q1xLmyq5PsvHOTDl85jb/ugW9NSaGzsiUIFQlGUSZnIggD4o7c0FbxORPi7W5Zx89eeJ+jz8b0/WY3PJ3zi7YvJZCcWiNJxgtQBv9XIbyiVHRGIkiBH+hKkszl2Hu7njy9t4jM3TG12OYxkilWMk43mbVfidP0d/XuYEY8Q8vvs1vFBFtSWMZTK5rV0dyyI+gksiMqo1awwlcnlJQ584KK5fPzHm1h/oIv9nYNctcSKXXz+1uXuuNwTjQqEoiiTEplEICZi+axy/vFd51BeEnQb5129ZPxGhQ7RcGELAiw3k1cgyqNBdh7pZ3fbAKlszi2GmyqL68uIRwLMrykt+L53Ax4vBuH3CY1VJVY2VDTE526eSzqTL4LO/A2nwLIQVmzE6oQ712OVXLbIGqb035taSWVHxOPKxcWbhaMCoSjKpOS7mKa/bYyukp4K82vKmFkeKVgoGC8JcKTP42IqCdE7nHaL3ZbPjI+5ZiKioQC/+8xbxw2aOy6mSNDnuqEKCeXcqij7OgapjNrZWaM8Vo6LqX4CFxNY1ojfJ3m1ITVlYZqqozxi10UsqCssZicSFQhFUSZlMhdTMZhRHmHdZ95a8D3HbeO4hCqiVn+mTc09lIb8bn+m6TA6ndaLk8pbFg54PnNsvMLpmVVVWvh3tHRmnDlVUVY0Thwb+fR1SxhOja2gPn9OJT97xaqan18zft3KiUIFQlGUSfEKxHhjUk8mTqB6JEht/f387k6WzSzPq3Y/ETgWRDQU4KwZMb582//f3t3GSHXVcRz//lgehZW2UijQCrRiA5i6hU1T29jQ1FraF92a0IgPiMYE09BoE42C0WjqGzWpTUywlkZSGlFEKZE0jdUSpWkihYWgFJC4obVd2xRMDUobkIe/L+7ZZWa5sw9h79xl5vd5szNn7p3935N78597zj3n3MCd8y+cXuX9KTHV6uyePnkCL3z99gH/36JZ+SPib5yVJYj3TRxb1S9SFCcIMxvQuIpn9et1B9GfniepJqeRz9PTlN6vvf0ud8wbuH9jqC7vXSJ2NJK4v8YAxo62GZw8fbZmX8bFWpieyupv1PtwcoIwswGNGqXeUccjIkGMPz/+AeBj86ay5YGP0HX0xKA6wIeqp5N64tjaA/sg6ydYdfuF04YMl+untfLe8aOrZpstkhOEmQ3K+DEtnDx97oJxEGXo6SjvaVqSxKJZV9RsmrlYlXcQZRrdMootD9zS+7hs0QodSS1piaTDkrokre5nu6WSQlJ7RdmatN9hSXcVGaeZDWzCmBbGjh5V9chrWVr73EEU/v/GZfM99R3VXYa501rr0v8ABd5BSGoB1gJ3At3Abknb+q4tLakV+DLwUkXZfGAZsACYATwv6YMRUcyqGGY2oAljWkZEBzVk4yheOfbOgI+LDhdJTG0dV3MqjkZV5P3STUBXRBwBkLQJ6AAO9tnue8APga9VlHUAmyLiFPCKpK70fX8uMF4z68f4MS30M3VSXV1/VSs/WHpDXf/nEyva69a0M1IU2cQ0E3i94n13Kusl6Ubgmoh4Zqj7pv1XSuqU1Hns2LHhidrMck0Y2zIi+h/KsmDG5LrdsYwURd5B5P3W6B13LmkU8Cjw+aHu21sQsQ5YB9De3t7/xC5mdlFW3nZt2SFYnRWZILqByoeFrwYq185rBT4E/ClN+XsVsE3SvYPY18zqrL/Fd6wxFdnEtBuYK2mOpLFknc7bej6MiOMRMSUiZkfEbGAncG9EdKbtlkkaJ2kOMBfYVWCsZmbWR2F3EBFxRtKDwHNAC7A+Ig5IehjojIht/ex7QNJmsg7tM8AqP8FkZlZf6rsY9qWqvb09Ojs7yw7DzOySImlPRLTnfeYlR83MLJcThJmZ5XKCMDOzXE4QZmaWywnCzMxyNcxTTJKOAf+4iK+YAvxrmMK51Lkuqrk+qrk+ql3q9TErIq7M+6BhEsTFktRZ61GvZuO6qOb6qOb6qNbI9eEmJjMzy+UEYWZmuZwgzltXdgAjiOuimuujmuujWsPWh/sgzMwsl+8gzMwslxOEmZnlavoEIWmJpMOSuiStLjueMkh6VdJ+SfskdaayKyT9QdLf09/Ly46zKJLWSzoq6eWKstzjV+bH6Xz5q6SF5UVejBr18V1J/0znyD5J91R8tibVx2FJd5UTdXEkXSPpj5IOSTog6SupvOHPkaZOEJJagLXA3cB84FOS5pcbVWluj4i2iue5VwPbI2IusD29b1RPAkv6lNU6/rvJFrCaC6wEHqtTjPX0JBfWB8Cj6Rxpi4hnAdL1sgxYkPb5SbquGskZ4KsRMQ+4GViVjrvhz5GmThDATUBXRByJiP8Bm4COkmMaKTqADen1BuC+EmMpVES8ALzdp7jW8XcAT0VmJ3CZpOn1ibQ+atRHLR3Apog4FRGvAF1k11XDiIg3I2Jvev1f4BAwkyY4R5o9QcwEXq94353Kmk0Av5e0R9LKVDYtIt6E7AIBppYWXTlqHX8znzMPpiaT9RVNjk1VH5JmAzcCL9EE50izJwjllDXjc7+3RsRCslvjVZJuKzugEaxZz5nHgOuANuBN4JFU3jT1IWkSsAV4KCL+09+mOWWXZJ00e4LoBq6peH818EZJsZQmIt5If48CW8maCN7quS1Of4+WF2Epah1/U54zEfFWRJyNiHPAE5xvRmqK+pA0hiw5bIyIp1Nxw58jzZ4gdgNzJc2RNJass21byTHVlaSJklp7XgMfB14mq4cVabMVwG/LibA0tY5/G/C59KTKzcDxnmaGRtanDf0TZOcIZPWxTNI4SXPIOmZ31Tu+IkkS8DPgUET8qOKjhj9HRpcdQJki4oykB4HngBZgfUQcKDmsepsGbM2uAUYDv4iI30naDWyW9EXgNeD+EmMslKRfAouBKZK6ge8A3yf/+J8F7iHrjH0X+ELdAy5YjfpYLKmNrKnkVeBLABFxQNJm4CDZ0z6rIuJsGXEX6FZgObBf0r5U9k2a4BzxVBtmZpar2ZuYzMysBicIMzPL5QRhZma5nCDMzCyXE4SZmeVygjAbASQtlvRM2XGYVXKCMDOzXE4QZkMg6bOSdqU1ER6X1CLphKRHJO2VtF3SlWnbNkk70wR3WyvWC/iApOcl/SXtc136+kmSfiPpb5I2phG8ZqVxgjAbJEnzgE+STW7YBpwFPgNMBPamCQ93kI08BngK+EZE3ADsryjfCKyNiA8Dt5BNfgfZLKEPka1Nci3ZCF6z0jT1VBtmQ3QHsAjYnX7cTyCboO0c8Ku0zc+BpyVNBi6LiB2pfAPw6zTv1cyI2AoQEScB0vftioju9H4fMBt4sfjDMsvnBGE2eAI2RMSaqkLp232262/+mv6ajU5VvD6Lr08rmZuYzAZvO7BU0lToXZN4Ftl1tDRt82ngxYg4Dvxb0kdT+XJgR1pHoFvSfek7xkl6T12PwmyQ/AvFbJAi4qCkb5GtvjcKOA2sAt4BFkjaAxwn66eAbAron6YEcITzs3ouBx6X9HD6joadKdcubZ7N1ewiSToREZPKjsNsuLmJyczMcvkOwszMcvkOwszMcjlBmJlZLicIMzPL5QRhZma5nCDMzCzX/wGszenYCiw2PQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/225\n",
      "1754/1754 [==============================] - 5s 3ms/step - loss: 0.4710 - binary_accuracy: 0.8218\n",
      "Epoch 2/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4499 - binary_accuracy: 0.8220\n",
      "Epoch 3/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4460 - binary_accuracy: 0.8218\n",
      "Epoch 4/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4458 - binary_accuracy: 0.8233\n",
      "Epoch 5/225\n",
      "1754/1754 [==============================] - 1s 832us/step - loss: 0.4330 - binary_accuracy: 0.8261\n",
      "Epoch 6/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4328 - binary_accuracy: 0.8222\n",
      "Epoch 7/225\n",
      "1754/1754 [==============================] - 1s 830us/step - loss: 0.4378 - binary_accuracy: 0.8228\n",
      "Epoch 8/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4361 - binary_accuracy: 0.8234\n",
      "Epoch 9/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4308 - binary_accuracy: 0.8255\n",
      "Epoch 10/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.4301 - binary_accuracy: 0.8220\n",
      "Epoch 11/225\n",
      "1754/1754 [==============================] - 1s 830us/step - loss: 0.4284 - binary_accuracy: 0.8255\n",
      "Epoch 12/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4325 - binary_accuracy: 0.8254\n",
      "Epoch 13/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.4244 - binary_accuracy: 0.8272\n",
      "Epoch 14/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4256 - binary_accuracy: 0.8240\n",
      "Epoch 15/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.4209 - binary_accuracy: 0.8294\n",
      "Epoch 16/225\n",
      "1754/1754 [==============================] - 1s 827us/step - loss: 0.4200 - binary_accuracy: 0.8284\n",
      "Epoch 17/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4289 - binary_accuracy: 0.8272\n",
      "Epoch 18/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4169 - binary_accuracy: 0.8271\n",
      "Epoch 19/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4304 - binary_accuracy: 0.8226\n",
      "Epoch 20/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4309 - binary_accuracy: 0.8208\n",
      "Epoch 21/225\n",
      "1754/1754 [==============================] - 1s 817us/step - loss: 0.4200 - binary_accuracy: 0.8301\n",
      "Epoch 22/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4242 - binary_accuracy: 0.8286\n",
      "Epoch 23/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4179 - binary_accuracy: 0.8310\n",
      "Epoch 24/225\n",
      "1754/1754 [==============================] - 1s 814us/step - loss: 0.4187 - binary_accuracy: 0.8319\n",
      "Epoch 25/225\n",
      "1754/1754 [==============================] - 1s 829us/step - loss: 0.4287 - binary_accuracy: 0.8207\n",
      "Epoch 26/225\n",
      "1754/1754 [==============================] - 1s 827us/step - loss: 0.4243 - binary_accuracy: 0.8287\n",
      "Epoch 27/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4207 - binary_accuracy: 0.8246\n",
      "Epoch 28/225\n",
      "1754/1754 [==============================] - 1s 821us/step - loss: 0.4179 - binary_accuracy: 0.8322\n",
      "Epoch 29/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4182 - binary_accuracy: 0.8291\n",
      "Epoch 30/225\n",
      "1754/1754 [==============================] - 1s 815us/step - loss: 0.4235 - binary_accuracy: 0.8275\n",
      "Epoch 31/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.4172 - binary_accuracy: 0.8287\n",
      "Epoch 32/225\n",
      "1754/1754 [==============================] - 1s 839us/step - loss: 0.4165 - binary_accuracy: 0.8272\n",
      "Epoch 33/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4196 - binary_accuracy: 0.8272\n",
      "Epoch 34/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4167 - binary_accuracy: 0.8334\n",
      "Epoch 35/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4158 - binary_accuracy: 0.8277\n",
      "Epoch 36/225\n",
      "1754/1754 [==============================] - 1s 822us/step - loss: 0.4220 - binary_accuracy: 0.8275\n",
      "Epoch 37/225\n",
      "1754/1754 [==============================] - 1s 816us/step - loss: 0.4141 - binary_accuracy: 0.8308\n",
      "Epoch 38/225\n",
      "1754/1754 [==============================] - 1s 816us/step - loss: 0.4115 - binary_accuracy: 0.8334\n",
      "Epoch 39/225\n",
      "1754/1754 [==============================] - 1s 828us/step - loss: 0.4222 - binary_accuracy: 0.8235\n",
      "Epoch 40/225\n",
      "1754/1754 [==============================] - 1s 836us/step - loss: 0.4216 - binary_accuracy: 0.8311\n",
      "Epoch 41/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4186 - binary_accuracy: 0.8254\n",
      "Epoch 42/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4144 - binary_accuracy: 0.8303\n",
      "Epoch 43/225\n",
      "1754/1754 [==============================] - 1s 830us/step - loss: 0.4183 - binary_accuracy: 0.8304\n",
      "Epoch 44/225\n",
      "1754/1754 [==============================] - 1s 827us/step - loss: 0.4174 - binary_accuracy: 0.8292\n",
      "Epoch 45/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4175 - binary_accuracy: 0.8301\n",
      "Epoch 46/225\n",
      "1754/1754 [==============================] - 1s 830us/step - loss: 0.4056 - binary_accuracy: 0.8361\n",
      "Epoch 47/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4186 - binary_accuracy: 0.8297\n",
      "Epoch 48/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4153 - binary_accuracy: 0.8313\n",
      "Epoch 49/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4123 - binary_accuracy: 0.8308\n",
      "Epoch 50/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4157 - binary_accuracy: 0.8267\n",
      "Epoch 51/225\n",
      "1754/1754 [==============================] - 1s 827us/step - loss: 0.4159 - binary_accuracy: 0.8288\n",
      "Epoch 52/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4204 - binary_accuracy: 0.8244\n",
      "Epoch 53/225\n",
      "1754/1754 [==============================] - 1s 827us/step - loss: 0.4200 - binary_accuracy: 0.8256\n",
      "Epoch 54/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4145 - binary_accuracy: 0.8312\n",
      "Epoch 55/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4175 - binary_accuracy: 0.8223\n",
      "Epoch 56/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4165 - binary_accuracy: 0.8272\n",
      "Epoch 57/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4254 - binary_accuracy: 0.8241\n",
      "Epoch 58/225\n",
      "1754/1754 [==============================] - 1s 828us/step - loss: 0.4165 - binary_accuracy: 0.8271\n",
      "Epoch 59/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4076 - binary_accuracy: 0.8280\n",
      "Epoch 60/225\n",
      "1754/1754 [==============================] - 1s 830us/step - loss: 0.4111 - binary_accuracy: 0.8294\n",
      "Epoch 61/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4120 - binary_accuracy: 0.8286\n",
      "Epoch 62/225\n",
      "1754/1754 [==============================] - 1s 822us/step - loss: 0.4110 - binary_accuracy: 0.8311\n",
      "Epoch 63/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4195 - binary_accuracy: 0.8247\n",
      "Epoch 64/225\n",
      "1754/1754 [==============================] - 1s 837us/step - loss: 0.4088 - binary_accuracy: 0.8248\n",
      "Epoch 65/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4149 - binary_accuracy: 0.8279\n",
      "Epoch 66/225\n",
      "1754/1754 [==============================] - 1s 831us/step - loss: 0.4117 - binary_accuracy: 0.8333\n",
      "Epoch 67/225\n",
      "1754/1754 [==============================] - 1s 828us/step - loss: 0.4119 - binary_accuracy: 0.8306\n",
      "Epoch 68/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4103 - binary_accuracy: 0.8316\n",
      "Epoch 69/225\n",
      "1754/1754 [==============================] - 1s 830us/step - loss: 0.4101 - binary_accuracy: 0.8356\n",
      "Epoch 70/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4129 - binary_accuracy: 0.8313\n",
      "Epoch 71/225\n",
      "1754/1754 [==============================] - 1s 817us/step - loss: 0.4119 - binary_accuracy: 0.8309\n",
      "Epoch 72/225\n",
      "1754/1754 [==============================] - 1s 816us/step - loss: 0.4066 - binary_accuracy: 0.8297\n",
      "Epoch 73/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4072 - binary_accuracy: 0.8327\n",
      "Epoch 74/225\n",
      "1754/1754 [==============================] - 1s 817us/step - loss: 0.4139 - binary_accuracy: 0.8303\n",
      "Epoch 75/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4095 - binary_accuracy: 0.8324\n",
      "Epoch 76/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4092 - binary_accuracy: 0.8347\n",
      "Epoch 77/225\n",
      "1754/1754 [==============================] - 1s 850us/step - loss: 0.4115 - binary_accuracy: 0.8312\n",
      "Epoch 78/225\n",
      "1754/1754 [==============================] - 1s 827us/step - loss: 0.4076 - binary_accuracy: 0.8311\n",
      "Epoch 79/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4114 - binary_accuracy: 0.8335\n",
      "Epoch 80/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4060 - binary_accuracy: 0.8312\n",
      "Epoch 81/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4067 - binary_accuracy: 0.8280\n",
      "Epoch 82/225\n",
      "1754/1754 [==============================] - 1s 831us/step - loss: 0.4132 - binary_accuracy: 0.8290\n",
      "Epoch 83/225\n",
      "1754/1754 [==============================] - 1s 830us/step - loss: 0.4109 - binary_accuracy: 0.8310\n",
      "Epoch 84/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4086 - binary_accuracy: 0.8260\n",
      "Epoch 85/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.4107 - binary_accuracy: 0.8306\n",
      "Epoch 86/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4115 - binary_accuracy: 0.8280\n",
      "Epoch 87/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4106 - binary_accuracy: 0.8334\n",
      "Epoch 88/225\n",
      "1754/1754 [==============================] - 1s 828us/step - loss: 0.4036 - binary_accuracy: 0.8315\n",
      "Epoch 89/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4072 - binary_accuracy: 0.8342\n",
      "Epoch 90/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.4118 - binary_accuracy: 0.8287\n",
      "Epoch 91/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4105 - binary_accuracy: 0.8318\n",
      "Epoch 92/225\n",
      "1754/1754 [==============================] - 1s 827us/step - loss: 0.4099 - binary_accuracy: 0.8300\n",
      "Epoch 93/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4124 - binary_accuracy: 0.8281\n",
      "Epoch 94/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4077 - binary_accuracy: 0.8295\n",
      "Epoch 95/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4115 - binary_accuracy: 0.8284\n",
      "Epoch 96/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4054 - binary_accuracy: 0.8326\n",
      "Epoch 97/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4107 - binary_accuracy: 0.8295\n",
      "Epoch 98/225\n",
      "1754/1754 [==============================] - 1s 836us/step - loss: 0.4170 - binary_accuracy: 0.8309\n",
      "Epoch 99/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4100 - binary_accuracy: 0.8297\n",
      "Epoch 100/225\n",
      "1754/1754 [==============================] - 1s 832us/step - loss: 0.4099 - binary_accuracy: 0.8271\n",
      "Epoch 101/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4083 - binary_accuracy: 0.8277\n",
      "Epoch 102/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4106 - binary_accuracy: 0.8284\n",
      "Epoch 103/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4087 - binary_accuracy: 0.8265\n",
      "Epoch 104/225\n",
      "1754/1754 [==============================] - 1s 832us/step - loss: 0.4088 - binary_accuracy: 0.8329\n",
      "Epoch 105/225\n",
      "1754/1754 [==============================] - 1s 827us/step - loss: 0.4112 - binary_accuracy: 0.8308\n",
      "Epoch 106/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.4206 - binary_accuracy: 0.8230\n",
      "Epoch 107/225\n",
      "1754/1754 [==============================] - 1s 828us/step - loss: 0.4057 - binary_accuracy: 0.8327\n",
      "Epoch 108/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4114 - binary_accuracy: 0.8335\n",
      "Epoch 109/225\n",
      "1754/1754 [==============================] - 1s 829us/step - loss: 0.4075 - binary_accuracy: 0.8310\n",
      "Epoch 110/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4079 - binary_accuracy: 0.8334\n",
      "Epoch 111/225\n",
      "1754/1754 [==============================] - 1s 822us/step - loss: 0.4085 - binary_accuracy: 0.8298\n",
      "Epoch 112/225\n",
      "1754/1754 [==============================] - 1s 827us/step - loss: 0.4035 - binary_accuracy: 0.8288\n",
      "Epoch 113/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4147 - binary_accuracy: 0.8298\n",
      "Epoch 114/225\n",
      "1754/1754 [==============================] - 1s 828us/step - loss: 0.4082 - binary_accuracy: 0.8305\n",
      "Epoch 115/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4092 - binary_accuracy: 0.8297\n",
      "Epoch 116/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.4166 - binary_accuracy: 0.8276\n",
      "Epoch 117/225\n",
      "1754/1754 [==============================] - 1s 832us/step - loss: 0.4051 - binary_accuracy: 0.8316\n",
      "Epoch 118/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4103 - binary_accuracy: 0.8286\n",
      "Epoch 119/225\n",
      "1754/1754 [==============================] - 1s 829us/step - loss: 0.4055 - binary_accuracy: 0.8335\n",
      "Epoch 120/225\n",
      "1754/1754 [==============================] - 1s 831us/step - loss: 0.4027 - binary_accuracy: 0.8302\n",
      "Epoch 121/225\n",
      "1754/1754 [==============================] - 1s 828us/step - loss: 0.4031 - binary_accuracy: 0.8341\n",
      "Epoch 122/225\n",
      "1754/1754 [==============================] - 1s 815us/step - loss: 0.4051 - binary_accuracy: 0.8295\n",
      "Epoch 123/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4020 - binary_accuracy: 0.8368\n",
      "Epoch 124/225\n",
      "1754/1754 [==============================] - 1s 829us/step - loss: 0.4046 - binary_accuracy: 0.8341\n",
      "Epoch 125/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4057 - binary_accuracy: 0.8293\n",
      "Epoch 126/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4022 - binary_accuracy: 0.8336\n",
      "Epoch 127/225\n",
      "1754/1754 [==============================] - 1s 827us/step - loss: 0.4030 - binary_accuracy: 0.8304\n",
      "Epoch 128/225\n",
      "1754/1754 [==============================] - 1s 817us/step - loss: 0.4029 - binary_accuracy: 0.8368\n",
      "Epoch 129/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4042 - binary_accuracy: 0.8343\n",
      "Epoch 130/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4076 - binary_accuracy: 0.8325\n",
      "Epoch 131/225\n",
      "1754/1754 [==============================] - 1s 831us/step - loss: 0.4007 - binary_accuracy: 0.8355\n",
      "Epoch 132/225\n",
      "1754/1754 [==============================] - 1s 828us/step - loss: 0.4045 - binary_accuracy: 0.8341\n",
      "Epoch 133/225\n",
      "1754/1754 [==============================] - 1s 843us/step - loss: 0.4108 - binary_accuracy: 0.8331\n",
      "Epoch 134/225\n",
      "1754/1754 [==============================] - 1s 850us/step - loss: 0.4036 - binary_accuracy: 0.8310\n",
      "Epoch 135/225\n",
      "1754/1754 [==============================] - 1s 853us/step - loss: 0.4094 - binary_accuracy: 0.8297\n",
      "Epoch 136/225\n",
      "1754/1754 [==============================] - 2s 872us/step - loss: 0.4065 - binary_accuracy: 0.8335\n",
      "Epoch 137/225\n",
      "1754/1754 [==============================] - 2s 881us/step - loss: 0.4029 - binary_accuracy: 0.8317\n",
      "Epoch 138/225\n",
      "1754/1754 [==============================] - 2s 867us/step - loss: 0.4062 - binary_accuracy: 0.8307\n",
      "Epoch 139/225\n",
      "1754/1754 [==============================] - 2s 860us/step - loss: 0.4108 - binary_accuracy: 0.8352\n",
      "Epoch 140/225\n",
      "1754/1754 [==============================] - 1s 828us/step - loss: 0.4049 - binary_accuracy: 0.8271\n",
      "Epoch 141/225\n",
      "1754/1754 [==============================] - 1s 836us/step - loss: 0.4064 - binary_accuracy: 0.8333\n",
      "Epoch 142/225\n",
      "1754/1754 [==============================] - 1s 832us/step - loss: 0.4056 - binary_accuracy: 0.8330\n",
      "Epoch 143/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4037 - binary_accuracy: 0.8363\n",
      "Epoch 144/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4002 - binary_accuracy: 0.8349\n",
      "Epoch 145/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1754/1754 [==============================] - 1s 836us/step - loss: 0.4082 - binary_accuracy: 0.8258\n",
      "Epoch 146/225\n",
      "1754/1754 [==============================] - 1s 830us/step - loss: 0.4028 - binary_accuracy: 0.8351\n",
      "Epoch 147/225\n",
      "1754/1754 [==============================] - 1s 829us/step - loss: 0.4073 - binary_accuracy: 0.8310\n",
      "Epoch 148/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4103 - binary_accuracy: 0.8318\n",
      "Epoch 149/225\n",
      "1754/1754 [==============================] - 1s 827us/step - loss: 0.4067 - binary_accuracy: 0.8287\n",
      "Epoch 150/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4046 - binary_accuracy: 0.8347\n",
      "Epoch 151/225\n",
      "1754/1754 [==============================] - 1s 841us/step - loss: 0.4008 - binary_accuracy: 0.8327\n",
      "Epoch 152/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4093 - binary_accuracy: 0.8289\n",
      "Epoch 153/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4010 - binary_accuracy: 0.8321\n",
      "Epoch 154/225\n",
      "1754/1754 [==============================] - 1s 831us/step - loss: 0.4110 - binary_accuracy: 0.8343\n",
      "Epoch 155/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4083 - binary_accuracy: 0.8311\n",
      "Epoch 156/225\n",
      "1754/1754 [==============================] - 1s 839us/step - loss: 0.4085 - binary_accuracy: 0.8298\n",
      "Epoch 157/225\n",
      "1754/1754 [==============================] - 1s 837us/step - loss: 0.4040 - binary_accuracy: 0.8326\n",
      "Epoch 158/225\n",
      "1754/1754 [==============================] - 1s 837us/step - loss: 0.4089 - binary_accuracy: 0.8295\n",
      "Epoch 159/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4080 - binary_accuracy: 0.8276\n",
      "Epoch 160/225\n",
      "1754/1754 [==============================] - 1s 828us/step - loss: 0.4054 - binary_accuracy: 0.8324\n",
      "Epoch 161/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4042 - binary_accuracy: 0.8323\n",
      "Epoch 162/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.3989 - binary_accuracy: 0.8346\n",
      "Epoch 163/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4058 - binary_accuracy: 0.8315\n",
      "Epoch 164/225\n",
      "1754/1754 [==============================] - 1s 832us/step - loss: 0.4068 - binary_accuracy: 0.8303\n",
      "Epoch 165/225\n",
      "1754/1754 [==============================] - 1s 830us/step - loss: 0.4084 - binary_accuracy: 0.8271\n",
      "Epoch 166/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4035 - binary_accuracy: 0.8355\n",
      "Epoch 167/225\n",
      "1754/1754 [==============================] - 1s 840us/step - loss: 0.4012 - binary_accuracy: 0.8356\n",
      "Epoch 168/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.3986 - binary_accuracy: 0.8381\n",
      "Epoch 169/225\n",
      "1754/1754 [==============================] - 1s 840us/step - loss: 0.4002 - binary_accuracy: 0.8368\n",
      "Epoch 170/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4087 - binary_accuracy: 0.8319\n",
      "Epoch 171/225\n",
      "1754/1754 [==============================] - 1s 841us/step - loss: 0.4053 - binary_accuracy: 0.8301\n",
      "Epoch 172/225\n",
      "1754/1754 [==============================] - 1s 850us/step - loss: 0.4042 - binary_accuracy: 0.8278\n",
      "Epoch 173/225\n",
      "1754/1754 [==============================] - 2s 881us/step - loss: 0.4077 - binary_accuracy: 0.8313\n",
      "Epoch 174/225\n",
      "1754/1754 [==============================] - 2s 876us/step - loss: 0.4069 - binary_accuracy: 0.8311\n",
      "Epoch 175/225\n",
      "1754/1754 [==============================] - 1s 850us/step - loss: 0.3998 - binary_accuracy: 0.8345\n",
      "Epoch 176/225\n",
      "1754/1754 [==============================] - 1s 844us/step - loss: 0.3996 - binary_accuracy: 0.8368\n",
      "Epoch 177/225\n",
      "1754/1754 [==============================] - 1s 853us/step - loss: 0.4070 - binary_accuracy: 0.8303\n",
      "Epoch 178/225\n",
      "1754/1754 [==============================] - 1s 844us/step - loss: 0.4054 - binary_accuracy: 0.8280\n",
      "Epoch 179/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4067 - binary_accuracy: 0.8248\n",
      "Epoch 180/225\n",
      "1754/1754 [==============================] - 1s 832us/step - loss: 0.4013 - binary_accuracy: 0.8353\n",
      "Epoch 181/225\n",
      "1754/1754 [==============================] - 1s 836us/step - loss: 0.4017 - binary_accuracy: 0.8325\n",
      "Epoch 182/225\n",
      "1754/1754 [==============================] - 1s 836us/step - loss: 0.3973 - binary_accuracy: 0.8390\n",
      "Epoch 183/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4040 - binary_accuracy: 0.8336\n",
      "Epoch 184/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4069 - binary_accuracy: 0.8268\n",
      "Epoch 185/225\n",
      "1754/1754 [==============================] - 1s 828us/step - loss: 0.4030 - binary_accuracy: 0.8339\n",
      "Epoch 186/225\n",
      "1754/1754 [==============================] - 1s 836us/step - loss: 0.4005 - binary_accuracy: 0.8331\n",
      "Epoch 187/225\n",
      "1754/1754 [==============================] - 1s 831us/step - loss: 0.3998 - binary_accuracy: 0.8359\n",
      "Epoch 188/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.3977 - binary_accuracy: 0.8353\n",
      "Epoch 189/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4014 - binary_accuracy: 0.8336\n",
      "Epoch 190/225\n",
      "1754/1754 [==============================] - 1s 840us/step - loss: 0.4059 - binary_accuracy: 0.8324\n",
      "Epoch 191/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4109 - binary_accuracy: 0.8275\n",
      "Epoch 192/225\n",
      "1754/1754 [==============================] - 1s 847us/step - loss: 0.4073 - binary_accuracy: 0.8305\n",
      "Epoch 193/225\n",
      "1754/1754 [==============================] - 1s 842us/step - loss: 0.4005 - binary_accuracy: 0.8342\n",
      "Epoch 194/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4037 - binary_accuracy: 0.8314\n",
      "Epoch 195/225\n",
      "1754/1754 [==============================] - 1s 840us/step - loss: 0.3960 - binary_accuracy: 0.8358\n",
      "Epoch 196/225\n",
      "1754/1754 [==============================] - 1s 839us/step - loss: 0.4003 - binary_accuracy: 0.8363\n",
      "Epoch 197/225\n",
      "1754/1754 [==============================] - 1s 850us/step - loss: 0.4039 - binary_accuracy: 0.8367\n",
      "Epoch 198/225\n",
      "1754/1754 [==============================] - 1s 844us/step - loss: 0.4037 - binary_accuracy: 0.8295\n",
      "Epoch 199/225\n",
      "1754/1754 [==============================] - 1s 851us/step - loss: 0.3969 - binary_accuracy: 0.8362\n",
      "Epoch 200/225\n",
      "1754/1754 [==============================] - 2s 863us/step - loss: 0.4059 - binary_accuracy: 0.8317\n",
      "Epoch 201/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4033 - binary_accuracy: 0.8343\n",
      "Epoch 202/225\n",
      "1754/1754 [==============================] - 1s 844us/step - loss: 0.4029 - binary_accuracy: 0.8313\n",
      "Epoch 203/225\n",
      "1754/1754 [==============================] - 1s 853us/step - loss: 0.4042 - binary_accuracy: 0.8341\n",
      "Epoch 204/225\n",
      "1754/1754 [==============================] - 1s 844us/step - loss: 0.4027 - binary_accuracy: 0.8316\n",
      "Epoch 205/225\n",
      "1754/1754 [==============================] - 1s 852us/step - loss: 0.4025 - binary_accuracy: 0.8301\n",
      "Epoch 206/225\n",
      "1754/1754 [==============================] - 2s 858us/step - loss: 0.4012 - binary_accuracy: 0.8335\n",
      "Epoch 207/225\n",
      "1754/1754 [==============================] - 2s 875us/step - loss: 0.4003 - binary_accuracy: 0.8292\n",
      "Epoch 208/225\n",
      "1754/1754 [==============================] - 2s 861us/step - loss: 0.4040 - binary_accuracy: 0.8310\n",
      "Epoch 209/225\n",
      "1754/1754 [==============================] - 2s 860us/step - loss: 0.3976 - binary_accuracy: 0.8335\n",
      "Epoch 210/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4034 - binary_accuracy: 0.8371\n",
      "Epoch 211/225\n",
      "1754/1754 [==============================] - 1s 844us/step - loss: 0.3984 - binary_accuracy: 0.8371\n",
      "Epoch 212/225\n",
      "1754/1754 [==============================] - 1s 836us/step - loss: 0.4037 - binary_accuracy: 0.8318\n",
      "Epoch 213/225\n",
      "1754/1754 [==============================] - 1s 831us/step - loss: 0.4029 - binary_accuracy: 0.8347\n",
      "Epoch 214/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.3991 - binary_accuracy: 0.8356\n",
      "Epoch 215/225\n",
      "1754/1754 [==============================] - 1s 842us/step - loss: 0.3953 - binary_accuracy: 0.8383\n",
      "Epoch 216/225\n",
      "1754/1754 [==============================] - 1s 847us/step - loss: 0.3969 - binary_accuracy: 0.8389\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 217/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.3984 - binary_accuracy: 0.8328\n",
      "Epoch 218/225\n",
      "1754/1754 [==============================] - 1s 832us/step - loss: 0.3979 - binary_accuracy: 0.8338\n",
      "Epoch 219/225\n",
      "1754/1754 [==============================] - 1s 829us/step - loss: 0.4018 - binary_accuracy: 0.8312\n",
      "Epoch 220/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4028 - binary_accuracy: 0.8297\n",
      "Epoch 221/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4026 - binary_accuracy: 0.8303\n",
      "Epoch 222/225\n",
      "1754/1754 [==============================] - 1s 828us/step - loss: 0.4017 - binary_accuracy: 0.8317\n",
      "Epoch 223/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4001 - binary_accuracy: 0.8300\n",
      "Epoch 224/225\n",
      "1754/1754 [==============================] - 1s 837us/step - loss: 0.4047 - binary_accuracy: 0.8301\n",
      "Epoch 225/225\n",
      "1754/1754 [==============================] - 1s 842us/step - loss: 0.3996 - binary_accuracy: 0.8308\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/225\n",
      "1754/1754 [==============================] - 5s 3ms/step - loss: 0.4683 - binary_accuracy: 0.8215\n",
      "Epoch 2/225\n",
      "1754/1754 [==============================] - 1s 850us/step - loss: 0.4526 - binary_accuracy: 0.8170\n",
      "Epoch 3/225\n",
      "1754/1754 [==============================] - 1s 852us/step - loss: 0.4507 - binary_accuracy: 0.8134\n",
      "Epoch 4/225\n",
      "1754/1754 [==============================] - 1s 852us/step - loss: 0.4507 - binary_accuracy: 0.8165\n",
      "Epoch 5/225\n",
      "1754/1754 [==============================] - 2s 857us/step - loss: 0.4322 - binary_accuracy: 0.8267\n",
      "Epoch 6/225\n",
      "1754/1754 [==============================] - 1s 847us/step - loss: 0.4401 - binary_accuracy: 0.8210\n",
      "Epoch 7/225\n",
      "1754/1754 [==============================] - 1s 851us/step - loss: 0.4387 - binary_accuracy: 0.8203\n",
      "Epoch 8/225\n",
      "1754/1754 [==============================] - 2s 863us/step - loss: 0.4451 - binary_accuracy: 0.8163\n",
      "Epoch 9/225\n",
      "1754/1754 [==============================] - 2s 865us/step - loss: 0.4305 - binary_accuracy: 0.8222\n",
      "Epoch 10/225\n",
      "1754/1754 [==============================] - 2s 855us/step - loss: 0.4357 - binary_accuracy: 0.8241\n",
      "Epoch 11/225\n",
      "1754/1754 [==============================] - 2s 867us/step - loss: 0.4287 - binary_accuracy: 0.8241\n",
      "Epoch 12/225\n",
      "1754/1754 [==============================] - 2s 859us/step - loss: 0.4350 - binary_accuracy: 0.8226\n",
      "Epoch 13/225\n",
      "1754/1754 [==============================] - 2s 859us/step - loss: 0.4347 - binary_accuracy: 0.8242\n",
      "Epoch 14/225\n",
      "1754/1754 [==============================] - 2s 858us/step - loss: 0.4327 - binary_accuracy: 0.8159\n",
      "Epoch 15/225\n",
      "1754/1754 [==============================] - 1s 849us/step - loss: 0.4268 - binary_accuracy: 0.8251\n",
      "Epoch 16/225\n",
      "1754/1754 [==============================] - 1s 846us/step - loss: 0.4353 - binary_accuracy: 0.8190\n",
      "Epoch 17/225\n",
      "1754/1754 [==============================] - 1s 849us/step - loss: 0.4303 - binary_accuracy: 0.8217\n",
      "Epoch 18/225\n",
      "1754/1754 [==============================] - 1s 850us/step - loss: 0.4287 - binary_accuracy: 0.8245\n",
      "Epoch 19/225\n",
      "1754/1754 [==============================] - 1s 854us/step - loss: 0.4290 - binary_accuracy: 0.8225\n",
      "Epoch 20/225\n",
      "1754/1754 [==============================] - 2s 856us/step - loss: 0.4327 - binary_accuracy: 0.8237\n",
      "Epoch 21/225\n",
      "1754/1754 [==============================] - 1s 850us/step - loss: 0.4279 - binary_accuracy: 0.8262\n",
      "Epoch 22/225\n",
      "1754/1754 [==============================] - 1s 849us/step - loss: 0.4207 - binary_accuracy: 0.8292\n",
      "Epoch 23/225\n",
      "1754/1754 [==============================] - 2s 864us/step - loss: 0.4285 - binary_accuracy: 0.8262\n",
      "Epoch 24/225\n",
      "1754/1754 [==============================] - 2s 866us/step - loss: 0.4297 - binary_accuracy: 0.8239\n",
      "Epoch 25/225\n",
      "1754/1754 [==============================] - 1s 854us/step - loss: 0.4285 - binary_accuracy: 0.8217\n",
      "Epoch 26/225\n",
      "1754/1754 [==============================] - 1s 855us/step - loss: 0.4345 - binary_accuracy: 0.8242\n",
      "Epoch 27/225\n",
      "1754/1754 [==============================] - 2s 872us/step - loss: 0.4271 - binary_accuracy: 0.8243\n",
      "Epoch 28/225\n",
      "1754/1754 [==============================] - 2s 871us/step - loss: 0.4253 - binary_accuracy: 0.8253\n",
      "Epoch 29/225\n",
      "1754/1754 [==============================] - 2s 886us/step - loss: 0.4233 - binary_accuracy: 0.8258\n",
      "Epoch 30/225\n",
      "1754/1754 [==============================] - 2s 871us/step - loss: 0.4270 - binary_accuracy: 0.8268\n",
      "Epoch 31/225\n",
      "1754/1754 [==============================] - 2s 875us/step - loss: 0.4271 - binary_accuracy: 0.8263\n",
      "Epoch 32/225\n",
      "1754/1754 [==============================] - 2s 878us/step - loss: 0.4228 - binary_accuracy: 0.8229\n",
      "Epoch 33/225\n",
      "1754/1754 [==============================] - 2s 866us/step - loss: 0.4223 - binary_accuracy: 0.8248\n",
      "Epoch 34/225\n",
      "1754/1754 [==============================] - 2s 873us/step - loss: 0.4298 - binary_accuracy: 0.8206\n",
      "Epoch 35/225\n",
      "1754/1754 [==============================] - 2s 861us/step - loss: 0.4314 - binary_accuracy: 0.8201\n",
      "Epoch 36/225\n",
      "1754/1754 [==============================] - 2s 875us/step - loss: 0.4280 - binary_accuracy: 0.8201\n",
      "Epoch 37/225\n",
      "1754/1754 [==============================] - 2s 877us/step - loss: 0.4242 - binary_accuracy: 0.8233\n",
      "Epoch 38/225\n",
      "1754/1754 [==============================] - 2s 862us/step - loss: 0.4212 - binary_accuracy: 0.8262\n",
      "Epoch 39/225\n",
      "1754/1754 [==============================] - 1s 851us/step - loss: 0.4194 - binary_accuracy: 0.8308\n",
      "Epoch 40/225\n",
      "1754/1754 [==============================] - 2s 869us/step - loss: 0.4221 - binary_accuracy: 0.8257\n",
      "Epoch 41/225\n",
      "1754/1754 [==============================] - 2s 864us/step - loss: 0.4301 - binary_accuracy: 0.8216\n",
      "Epoch 42/225\n",
      "1754/1754 [==============================] - 2s 867us/step - loss: 0.4184 - binary_accuracy: 0.8275\n",
      "Epoch 43/225\n",
      "1754/1754 [==============================] - 2s 864us/step - loss: 0.4276 - binary_accuracy: 0.8212\n",
      "Epoch 44/225\n",
      "1754/1754 [==============================] - 2s 857us/step - loss: 0.4210 - binary_accuracy: 0.8240\n",
      "Epoch 45/225\n",
      "1754/1754 [==============================] - 1s 853us/step - loss: 0.4274 - binary_accuracy: 0.8295\n",
      "Epoch 46/225\n",
      "1754/1754 [==============================] - 1s 852us/step - loss: 0.4311 - binary_accuracy: 0.8191\n",
      "Epoch 47/225\n",
      "1754/1754 [==============================] - 2s 912us/step - loss: 0.4203 - binary_accuracy: 0.8231\n",
      "Epoch 48/225\n",
      "1754/1754 [==============================] - 2s 905us/step - loss: 0.4207 - binary_accuracy: 0.8259\n",
      "Epoch 49/225\n",
      "1754/1754 [==============================] - 1s 855us/step - loss: 0.4220 - binary_accuracy: 0.8265\n",
      "Epoch 50/225\n",
      "1754/1754 [==============================] - 2s 856us/step - loss: 0.4239 - binary_accuracy: 0.8264\n",
      "Epoch 51/225\n",
      "1754/1754 [==============================] - 2s 883us/step - loss: 0.4203 - binary_accuracy: 0.8260\n",
      "Epoch 52/225\n",
      "1754/1754 [==============================] - 2s 861us/step - loss: 0.4272 - binary_accuracy: 0.8197\n",
      "Epoch 53/225\n",
      "1754/1754 [==============================] - 2s 882us/step - loss: 0.4274 - binary_accuracy: 0.8211\n",
      "Epoch 54/225\n",
      "1754/1754 [==============================] - 2s 892us/step - loss: 0.4179 - binary_accuracy: 0.8281\n",
      "Epoch 55/225\n",
      "1754/1754 [==============================] - 2s 863us/step - loss: 0.4243 - binary_accuracy: 0.8242\n",
      "Epoch 56/225\n",
      "1754/1754 [==============================] - 1s 854us/step - loss: 0.4170 - binary_accuracy: 0.8233\n",
      "Epoch 57/225\n",
      "1754/1754 [==============================] - 2s 869us/step - loss: 0.4246 - binary_accuracy: 0.8259\n",
      "Epoch 58/225\n",
      "1754/1754 [==============================] - 2s 867us/step - loss: 0.4241 - binary_accuracy: 0.8241\n",
      "Epoch 59/225\n",
      "1754/1754 [==============================] - 2s 860us/step - loss: 0.4201 - binary_accuracy: 0.8258\n",
      "Epoch 60/225\n",
      "1754/1754 [==============================] - 2s 877us/step - loss: 0.4229 - binary_accuracy: 0.8253\n",
      "Epoch 61/225\n",
      "1754/1754 [==============================] - 2s 863us/step - loss: 0.4214 - binary_accuracy: 0.8246\n",
      "Epoch 62/225\n",
      "1754/1754 [==============================] - 2s 873us/step - loss: 0.4178 - binary_accuracy: 0.8256\n",
      "Epoch 63/225\n",
      "1754/1754 [==============================] - 2s 875us/step - loss: 0.4231 - binary_accuracy: 0.8223\n",
      "Epoch 64/225\n",
      "1754/1754 [==============================] - 1s 846us/step - loss: 0.4175 - binary_accuracy: 0.8242\n",
      "Epoch 65/225\n",
      "1754/1754 [==============================] - 1s 846us/step - loss: 0.4219 - binary_accuracy: 0.8261\n",
      "Epoch 66/225\n",
      "1754/1754 [==============================] - 1s 839us/step - loss: 0.4211 - binary_accuracy: 0.8241\n",
      "Epoch 67/225\n",
      "1754/1754 [==============================] - 1s 841us/step - loss: 0.4270 - binary_accuracy: 0.8248\n",
      "Epoch 68/225\n",
      "1754/1754 [==============================] - 1s 846us/step - loss: 0.4179 - binary_accuracy: 0.8247\n",
      "Epoch 69/225\n",
      "1754/1754 [==============================] - 1s 845us/step - loss: 0.4198 - binary_accuracy: 0.8269\n",
      "Epoch 70/225\n",
      "1754/1754 [==============================] - 1s 843us/step - loss: 0.4201 - binary_accuracy: 0.8219\n",
      "Epoch 71/225\n",
      "1754/1754 [==============================] - 1s 854us/step - loss: 0.4148 - binary_accuracy: 0.8322\n",
      "Epoch 72/225\n",
      "1754/1754 [==============================] - 1s 849us/step - loss: 0.4190 - binary_accuracy: 0.8248\n",
      "Epoch 73/225\n",
      "1754/1754 [==============================] - 1s 844us/step - loss: 0.4173 - binary_accuracy: 0.8259\n",
      "Epoch 74/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4231 - binary_accuracy: 0.8261\n",
      "Epoch 75/225\n",
      "1754/1754 [==============================] - 1s 830us/step - loss: 0.4205 - binary_accuracy: 0.8250\n",
      "Epoch 76/225\n",
      "1754/1754 [==============================] - 1s 836us/step - loss: 0.4180 - binary_accuracy: 0.8293\n",
      "Epoch 77/225\n",
      "1754/1754 [==============================] - 1s 830us/step - loss: 0.4213 - binary_accuracy: 0.8256\n",
      "Epoch 78/225\n",
      "1754/1754 [==============================] - 1s 831us/step - loss: 0.4159 - binary_accuracy: 0.8277\n",
      "Epoch 79/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.4162 - binary_accuracy: 0.8256\n",
      "Epoch 80/225\n",
      "1754/1754 [==============================] - 1s 829us/step - loss: 0.4178 - binary_accuracy: 0.8259\n",
      "Epoch 81/225\n",
      "1754/1754 [==============================] - 1s 852us/step - loss: 0.4204 - binary_accuracy: 0.8237\n",
      "Epoch 82/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4124 - binary_accuracy: 0.8249\n",
      "Epoch 83/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4178 - binary_accuracy: 0.8269\n",
      "Epoch 84/225\n",
      "1754/1754 [==============================] - 1s 853us/step - loss: 0.4111 - binary_accuracy: 0.8331\n",
      "Epoch 85/225\n",
      "1754/1754 [==============================] - 1s 843us/step - loss: 0.4142 - binary_accuracy: 0.8284\n",
      "Epoch 86/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4099 - binary_accuracy: 0.8308\n",
      "Epoch 87/225\n",
      "1754/1754 [==============================] - 1s 840us/step - loss: 0.4122 - binary_accuracy: 0.8296\n",
      "Epoch 88/225\n",
      "1754/1754 [==============================] - 1s 842us/step - loss: 0.4152 - binary_accuracy: 0.8283\n",
      "Epoch 89/225\n",
      "1754/1754 [==============================] - 1s 841us/step - loss: 0.4181 - binary_accuracy: 0.8243\n",
      "Epoch 90/225\n",
      "1754/1754 [==============================] - 1s 847us/step - loss: 0.4181 - binary_accuracy: 0.8260\n",
      "Epoch 91/225\n",
      "1754/1754 [==============================] - 1s 832us/step - loss: 0.4248 - binary_accuracy: 0.8230\n",
      "Epoch 92/225\n",
      "1754/1754 [==============================] - 1s 843us/step - loss: 0.4232 - binary_accuracy: 0.8223\n",
      "Epoch 93/225\n",
      "1754/1754 [==============================] - 1s 836us/step - loss: 0.4151 - binary_accuracy: 0.8245\n",
      "Epoch 94/225\n",
      "1754/1754 [==============================] - 1s 840us/step - loss: 0.4208 - binary_accuracy: 0.8215\n",
      "Epoch 95/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4221 - binary_accuracy: 0.8241\n",
      "Epoch 96/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4137 - binary_accuracy: 0.8278\n",
      "Epoch 97/225\n",
      "1754/1754 [==============================] - 1s 831us/step - loss: 0.4231 - binary_accuracy: 0.8265\n",
      "Epoch 98/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4119 - binary_accuracy: 0.8301\n",
      "Epoch 99/225\n",
      "1754/1754 [==============================] - 1s 841us/step - loss: 0.4178 - binary_accuracy: 0.8272\n",
      "Epoch 100/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4215 - binary_accuracy: 0.8273\n",
      "Epoch 101/225\n",
      "1754/1754 [==============================] - 1s 831us/step - loss: 0.4113 - binary_accuracy: 0.8303\n",
      "Epoch 102/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4162 - binary_accuracy: 0.8275\n",
      "Epoch 103/225\n",
      "1754/1754 [==============================] - 1s 837us/step - loss: 0.4140 - binary_accuracy: 0.8269\n",
      "Epoch 104/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4192 - binary_accuracy: 0.8266\n",
      "Epoch 105/225\n",
      "1754/1754 [==============================] - 1s 831us/step - loss: 0.4176 - binary_accuracy: 0.8260\n",
      "Epoch 106/225\n",
      "1754/1754 [==============================] - 1s 832us/step - loss: 0.4135 - binary_accuracy: 0.8270\n",
      "Epoch 107/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4133 - binary_accuracy: 0.8298\n",
      "Epoch 108/225\n",
      "1754/1754 [==============================] - 1s 842us/step - loss: 0.4141 - binary_accuracy: 0.8317\n",
      "Epoch 109/225\n",
      "1754/1754 [==============================] - 1s 832us/step - loss: 0.4146 - binary_accuracy: 0.8288\n",
      "Epoch 110/225\n",
      "1754/1754 [==============================] - 1s 839us/step - loss: 0.4110 - binary_accuracy: 0.8297\n",
      "Epoch 111/225\n",
      "1754/1754 [==============================] - 1s 841us/step - loss: 0.4205 - binary_accuracy: 0.8199\n",
      "Epoch 112/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4147 - binary_accuracy: 0.8227\n",
      "Epoch 113/225\n",
      "1754/1754 [==============================] - 1s 840us/step - loss: 0.4168 - binary_accuracy: 0.8233\n",
      "Epoch 114/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4115 - binary_accuracy: 0.8296\n",
      "Epoch 115/225\n",
      "1754/1754 [==============================] - 1s 842us/step - loss: 0.4116 - binary_accuracy: 0.8293\n",
      "Epoch 116/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4228 - binary_accuracy: 0.8168\n",
      "Epoch 117/225\n",
      "1754/1754 [==============================] - 1s 841us/step - loss: 0.4200 - binary_accuracy: 0.8252\n",
      "Epoch 118/225\n",
      "1754/1754 [==============================] - 1s 837us/step - loss: 0.4193 - binary_accuracy: 0.8294\n",
      "Epoch 119/225\n",
      "1754/1754 [==============================] - 1s 845us/step - loss: 0.4123 - binary_accuracy: 0.8264\n",
      "Epoch 120/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4110 - binary_accuracy: 0.8238\n",
      "Epoch 121/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4172 - binary_accuracy: 0.8279\n",
      "Epoch 122/225\n",
      "1754/1754 [==============================] - 1s 840us/step - loss: 0.4179 - binary_accuracy: 0.8271\n",
      "Epoch 123/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4132 - binary_accuracy: 0.8257\n",
      "Epoch 124/225\n",
      "1754/1754 [==============================] - 1s 840us/step - loss: 0.4108 - binary_accuracy: 0.8301\n",
      "Epoch 125/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4104 - binary_accuracy: 0.8300\n",
      "Epoch 126/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4152 - binary_accuracy: 0.8264\n",
      "Epoch 127/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4131 - binary_accuracy: 0.8249\n",
      "Epoch 128/225\n",
      "1754/1754 [==============================] - 1s 841us/step - loss: 0.4109 - binary_accuracy: 0.8303\n",
      "Epoch 129/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4169 - binary_accuracy: 0.8252\n",
      "Epoch 130/225\n",
      "1754/1754 [==============================] - 1s 836us/step - loss: 0.4156 - binary_accuracy: 0.8278\n",
      "Epoch 131/225\n",
      "1754/1754 [==============================] - 1s 831us/step - loss: 0.4187 - binary_accuracy: 0.8186\n",
      "Epoch 132/225\n",
      "1754/1754 [==============================] - 1s 837us/step - loss: 0.4177 - binary_accuracy: 0.8272\n",
      "Epoch 133/225\n",
      "1754/1754 [==============================] - 1s 836us/step - loss: 0.4186 - binary_accuracy: 0.8274\n",
      "Epoch 134/225\n",
      "1754/1754 [==============================] - 1s 839us/step - loss: 0.4130 - binary_accuracy: 0.8287\n",
      "Epoch 135/225\n",
      "1754/1754 [==============================] - 1s 841us/step - loss: 0.4140 - binary_accuracy: 0.8271\n",
      "Epoch 136/225\n",
      "1754/1754 [==============================] - 1s 844us/step - loss: 0.4136 - binary_accuracy: 0.8300\n",
      "Epoch 137/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4128 - binary_accuracy: 0.8251\n",
      "Epoch 138/225\n",
      "1754/1754 [==============================] - 1s 836us/step - loss: 0.4175 - binary_accuracy: 0.8254\n",
      "Epoch 139/225\n",
      "1754/1754 [==============================] - 1s 831us/step - loss: 0.4155 - binary_accuracy: 0.8244\n",
      "Epoch 140/225\n",
      "1754/1754 [==============================] - 1s 841us/step - loss: 0.4159 - binary_accuracy: 0.8251\n",
      "Epoch 141/225\n",
      "1754/1754 [==============================] - 1s 836us/step - loss: 0.4134 - binary_accuracy: 0.8292\n",
      "Epoch 142/225\n",
      "1754/1754 [==============================] - 1s 839us/step - loss: 0.4107 - binary_accuracy: 0.8303\n",
      "Epoch 143/225\n",
      "1754/1754 [==============================] - 1s 840us/step - loss: 0.4090 - binary_accuracy: 0.8297\n",
      "Epoch 144/225\n",
      "1754/1754 [==============================] - 1s 836us/step - loss: 0.4140 - binary_accuracy: 0.8275\n",
      "Epoch 145/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1754/1754 [==============================] - 1s 839us/step - loss: 0.4177 - binary_accuracy: 0.8237\n",
      "Epoch 146/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4137 - binary_accuracy: 0.8259\n",
      "Epoch 147/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4059 - binary_accuracy: 0.8290\n",
      "Epoch 148/225\n",
      "1754/1754 [==============================] - 1s 839us/step - loss: 0.4206 - binary_accuracy: 0.8293\n",
      "Epoch 149/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4097 - binary_accuracy: 0.8289\n",
      "Epoch 150/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4128 - binary_accuracy: 0.8248\n",
      "Epoch 151/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4097 - binary_accuracy: 0.8291\n",
      "Epoch 152/225\n",
      "1754/1754 [==============================] - 1s 837us/step - loss: 0.4125 - binary_accuracy: 0.8240\n",
      "Epoch 153/225\n",
      "1754/1754 [==============================] - 1s 837us/step - loss: 0.4071 - binary_accuracy: 0.8325\n",
      "Epoch 154/225\n",
      "1754/1754 [==============================] - 1s 839us/step - loss: 0.4173 - binary_accuracy: 0.8249\n",
      "Epoch 155/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4095 - binary_accuracy: 0.8245\n",
      "Epoch 156/225\n",
      "1754/1754 [==============================] - 1s 849us/step - loss: 0.4109 - binary_accuracy: 0.8302\n",
      "Epoch 157/225\n",
      "1754/1754 [==============================] - 1s 839us/step - loss: 0.4203 - binary_accuracy: 0.8226\n",
      "Epoch 158/225\n",
      "1754/1754 [==============================] - 1s 837us/step - loss: 0.4088 - binary_accuracy: 0.8290\n",
      "Epoch 159/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4086 - binary_accuracy: 0.8342\n",
      "Epoch 160/225\n",
      "1754/1754 [==============================] - 1s 846us/step - loss: 0.4077 - binary_accuracy: 0.8294\n",
      "Epoch 161/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4048 - binary_accuracy: 0.8314\n",
      "Epoch 162/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4156 - binary_accuracy: 0.8277\n",
      "Epoch 163/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4056 - binary_accuracy: 0.8302\n",
      "Epoch 164/225\n",
      "1754/1754 [==============================] - 1s 831us/step - loss: 0.4087 - binary_accuracy: 0.8266\n",
      "Epoch 165/225\n",
      "1754/1754 [==============================] - 1s 832us/step - loss: 0.4161 - binary_accuracy: 0.8275\n",
      "Epoch 166/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4113 - binary_accuracy: 0.8264\n",
      "Epoch 167/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4111 - binary_accuracy: 0.8271\n",
      "Epoch 168/225\n",
      "1754/1754 [==============================] - 1s 832us/step - loss: 0.4132 - binary_accuracy: 0.8284\n",
      "Epoch 169/225\n",
      "1754/1754 [==============================] - 1s 842us/step - loss: 0.4076 - binary_accuracy: 0.8306\n",
      "Epoch 170/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4123 - binary_accuracy: 0.8251\n",
      "Epoch 171/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4113 - binary_accuracy: 0.8267\n",
      "Epoch 172/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4119 - binary_accuracy: 0.8231\n",
      "Epoch 173/225\n",
      "1754/1754 [==============================] - 1s 841us/step - loss: 0.4109 - binary_accuracy: 0.8293\n",
      "Epoch 174/225\n",
      "1754/1754 [==============================] - 1s 844us/step - loss: 0.4190 - binary_accuracy: 0.8276\n",
      "Epoch 175/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4071 - binary_accuracy: 0.8259\n",
      "Epoch 176/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4096 - binary_accuracy: 0.8283\n",
      "Epoch 177/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4155 - binary_accuracy: 0.8240\n",
      "Epoch 178/225\n",
      "1754/1754 [==============================] - 1s 836us/step - loss: 0.4115 - binary_accuracy: 0.8267\n",
      "Epoch 179/225\n",
      "1754/1754 [==============================] - 1s 844us/step - loss: 0.4145 - binary_accuracy: 0.8258\n",
      "Epoch 180/225\n",
      "1754/1754 [==============================] - 1s 842us/step - loss: 0.4100 - binary_accuracy: 0.8195\n",
      "Epoch 181/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4174 - binary_accuracy: 0.8264\n",
      "Epoch 182/225\n",
      "1754/1754 [==============================] - 1s 836us/step - loss: 0.4083 - binary_accuracy: 0.8312\n",
      "Epoch 183/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4076 - binary_accuracy: 0.8295\n",
      "Epoch 184/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4071 - binary_accuracy: 0.8279\n",
      "Epoch 185/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4102 - binary_accuracy: 0.8268\n",
      "Epoch 186/225\n",
      "1754/1754 [==============================] - 1s 845us/step - loss: 0.4086 - binary_accuracy: 0.8325\n",
      "Epoch 187/225\n",
      "1754/1754 [==============================] - 1s 842us/step - loss: 0.4116 - binary_accuracy: 0.8284\n",
      "Epoch 188/225\n",
      "1754/1754 [==============================] - 1s 840us/step - loss: 0.4156 - binary_accuracy: 0.8252\n",
      "Epoch 189/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4092 - binary_accuracy: 0.8277\n",
      "Epoch 190/225\n",
      "1754/1754 [==============================] - 1s 831us/step - loss: 0.4092 - binary_accuracy: 0.8276\n",
      "Epoch 191/225\n",
      "1754/1754 [==============================] - 1s 837us/step - loss: 0.4119 - binary_accuracy: 0.8285\n",
      "Epoch 192/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4175 - binary_accuracy: 0.8231\n",
      "Epoch 193/225\n",
      "1754/1754 [==============================] - 1s 837us/step - loss: 0.4123 - binary_accuracy: 0.8263\n",
      "Epoch 194/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4114 - binary_accuracy: 0.8317\n",
      "Epoch 195/225\n",
      "1754/1754 [==============================] - 1s 843us/step - loss: 0.4039 - binary_accuracy: 0.8253\n",
      "Epoch 196/225\n",
      "1754/1754 [==============================] - 1s 840us/step - loss: 0.4072 - binary_accuracy: 0.8292\n",
      "Epoch 197/225\n",
      "1754/1754 [==============================] - 1s 842us/step - loss: 0.4111 - binary_accuracy: 0.8254\n",
      "Epoch 198/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4085 - binary_accuracy: 0.8280\n",
      "Epoch 199/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4110 - binary_accuracy: 0.8286\n",
      "Epoch 200/225\n",
      "1754/1754 [==============================] - 1s 830us/step - loss: 0.4120 - binary_accuracy: 0.8300\n",
      "Epoch 201/225\n",
      "1754/1754 [==============================] - 1s 832us/step - loss: 0.4074 - binary_accuracy: 0.8285\n",
      "Epoch 202/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4105 - binary_accuracy: 0.8282\n",
      "Epoch 203/225\n",
      "1754/1754 [==============================] - 1s 832us/step - loss: 0.4037 - binary_accuracy: 0.8314\n",
      "Epoch 204/225\n",
      "1754/1754 [==============================] - 1s 834us/step - loss: 0.4122 - binary_accuracy: 0.8293\n",
      "Epoch 205/225\n",
      "1754/1754 [==============================] - 1s 836us/step - loss: 0.4087 - binary_accuracy: 0.8284\n",
      "Epoch 206/225\n",
      "1754/1754 [==============================] - 1s 839us/step - loss: 0.4125 - binary_accuracy: 0.8294\n",
      "Epoch 207/225\n",
      "1754/1754 [==============================] - 1s 841us/step - loss: 0.4113 - binary_accuracy: 0.8276\n",
      "Epoch 208/225\n",
      "1754/1754 [==============================] - 1s 840us/step - loss: 0.4057 - binary_accuracy: 0.8302\n",
      "Epoch 209/225\n",
      "1754/1754 [==============================] - 1s 832us/step - loss: 0.4083 - binary_accuracy: 0.8287\n",
      "Epoch 210/225\n",
      "1754/1754 [==============================] - 1s 836us/step - loss: 0.4104 - binary_accuracy: 0.8303\n",
      "Epoch 211/225\n",
      "1754/1754 [==============================] - 1s 837us/step - loss: 0.4105 - binary_accuracy: 0.8243\n",
      "Epoch 212/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4110 - binary_accuracy: 0.8270\n",
      "Epoch 213/225\n",
      "1754/1754 [==============================] - 1s 850us/step - loss: 0.4113 - binary_accuracy: 0.8230\n",
      "Epoch 214/225\n",
      "1754/1754 [==============================] - 1s 852us/step - loss: 0.4127 - binary_accuracy: 0.8262\n",
      "Epoch 215/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.3994 - binary_accuracy: 0.8352\n",
      "Epoch 216/225\n",
      "1754/1754 [==============================] - 1s 837us/step - loss: 0.4121 - binary_accuracy: 0.8238\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 217/225\n",
      "1754/1754 [==============================] - 1s 832us/step - loss: 0.4167 - binary_accuracy: 0.8263\n",
      "Epoch 218/225\n",
      "1754/1754 [==============================] - 1s 830us/step - loss: 0.4136 - binary_accuracy: 0.8263\n",
      "Epoch 219/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4134 - binary_accuracy: 0.8275\n",
      "Epoch 220/225\n",
      "1754/1754 [==============================] - 1s 840us/step - loss: 0.4155 - binary_accuracy: 0.8264\n",
      "Epoch 221/225\n",
      "1754/1754 [==============================] - 1s 835us/step - loss: 0.4056 - binary_accuracy: 0.8327\n",
      "Epoch 222/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4097 - binary_accuracy: 0.8310\n",
      "Epoch 223/225\n",
      "1754/1754 [==============================] - 1s 840us/step - loss: 0.4076 - binary_accuracy: 0.8301\n",
      "Epoch 224/225\n",
      "1754/1754 [==============================] - 1s 837us/step - loss: 0.4001 - binary_accuracy: 0.8320\n",
      "Epoch 225/225\n",
      "1754/1754 [==============================] - 1s 841us/step - loss: 0.4059 - binary_accuracy: 0.8273\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/225\n",
      "1754/1754 [==============================] - 5s 3ms/step - loss: 0.4763 - binary_accuracy: 0.8161\n",
      "Epoch 2/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4518 - binary_accuracy: 0.8203\n",
      "Epoch 3/225\n",
      "1754/1754 [==============================] - 1s 821us/step - loss: 0.4489 - binary_accuracy: 0.8185\n",
      "Epoch 4/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4450 - binary_accuracy: 0.8184\n",
      "Epoch 5/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4467 - binary_accuracy: 0.8132\n",
      "Epoch 6/225\n",
      "1754/1754 [==============================] - 1s 830us/step - loss: 0.4419 - binary_accuracy: 0.8193\n",
      "Epoch 7/225\n",
      "1754/1754 [==============================] - 1s 827us/step - loss: 0.4389 - binary_accuracy: 0.8186\n",
      "Epoch 8/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4342 - binary_accuracy: 0.8225\n",
      "Epoch 9/225\n",
      "1754/1754 [==============================] - 1s 838us/step - loss: 0.4334 - binary_accuracy: 0.8190\n",
      "Epoch 10/225\n",
      "1754/1754 [==============================] - 1s 822us/step - loss: 0.4417 - binary_accuracy: 0.8184\n",
      "Epoch 11/225\n",
      "1754/1754 [==============================] - 1s 828us/step - loss: 0.4335 - binary_accuracy: 0.8215\n",
      "Epoch 12/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4336 - binary_accuracy: 0.8232\n",
      "Epoch 13/225\n",
      "1754/1754 [==============================] - 1s 830us/step - loss: 0.4352 - binary_accuracy: 0.8204\n",
      "Epoch 14/225\n",
      "1754/1754 [==============================] - 1s 836us/step - loss: 0.4373 - binary_accuracy: 0.8177\n",
      "Epoch 15/225\n",
      "1754/1754 [==============================] - 1s 830us/step - loss: 0.4364 - binary_accuracy: 0.8194\n",
      "Epoch 16/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4298 - binary_accuracy: 0.8219\n",
      "Epoch 17/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.4331 - binary_accuracy: 0.8209\n",
      "Epoch 18/225\n",
      "1754/1754 [==============================] - 1s 828us/step - loss: 0.4304 - binary_accuracy: 0.8248\n",
      "Epoch 19/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.4267 - binary_accuracy: 0.8244\n",
      "Epoch 20/225\n",
      "1754/1754 [==============================] - 1s 842us/step - loss: 0.4308 - binary_accuracy: 0.8246\n",
      "Epoch 21/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4321 - binary_accuracy: 0.8203\n",
      "Epoch 22/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4270 - binary_accuracy: 0.8261\n",
      "Epoch 23/225\n",
      "1754/1754 [==============================] - 1s 832us/step - loss: 0.4353 - binary_accuracy: 0.8189\n",
      "Epoch 24/225\n",
      "1754/1754 [==============================] - 1s 827us/step - loss: 0.4276 - binary_accuracy: 0.8246\n",
      "Epoch 25/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4321 - binary_accuracy: 0.8194\n",
      "Epoch 26/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4281 - binary_accuracy: 0.8229\n",
      "Epoch 27/225\n",
      "1754/1754 [==============================] - 1s 821us/step - loss: 0.4382 - binary_accuracy: 0.8125\n",
      "Epoch 28/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.4246 - binary_accuracy: 0.8217\n",
      "Epoch 29/225\n",
      "1754/1754 [==============================] - 1s 822us/step - loss: 0.4249 - binary_accuracy: 0.8238\n",
      "Epoch 30/225\n",
      "1754/1754 [==============================] - 1s 821us/step - loss: 0.4270 - binary_accuracy: 0.8209\n",
      "Epoch 31/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4262 - binary_accuracy: 0.8283\n",
      "Epoch 32/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.4236 - binary_accuracy: 0.8185\n",
      "Epoch 33/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4319 - binary_accuracy: 0.8172\n",
      "Epoch 34/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4244 - binary_accuracy: 0.8253\n",
      "Epoch 35/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4261 - binary_accuracy: 0.8202\n",
      "Epoch 36/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4236 - binary_accuracy: 0.8213\n",
      "Epoch 37/225\n",
      "1754/1754 [==============================] - 1s 832us/step - loss: 0.4248 - binary_accuracy: 0.8222\n",
      "Epoch 38/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4321 - binary_accuracy: 0.8217\n",
      "Epoch 39/225\n",
      "1754/1754 [==============================] - 1s 828us/step - loss: 0.4275 - binary_accuracy: 0.8198\n",
      "Epoch 40/225\n",
      "1754/1754 [==============================] - 1s 832us/step - loss: 0.4242 - binary_accuracy: 0.8214\n",
      "Epoch 41/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4298 - binary_accuracy: 0.8196\n",
      "Epoch 42/225\n",
      "1754/1754 [==============================] - 1s 817us/step - loss: 0.4258 - binary_accuracy: 0.8276\n",
      "Epoch 43/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4262 - binary_accuracy: 0.8274\n",
      "Epoch 44/225\n",
      "1754/1754 [==============================] - 1s 822us/step - loss: 0.4181 - binary_accuracy: 0.8250\n",
      "Epoch 45/225\n",
      "1754/1754 [==============================] - 1s 822us/step - loss: 0.4298 - binary_accuracy: 0.8194\n",
      "Epoch 46/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4181 - binary_accuracy: 0.8252\n",
      "Epoch 47/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4209 - binary_accuracy: 0.8237\n",
      "Epoch 48/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4227 - binary_accuracy: 0.8231\n",
      "Epoch 49/225\n",
      "1754/1754 [==============================] - 1s 833us/step - loss: 0.4223 - binary_accuracy: 0.8229\n",
      "Epoch 50/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4227 - binary_accuracy: 0.8291\n",
      "Epoch 51/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4215 - binary_accuracy: 0.8244\n",
      "Epoch 52/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4193 - binary_accuracy: 0.8231\n",
      "Epoch 53/225\n",
      "1754/1754 [==============================] - 1s 828us/step - loss: 0.4177 - binary_accuracy: 0.8273\n",
      "Epoch 54/225\n",
      "1754/1754 [==============================] - 1s 827us/step - loss: 0.4204 - binary_accuracy: 0.8274\n",
      "Epoch 55/225\n",
      "1754/1754 [==============================] - 1s 828us/step - loss: 0.4199 - binary_accuracy: 0.8227\n",
      "Epoch 56/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4229 - binary_accuracy: 0.8237\n",
      "Epoch 57/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4300 - binary_accuracy: 0.8171\n",
      "Epoch 58/225\n",
      "1754/1754 [==============================] - 1s 821us/step - loss: 0.4256 - binary_accuracy: 0.8224\n",
      "Epoch 59/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4199 - binary_accuracy: 0.8266\n",
      "Epoch 60/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4229 - binary_accuracy: 0.8264\n",
      "Epoch 61/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4200 - binary_accuracy: 0.8254\n",
      "Epoch 62/225\n",
      "1754/1754 [==============================] - 1s 821us/step - loss: 0.4193 - binary_accuracy: 0.8186\n",
      "Epoch 63/225\n",
      "1754/1754 [==============================] - 1s 815us/step - loss: 0.4196 - binary_accuracy: 0.8238\n",
      "Epoch 64/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4217 - binary_accuracy: 0.8213\n",
      "Epoch 65/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4237 - binary_accuracy: 0.8253\n",
      "Epoch 66/225\n",
      "1754/1754 [==============================] - 1s 822us/step - loss: 0.4133 - binary_accuracy: 0.8311\n",
      "Epoch 67/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4203 - binary_accuracy: 0.8255\n",
      "Epoch 68/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4207 - binary_accuracy: 0.8211\n",
      "Epoch 69/225\n",
      "1754/1754 [==============================] - 1s 822us/step - loss: 0.4221 - binary_accuracy: 0.8214\n",
      "Epoch 70/225\n",
      "1754/1754 [==============================] - 1s 821us/step - loss: 0.4235 - binary_accuracy: 0.8215\n",
      "Epoch 71/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4141 - binary_accuracy: 0.8263\n",
      "Epoch 72/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4223 - binary_accuracy: 0.8212\n",
      "Epoch 73/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4188 - binary_accuracy: 0.8284\n",
      "Epoch 74/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4157 - binary_accuracy: 0.8293\n",
      "Epoch 75/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4155 - binary_accuracy: 0.8267\n",
      "Epoch 76/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4157 - binary_accuracy: 0.8279\n",
      "Epoch 77/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4167 - binary_accuracy: 0.8223\n",
      "Epoch 78/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4132 - binary_accuracy: 0.8256\n",
      "Epoch 79/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4133 - binary_accuracy: 0.8252\n",
      "Epoch 80/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4203 - binary_accuracy: 0.8171\n",
      "Epoch 81/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4195 - binary_accuracy: 0.8196\n",
      "Epoch 82/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4219 - binary_accuracy: 0.8267\n",
      "Epoch 83/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4156 - binary_accuracy: 0.8240\n",
      "Epoch 84/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4143 - binary_accuracy: 0.8299\n",
      "Epoch 85/225\n",
      "1754/1754 [==============================] - 1s 822us/step - loss: 0.4142 - binary_accuracy: 0.8260\n",
      "Epoch 86/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4141 - binary_accuracy: 0.8283\n",
      "Epoch 87/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4223 - binary_accuracy: 0.8263\n",
      "Epoch 88/225\n",
      "1754/1754 [==============================] - 1s 827us/step - loss: 0.4162 - binary_accuracy: 0.8268\n",
      "Epoch 89/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4171 - binary_accuracy: 0.8269\n",
      "Epoch 90/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4138 - binary_accuracy: 0.8264\n",
      "Epoch 91/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4225 - binary_accuracy: 0.8235\n",
      "Epoch 92/225\n",
      "1754/1754 [==============================] - 1s 821us/step - loss: 0.4213 - binary_accuracy: 0.8253\n",
      "Epoch 93/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4129 - binary_accuracy: 0.8328\n",
      "Epoch 94/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4146 - binary_accuracy: 0.8295\n",
      "Epoch 95/225\n",
      "1754/1754 [==============================] - 1s 849us/step - loss: 0.4206 - binary_accuracy: 0.8237\n",
      "Epoch 96/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4207 - binary_accuracy: 0.8240\n",
      "Epoch 97/225\n",
      "1754/1754 [==============================] - 1s 817us/step - loss: 0.4179 - binary_accuracy: 0.8253\n",
      "Epoch 98/225\n",
      "1754/1754 [==============================] - 1s 822us/step - loss: 0.4181 - binary_accuracy: 0.8242\n",
      "Epoch 99/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4167 - binary_accuracy: 0.8263\n",
      "Epoch 100/225\n",
      "1754/1754 [==============================] - 1s 814us/step - loss: 0.4192 - binary_accuracy: 0.8221\n",
      "Epoch 101/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.4170 - binary_accuracy: 0.8259\n",
      "Epoch 102/225\n",
      "1754/1754 [==============================] - 1s 811us/step - loss: 0.4157 - binary_accuracy: 0.8271\n",
      "Epoch 103/225\n",
      "1754/1754 [==============================] - 1s 817us/step - loss: 0.4177 - binary_accuracy: 0.8182\n",
      "Epoch 104/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4133 - binary_accuracy: 0.8262\n",
      "Epoch 105/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4178 - binary_accuracy: 0.8250\n",
      "Epoch 106/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4191 - binary_accuracy: 0.8270\n",
      "Epoch 107/225\n",
      "1754/1754 [==============================] - 1s 828us/step - loss: 0.4151 - binary_accuracy: 0.8276\n",
      "Epoch 108/225\n",
      "1754/1754 [==============================] - 1s 816us/step - loss: 0.4070 - binary_accuracy: 0.8269\n",
      "Epoch 109/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4202 - binary_accuracy: 0.8205\n",
      "Epoch 110/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4127 - binary_accuracy: 0.8287\n",
      "Epoch 111/225\n",
      "1754/1754 [==============================] - 1s 816us/step - loss: 0.4167 - binary_accuracy: 0.8327\n",
      "Epoch 112/225\n",
      "1754/1754 [==============================] - 1s 817us/step - loss: 0.4142 - binary_accuracy: 0.8245\n",
      "Epoch 113/225\n",
      "1754/1754 [==============================] - 1s 821us/step - loss: 0.4139 - binary_accuracy: 0.8289\n",
      "Epoch 114/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4137 - binary_accuracy: 0.8289\n",
      "Epoch 115/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4089 - binary_accuracy: 0.8309\n",
      "Epoch 116/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.4090 - binary_accuracy: 0.8278\n",
      "Epoch 117/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.4124 - binary_accuracy: 0.8246\n",
      "Epoch 118/225\n",
      "1754/1754 [==============================] - 1s 821us/step - loss: 0.4163 - binary_accuracy: 0.8293\n",
      "Epoch 119/225\n",
      "1754/1754 [==============================] - 1s 816us/step - loss: 0.4136 - binary_accuracy: 0.8276\n",
      "Epoch 120/225\n",
      "1754/1754 [==============================] - 1s 821us/step - loss: 0.4121 - binary_accuracy: 0.8239\n",
      "Epoch 121/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4207 - binary_accuracy: 0.8239\n",
      "Epoch 122/225\n",
      "1754/1754 [==============================] - 1s 816us/step - loss: 0.4091 - binary_accuracy: 0.8286\n",
      "Epoch 123/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4188 - binary_accuracy: 0.8247\n",
      "Epoch 124/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4152 - binary_accuracy: 0.8250\n",
      "Epoch 125/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4138 - binary_accuracy: 0.8266\n",
      "Epoch 126/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4161 - binary_accuracy: 0.8284\n",
      "Epoch 127/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4060 - binary_accuracy: 0.8326\n",
      "Epoch 128/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4207 - binary_accuracy: 0.8237\n",
      "Epoch 129/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4126 - binary_accuracy: 0.8291\n",
      "Epoch 130/225\n",
      "1754/1754 [==============================] - 1s 821us/step - loss: 0.4126 - binary_accuracy: 0.8256\n",
      "Epoch 131/225\n",
      "1754/1754 [==============================] - 1s 822us/step - loss: 0.4221 - binary_accuracy: 0.8196\n",
      "Epoch 132/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4119 - binary_accuracy: 0.8292\n",
      "Epoch 133/225\n",
      "1754/1754 [==============================] - 1s 815us/step - loss: 0.4160 - binary_accuracy: 0.8290\n",
      "Epoch 134/225\n",
      "1754/1754 [==============================] - 1s 817us/step - loss: 0.4128 - binary_accuracy: 0.8289\n",
      "Epoch 135/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4169 - binary_accuracy: 0.8192\n",
      "Epoch 136/225\n",
      "1754/1754 [==============================] - 1s 814us/step - loss: 0.4161 - binary_accuracy: 0.8230\n",
      "Epoch 137/225\n",
      "1754/1754 [==============================] - 1s 813us/step - loss: 0.4172 - binary_accuracy: 0.8232\n",
      "Epoch 138/225\n",
      "1754/1754 [==============================] - 1s 821us/step - loss: 0.4140 - binary_accuracy: 0.8272\n",
      "Epoch 139/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4097 - binary_accuracy: 0.8282\n",
      "Epoch 140/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4173 - binary_accuracy: 0.8223\n",
      "Epoch 141/225\n",
      "1754/1754 [==============================] - 1s 815us/step - loss: 0.4165 - binary_accuracy: 0.8253\n",
      "Epoch 142/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4109 - binary_accuracy: 0.8309\n",
      "Epoch 143/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4136 - binary_accuracy: 0.8263\n",
      "Epoch 144/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4076 - binary_accuracy: 0.8304\n",
      "Epoch 145/225\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4177 - binary_accuracy: 0.8250\n",
      "Epoch 146/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4129 - binary_accuracy: 0.8294\n",
      "Epoch 147/225\n",
      "1754/1754 [==============================] - 1s 817us/step - loss: 0.4088 - binary_accuracy: 0.8296\n",
      "Epoch 148/225\n",
      "1754/1754 [==============================] - 1s 816us/step - loss: 0.4148 - binary_accuracy: 0.8269\n",
      "Epoch 149/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4163 - binary_accuracy: 0.8220\n",
      "Epoch 150/225\n",
      "1754/1754 [==============================] - 1s 816us/step - loss: 0.4123 - binary_accuracy: 0.8257\n",
      "Epoch 151/225\n",
      "1754/1754 [==============================] - 1s 822us/step - loss: 0.4090 - binary_accuracy: 0.8270\n",
      "Epoch 152/225\n",
      "1754/1754 [==============================] - 1s 817us/step - loss: 0.4136 - binary_accuracy: 0.8288\n",
      "Epoch 153/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4064 - binary_accuracy: 0.8289\n",
      "Epoch 154/225\n",
      "1754/1754 [==============================] - 1s 817us/step - loss: 0.4134 - binary_accuracy: 0.8260\n",
      "Epoch 155/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4132 - binary_accuracy: 0.8248\n",
      "Epoch 156/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4121 - binary_accuracy: 0.8271\n",
      "Epoch 157/225\n",
      "1754/1754 [==============================] - 1s 830us/step - loss: 0.4113 - binary_accuracy: 0.8303\n",
      "Epoch 158/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4158 - binary_accuracy: 0.8273\n",
      "Epoch 159/225\n",
      "1754/1754 [==============================] - 1s 816us/step - loss: 0.4121 - binary_accuracy: 0.8250\n",
      "Epoch 160/225\n",
      "1754/1754 [==============================] - 1s 815us/step - loss: 0.4154 - binary_accuracy: 0.8263\n",
      "Epoch 161/225\n",
      "1754/1754 [==============================] - 1s 816us/step - loss: 0.4137 - binary_accuracy: 0.8232\n",
      "Epoch 162/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4123 - binary_accuracy: 0.8282\n",
      "Epoch 163/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4111 - binary_accuracy: 0.8298\n",
      "Epoch 164/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4109 - binary_accuracy: 0.8273\n",
      "Epoch 165/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4216 - binary_accuracy: 0.8247\n",
      "Epoch 166/225\n",
      "1754/1754 [==============================] - 1s 822us/step - loss: 0.4124 - binary_accuracy: 0.8253\n",
      "Epoch 167/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4134 - binary_accuracy: 0.8238\n",
      "Epoch 168/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4100 - binary_accuracy: 0.8280\n",
      "Epoch 169/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4124 - binary_accuracy: 0.8290\n",
      "Epoch 170/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4093 - binary_accuracy: 0.8275\n",
      "Epoch 171/225\n",
      "1754/1754 [==============================] - 1s 813us/step - loss: 0.4105 - binary_accuracy: 0.8302\n",
      "Epoch 172/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4098 - binary_accuracy: 0.8258\n",
      "Epoch 173/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4069 - binary_accuracy: 0.8288\n",
      "Epoch 174/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4120 - binary_accuracy: 0.8257\n",
      "Epoch 175/225\n",
      "1754/1754 [==============================] - 1s 830us/step - loss: 0.4125 - binary_accuracy: 0.8243\n",
      "Epoch 176/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4128 - binary_accuracy: 0.8255\n",
      "Epoch 177/225\n",
      "1754/1754 [==============================] - 1s 822us/step - loss: 0.4077 - binary_accuracy: 0.8281\n",
      "Epoch 178/225\n",
      "1754/1754 [==============================] - 1s 817us/step - loss: 0.4088 - binary_accuracy: 0.8282\n",
      "Epoch 179/225\n",
      "1754/1754 [==============================] - 1s 817us/step - loss: 0.4095 - binary_accuracy: 0.8259\n",
      "Epoch 180/225\n",
      "1754/1754 [==============================] - 1s 829us/step - loss: 0.4085 - binary_accuracy: 0.8286\n",
      "Epoch 181/225\n",
      "1754/1754 [==============================] - 1s 831us/step - loss: 0.4143 - binary_accuracy: 0.8218\n",
      "Epoch 182/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4133 - binary_accuracy: 0.8264\n",
      "Epoch 183/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4205 - binary_accuracy: 0.8250\n",
      "Epoch 184/225\n",
      "1754/1754 [==============================] - 1s 815us/step - loss: 0.4083 - binary_accuracy: 0.8279\n",
      "Epoch 185/225\n",
      "1754/1754 [==============================] - 1s 816us/step - loss: 0.4098 - binary_accuracy: 0.8293\n",
      "Epoch 186/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4054 - binary_accuracy: 0.8289\n",
      "Epoch 187/225\n",
      "1754/1754 [==============================] - 1s 814us/step - loss: 0.4070 - binary_accuracy: 0.8269\n",
      "Epoch 188/225\n",
      "1754/1754 [==============================] - 1s 815us/step - loss: 0.4080 - binary_accuracy: 0.8264\n",
      "Epoch 189/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4110 - binary_accuracy: 0.8300\n",
      "Epoch 190/225\n",
      "1754/1754 [==============================] - 1s 829us/step - loss: 0.4007 - binary_accuracy: 0.8329\n",
      "Epoch 191/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4130 - binary_accuracy: 0.8252\n",
      "Epoch 192/225\n",
      "1754/1754 [==============================] - 1s 817us/step - loss: 0.4022 - binary_accuracy: 0.8348\n",
      "Epoch 193/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.4121 - binary_accuracy: 0.8217\n",
      "Epoch 194/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.4096 - binary_accuracy: 0.8272\n",
      "Epoch 195/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4136 - binary_accuracy: 0.8236\n",
      "Epoch 196/225\n",
      "1754/1754 [==============================] - 1s 817us/step - loss: 0.4147 - binary_accuracy: 0.8275\n",
      "Epoch 197/225\n",
      "1754/1754 [==============================] - 1s 815us/step - loss: 0.4056 - binary_accuracy: 0.8277\n",
      "Epoch 198/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4082 - binary_accuracy: 0.8294\n",
      "Epoch 199/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4054 - binary_accuracy: 0.8284\n",
      "Epoch 200/225\n",
      "1754/1754 [==============================] - 1s 824us/step - loss: 0.4131 - binary_accuracy: 0.8251\n",
      "Epoch 201/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.4009 - binary_accuracy: 0.8325\n",
      "Epoch 202/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4047 - binary_accuracy: 0.8304\n",
      "Epoch 203/225\n",
      "1754/1754 [==============================] - 1s 822us/step - loss: 0.4131 - binary_accuracy: 0.8236\n",
      "Epoch 204/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4113 - binary_accuracy: 0.8258\n",
      "Epoch 205/225\n",
      "1754/1754 [==============================] - 1s 823us/step - loss: 0.4079 - binary_accuracy: 0.8300\n",
      "Epoch 206/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4075 - binary_accuracy: 0.8232\n",
      "Epoch 207/225\n",
      "1754/1754 [==============================] - 1s 817us/step - loss: 0.4136 - binary_accuracy: 0.8233\n",
      "Epoch 208/225\n",
      "1754/1754 [==============================] - 1s 816us/step - loss: 0.4120 - binary_accuracy: 0.8242\n",
      "Epoch 209/225\n",
      "1754/1754 [==============================] - 1s 817us/step - loss: 0.4065 - binary_accuracy: 0.8269\n",
      "Epoch 210/225\n",
      "1754/1754 [==============================] - 1s 825us/step - loss: 0.4066 - binary_accuracy: 0.8313\n",
      "Epoch 211/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4149 - binary_accuracy: 0.8190\n",
      "Epoch 212/225\n",
      "1754/1754 [==============================] - 1s 820us/step - loss: 0.4068 - binary_accuracy: 0.8268\n",
      "Epoch 213/225\n",
      "1754/1754 [==============================] - 1s 818us/step - loss: 0.4144 - binary_accuracy: 0.8251\n",
      "Epoch 214/225\n",
      "1754/1754 [==============================] - 1s 822us/step - loss: 0.4100 - binary_accuracy: 0.8248\n",
      "Epoch 215/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4155 - binary_accuracy: 0.8243\n",
      "Epoch 216/225\n",
      "1754/1754 [==============================] - 1s 813us/step - loss: 0.4076 - binary_accuracy: 0.8312\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 217/225\n",
      "1754/1754 [==============================] - 1s 821us/step - loss: 0.4095 - binary_accuracy: 0.8283\n",
      "Epoch 218/225\n",
      "1754/1754 [==============================] - 1s 822us/step - loss: 0.4106 - binary_accuracy: 0.8270\n",
      "Epoch 219/225\n",
      "1754/1754 [==============================] - 1s 822us/step - loss: 0.4101 - binary_accuracy: 0.8251\n",
      "Epoch 220/225\n",
      "1754/1754 [==============================] - 1s 815us/step - loss: 0.4150 - binary_accuracy: 0.8278\n",
      "Epoch 221/225\n",
      "1754/1754 [==============================] - 1s 817us/step - loss: 0.4092 - binary_accuracy: 0.8319\n",
      "Epoch 222/225\n",
      "1754/1754 [==============================] - 1s 819us/step - loss: 0.4029 - binary_accuracy: 0.8295\n",
      "Epoch 223/225\n",
      "1754/1754 [==============================] - 1s 822us/step - loss: 0.4120 - binary_accuracy: 0.8262\n",
      "Epoch 224/225\n",
      "1754/1754 [==============================] - 1s 821us/step - loss: 0.4085 - binary_accuracy: 0.8274\n",
      "Epoch 225/225\n",
      "1754/1754 [==============================] - 1s 826us/step - loss: 0.4079 - binary_accuracy: 0.8259\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Run 5-fold cross-validation to get average performance characteristics\n",
    "unit = 32\n",
    "ratio = 0.5\n",
    "epoch = 225\n",
    "(auroc_list, other_metric_list) = kfold_validation(unit, ratio, epoch, X_train_transformed, y_train, 'patient-id', verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "average precision:0.36352633931786027\n",
      "average recall:0.5886624350134049\n",
      "average specificity:0.7720560876428268\n",
      "average F1:0.44888757540286245\n"
     ]
    }
   ],
   "source": [
    "## other_metric_list has precision, recall, accuracy, specificity, f1, threshold for metrics\n",
    "ave_prec = np.mean([x[0] for x in other_metric_list])\n",
    "print(\"average precision:\" + str(ave_prec))\n",
    "\n",
    "ave_recall = np.mean([x[1] for x in other_metric_list])\n",
    "print(\"average recall:\" + str(ave_recall))\n",
    "\n",
    "ave_specificity = np.mean([x[3] for x in other_metric_list])\n",
    "print(\"average specificity:\" + str(ave_specificity))\n",
    "\n",
    "ave_f1 = np.mean([x[4] for x in other_metric_list])\n",
    "print(\"average F1:\" + str(ave_f1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "max F1: 0.48940269749518306\n"
     ]
    }
   ],
   "source": [
    "max_f1 = np.max([x[4] for x in other_metric_list])\n",
    "print(\"max F1: \" + str(max_f1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "unit: 32\tepoch: 225\tauroc_mean: 0.728596\tauroc_std: 0.014336\n"
     ]
    }
   ],
   "source": [
    "## Average metrics performance\n",
    "print(\"unit: %d\\tepoch: %d\\tauroc_mean: %f\\tauroc_std: %f\"  % (unit, epoch, np.mean(auroc_list), np.std(auroc_list)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.7077088055372767,\n",
       " 0.7433450644628484,\n",
       " 0.7229418636613524,\n",
       " 0.7462190124409652,\n",
       " 0.7227889553045588]"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "auroc_list"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}