pneumonia_vs_normal.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "_dGgVb8DPZMp",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "Bad key \"text.kerning_factor\" on line 4 in\n",
      "C:\\Users\\jrizz_000\\anaconda3\\anaconda3\\envs\\learn-env\\lib\\site-packages\\matplotlib\\mpl-data\\stylelib\\_classic_test_patch.mplstyle.\n",
      "You probably need to get an updated matplotlibrc file from\n",
      "http://github.com/matplotlib/matplotlib/blob/master/matplotlibrc.template\n",
      "or from the matplotlib source distribution\n"
     ]
    }
   ],
   "source": [
    "# importing neccessary libraries\n",
    "from glob import glob\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import random\n",
    "from skimage.io import imread\n",
    "import os, shutil\n",
    "import tensorflow as tf\n",
    "import keras\n",
    "from tensorflow.keras.layers import Input, MaxPooling2D\n",
    "from collections import Counter\n",
    "from scikitplot.metrics import plot_roc\n",
    "from sklearn import metrics\n",
    "import matplotlib.image as mpimg\n",
    "import matplotlib.pyplot as plt\n",
    "from PIL import Image\n",
    "import seaborn as sns\n",
    "from sklearn.preprocessing import LabelBinarizer\n",
    "import scikitplot\n",
    "import sklearn\n",
    "import cv2\n",
    "from tensorflow.keras.utils import to_categorical\n",
    "from sklearn.linear_model import Ridge\n",
    "from sklearn.model_selection import train_test_split, GridSearchCV\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.preprocessing import MinMaxScaler, StandardScaler\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.ensemble import VotingClassifier,AdaBoostClassifier,RandomForestClassifier,ExtraTreesClassifier,GradientBoostingClassifier,GradientBoostingRegressor\n",
    "from sklearn.tree import DecisionTreeClassifier, export_graphviz\n",
    "from sklearn.feature_selection import VarianceThreshold\n",
    "from sklearn import datasets \n",
    "from tensorflow.keras.applications import VGG16\n",
    "from tensorflow.keras.layers import AveragePooling2D\n",
    "from sklearn import tree\n",
    "from sklearn.naive_bayes import GaussianNB \n",
    "from sklearn.model_selection import validation_curve\n",
    "import imblearn\n",
    "from sklearn.decomposition import TruncatedSVD\n",
    "from sklearn.metrics import precision_score, precision_recall_curve, average_precision_score, recall_score, accuracy_score, f1_score, make_scorer, classification_report, auc, roc_curve, roc_auc_score, confusion_matrix\n",
    "from imblearn.pipeline import make_pipeline as make_pipeline_imb\n",
    "from imblearn.under_sampling import NearMiss\n",
    "from imblearn.metrics import classification_report_imbalanced\n",
    "from imblearn.over_sampling import SMOTE\n",
    "from keras.applications import MobileNet, DenseNet201, DenseNet121, VGG19, InceptionV3, MobileNetV2, Xception, ResNet152V2, NASNetLarge\n",
    "from keras.models import Sequential, Model\n",
    "from keras.layers import Dense, Dropout, Activation, Flatten, GlobalAveragePooling2D\n",
    "from keras.layers import Conv2D, MaxPool2D, ZeroPadding2D\n",
    "from keras.layers.normalization import BatchNormalization\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from keras.applications.resnet50 import preprocess_input, decode_predictions\n",
    "from tensorflow.python.keras.preprocessing.image import ImageDataGenerator, img_to_array, load_img\n",
    "from keras.optimizers import Adam, SGD\n",
    "from keras.callbacks import ModelCheckpoint, EarlyStopping, ReduceLROnPlateau\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data Processessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 340
    },
    "colab_type": "code",
    "id": "6pw0aLmIeoe0",
    "outputId": "32f6f93a-4837-49b5-f609-49000c28a1f6",
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total\n",
      "---------------------\n",
      "normal:  863\n",
      "normal no val:  206\n",
      "normal no test:  272\n",
      "---------------------\n",
      "pneumonia no train:  845\n",
      "pneumonia no val:  226\n",
      "pneumonia no test:  270\n",
      "\n",
      "Total\n",
      "---------------------\n",
      "total train:  1708\n",
      "total val:  432\n",
      "total test:  542\n"
     ]
    }
   ],
   "source": [
    "# counting the amount of x rays images\n",
    "len_normal_train = len([iq for iq in os.scandir('xrays/normalvspneumonia/train/normal')])\n",
    "len_normal_test = len([iq for iq in os.scandir('xrays/normalvspneumonia/test/normal')])\n",
    "len_normal_val = len([iq for iq in os.scandir('xrays/normalvspneumonia/val/normal')])\n",
    "\n",
    "len_pneumonia_train = len([iq for iq in os.scandir('xrays/normalvspneumonia/train/pneumonia')])\n",
    "len_pneumonia_val = len([iq for iq in os.scandir('xrays/normalvspneumonia/val/pneumonia')])\n",
    "len_pneumonia_test = len([iq for iq in os.scandir('xrays/normalvspneumonia/test/pneumonia')])\n",
    "\n",
    "len_train_total = len_normal_train + len_pneumonia_train\n",
    "len_val_total = len_pneumonia_val + len_normal_val\n",
    "len_test_total = len_normal_test + len_pneumonia_test\n",
    "\n",
    "print(\"Total\")\n",
    "print(\"---------------------\")\n",
    "print (\"normal: \", len_normal_train)\n",
    "print (\"normal no val: \", len_normal_val)\n",
    "print (\"normal no test: \", len_normal_test)\n",
    "print(\"---------------------\")\n",
    "print (\"pneumonia no train: \", len_pneumonia_train)\n",
    "print (\"pneumonia no val: \", len_pneumonia_val)\n",
    "print (\"pneumonia no test: \", len_pneumonia_test)\n",
    "print()\n",
    "print(\"Total\")\n",
    "print(\"---------------------\")\n",
    "print(\"total train: \", len_train_total)\n",
    "print(\"total val: \", len_val_total)\n",
    "print(\"total test: \", len_test_total)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# extracting the images\n",
    "DIR_NAME = 'xrays/normalvspneumonia/'\n",
    "\n",
    "imagePaths=[]\n",
    "for dirname, _, filenames in os.walk(DIR_NAME):\n",
    "    for filename in filenames:\n",
    "        imagePaths.append(os.path.join(dirname, filename))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (432).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (433).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (433)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (434).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (434)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (435).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (435)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (436).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (436)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (437).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (437)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (438).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (438)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (439).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (439)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (440).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (440)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (441).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (441)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (442).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (442)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (443).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (443)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (444).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (444)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (445).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (445)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (446).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (446)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (447).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (447)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (448).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (448)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (449).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (449)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (450).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (450)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (451).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (451)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (452).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (452)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (453).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (453)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (454).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (454)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (455).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (455)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (456).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (456)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (457).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (457)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (458).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (458)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (459).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (459)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (460).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (460)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (461).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (461)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (462).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (462)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (463).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (463)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (464).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (464)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (465).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (465)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (466).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (466)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (467).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (467)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (468).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (468)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (469).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (469)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (470).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (470)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (471).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (471)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (472).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (472)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (473).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (473)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (474).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (474)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (475).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (475)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (476).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (476)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (477).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (477)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (478).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (478)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (479).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (479)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (480).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (480)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (481).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (481)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (482).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (482)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (483).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (483)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (484).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (484)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (485).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (485)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (486).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (486)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (487).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (487)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (488).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (488)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (489).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (489)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (490).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (490)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (491).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (491)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (492).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (492)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (493).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (493)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (494).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (494)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (495).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (495)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (496).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (496)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (497).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (497)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (498).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (498)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (499).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (499)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (500).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (500)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (501).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (501)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (502).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (502)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (503).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (503)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (504).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (504)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (505).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (505)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (506).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (506)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (507).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (507)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (508).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (508)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (509).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (509)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (510).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (510)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (511).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (511)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (512).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (512)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (513).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (513)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (514).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (514)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (515).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (515)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (516).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (516)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (517).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (517)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (518).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (518)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (519).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (519)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (520).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (520)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (521).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (521)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (522).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (522)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (523).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (523)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (524).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (524)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (525).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (525)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (526).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (526)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (527).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (527)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (528).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (528)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (529).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (529)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (530).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (530)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (531).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (531)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (532).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (532)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (533).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (533)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (534).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (534)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (535).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (535)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (536).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (536)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (537).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (537)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (538).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (538)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (539).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (539)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (540).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (540)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (541).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (541)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (542).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (542)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (543).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (543)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (544).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (544)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (545).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (545)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (546).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (546)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (547).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (547)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (548).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (548)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (549).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (549)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (550).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (550)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (551).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (551)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (552).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (552)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (553).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (553)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (554).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (554)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (555).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (555)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (556).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (556)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (557).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (557)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (558).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (558)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (559).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (559)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (560).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (560)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (561).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (561)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (562).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (562)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (563).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (563)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (564).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (564)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (565).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (565)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (566).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (566)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (567).png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (567)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\normal\\\\NORMAL (568).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (538)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (539).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (539)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (540).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (540)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (541).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (541)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (542).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (542)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (543).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (543)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (544).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (544)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (545).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (545)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (546).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (546)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (547).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (547)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (548).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (548)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (549).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (549)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (550).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (550)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (551).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (551)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (552).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (552)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (553).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (553)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (554).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (554)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (555).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (555)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (556).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (556)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (557).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (557)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (558).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (558)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (559).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (559)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (560).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (560)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (561).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (561)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (562).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (562)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (563).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (563)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (564).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (564)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (565).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (565)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (566).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (566)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (567).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (567)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (568).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (568)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (569).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (569)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (570).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (570)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (571).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (571)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (572).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (572)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (573).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (573)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (574).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (574)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (575).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (575)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (576).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (576)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (577).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (577)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (578).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (578)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (579).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (579)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (580).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (580)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (581).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (581)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (582).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (582)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (583).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (583)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (584).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (584)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (585).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (585)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (586).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (586)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (587).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (587)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (588).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (588)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (589).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (589)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (590).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (590)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (591).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (591)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (592).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (592)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (593).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (593)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (594).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (594)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (595).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (595)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (596).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (596)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (597).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (597)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (598).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (598)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (599).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (599)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (600).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (600)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (601).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (601)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (602).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (602)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (603).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (603)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (604).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (604)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (605).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (605)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (606).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (606)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (607).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (607)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (608).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (608)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (609).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (609)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (610).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (610)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (611).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (611)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (612).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (612)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (613).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (613)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (614).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (614)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (615).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (615)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (616).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (616)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (617).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (617)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (618).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (618)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (619).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (619)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (620).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (620)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (621).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (621)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (622).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (622)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (623).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (623)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (624).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (624)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (625).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (625)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (626).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (626)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (627).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (627)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (628).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (628)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (629).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (629)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (630).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (630)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (631).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (631)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (632).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (632)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (633).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (633)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (634).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (634)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (635).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (635)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (636).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (636)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (637).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (637)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (638).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (638)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (639).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (639)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (640).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (640)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (641).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (641)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (642).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (642)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (643).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (643)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (644).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (644)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (645).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (645)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (646).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (646)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (647).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (647)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (648).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (648)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (649).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (649)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (650).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (650)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (651).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (651)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (652).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (652)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (653).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (653)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (654).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (654)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (655).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (655)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (656).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (656)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (657).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (657)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (658).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (658)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (659).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (659)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (660).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (660)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (661).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (661)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (662).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (662)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (663).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (663)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (664).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (664)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (665).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (665)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (666).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (666)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (667).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (667)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (668).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (668)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (669).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (669)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (670).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (670)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (671).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (671)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (672).png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (672)_1.png',\n",
       " 'xrays/normalvspneumonia/test\\\\pneumonia\\\\Viral Pneumonia (673).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (1).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (1)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (10).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (10)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (100).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (100)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (101).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (101)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (102).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (102)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (103).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (103)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (104).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (104)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (105).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (105)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (106).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (106)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (107).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (107)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (108).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (108)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (109).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (109)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (11).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (11)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (110).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (110)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (111).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (111)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (112).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (112)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (113).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (113)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (114).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (114)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (115).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (115)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (116).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (116)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (117).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (117)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (118).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (118)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (119).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (119)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (12).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (12)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (120).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (120)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (121).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (121)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (122).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (122)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (123).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (123)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (124).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (124)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (125).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (125)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (126).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (126)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (127).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (127)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (128).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (128)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (129).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (129)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (13).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (13)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (130).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (130)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (131).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (131)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (132).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (132)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (133).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (133)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (134).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (134)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (135).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (135)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (136).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (136)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (137).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (137)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (138).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (138)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (139).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (139)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (14).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (14)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (140).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (140)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (141).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (141)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (142).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (142)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (143).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (143)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (144).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (144)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (145).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (145)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (146).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (146)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (147).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (147)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (148).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (148)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (149).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (149)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (15).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (15)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (150).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (150)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (151).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (151)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (152).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (152)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (153).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (153)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (154).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (154)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (155).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (155)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (156).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (156)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (157).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (157)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (158).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (158)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (159).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (159)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (16).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (16)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (160).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (160)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (161).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (161)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (162).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (162)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (163).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (163)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (164).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (164)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (165).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (165)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (166).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (166)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (167).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (167)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (168).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (168)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (169).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (169)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (17).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (17)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (170).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (170)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (171).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (171)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (172).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (172)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (173).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (173)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (174).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (174)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (175).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (175)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (176).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (176)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (177).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (177)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (178).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (178)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (179).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (179)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (18).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (18)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (180).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (180)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (181).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (181)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (182).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (182)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (183).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (183)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (184).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (184)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (185).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (185)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (186).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (186)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (187).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (187)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (188).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (188)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (189).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (189)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (19).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (19)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (190).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (190)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (191).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (191)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (192).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (192)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (193).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (193)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (194).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (194)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (195).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (195)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (196).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (196)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (197).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (197)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (198).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (198)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (199).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (199)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (2).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (2)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (20).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (20)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (200).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (200)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (201).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (201)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (202).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (202)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (203).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (203)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (204).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (204)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (205).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (205)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (206).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (206)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (207).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (207)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (208).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (208)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (209).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (209)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (21).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (21)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (210).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (210)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (211).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (211)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (212).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (212)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (213).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (213)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (214).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (214)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (215).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (215)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (216).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (216)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (217).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (217)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (218).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (218)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (219).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (219)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (22).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (22)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (220).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (220)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (221).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (221)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (222).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (222)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (223).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (223)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (224).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (224)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (225).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (225)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (226).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (226)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (227).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (227)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (228).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (228)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (229).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (229)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (23).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (23)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (230).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (230)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (231).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (231)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (232).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (232)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (233).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (233)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (234).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (234)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (235).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (235)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (236).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (236)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (237).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (237)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (238).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (238)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (239).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (239)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (24).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (24)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (240).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (240)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (241).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (241)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (242).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (242)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (243).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (243)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (244).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (244)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (245).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (245)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (246).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (246)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (247).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (247)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (248).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (248)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (249).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (249)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (25).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (25)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (250).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (250)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (251).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (251)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (252).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (252)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (253).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (253)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (254).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (254)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (255).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (255)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (256).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (256)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (257).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (257)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (258).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (258)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (259).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (259)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (26).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (26)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (260).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (260)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (261).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (261)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (262).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (262)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (263).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (263)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (264).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (264)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (265).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (265)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (266).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (266)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (267).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (267)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (268).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (268)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (269).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (269)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (27).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (27)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (270).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (270)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (271).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (271)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (272).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (272)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (273).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (273)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (274).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (274)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (275).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (275)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (276).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (276)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (277).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (277)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (278).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (278)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (279).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (279)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (28).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (28)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (280).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (280)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (281).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (281)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (282).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (282)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (283).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (283)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (284).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (284)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (285).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (285)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (286).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (286)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (287).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (287)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (288).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (288)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (289).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (289)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (29).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (29)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (290).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (290)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (291).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (291)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (292).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (292)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (293).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (293)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (294).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (294)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (295).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (295)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (296).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (296)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (297).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (297)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (298).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (298)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (299).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (299)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (3).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (3)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (30).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (30)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (300).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (300)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (301).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (301)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (302).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (302)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (303).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (303)_1.png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (304).png',\n",
       " 'xrays/normalvspneumonia/train\\\\normal\\\\NORMAL (304)_1.png',\n",
       " ...]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# verifying if the images have been extracted\n",
    "imagePaths"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# assigining the labels to the images\n",
    "data = []\n",
    "labels = []\n",
    "# loop over the image paths\n",
    "for imagePath in imagePaths:\n",
    "    # extract the class label from the filename\n",
    "    label = imagePath.split(os.path.sep)[-2]\n",
    "    # load the image, swap color channels, and resize it to be a fixed\n",
    "    # 224x224 pixels while ignoring aspect ratio\n",
    "    image = cv2.imread(imagePath)\n",
    "    image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)\n",
    "    image = cv2.resize(image, (64, 64))\n",
    "    # update the data and labels lists, respectively\n",
    "    data.append(image)\n",
    "    labels.append(label)\n",
    "# convert the data and labels to NumPy arrays while scaling the pixel\n",
    "# intensities to the range [0, 1]\n",
    "data = np.array(data) / 255.0\n",
    "labels = np.array(labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2682, 64, 64, 3)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# verifying the shape\n",
    "data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['normal', 'normal', 'normal', ..., 'pneumonia', 'pneumonia',\n",
       "       'pneumonia'], dtype='<U9')"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# verifying the labels\n",
    "labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "# perform one-hot encoding on the labels\n",
    "lb = LabelBinarizer()\n",
    "labels = lb.fit_transform(labels)\n",
    "labels =to_categorical(labels)\n",
    "# partition the data into training and testing splits using 80% of\n",
    "# the data for training and the remaining 20% for testing\n",
    "(x_train, x_test, y_train, y_test) = train_test_split(data, labels,\n",
    "    test_size=0.20, stratify=labels,shuffle= True, random_state=42)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[[[0.03529412, 0.03529412, 0.03529412],\n",
       "         [0.15294118, 0.15294118, 0.15294118],\n",
       "         [0.26666667, 0.26666667, 0.26666667],\n",
       "         ...,\n",
       "         [0.23137255, 0.23137255, 0.23137255],\n",
       "         [0.17647059, 0.17647059, 0.17647059],\n",
       "         [0.07843137, 0.07843137, 0.07843137]],\n",
       "\n",
       "        [[0.01960784, 0.01960784, 0.01960784],\n",
       "         [0.15294118, 0.15294118, 0.15294118],\n",
       "         [0.23529412, 0.23529412, 0.23529412],\n",
       "         ...,\n",
       "         [0.22745098, 0.22745098, 0.22745098],\n",
       "         [0.16470588, 0.16470588, 0.16470588],\n",
       "         [0.0627451 , 0.0627451 , 0.0627451 ]],\n",
       "\n",
       "        [[0.00392157, 0.00392157, 0.00392157],\n",
       "         [0.12941176, 0.12941176, 0.12941176],\n",
       "         [0.23137255, 0.23137255, 0.23137255],\n",
       "         ...,\n",
       "         [0.22745098, 0.22745098, 0.22745098],\n",
       "         [0.1372549 , 0.1372549 , 0.1372549 ],\n",
       "         [0.05098039, 0.05098039, 0.05098039]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         ...,\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ]],\n",
       "\n",
       "        [[0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         ...,\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ]],\n",
       "\n",
       "        [[0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         ...,\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ]]],\n",
       "\n",
       "\n",
       "       [[[0.        , 0.        , 0.        ],\n",
       "         [0.0745098 , 0.0745098 , 0.0745098 ],\n",
       "         [0.21176471, 0.21176471, 0.21176471],\n",
       "         ...,\n",
       "         [0.35294118, 0.35294118, 0.35294118],\n",
       "         [0.31764706, 0.31764706, 0.31764706],\n",
       "         [0.25882353, 0.25882353, 0.25882353]],\n",
       "\n",
       "        [[0.        , 0.        , 0.        ],\n",
       "         [0.06666667, 0.06666667, 0.06666667],\n",
       "         [0.19607843, 0.19607843, 0.19607843],\n",
       "         ...,\n",
       "         [0.36470588, 0.36470588, 0.36470588],\n",
       "         [0.32156863, 0.32156863, 0.32156863],\n",
       "         [0.24705882, 0.24705882, 0.24705882]],\n",
       "\n",
       "        [[0.        , 0.        , 0.        ],\n",
       "         [0.05098039, 0.05098039, 0.05098039],\n",
       "         [0.18039216, 0.18039216, 0.18039216],\n",
       "         ...,\n",
       "         [0.36862745, 0.36862745, 0.36862745],\n",
       "         [0.32156863, 0.32156863, 0.32156863],\n",
       "         [0.23137255, 0.23137255, 0.23137255]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         ...,\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ]],\n",
       "\n",
       "        [[0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         ...,\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ]],\n",
       "\n",
       "        [[0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         ...,\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ]]],\n",
       "\n",
       "\n",
       "       [[[0.28235294, 0.28235294, 0.28235294],\n",
       "         [0.31764706, 0.31764706, 0.31764706],\n",
       "         [0.37254902, 0.37254902, 0.37254902],\n",
       "         ...,\n",
       "         [0.30588235, 0.30588235, 0.30588235],\n",
       "         [0.25882353, 0.25882353, 0.25882353],\n",
       "         [0.18039216, 0.18039216, 0.18039216]],\n",
       "\n",
       "        [[0.27058824, 0.27058824, 0.27058824],\n",
       "         [0.31372549, 0.31372549, 0.31372549],\n",
       "         [0.34901961, 0.34901961, 0.34901961],\n",
       "         ...,\n",
       "         [0.27058824, 0.27058824, 0.27058824],\n",
       "         [0.21568627, 0.21568627, 0.21568627],\n",
       "         [0.16470588, 0.16470588, 0.16470588]],\n",
       "\n",
       "        [[0.25098039, 0.25098039, 0.25098039],\n",
       "         [0.31764706, 0.31764706, 0.31764706],\n",
       "         [0.34901961, 0.34901961, 0.34901961],\n",
       "         ...,\n",
       "         [0.24705882, 0.24705882, 0.24705882],\n",
       "         [0.18823529, 0.18823529, 0.18823529],\n",
       "         [0.14901961, 0.14901961, 0.14901961]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[0.        , 0.        , 0.        ],\n",
       "         [0.00392157, 0.00392157, 0.00392157],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         ...,\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ]],\n",
       "\n",
       "        [[0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         ...,\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ]],\n",
       "\n",
       "        [[0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         ...,\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ]]],\n",
       "\n",
       "\n",
       "       ...,\n",
       "\n",
       "\n",
       "       [[[0.71372549, 0.71372549, 0.71372549],\n",
       "         [0.75294118, 0.75294118, 0.75294118],\n",
       "         [0.89803922, 0.89803922, 0.89803922],\n",
       "         ...,\n",
       "         [0.41960784, 0.41960784, 0.41960784],\n",
       "         [0.41960784, 0.41960784, 0.41960784],\n",
       "         [0.4       , 0.4       , 0.4       ]],\n",
       "\n",
       "        [[0.58039216, 0.58039216, 0.58039216],\n",
       "         [0.70980392, 0.70980392, 0.70980392],\n",
       "         [0.82352941, 0.82352941, 0.82352941],\n",
       "         ...,\n",
       "         [0.39607843, 0.39607843, 0.39607843],\n",
       "         [0.39215686, 0.39215686, 0.39215686],\n",
       "         [0.35686275, 0.35686275, 0.35686275]],\n",
       "\n",
       "        [[0.50196078, 0.50196078, 0.50196078],\n",
       "         [0.60784314, 0.60784314, 0.60784314],\n",
       "         [0.72941176, 0.72941176, 0.72941176],\n",
       "         ...,\n",
       "         [0.38039216, 0.38039216, 0.38039216],\n",
       "         [0.35686275, 0.35686275, 0.35686275],\n",
       "         [0.36862745, 0.36862745, 0.36862745]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         ...,\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ]],\n",
       "\n",
       "        [[0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         ...,\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ]],\n",
       "\n",
       "        [[0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         ...,\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ]]],\n",
       "\n",
       "\n",
       "       [[[0.00392157, 0.00392157, 0.00392157],\n",
       "         [0.11372549, 0.11372549, 0.11372549],\n",
       "         [0.20784314, 0.20784314, 0.20784314],\n",
       "         ...,\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.02352941, 0.02352941, 0.02352941]],\n",
       "\n",
       "        [[0.        , 0.        , 0.        ],\n",
       "         [0.09411765, 0.09411765, 0.09411765],\n",
       "         [0.2       , 0.2       , 0.2       ],\n",
       "         ...,\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.01960784, 0.01960784, 0.01960784]],\n",
       "\n",
       "        [[0.        , 0.        , 0.        ],\n",
       "         [0.07843137, 0.07843137, 0.07843137],\n",
       "         [0.16470588, 0.16470588, 0.16470588],\n",
       "         ...,\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.02352941, 0.02352941, 0.02352941]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         ...,\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ]],\n",
       "\n",
       "        [[0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         ...,\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ]],\n",
       "\n",
       "        [[0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         ...,\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ],\n",
       "         [0.        , 0.        , 0.        ]]],\n",
       "\n",
       "\n",
       "       [[[0.41176471, 0.41176471, 0.41176471],\n",
       "         [0.38039216, 0.38039216, 0.38039216],\n",
       "         [0.41568627, 0.41568627, 0.41568627],\n",
       "         ...,\n",
       "         [0.59215686, 0.59215686, 0.59215686],\n",
       "         [0.56470588, 0.56470588, 0.56470588],\n",
       "         [0.57254902, 0.57254902, 0.57254902]],\n",
       "\n",
       "        [[0.39215686, 0.39215686, 0.39215686],\n",
       "         [0.39215686, 0.39215686, 0.39215686],\n",
       "         [0.40784314, 0.40784314, 0.40784314],\n",
       "         ...,\n",
       "         [0.6       , 0.6       , 0.6       ],\n",
       "         [0.56470588, 0.56470588, 0.56470588],\n",
       "         [0.55294118, 0.55294118, 0.55294118]],\n",
       "\n",
       "        [[0.38431373, 0.38431373, 0.38431373],\n",
       "         [0.4       , 0.4       , 0.4       ],\n",
       "         [0.43921569, 0.43921569, 0.43921569],\n",
       "         ...,\n",
       "         [0.59607843, 0.59607843, 0.59607843],\n",
       "         [0.55686275, 0.55686275, 0.55686275],\n",
       "         [0.57254902, 0.57254902, 0.57254902]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[0.17647059, 0.17647059, 0.17647059],\n",
       "         [0.23921569, 0.23921569, 0.23921569],\n",
       "         [0.31764706, 0.31764706, 0.31764706],\n",
       "         ...,\n",
       "         [0.3372549 , 0.3372549 , 0.3372549 ],\n",
       "         [0.22352941, 0.22352941, 0.22352941],\n",
       "         [0.07058824, 0.07058824, 0.07058824]],\n",
       "\n",
       "        [[0.18431373, 0.18431373, 0.18431373],\n",
       "         [0.24705882, 0.24705882, 0.24705882],\n",
       "         [0.34901961, 0.34901961, 0.34901961],\n",
       "         ...,\n",
       "         [0.33333333, 0.33333333, 0.33333333],\n",
       "         [0.24313725, 0.24313725, 0.24313725],\n",
       "         [0.05882353, 0.05882353, 0.05882353]],\n",
       "\n",
       "        [[0.21176471, 0.21176471, 0.21176471],\n",
       "         [0.29803922, 0.29803922, 0.29803922],\n",
       "         [0.41568627, 0.41568627, 0.41568627],\n",
       "         ...,\n",
       "         [0.35294118, 0.35294118, 0.35294118],\n",
       "         [0.25490196, 0.25490196, 0.25490196],\n",
       "         [0.06666667, 0.06666667, 0.06666667]]]])"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_train"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data Augmentation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Data Augmentation\n",
    "def process_data(x_train,y_train, x_test,y_test, batch_size):\n",
    "    #to prevent overfitting\n",
    "    train_datagen = ImageDataGenerator(shear_range=0.2,\n",
    "                                   rotation_range=20,\n",
    "                                   width_shift_range=0.2,\n",
    "                                   height_shift_range=0.2,\n",
    "                                   horizontal_flip=True,\n",
    "                                   vertical_flip=False,\n",
    "                                         zoom_range=0.2)\n",
    "\n",
    "\n",
    "    validation_datagen = ImageDataGenerator()\n",
    "\n",
    "    train_generator = train_datagen.flow(x_train,y_train,\n",
    "                                                        batch_size=batch_size)\n",
    "\n",
    "    validation_generator = validation_datagen.flow(x_test,y_test,\n",
    "                                                  batch_size=batch_size)\n",
    "\n",
    "\n",
    "\n",
    "    return train_generator, validation_generator "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Transfer Learning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "# build sequencial model\n",
    "def build_model(bottom_model, num_classes):\n",
    "    top_model = bottom_model.output\n",
    "    top_model = MaxPooling2D(pool_size=(2,2), strides=2)(top_model)\n",
    "    top_model = Flatten(name=\"flatten\")(top_model)\n",
    "    top_model = Dense(2, activation='sigmoid')(top_model)\n",
    "    model = Model(inputs=bottom_model.input, outputs=top_model)\n",
    "    \n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# adding top layer\n",
    "def addTopModel(bottom_model, num_classes):\n",
    "    top_model = bottom_model.output\n",
    "    top_model = MaxPooling2D(pool_size=(2,2), strides=2)(top_model)\n",
    "    top_model = Flatten(name=\"flatten\")(top_model)\n",
    "    top_model = Dense(512, activation=\"relu\")(top_model)\n",
    "    top_model = Dropout(0.5)(top_model)\n",
    "    top_model = Dense(2, activation='sigmoid')(top_model)\n",
    "    model = Model(inputs=bottom_model.input, outputs=top_model)\n",
    "    \n",
    "    return model\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Confusion Matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# plotting confusion matrix for testing\n",
    "def plot_confusion_matrix(model, x_test, y_test):\n",
    "    fig, ax = plt.subplots(figsize=(8,6))\n",
    "\n",
    "    classes = ['PNEUMONIA','NORMAL']\n",
    "    \n",
    "    y_pred = model.predict(x_test, batch_size=batch_size)\n",
    "    y_pred = np.argmax(y_pred, axis=1)\n",
    "    y_test = np.argmax(y_test, axis=1) \n",
    "    \n",
    "    print('Confusion Matrix')\n",
    "    cm = confusion_matrix(y_test, y_pred, normalize='true')\n",
    "    #cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    " \n",
    "    sns.heatmap(cm,cmap='Purples',fmt='g', annot=True)\n",
    "    tick_marks = [0.5,1.5]\n",
    "    plt.xticks(tick_marks, classes)\n",
    "    plt.yticks(tick_marks, classes)\n",
    "\n",
    "    plt.title('Confusion Matrix - Normalized')\n",
    "    plt.ylabel('True label')\n",
    "    plt.xlabel('Predicted label')\n",
    "    bottom, top = ax.get_ylim()\n",
    "    \n",
    "    ax.set_ylim(bottom + 0.5, top - 0.5)\n",
    "    \n",
    "    plt.show()\n",
    "    \n",
    "    acc = accuracy_score(y_test, y_pred)\n",
    "    precision = precision_score(y_test, y_pred, average=None)\n",
    "    recall = recall_score(y_test, y_pred, average=None)\n",
    "    f1 = f1_score(y_test, y_pred, average=None)\n",
    "    \n",
    "    print(\"Precision Score: {}\".format(precision))\n",
    "    print(\"Recall Score: {}\".format(recall))\n",
    "    print(\"F1 Score: {}\".format(f1))\n",
    "    print(\"Accuracy Score: {}\".format(acc))\n",
    "\n",
    "    return plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# ROC AUC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "#ROC curve\n",
    "def multiclass_roc_auc_score(x_test, y_test, model, average=\"micro\"):\n",
    "    \n",
    "\n",
    "    y_pred = model.predict(x_test, batch_size=batch_size)\n",
    "    \n",
    "    # Convert to Binary classes\n",
    "    y_pred_bin = np.argmax(y_pred, axis=1)\n",
    "    y_test_bin = np.argmax(y_test, axis=1)\n",
    "\n",
    "    \n",
    "    fpr, tpr, thresholds = roc_curve(y_test_bin, y_pred_bin)\n",
    "    \n",
    "    auc_keras = auc(fpr, tpr)\n",
    "    \n",
    "    print('AUC: {}'.format(auc_keras))    \n",
    "    print('Log Loss: {}'.format(metrics.log_loss(y_test.argmax(axis=1), y_pred)))\n",
    "\n",
    "        \n",
    "    plt.plot(fpr, tpr,  color='darkorange', label='ROC curve (area = %0.2f)' % auc_keras)\n",
    "    plt.plot([0, 1], [0, 1], color='navy', lw=2, linestyle='--')\n",
    "    plt.xlim([0.0, 1.0])\n",
    "    plt.ylim([0.0, 1.0])\n",
    "    plt.rcParams['font.size'] = 12\n",
    "    plt.title('ROC curve for our model')\n",
    "    plt.xlabel('False Positive Rate')\n",
    "    plt.ylabel('True Positive Rate')\n",
    "    plt.legend(loc='best', fancybox=True)\n",
    "    plt.grid(True)\n",
    "\n",
    "\n",
    "    return plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Model Loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "#model loss\n",
    "def plot_learning_curves(r):\n",
    "    plt.figure(figsize=(12,4))\n",
    "\n",
    "    plt.subplot(1,2,1)\n",
    "    plt.plot(r.history['loss'])\n",
    "    plt.plot(r.history['val_loss'])\n",
    "    plt.title('model loss')\n",
    "    plt.ylabel('loss')\n",
    "    plt.xlabel('epoch')\n",
    "    plt.grid(True)\n",
    "    \n",
    "    plt.subplot(1,2,2)\n",
    "    plt.plot(r.history['accuracy'])\n",
    "    plt.plot(r.history['val_accuracy'])\n",
    "    plt.title('model accuracy')\n",
    "    plt.ylabel('accuracy')\n",
    "    plt.xlabel('epoch')\n",
    "    plt.legend(['train', 'test'], loc='upper left')\n",
    "    plt.grid(True)\n",
    "    plt.tight_layout()\n",
    "    \n",
    "    return plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Assigning Parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Moo8-9Hedbty"
   },
   "outputs": [],
   "source": [
    "#parameters\n",
    "data_dir = '/'\n",
    "version = '-v5-'\n",
    "num_classes = 2\n",
    "ig_size= 64\n",
    "epochs = 100\n",
    "batch_size = 32\n",
    "\n",
    "models_loss = 'binary_crossentropy'\n",
    "models_opt = 'adam' #Adam(lr=0.001, decay=0.001/600) #SGD(learning_rate=0.001, momentum=0.9) #ADAM(lr=0.001)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Modeling - Base CNN Model\n",
    "\n",
    "First, I used the sequential keras model which is th easiest way to build a model in keras. It allows one to build a model, layer by layer. My first layer will be a Conv2D layer that will deal with my input x-ray images of shape 64, 64, 3 using 64 nodes. The 3 in the shape sigifying that the images are in RGB. The kernel size is set to 3, which means the filter that would be used to scan through every pixel will be a matrix of 3x3. \n",
    "\n",
    "Next, a MaxPooling2D layer was utilized in order to reduce overfitting and dimentionality i.e in order to reduce the number of parameters to learn from and the amount of comuputation performed in the network. I set it to divide the dimentions by a matrix of 2x2 while moving 2 strides after choosing the maximum value of each patch. After this, I flattened the output from the MaxPooling layer in order to make it linear and pass it to the Dense layer implemented, which uses a node of 2 which represents the number of output classes to be predicted."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_5\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_3_input (InputLayer)  [(None, 64, 64, 3)]       0         \n",
      "_________________________________________________________________\n",
      "conv2d_3 (Conv2D)            (None, 62, 62, 64)        1792      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_4 (MaxPooling2 (None, 31, 31, 64)        0         \n",
      "_________________________________________________________________\n",
      "flatten (Flatten)            (None, 61504)             0         \n",
      "_________________________________________________________________\n",
      "dense_11 (Dense)             (None, 2)                 123010    \n",
      "=================================================================\n",
      "Total params: 124,802\n",
      "Trainable params: 123,010\n",
      "Non-trainable params: 1,792\n",
      "_________________________________________________________________\n",
      "Epoch 1/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.7545 - accuracy: 0.5692 ETA: 0s - loss: 0.774\n",
      "Epoch 00001: val_loss improved from inf to 0.73971, saving model to /sequentialpnd.h5\n",
      "68/67 [==============================] - 7s 101ms/step - loss: 0.7545 - accuracy: 0.5692 - val_loss: 0.7397 - val_accuracy: 0.5140 - lr: 0.0010\n",
      "Epoch 2/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.6036 - accuracy: 0.6681\n",
      "Epoch 00002: val_loss improved from 0.73971 to 0.45842, saving model to /sequentialpnd.h5\n",
      "68/67 [==============================] - 6s 91ms/step - loss: 0.6036 - accuracy: 0.6681 - val_loss: 0.4584 - val_accuracy: 0.8585 - lr: 0.0010\n",
      "Epoch 3/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.5263 - accuracy: 0.7604\n",
      "Epoch 00003: val_loss improved from 0.45842 to 0.44041, saving model to /sequentialpnd.h5\n",
      "68/67 [==============================] - 6s 89ms/step - loss: 0.5263 - accuracy: 0.7604 - val_loss: 0.4404 - val_accuracy: 0.8492 - lr: 0.0010\n",
      "Epoch 4/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.5120 - accuracy: 0.7501\n",
      "Epoch 00004: val_loss did not improve from 0.44041\n",
      "68/67 [==============================] - 6s 88ms/step - loss: 0.5120 - accuracy: 0.7501 - val_loss: 0.4478 - val_accuracy: 0.8175 - lr: 0.0010\n",
      "Epoch 5/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4966 - accuracy: 0.7739\n",
      "Epoch 00005: val_loss did not improve from 0.44041\n",
      "68/67 [==============================] - 6s 90ms/step - loss: 0.4966 - accuracy: 0.7739 - val_loss: 0.5296 - val_accuracy: 0.7356 - lr: 0.0010\n",
      "Epoch 6/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.6015 - accuracy: 0.7282\n",
      "Epoch 00006: val_loss improved from 0.44041 to 0.35214, saving model to /sequentialpnd.h5\n",
      "68/67 [==============================] - 6s 91ms/step - loss: 0.6015 - accuracy: 0.7282 - val_loss: 0.3521 - val_accuracy: 0.8845 - lr: 0.0010\n",
      "Epoch 7/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.5042 - accuracy: 0.7534\n",
      "Epoch 00007: val_loss did not improve from 0.35214\n",
      "68/67 [==============================] - 6s 88ms/step - loss: 0.5042 - accuracy: 0.7534 - val_loss: 0.6448 - val_accuracy: 0.6816 - lr: 0.0010\n",
      "Epoch 8/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4918 - accuracy: 0.7655\n",
      "Epoch 00008: val_loss did not improve from 0.35214\n",
      "68/67 [==============================] - 6s 95ms/step - loss: 0.4918 - accuracy: 0.7655 - val_loss: 0.5846 - val_accuracy: 0.7225 - lr: 0.0010\n",
      "Epoch 9/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4740 - accuracy: 0.7790\n",
      "Epoch 00009: val_loss did not improve from 0.35214\n",
      "68/67 [==============================] - 6s 90ms/step - loss: 0.4740 - accuracy: 0.7790 - val_loss: 0.5195 - val_accuracy: 0.7691 - lr: 0.0010\n",
      "Epoch 10/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4952 - accuracy: 0.7613\n",
      "Epoch 00010: val_loss did not improve from 0.35214\n",
      "68/67 [==============================] - 6s 88ms/step - loss: 0.4952 - accuracy: 0.7613 - val_loss: 0.4859 - val_accuracy: 0.8101 - lr: 0.0010\n",
      "Epoch 11/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4417 - accuracy: 0.7869\n",
      "Epoch 00011: val_loss did not improve from 0.35214\n",
      "68/67 [==============================] - 6s 92ms/step - loss: 0.4417 - accuracy: 0.7869 - val_loss: 0.4880 - val_accuracy: 0.8156 - lr: 0.0010\n",
      "Epoch 12/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4445 - accuracy: 0.7944\n",
      "Epoch 00012: val_loss did not improve from 0.35214\n",
      "68/67 [==============================] - 6s 88ms/step - loss: 0.4445 - accuracy: 0.7944 - val_loss: 0.8577 - val_accuracy: 0.6145 - lr: 0.0010\n",
      "Epoch 13/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4297 - accuracy: 0.8126\n",
      "Epoch 00013: val_loss did not improve from 0.35214\n",
      "68/67 [==============================] - 6s 88ms/step - loss: 0.4297 - accuracy: 0.8126 - val_loss: 0.8215 - val_accuracy: 0.6276 - lr: 0.0010\n",
      "Epoch 14/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4500 - accuracy: 0.7902\n",
      "Epoch 00014: val_loss did not improve from 0.35214\n",
      "68/67 [==============================] - 6s 88ms/step - loss: 0.4500 - accuracy: 0.7902 - val_loss: 0.4625 - val_accuracy: 0.8305 - lr: 0.0010\n",
      "Epoch 15/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4146 - accuracy: 0.8163\n",
      "Epoch 00015: val_loss did not improve from 0.35214\n",
      "68/67 [==============================] - 6s 88ms/step - loss: 0.4146 - accuracy: 0.8163 - val_loss: 0.6669 - val_accuracy: 0.7020 - lr: 0.0010\n",
      "Epoch 16/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4175 - accuracy: 0.8219\n",
      "Epoch 00016: val_loss did not improve from 0.35214\n",
      "68/67 [==============================] - 6s 88ms/step - loss: 0.4175 - accuracy: 0.8219 - val_loss: 0.4310 - val_accuracy: 0.8380 - lr: 0.0010\n",
      "Epoch 17/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4097 - accuracy: 0.8182\n",
      "Epoch 00017: val_loss did not improve from 0.35214\n",
      "68/67 [==============================] - 6s 89ms/step - loss: 0.4097 - accuracy: 0.8182 - val_loss: 0.5244 - val_accuracy: 0.7896 - lr: 0.0010\n",
      "Epoch 18/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4179 - accuracy: 0.8168\n",
      "Epoch 00018: val_loss did not improve from 0.35214\n",
      "68/67 [==============================] - 6s 88ms/step - loss: 0.4179 - accuracy: 0.8168 - val_loss: 0.6127 - val_accuracy: 0.7449 - lr: 0.0010\n",
      "Epoch 19/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4046 - accuracy: 0.8219\n",
      "Epoch 00019: val_loss did not improve from 0.35214\n",
      "68/67 [==============================] - 6s 88ms/step - loss: 0.4046 - accuracy: 0.8219 - val_loss: 0.8471 - val_accuracy: 0.6425 - lr: 0.0010\n",
      "Epoch 20/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4048 - accuracy: 0.8191\n",
      "Epoch 00020: val_loss did not improve from 0.35214\n",
      "68/67 [==============================] - 6s 89ms/step - loss: 0.4048 - accuracy: 0.8191 - val_loss: 0.5269 - val_accuracy: 0.7970 - lr: 0.0010\n",
      "Epoch 21/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4146 - accuracy: 0.8070\n",
      "Epoch 00021: val_loss did not improve from 0.35214\n",
      "68/67 [==============================] - 6s 88ms/step - loss: 0.4146 - accuracy: 0.8070 - val_loss: 0.6044 - val_accuracy: 0.7523 - lr: 0.0010\n",
      "Epoch 22/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.5257 - accuracy: 0.7678 ETA: 1s - loss:\n",
      "Epoch 00022: val_loss improved from 0.35214 to 0.32315, saving model to /sequentialpnd.h5\n",
      "68/67 [==============================] - 6s 96ms/step - loss: 0.5257 - accuracy: 0.7678 - val_loss: 0.3231 - val_accuracy: 0.9032 - lr: 0.0010\n",
      "Epoch 23/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4304 - accuracy: 0.8121\n",
      "Epoch 00023: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 89ms/step - loss: 0.4304 - accuracy: 0.8121 - val_loss: 0.5442 - val_accuracy: 0.7877 - lr: 0.0010\n",
      "Epoch 24/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3905 - accuracy: 0.8312\n",
      "Epoch 00024: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 88ms/step - loss: 0.3905 - accuracy: 0.8312 - val_loss: 0.7062 - val_accuracy: 0.7151 - lr: 0.0010\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 25/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4151 - accuracy: 0.8126\n",
      "Epoch 00025: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 87ms/step - loss: 0.4151 - accuracy: 0.8126 - val_loss: 0.5793 - val_accuracy: 0.7914 - lr: 0.0010\n",
      "Epoch 26/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3921 - accuracy: 0.8340\n",
      "Epoch 00026: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 85ms/step - loss: 0.3921 - accuracy: 0.8340 - val_loss: 0.7882 - val_accuracy: 0.6816 - lr: 0.0010\n",
      "Epoch 27/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4137 - accuracy: 0.8191\n",
      "Epoch 00027: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 85ms/step - loss: 0.4137 - accuracy: 0.8191 - val_loss: 0.4644 - val_accuracy: 0.8287 - lr: 0.0010\n",
      "Epoch 28/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4117 - accuracy: 0.8224\n",
      "Epoch 00028: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 87ms/step - loss: 0.4117 - accuracy: 0.8224 - val_loss: 0.8086 - val_accuracy: 0.6778 - lr: 0.0010\n",
      "Epoch 29/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4456 - accuracy: 0.8005\n",
      "Epoch 00029: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 87ms/step - loss: 0.4456 - accuracy: 0.8005 - val_loss: 0.4446 - val_accuracy: 0.8361 - lr: 0.0010\n",
      "Epoch 30/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3977 - accuracy: 0.8354\n",
      "Epoch 00030: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 86ms/step - loss: 0.3977 - accuracy: 0.8354 - val_loss: 0.5492 - val_accuracy: 0.8082 - lr: 0.0010\n",
      "Epoch 31/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4183 - accuracy: 0.8131\n",
      "Epoch 00031: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 85ms/step - loss: 0.4183 - accuracy: 0.8131 - val_loss: 0.7922 - val_accuracy: 0.6946 - lr: 0.0010\n",
      "Epoch 32/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3881 - accuracy: 0.8280\n",
      "Epoch 00032: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 86ms/step - loss: 0.3881 - accuracy: 0.8280 - val_loss: 1.0875 - val_accuracy: 0.5978 - lr: 0.0010\n",
      "Epoch 33/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3921 - accuracy: 0.8284\n",
      "Epoch 00033: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 91ms/step - loss: 0.3921 - accuracy: 0.8284 - val_loss: 0.6877 - val_accuracy: 0.7374 - lr: 0.0010\n",
      "Epoch 34/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4074 - accuracy: 0.8214\n",
      "Epoch 00034: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 86ms/step - loss: 0.4074 - accuracy: 0.8214 - val_loss: 0.7478 - val_accuracy: 0.7244 - lr: 0.0010\n",
      "Epoch 35/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3883 - accuracy: 0.8322\n",
      "Epoch 00035: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 86ms/step - loss: 0.3883 - accuracy: 0.8322 - val_loss: 0.3590 - val_accuracy: 0.8939 - lr: 0.0010\n",
      "Epoch 36/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4298 - accuracy: 0.8014\n",
      "Epoch 00036: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 85ms/step - loss: 0.4298 - accuracy: 0.8014 - val_loss: 0.5664 - val_accuracy: 0.8063 - lr: 0.0010\n",
      "Epoch 37/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3975 - accuracy: 0.8284\n",
      "Epoch 00037: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 86ms/step - loss: 0.3975 - accuracy: 0.8284 - val_loss: 0.4494 - val_accuracy: 0.8492 - lr: 0.0010\n",
      "Epoch 38/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3704 - accuracy: 0.8434\n",
      "Epoch 00038: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 86ms/step - loss: 0.3704 - accuracy: 0.8434 - val_loss: 0.7133 - val_accuracy: 0.7318 - lr: 0.0010\n",
      "Epoch 39/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3990 - accuracy: 0.8238\n",
      "Epoch 00039: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 86ms/step - loss: 0.3990 - accuracy: 0.8238 - val_loss: 0.9035 - val_accuracy: 0.6518 - lr: 0.0010\n",
      "Epoch 40/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3758 - accuracy: 0.8378\n",
      "Epoch 00040: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 91ms/step - loss: 0.3758 - accuracy: 0.8378 - val_loss: 0.5250 - val_accuracy: 0.8212 - lr: 0.0010\n",
      "Epoch 41/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4141 - accuracy: 0.8163\n",
      "Epoch 00041: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 85ms/step - loss: 0.4141 - accuracy: 0.8163 - val_loss: 0.5091 - val_accuracy: 0.8175 - lr: 0.0010\n",
      "Epoch 42/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4289 - accuracy: 0.8145\n",
      "Epoch 00042: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 86ms/step - loss: 0.4289 - accuracy: 0.8145 - val_loss: 0.9969 - val_accuracy: 0.6387 - lr: 0.0010\n",
      "Epoch 43/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4335 - accuracy: 0.8103\n",
      "Epoch 00043: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 86ms/step - loss: 0.4335 - accuracy: 0.8103 - val_loss: 0.7198 - val_accuracy: 0.7337 - lr: 0.0010\n",
      "Epoch 44/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4213 - accuracy: 0.8210\n",
      "Epoch 00044: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 90ms/step - loss: 0.4213 - accuracy: 0.8210 - val_loss: 0.5988 - val_accuracy: 0.7840 - lr: 0.0010\n",
      "Epoch 45/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3913 - accuracy: 0.8266\n",
      "Epoch 00045: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 88ms/step - loss: 0.3913 - accuracy: 0.8266 - val_loss: 0.4633 - val_accuracy: 0.8492 - lr: 0.0010\n",
      "Epoch 46/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4001 - accuracy: 0.8368\n",
      "Epoch 00046: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 89ms/step - loss: 0.4001 - accuracy: 0.8368 - val_loss: 0.5623 - val_accuracy: 0.7989 - lr: 0.0010\n",
      "Epoch 47/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3764 - accuracy: 0.8401\n",
      "Epoch 00047: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 90ms/step - loss: 0.3764 - accuracy: 0.8401 - val_loss: 0.7267 - val_accuracy: 0.7318 - lr: 0.0010\n",
      "Epoch 48/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3700 - accuracy: 0.8396\n",
      "Epoch 00048: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 88ms/step - loss: 0.3700 - accuracy: 0.8396 - val_loss: 0.6834 - val_accuracy: 0.7412 - lr: 0.0010\n",
      "Epoch 49/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3507 - accuracy: 0.8531\n",
      "Epoch 00049: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 90ms/step - loss: 0.3507 - accuracy: 0.8531 - val_loss: 0.4747 - val_accuracy: 0.8361 - lr: 0.0010\n",
      "Epoch 50/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3635 - accuracy: 0.8513\n",
      "Epoch 00050: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 87ms/step - loss: 0.3635 - accuracy: 0.8513 - val_loss: 0.4582 - val_accuracy: 0.8417 - lr: 0.0010\n",
      "Epoch 51/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3771 - accuracy: 0.8350\n",
      "Epoch 00051: val_loss did not improve from 0.32315\n",
      "68/67 [==============================] - 6s 87ms/step - loss: 0.3771 - accuracy: 0.8350 - val_loss: 0.6814 - val_accuracy: 0.7467 - lr: 0.0010\n",
      "Epoch 52/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3944 - accuracy: 0.8289\n",
      "Epoch 00052: val_loss did not improve from 0.32315\n",
      "Restoring model weights from the end of the best epoch.\n",
      "\n",
      "Epoch 00052: ReduceLROnPlateau reducing learning rate to 0.000800000037997961.\n",
      "68/67 [==============================] - 6s 86ms/step - loss: 0.3944 - accuracy: 0.8289 - val_loss: 0.6115 - val_accuracy: 0.7672 - lr: 0.0010\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 00052: early stopping\n"
     ]
    }
   ],
   "source": [
    "# calling thesequential model\n",
    "m = Sequential(Conv2D(ig_size, kernel_size=3, activation='relu', \n",
    "                      input_shape=(64,64,3)))\n",
    "\n",
    "model = build_model(m, num_classes) \n",
    "\n",
    "for layer in m.layers:\n",
    "    layer.trainable = False\n",
    " \n",
    "\n",
    " #data augmentation\n",
    "train_generator, validation_generator = process_data(x_train,y_train,\n",
    "                                                 x_test,y_test, batch_size)\n",
    "model.summary()\n",
    "\n",
    "checkpoint = ModelCheckpoint(data_dir+'sequentialpnd' + '.h5',\n",
    "                             monitor='val_loss',\n",
    "                             mode=\"min\",\n",
    "                             save_best_only=True,\n",
    "                             verbose=1)\n",
    "\n",
    "earlystop = EarlyStopping(monitor='val_loss',\n",
    "                          min_delta=0,\n",
    "                          patience=30,\n",
    "                          verbose=1,\n",
    "                          restore_best_weights=True)\n",
    "\n",
    "learning_rate_reduction = ReduceLROnPlateau(monitor='val_loss',\n",
    "                                            patience=30,\n",
    "                                            verbose=1,\n",
    "                                            factor=0.8,\n",
    "                                            min_lr=0.0001,\n",
    "                                            mode=\"auto\",\n",
    "                                            min_delta=0.0001,\n",
    "                                            cooldown=5)\n",
    "\n",
    "\n",
    "callbacks = [checkpoint, earlystop, learning_rate_reduction]\n",
    "\n",
    "model.compile(loss=models_loss,\n",
    "              optimizer=models_opt,\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "\n",
    "history = model.fit(train_generator,\n",
    "                             epochs=epochs,\n",
    "                               callbacks=callbacks,\n",
    "                               validation_data=validation_generator,\n",
    "                              validation_steps=len(x_test) / batch_size,\n",
    "                              steps_per_epoch=len(x_train) / batch_size)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Testing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision Score: [0.86241611 0.9539749 ]\n",
      "Recall Score: [0.95895522 0.84758364]\n",
      "F1 Score: [0.90812721 0.8976378 ]\n",
      "Accuracy Score: 0.9031657355679702\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AUC: 0.9032694335016369\n",
      "Log Loss: 0.32311943288617945\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": [
    "## Confusion matrix\n",
    "plot_confusion_matrix(model, x_test, y_test)\n",
    "## Learning curve\n",
    "plot_learning_curves(history)\n",
    "## ROC AUC\n",
    "multiclass_roc_auc_score(x_test, y_test, model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# saving model\n",
    "model.save(\"sequentialpnd.h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import load_model\n",
    "model = load_model('./sequentialpnd.h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Modeling - Using VGG19\n",
    "\n",
    "VGG-19 is a transfer learning alorithm which means its an algorithm with 16 convolutional layers that focuses on storing knowledge that can be applied to different but relaed problems."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_2\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_2 (InputLayer)         [(None, 64, 64, 3)]       0         \n",
      "_________________________________________________________________\n",
      "block1_conv1 (Conv2D)        (None, 64, 64, 64)        1792      \n",
      "_________________________________________________________________\n",
      "block1_conv2 (Conv2D)        (None, 64, 64, 64)        36928     \n",
      "_________________________________________________________________\n",
      "block1_pool (MaxPooling2D)   (None, 32, 32, 64)        0         \n",
      "_________________________________________________________________\n",
      "block2_conv1 (Conv2D)        (None, 32, 32, 128)       73856     \n",
      "_________________________________________________________________\n",
      "block2_conv2 (Conv2D)        (None, 32, 32, 128)       147584    \n",
      "_________________________________________________________________\n",
      "block2_pool (MaxPooling2D)   (None, 16, 16, 128)       0         \n",
      "_________________________________________________________________\n",
      "block3_conv1 (Conv2D)        (None, 16, 16, 256)       295168    \n",
      "_________________________________________________________________\n",
      "block3_conv2 (Conv2D)        (None, 16, 16, 256)       590080    \n",
      "_________________________________________________________________\n",
      "block3_conv3 (Conv2D)        (None, 16, 16, 256)       590080    \n",
      "_________________________________________________________________\n",
      "block3_conv4 (Conv2D)        (None, 16, 16, 256)       590080    \n",
      "_________________________________________________________________\n",
      "block3_pool (MaxPooling2D)   (None, 8, 8, 256)         0         \n",
      "_________________________________________________________________\n",
      "block4_conv1 (Conv2D)        (None, 8, 8, 512)         1180160   \n",
      "_________________________________________________________________\n",
      "block4_conv2 (Conv2D)        (None, 8, 8, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block4_conv3 (Conv2D)        (None, 8, 8, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block4_conv4 (Conv2D)        (None, 8, 8, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block4_pool (MaxPooling2D)   (None, 4, 4, 512)         0         \n",
      "_________________________________________________________________\n",
      "block5_conv1 (Conv2D)        (None, 4, 4, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv2 (Conv2D)        (None, 4, 4, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv3 (Conv2D)        (None, 4, 4, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv4 (Conv2D)        (None, 4, 4, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_pool (MaxPooling2D)   (None, 2, 2, 512)         0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 1, 1, 512)         0         \n",
      "_________________________________________________________________\n",
      "flatten (Flatten)            (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 512)               262656    \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dense_6 (Dense)              (None, 2)                 1026      \n",
      "=================================================================\n",
      "Total params: 20,288,066\n",
      "Trainable params: 263,682\n",
      "Non-trainable params: 20,024,384\n",
      "_________________________________________________________________\n",
      "Epoch 1/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.5661 - accuracy: 0.7058\n",
      "Epoch 00001: val_loss improved from inf to 0.50812, saving model to /modelpnd.h5\n",
      "68/67 [==============================] - 88s 1s/step - loss: 0.5661 - accuracy: 0.7058 - val_loss: 0.5081 - val_accuracy: 0.7523 - lr: 0.0010\n",
      "Epoch 2/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4542 - accuracy: 0.7977\n",
      "Epoch 00002: val_loss did not improve from 0.50812\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.4542 - accuracy: 0.7977 - val_loss: 0.6877 - val_accuracy: 0.6667 - lr: 0.0010\n",
      "Epoch 3/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.4106 - accuracy: 0.8145\n",
      "Epoch 00003: val_loss improved from 0.50812 to 0.45546, saving model to /modelpnd.h5\n",
      "68/67 [==============================] - 86s 1s/step - loss: 0.4106 - accuracy: 0.8145 - val_loss: 0.4555 - val_accuracy: 0.7747 - lr: 0.0010\n",
      "Epoch 4/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3865 - accuracy: 0.8340\n",
      "Epoch 00004: val_loss improved from 0.45546 to 0.35751, saving model to /modelpnd.h5\n",
      "68/67 [==============================] - 87s 1s/step - loss: 0.3865 - accuracy: 0.8340 - val_loss: 0.3575 - val_accuracy: 0.8268 - lr: 0.0010\n",
      "Epoch 5/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3489 - accuracy: 0.8513\n",
      "Epoch 00005: val_loss did not improve from 0.35751\n",
      "68/67 [==============================] - 85s 1s/step - loss: 0.3489 - accuracy: 0.8513 - val_loss: 0.4475 - val_accuracy: 0.7803 - lr: 0.0010\n",
      "Epoch 6/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3301 - accuracy: 0.8653\n",
      "Epoch 00006: val_loss improved from 0.35751 to 0.34499, saving model to /modelpnd.h5\n",
      "68/67 [==============================] - 85s 1s/step - loss: 0.3301 - accuracy: 0.8653 - val_loss: 0.3450 - val_accuracy: 0.8324 - lr: 0.0010\n",
      "Epoch 7/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3211 - accuracy: 0.8690\n",
      "Epoch 00007: val_loss did not improve from 0.34499\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.3211 - accuracy: 0.8690 - val_loss: 0.4044 - val_accuracy: 0.8082 - lr: 0.0010\n",
      "Epoch 8/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3271 - accuracy: 0.8639\n",
      "Epoch 00008: val_loss improved from 0.34499 to 0.26764, saving model to /modelpnd.h5\n",
      "68/67 [==============================] - 85s 1s/step - loss: 0.3271 - accuracy: 0.8639 - val_loss: 0.2676 - val_accuracy: 0.8864 - lr: 0.0010\n",
      "Epoch 9/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3074 - accuracy: 0.8727\n",
      "Epoch 00009: val_loss did not improve from 0.26764\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.3074 - accuracy: 0.8727 - val_loss: 0.4222 - val_accuracy: 0.8063 - lr: 0.0010\n",
      "Epoch 10/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2923 - accuracy: 0.8895\n",
      "Epoch 00010: val_loss did not improve from 0.26764\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2923 - accuracy: 0.8895 - val_loss: 0.4383 - val_accuracy: 0.7989 - lr: 0.0010\n",
      "Epoch 11/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2967 - accuracy: 0.8783\n",
      "Epoch 00011: val_loss did not improve from 0.26764\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2967 - accuracy: 0.8783 - val_loss: 0.4408 - val_accuracy: 0.7858 - lr: 0.0010\n",
      "Epoch 12/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3136 - accuracy: 0.8685\n",
      "Epoch 00012: val_loss did not improve from 0.26764\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.3136 - accuracy: 0.8685 - val_loss: 0.3202 - val_accuracy: 0.8510 - lr: 0.0010\n",
      "Epoch 13/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3120 - accuracy: 0.8704\n",
      "Epoch 00013: val_loss did not improve from 0.26764\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.3120 - accuracy: 0.8704 - val_loss: 0.3322 - val_accuracy: 0.8361 - lr: 0.0010\n",
      "Epoch 14/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3157 - accuracy: 0.8667\n",
      "Epoch 00014: val_loss did not improve from 0.26764\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.3157 - accuracy: 0.8667 - val_loss: 0.3904 - val_accuracy: 0.8138 - lr: 0.0010\n",
      "Epoch 15/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3039 - accuracy: 0.8755\n",
      "Epoch 00015: val_loss did not improve from 0.26764\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.3039 - accuracy: 0.8755 - val_loss: 0.4074 - val_accuracy: 0.8156 - lr: 0.0010\n",
      "Epoch 16/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2831 - accuracy: 0.8788\n",
      "Epoch 00016: val_loss did not improve from 0.26764\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2831 - accuracy: 0.8788 - val_loss: 0.2808 - val_accuracy: 0.8678 - lr: 0.0010\n",
      "Epoch 17/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2911 - accuracy: 0.8830\n",
      "Epoch 00017: val_loss did not improve from 0.26764\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2911 - accuracy: 0.8830 - val_loss: 0.3236 - val_accuracy: 0.8529 - lr: 0.0010\n",
      "Epoch 18/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2864 - accuracy: 0.8848\n",
      "Epoch 00018: val_loss did not improve from 0.26764\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2864 - accuracy: 0.8848 - val_loss: 0.4604 - val_accuracy: 0.7486 - lr: 0.0010\n",
      "Epoch 19/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2928 - accuracy: 0.8811\n",
      "Epoch 00019: val_loss did not improve from 0.26764\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2928 - accuracy: 0.8811 - val_loss: 0.3273 - val_accuracy: 0.8417 - lr: 0.0010\n",
      "Epoch 20/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2816 - accuracy: 0.8816\n",
      "Epoch 00020: val_loss did not improve from 0.26764\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2816 - accuracy: 0.8816 - val_loss: 0.3770 - val_accuracy: 0.8250 - lr: 0.0010\n",
      "Epoch 21/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2861 - accuracy: 0.8937\n",
      "Epoch 00021: val_loss did not improve from 0.26764\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2861 - accuracy: 0.8937 - val_loss: 0.3449 - val_accuracy: 0.8399 - lr: 0.0010\n",
      "Epoch 22/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2756 - accuracy: 0.8918\n",
      "Epoch 00022: val_loss did not improve from 0.26764\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2756 - accuracy: 0.8918 - val_loss: 0.3730 - val_accuracy: 0.8175 - lr: 0.0010\n",
      "Epoch 23/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2927 - accuracy: 0.8816\n",
      "Epoch 00023: val_loss did not improve from 0.26764\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2927 - accuracy: 0.8816 - val_loss: 0.2855 - val_accuracy: 0.8752 - lr: 0.0010\n",
      "Epoch 24/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2615 - accuracy: 0.8937\n",
      "Epoch 00024: val_loss did not improve from 0.26764\n",
      "68/67 [==============================] - 85s 1s/step - loss: 0.2615 - accuracy: 0.8937 - val_loss: 0.3414 - val_accuracy: 0.8436 - lr: 0.0010\n",
      "Epoch 25/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2832 - accuracy: 0.8895\n",
      "Epoch 00025: val_loss did not improve from 0.26764\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2832 - accuracy: 0.8895 - val_loss: 0.4847 - val_accuracy: 0.7635 - lr: 0.0010\n",
      "Epoch 26/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2895 - accuracy: 0.8821\n",
      "Epoch 00026: val_loss did not improve from 0.26764\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2895 - accuracy: 0.8821 - val_loss: 0.3305 - val_accuracy: 0.8529 - lr: 0.0010\n",
      "Epoch 27/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2767 - accuracy: 0.8909\n",
      "Epoch 00027: val_loss did not improve from 0.26764\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2767 - accuracy: 0.8909 - val_loss: 0.3671 - val_accuracy: 0.8194 - lr: 0.0010\n",
      "Epoch 28/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2805 - accuracy: 0.8900\n",
      "Epoch 00028: val_loss improved from 0.26764 to 0.22394, saving model to /modelpnd.h5\n",
      "68/67 [==============================] - 85s 1s/step - loss: 0.2805 - accuracy: 0.8900 - val_loss: 0.2239 - val_accuracy: 0.9162 - lr: 0.0010\n",
      "Epoch 29/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2804 - accuracy: 0.8830\n",
      "Epoch 00029: val_loss did not improve from 0.22394\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2804 - accuracy: 0.8830 - val_loss: 0.2941 - val_accuracy: 0.8808 - lr: 0.0010\n",
      "Epoch 30/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2619 - accuracy: 0.8984\n",
      "Epoch 00030: val_loss did not improve from 0.22394\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2619 - accuracy: 0.8984 - val_loss: 0.2472 - val_accuracy: 0.9106 - lr: 0.0010\n",
      "Epoch 31/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2854 - accuracy: 0.8802\n",
      "Epoch 00031: val_loss did not improve from 0.22394\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2854 - accuracy: 0.8802 - val_loss: 0.2593 - val_accuracy: 0.8920 - lr: 0.0010\n",
      "Epoch 32/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2724 - accuracy: 0.8872\n",
      "Epoch 00032: val_loss did not improve from 0.22394\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2724 - accuracy: 0.8872 - val_loss: 0.3155 - val_accuracy: 0.8603 - lr: 0.0010\n",
      "Epoch 33/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2609 - accuracy: 0.8951\n",
      "Epoch 00033: val_loss did not improve from 0.22394\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2609 - accuracy: 0.8951 - val_loss: 0.3545 - val_accuracy: 0.8454 - lr: 0.0010\n",
      "Epoch 34/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2693 - accuracy: 0.8862\n",
      "Epoch 00034: val_loss did not improve from 0.22394\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2693 - accuracy: 0.8862 - val_loss: 0.2293 - val_accuracy: 0.9069 - lr: 0.0010\n",
      "Epoch 35/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2678 - accuracy: 0.8834\n",
      "Epoch 00035: val_loss improved from 0.22394 to 0.19251, saving model to /modelpnd.h5\n",
      "68/67 [==============================] - 85s 1s/step - loss: 0.2678 - accuracy: 0.8834 - val_loss: 0.1925 - val_accuracy: 0.9274 - lr: 0.0010\n",
      "Epoch 36/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2512 - accuracy: 0.8993\n",
      "Epoch 00036: val_loss did not improve from 0.19251\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2512 - accuracy: 0.8993 - val_loss: 0.2802 - val_accuracy: 0.8790 - lr: 0.0010\n",
      "Epoch 37/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2526 - accuracy: 0.9026\n",
      "Epoch 00037: val_loss did not improve from 0.19251\n",
      "68/67 [==============================] - 85s 1s/step - loss: 0.2526 - accuracy: 0.9026 - val_loss: 0.2559 - val_accuracy: 0.8976 - lr: 0.0010\n",
      "Epoch 38/100\n",
      "68/67 [==============================] - ETA: -1s - loss: 0.2937 - accuracy: 0.8825\n",
      "Epoch 00038: val_loss did not improve from 0.19251\n",
      "68/67 [==============================] - 91s 1s/step - loss: 0.2937 - accuracy: 0.8825 - val_loss: 0.2653 - val_accuracy: 0.8845 - lr: 0.0010\n",
      "Epoch 39/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2477 - accuracy: 0.9040\n",
      "Epoch 00039: val_loss did not improve from 0.19251\n",
      "68/67 [==============================] - 85s 1s/step - loss: 0.2477 - accuracy: 0.9040 - val_loss: 0.3792 - val_accuracy: 0.8175 - lr: 0.0010\n",
      "Epoch 40/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2685 - accuracy: 0.8970\n",
      "Epoch 00040: val_loss did not improve from 0.19251\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2685 - accuracy: 0.8970 - val_loss: 0.2239 - val_accuracy: 0.9143 - lr: 0.0010\n",
      "Epoch 41/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2513 - accuracy: 0.8951\n",
      "Epoch 00041: val_loss did not improve from 0.19251\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2513 - accuracy: 0.8951 - val_loss: 0.3369 - val_accuracy: 0.8436 - lr: 0.0010\n",
      "Epoch 42/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2586 - accuracy: 0.9040\n",
      "Epoch 00042: val_loss did not improve from 0.19251\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2586 - accuracy: 0.9040 - val_loss: 0.2206 - val_accuracy: 0.9050 - lr: 0.0010\n",
      "Epoch 43/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "68/67 [==============================] - ETA: 0s - loss: 0.2632 - accuracy: 0.8984\n",
      "Epoch 00043: val_loss did not improve from 0.19251\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2632 - accuracy: 0.8984 - val_loss: 0.3205 - val_accuracy: 0.8566 - lr: 0.0010\n",
      "Epoch 44/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2572 - accuracy: 0.8946\n",
      "Epoch 00044: val_loss did not improve from 0.19251\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2572 - accuracy: 0.8946 - val_loss: 0.2541 - val_accuracy: 0.8901 - lr: 0.0010\n",
      "Epoch 45/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2468 - accuracy: 0.9086\n",
      "Epoch 00045: val_loss did not improve from 0.19251\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2468 - accuracy: 0.9086 - val_loss: 0.2644 - val_accuracy: 0.8827 - lr: 0.0010\n",
      "Epoch 46/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2475 - accuracy: 0.8993\n",
      "Epoch 00046: val_loss did not improve from 0.19251\n",
      "68/67 [==============================] - 85s 1s/step - loss: 0.2475 - accuracy: 0.8993 - val_loss: 0.2484 - val_accuracy: 0.8976 - lr: 0.0010\n",
      "Epoch 47/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2634 - accuracy: 0.8965\n",
      "Epoch 00047: val_loss did not improve from 0.19251\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2634 - accuracy: 0.8965 - val_loss: 0.2740 - val_accuracy: 0.8845 - lr: 0.0010\n",
      "Epoch 48/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2526 - accuracy: 0.8974\n",
      "Epoch 00048: val_loss did not improve from 0.19251\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2526 - accuracy: 0.8974 - val_loss: 0.2616 - val_accuracy: 0.8883 - lr: 0.0010\n",
      "Epoch 49/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2428 - accuracy: 0.9030\n",
      "Epoch 00049: val_loss did not improve from 0.19251\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2428 - accuracy: 0.9030 - val_loss: 0.2326 - val_accuracy: 0.9032 - lr: 0.0010\n",
      "Epoch 50/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2421 - accuracy: 0.9021\n",
      "Epoch 00050: val_loss did not improve from 0.19251\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2421 - accuracy: 0.9021 - val_loss: 0.3199 - val_accuracy: 0.8603 - lr: 0.0010\n",
      "Epoch 51/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2455 - accuracy: 0.8998\n",
      "Epoch 00051: val_loss improved from 0.19251 to 0.16800, saving model to /modelpnd.h5\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2455 - accuracy: 0.8998 - val_loss: 0.1680 - val_accuracy: 0.9460 - lr: 0.0010\n",
      "Epoch 52/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2680 - accuracy: 0.8886\n",
      "Epoch 00052: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2680 - accuracy: 0.8886 - val_loss: 0.2356 - val_accuracy: 0.9032 - lr: 0.0010\n",
      "Epoch 53/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.3032 - accuracy: 0.8695\n",
      "Epoch 00053: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 85s 1s/step - loss: 0.3032 - accuracy: 0.8695 - val_loss: 0.2940 - val_accuracy: 0.8715 - lr: 0.0010\n",
      "Epoch 54/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2521 - accuracy: 0.9077\n",
      "Epoch 00054: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2521 - accuracy: 0.9077 - val_loss: 0.2425 - val_accuracy: 0.9013 - lr: 0.0010\n",
      "Epoch 55/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2599 - accuracy: 0.8946\n",
      "Epoch 00055: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2599 - accuracy: 0.8946 - val_loss: 0.2340 - val_accuracy: 0.9125 - lr: 0.0010\n",
      "Epoch 56/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2459 - accuracy: 0.8988\n",
      "Epoch 00056: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2459 - accuracy: 0.8988 - val_loss: 0.1998 - val_accuracy: 0.9274 - lr: 0.0010\n",
      "Epoch 57/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2593 - accuracy: 0.8946\n",
      "Epoch 00057: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2593 - accuracy: 0.8946 - val_loss: 0.2027 - val_accuracy: 0.9218 - lr: 0.0010\n",
      "Epoch 58/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2511 - accuracy: 0.8946\n",
      "Epoch 00058: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2511 - accuracy: 0.8946 - val_loss: 0.2156 - val_accuracy: 0.9162 - lr: 0.0010\n",
      "Epoch 59/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2443 - accuracy: 0.9049\n",
      "Epoch 00059: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2443 - accuracy: 0.9049 - val_loss: 0.3089 - val_accuracy: 0.8547 - lr: 0.0010\n",
      "Epoch 60/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2485 - accuracy: 0.8984\n",
      "Epoch 00060: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2485 - accuracy: 0.8984 - val_loss: 0.2192 - val_accuracy: 0.9106 - lr: 0.0010\n",
      "Epoch 61/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2368 - accuracy: 0.9063\n",
      "Epoch 00061: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2368 - accuracy: 0.9063 - val_loss: 0.2106 - val_accuracy: 0.9199 - lr: 0.0010\n",
      "Epoch 62/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2449 - accuracy: 0.9030\n",
      "Epoch 00062: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2449 - accuracy: 0.9030 - val_loss: 0.1927 - val_accuracy: 0.9274 - lr: 0.0010\n",
      "Epoch 63/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2493 - accuracy: 0.9016\n",
      "Epoch 00063: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2493 - accuracy: 0.9016 - val_loss: 0.3681 - val_accuracy: 0.8287 - lr: 0.0010\n",
      "Epoch 64/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2476 - accuracy: 0.9054\n",
      "Epoch 00064: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2476 - accuracy: 0.9054 - val_loss: 0.2312 - val_accuracy: 0.9050 - lr: 0.0010\n",
      "Epoch 65/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2429 - accuracy: 0.8974\n",
      "Epoch 00065: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2429 - accuracy: 0.8974 - val_loss: 0.2179 - val_accuracy: 0.9050 - lr: 0.0010\n",
      "Epoch 66/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2547 - accuracy: 0.9002\n",
      "Epoch 00066: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2547 - accuracy: 0.9002 - val_loss: 0.3327 - val_accuracy: 0.8436 - lr: 0.0010\n",
      "Epoch 67/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2587 - accuracy: 0.8979\n",
      "Epoch 00067: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2587 - accuracy: 0.8979 - val_loss: 0.3005 - val_accuracy: 0.8603 - lr: 0.0010\n",
      "Epoch 68/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2467 - accuracy: 0.9063\n",
      "Epoch 00068: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2467 - accuracy: 0.9063 - val_loss: 0.2000 - val_accuracy: 0.9218 - lr: 0.0010\n",
      "Epoch 69/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2236 - accuracy: 0.9077\n",
      "Epoch 00069: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2236 - accuracy: 0.9077 - val_loss: 0.2998 - val_accuracy: 0.8603 - lr: 0.0010\n",
      "Epoch 70/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2657 - accuracy: 0.8951\n",
      "Epoch 00070: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 85s 1s/step - loss: 0.2657 - accuracy: 0.8951 - val_loss: 0.2679 - val_accuracy: 0.8920 - lr: 0.0010\n",
      "Epoch 71/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2423 - accuracy: 0.9016\n",
      "Epoch 00071: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2423 - accuracy: 0.9016 - val_loss: 0.3069 - val_accuracy: 0.8454 - lr: 0.0010\n",
      "Epoch 72/100\n",
      "68/67 [==============================] - ETA: -1s - loss: 0.2363 - accuracy: 0.9068\n",
      "Epoch 00072: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 89s 1s/step - loss: 0.2363 - accuracy: 0.9068 - val_loss: 0.3057 - val_accuracy: 0.8547 - lr: 0.0010\n",
      "Epoch 73/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2401 - accuracy: 0.9035\n",
      "Epoch 00073: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2401 - accuracy: 0.9035 - val_loss: 0.2754 - val_accuracy: 0.8715 - lr: 0.0010\n",
      "Epoch 74/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2382 - accuracy: 0.9077\n",
      "Epoch 00074: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2382 - accuracy: 0.9077 - val_loss: 0.3172 - val_accuracy: 0.8529 - lr: 0.0010\n",
      "Epoch 75/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2461 - accuracy: 0.8979\n",
      "Epoch 00075: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2461 - accuracy: 0.8979 - val_loss: 0.2189 - val_accuracy: 0.9050 - lr: 0.0010\n",
      "Epoch 76/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2334 - accuracy: 0.9100\n",
      "Epoch 00076: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2334 - accuracy: 0.9100 - val_loss: 0.2635 - val_accuracy: 0.8827 - lr: 0.0010\n",
      "Epoch 77/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2554 - accuracy: 0.8932\n",
      "Epoch 00077: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2554 - accuracy: 0.8932 - val_loss: 0.3696 - val_accuracy: 0.8268 - lr: 0.0010\n",
      "Epoch 78/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2492 - accuracy: 0.8965\n",
      "Epoch 00078: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 84s 1s/step - loss: 0.2492 - accuracy: 0.8965 - val_loss: 0.2205 - val_accuracy: 0.9050 - lr: 0.0010\n",
      "Epoch 79/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2326 - accuracy: 0.9035\n",
      "Epoch 00079: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2326 - accuracy: 0.9035 - val_loss: 0.3349 - val_accuracy: 0.8417 - lr: 0.0010\n",
      "Epoch 80/100\n",
      "68/67 [==============================] - ETA: -1s - loss: 0.2358 - accuracy: 0.9077\n",
      "Epoch 00080: val_loss did not improve from 0.16800\n",
      "68/67 [==============================] - 94s 1s/step - loss: 0.2358 - accuracy: 0.9077 - val_loss: 0.2722 - val_accuracy: 0.8752 - lr: 0.0010\n",
      "Epoch 81/100\n",
      "68/67 [==============================] - ETA: 0s - loss: 0.2423 - accuracy: 0.9040\n",
      "Epoch 00081: val_loss did not improve from 0.16800\n",
      "Restoring model weights from the end of the best epoch.\n",
      "\n",
      "Epoch 00081: ReduceLROnPlateau reducing learning rate to 0.000800000037997961.\n",
      "68/67 [==============================] - 83s 1s/step - loss: 0.2423 - accuracy: 0.9040 - val_loss: 0.2149 - val_accuracy: 0.9050 - lr: 0.0010\n",
      "Epoch 00081: early stopping\n"
     ]
    }
   ],
   "source": [
    "# assiging the VGG19 model\n",
    "m = VGG19(weights=\"imagenet\", include_top=False, \n",
    "          input_tensor=Input(shape=(64, 64, 3)))\n",
    "\n",
    "model = addTopModel(m, num_classes) \n",
    "\n",
    "for layer in m.layers:\n",
    "    layer.trainable = False\n",
    " \n",
    "\n",
    " #data augmentation\n",
    "train_generator, validation_generator = process_data(x_train,y_train,\n",
    "                                                 x_test,y_test, batch_size)\n",
    "\n",
    "\n",
    "model.summary()\n",
    "\n",
    "checkpoint = ModelCheckpoint(data_dir+'modelpnd' + '.h5',\n",
    "                             monitor='val_loss',\n",
    "                             mode=\"min\",\n",
    "                             save_best_only=True,\n",
    "                             verbose=1)\n",
    "\n",
    "earlystop = EarlyStopping(monitor='val_loss',\n",
    "                          min_delta=0,\n",
    "                          patience=30,\n",
    "                          verbose=1,\n",
    "                          restore_best_weights=True)\n",
    "\n",
    "learning_rate_reduction = ReduceLROnPlateau(monitor='val_loss',\n",
    "                                            patience=30,\n",
    "                                            verbose=1,\n",
    "                                            factor=0.8,\n",
    "                                            min_lr=0.0001,\n",
    "                                            mode=\"auto\",\n",
    "                                            min_delta=0.0001,\n",
    "                                            cooldown=5)\n",
    "\n",
    "\n",
    "callbacks = [checkpoint, earlystop, learning_rate_reduction]\n",
    "\n",
    "model.compile(loss=models_loss,\n",
    "              optimizer=models_opt,\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "\n",
    "history = model.fit(train_generator,\n",
    "                             epochs=epochs,\n",
    "                               callbacks=callbacks,\n",
    "                               validation_data=validation_generator,\n",
    "                              validation_steps=len(x_test) / batch_size,\n",
    "                              steps_per_epoch=len(x_train) / batch_size)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Testing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision Score: [0.92831541 0.96511628]\n",
      "Recall Score: [0.96641791 0.92565056]\n",
      "F1 Score: [0.94698355 0.94497154]\n",
      "Accuracy Score: 0.9459962756052142\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AUC: 0.9460342340342894\n",
      "Log Loss: 0.16809428236942026\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": [
    "## Confusion matrix\n",
    "plot_confusion_matrix(model, x_test, y_test)\n",
    "## Learning curve\n",
    "plot_learning_curves(history)\n",
    "## ROC AUC\n",
    "multiclass_roc_auc_score(x_test, y_test, model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# saving model\n",
    "model.save(\"modelpnd.h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import load_model\n",
    "model = load_model('./modelpnd.h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Conclusion\n",
    "The accuracy is 95 % and this is the amount of time the predicted result is actually correct.\n",
    "\n",
    "The recall percentage is 95% and this is the probability of the model diagnosing a correct positive diagnosis out of all the times it diagnosed positive. This would be the best metric in this case as we would rather give a wrong positive diagnosis than give a wrong negative diagnosis.\n",
    "\n",
    "### Recommendation\n",
    "* Reshape the x-ray sizes to 64 x 64 to reduce the amount of computational time and power consumed when using the VGG19 model.\n",
    "\n",
    "* Use a maxpooling layer before flatten the model and using a dense layer of 512 and also a learning rate of 0.000800000037997961 in order to get a better performing model when using the VGG19 model.\n",
    "\n",
    "* Add a dropout layer before the final dense layer in order to reduce overfitting when using the VGG19 model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_model(m,x_train,y_train, x_test,y_test):\n",
    "\n",
    "    model = addTopModel(m, num_classes) \n",
    "\n",
    "    for layer in m.layers:\n",
    "        layer.trainable = False\n",
    "\n",
    "\n",
    "        #data augmentation\n",
    "    train_generator, validation_generator = process_data(x_train,y_train,\n",
    "                                                     x_test,y_test, batch_size)\n",
    "\n",
    "    model.summary()\n",
    "\n",
    "    checkpoint = ModelCheckpoint(data_dir+'modelpnd' + '.h5',\n",
    "                                 monitor='val_loss',\n",
    "                                 mode=\"min\",\n",
    "                                 save_best_only=True,\n",
    "                                 verbose=1)\n",
    "\n",
    "    earlystop = EarlyStopping(monitor='val_loss',\n",
    "                              min_delta=0,\n",
    "                              patience=30,\n",
    "                              verbose=1,\n",
    "                              restore_best_weights=True)\n",
    "\n",
    "    learning_rate_reduction = ReduceLROnPlateau(monitor='val_loss',\n",
    "                                                patience=30,\n",
    "                                                verbose=1,\n",
    "                                                factor=0.8,\n",
    "                                                min_lr=0.0001,\n",
    "                                                mode=\"auto\",\n",
    "                                                min_delta=0.0001,\n",
    "                                                cooldown=5)\n",
    "\n",
    "\n",
    "    callbacks = [checkpoint, earlystop, learning_rate_reduction]\n",
    "\n",
    "    model.compile(loss=models_loss,\n",
    "                  optimizer=models_opt,\n",
    "                  metrics=['accuracy'])\n",
    "\n",
    "\n",
    "    history = model.fit(train_generator,\n",
    "                                 epochs=epochs,\n",
    "                                   callbacks=callbacks,\n",
    "                                   validation_data=validation_generator,\n",
    "                                  validation_steps=len(x_test) / batch_size,\n",
    "                                  steps_per_epoch=len(x_train) / batch_size)\n",
    "    return model, history"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "authorship_tag": "ABX9TyOBK/qqRk5TcvVbe4on/F9j",
   "collapsed_sections": [
    "CIQEU2HjTPzF",
    "4Zlv5kGh3LvC",
    "71bunMrmSEGF",
    "7TopSN6fOcmf",
    "ZyYB-p4qOTNx",
    "_wBgbcDEOqJd",
    "wYCX523UOW5K",
    "TalwgYalOdPd",
    "Cf2dxyj6O3yt",
    "BXzugA8pOdlv"
   ],
   "include_colab_link": true,
   "mount_file_id": "1eAzBe9FO37HcJfumZGrqDBb7IyNw7AuW",
   "name": "v7_3 classes_bin_covid19_multiple_models.ipynb",
   "provenance": [],
   "toc_visible": true
  },
  "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": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}