01. train_mask_detector.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "6ad01198",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
    "from tensorflow.keras.applications import MobileNetV2\n",
    "from tensorflow.keras.layers import AveragePooling2D\n",
    "from tensorflow.keras.layers import Dropout\n",
    "from tensorflow.keras.layers import Flatten\n",
    "from tensorflow.keras.layers import Dense\n",
    "from tensorflow.keras.layers import Input\n",
    "from tensorflow.keras.models import Model\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "from tensorflow.keras.applications.mobilenet_v2 import preprocess_input\n",
    "from tensorflow.keras.preprocessing.image import img_to_array\n",
    "from tensorflow.keras.preprocessing.image import load_img\n",
    "from tensorflow.keras.utils import to_categorical\n",
    "from sklearn.preprocessing import LabelBinarizer\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import classification_report\n",
    "from imutils import paths\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import argparse\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "6af5b154",
   "metadata": {},
   "outputs": [],
   "source": [
    "#\n",
    "INIT_LR = 1e-4  #0.0004\n",
    "EPOCHS = 20\n",
    "BS = 32"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "de3eeff2",
   "metadata": {},
   "outputs": [],
   "source": [
    "DIRECTORY = 'dataset'\n",
    "CATEGORIES = [\"with mask\", \"without mask\"]\n",
    "#CATEGORIES[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "356b94f0",
   "metadata": {},
   "outputs": [],
   "source": [
    "data = []\n",
    "targets = []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "4254a9e2",
   "metadata": {},
   "outputs": [],
   "source": [
    "for category in CATEGORIES:\n",
    "    path = os.path.join(DIRECTORY, category)\n",
    "    for img in os.listdir(path):\n",
    "    \timg_path = os.path.join(path, img)\n",
    "    \timage = load_img(img_path, target_size=(224, 224))\n",
    "    \timage = img_to_array(image)\n",
    "    \timage = preprocess_input(image)\n",
    "\n",
    "    \tdata.append(image)\n",
    "    \ttargets.append(category)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "3fa2d353",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(224, 224, 3)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(data[6].shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "52d815c7",
   "metadata": {},
   "outputs": [],
   "source": [
    "lb = LabelBinarizer()\n",
    "\n",
    "# transform multi-class labels to binary labels. \n",
    "targets = lb.fit_transform(targets)\n",
    "\n",
    "#returns a numpy array having columns equal to the number of categories in the data.\n",
    "# 'with mask' and 'without mask'\n",
    "targets = to_categorical(targets)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "74a4e516",
   "metadata": {},
   "outputs": [],
   "source": [
    "data = np.array(data, dtype=\"float32\")\n",
    "targets = np.array(targets)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "b3b7baad",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[1. 0.]\n",
      " [1. 0.]\n",
      " [1. 0.]\n",
      " ...\n",
      " [0. 1.]\n",
      " [0. 1.]\n",
      " [0. 1.]]\n"
     ]
    }
   ],
   "source": [
    "print(targets)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "c9b375fd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[[[ 0.7647059   0.6         0.18431377]\n",
      "   [ 0.67058825  0.54509807  0.13725495]\n",
      "   [ 0.8745098   0.79607844  0.39607847]\n",
      "   ...\n",
      "   [ 0.73333335  0.6627451   0.5294118 ]\n",
      "   [ 0.8980392   0.81960785  0.7411765 ]\n",
      "   [-0.46666664 -0.5686275  -0.6392157 ]]\n",
      "\n",
      "  [[ 0.62352943  0.45882356  0.12156868]\n",
      "   [ 0.7254902   0.5764706   0.24705887]\n",
      "   [ 0.8666667   0.7176471   0.4039216 ]\n",
      "   ...\n",
      "   [ 0.5686275   0.49803925  0.27058828]\n",
      "   [ 0.8901961   0.8901961   0.6862745 ]\n",
      "   [-0.5686275  -0.5058824  -0.67058825]]\n",
      "\n",
      "  [[ 0.6627451   0.52156866  0.20784318]\n",
      "   [ 0.7254902   0.5529412   0.24705887]\n",
      "   [ 0.8980392   0.67058825  0.4039216 ]\n",
      "   ...\n",
      "   [ 0.827451    0.75686276  0.43529415]\n",
      "   [ 0.62352943  0.6627451   0.34901965]\n",
      "   [-0.6862745  -0.5686275  -0.79607844]]\n",
      "\n",
      "  ...\n",
      "\n",
      "  [[-0.8901961  -0.8117647  -1.        ]\n",
      "   [-0.9607843  -0.92941177 -1.        ]\n",
      "   [-0.58431375 -0.5686275  -0.654902  ]\n",
      "   ...\n",
      "   [ 0.94509804  0.58431375  0.6       ]\n",
      "   [ 0.85882354  0.64705884  0.5137255 ]\n",
      "   [-0.41176468 -0.44313723 -0.654902  ]]\n",
      "\n",
      "  [[-0.70980394 -0.6        -0.8745098 ]\n",
      "   [-0.79607844 -0.7019608  -0.9843137 ]\n",
      "   [-0.67058825 -0.6313726  -0.7882353 ]\n",
      "   ...\n",
      "   [ 0.8901961   0.64705884  0.69411767]\n",
      "   [ 0.60784316  0.48235297  0.36470592]\n",
      "   [-0.70980394 -0.77254903 -0.9372549 ]]\n",
      "\n",
      "  [[-0.4588235  -0.31764704 -0.64705884]\n",
      "   [-0.62352943 -0.4823529  -0.8901961 ]\n",
      "   [-0.70980394 -0.6313726  -0.84313726]\n",
      "   ...\n",
      "   [ 0.84313726  0.6784314   0.79607844]\n",
      "   [ 0.6313726   0.56078434  0.52156866]\n",
      "   [-0.6627451  -0.73333335 -0.7882353 ]]]\n",
      "\n",
      "\n",
      " [[[ 0.05882359 -0.08235294 -0.25490195]\n",
      "   [ 0.06666672 -0.0745098  -0.23137254]\n",
      "   [ 0.082353   -0.08235294 -0.21568626]\n",
      "   ...\n",
      "   [-0.85882354 -0.827451   -0.7411765 ]\n",
      "   [-0.85882354 -0.827451   -0.7411765 ]\n",
      "   [-0.85882354 -0.827451   -0.7411765 ]]\n",
      "\n",
      "  [[ 0.04313731 -0.09019607 -0.24705881]\n",
      "   [ 0.04313731 -0.09019607 -0.23137254]\n",
      "   [ 0.05882359 -0.09019607 -0.2235294 ]\n",
      "   ...\n",
      "   [-0.85882354 -0.827451   -0.7411765 ]\n",
      "   [-0.85882354 -0.827451   -0.7411765 ]\n",
      "   [-0.85882354 -0.827451   -0.7411765 ]]\n",
      "\n",
      "  [[ 0.01176476 -0.11372548 -0.23921567]\n",
      "   [ 0.0196079  -0.10588235 -0.23137254]\n",
      "   [ 0.02745104 -0.11372548 -0.2235294 ]\n",
      "   ...\n",
      "   [-0.8352941  -0.81960785 -0.7254902 ]\n",
      "   [-0.84313726 -0.827451   -0.73333335]\n",
      "   [-0.8352941  -0.81960785 -0.7254902 ]]\n",
      "\n",
      "  ...\n",
      "\n",
      "  [[ 0.19215691  0.6313726   0.60784316]\n",
      "   [ 0.10588241  0.6156863   0.58431375]\n",
      "   [ 0.04313731  0.6         0.5686275 ]\n",
      "   ...\n",
      "   [ 0.5529412   0.75686276  0.73333335]\n",
      "   [ 0.5764706   0.77254903  0.7490196 ]\n",
      "   [ 0.6         0.77254903  0.75686276]]\n",
      "\n",
      "  [[ 0.11372554  0.58431375  0.5686275 ]\n",
      "   [ 0.04313731  0.58431375  0.56078434]\n",
      "   [-0.01176471  0.5764706   0.54509807]\n",
      "   ...\n",
      "   [ 0.49803925  0.7254902   0.6862745 ]\n",
      "   [ 0.5294118   0.7411765   0.7019608 ]\n",
      "   [ 0.54509807  0.7411765   0.70980394]]\n",
      "\n",
      "  [[ 0.07450986  0.58431375  0.5529412 ]\n",
      "   [ 0.01176476  0.5764706   0.54509807]\n",
      "   [-0.02745098  0.5686275   0.5372549 ]\n",
      "   ...\n",
      "   [ 0.45098042  0.69411767  0.64705884]\n",
      "   [ 0.47450984  0.7019608   0.6627451 ]\n",
      "   [ 0.5058824   0.7019608   0.67058825]]]\n",
      "\n",
      "\n",
      " [[[-0.85882354 -0.827451   -0.7411765 ]\n",
      "   [-0.8745098  -0.84313726 -0.75686276]\n",
      "   [-0.8745098  -0.84313726 -0.75686276]\n",
      "   ...\n",
      "   [ 0.09019613 -0.0745098  -0.20784312]\n",
      "   [ 0.09019613 -0.0745098  -0.20784312]\n",
      "   [ 0.05882359 -0.08235294 -0.25490195]]\n",
      "\n",
      "  [[-0.85882354 -0.827451   -0.7411765 ]\n",
      "   [-0.8745098  -0.84313726 -0.75686276]\n",
      "   [-0.8745098  -0.84313726 -0.75686276]\n",
      "   ...\n",
      "   [ 0.05882359 -0.09019607 -0.20784312]\n",
      "   [ 0.05882359 -0.09019607 -0.2235294 ]\n",
      "   [ 0.04313731 -0.09019607 -0.24705881]]\n",
      "\n",
      "  [[-0.84313726 -0.827451   -0.73333335]\n",
      "   [-0.84313726 -0.827451   -0.73333335]\n",
      "   [-0.8352941  -0.81960785 -0.7254902 ]\n",
      "   ...\n",
      "   [ 0.0196079  -0.12156862 -0.23137254]\n",
      "   [ 0.0196079  -0.12156862 -0.23137254]\n",
      "   [ 0.02745104 -0.09803921 -0.2235294 ]]\n",
      "\n",
      "  ...\n",
      "\n",
      "  [[ 0.5921569   0.7647059   0.7490196 ]\n",
      "   [ 0.6392157   0.8352941   0.8117647 ]\n",
      "   [ 0.6392157   0.8352941   0.8117647 ]\n",
      "   ...\n",
      "   [ 0.082353    0.67058825  0.6392157 ]\n",
      "   [ 0.18431377  0.7254902   0.7019608 ]\n",
      "   [ 0.11372554  0.54509807  0.52156866]]\n",
      "\n",
      "  [[ 0.5764706   0.7490196   0.7254902 ]\n",
      "   [ 0.56078434  0.75686276  0.7254902 ]\n",
      "   [ 0.52156866  0.73333335  0.69411767]\n",
      "   ...\n",
      "   [-0.09803921  0.52156866  0.49803925]\n",
      "   [ 0.06666672  0.6313726   0.60784316]\n",
      "   [ 0.17647064  0.64705884  0.6313726 ]]\n",
      "\n",
      "  [[ 0.45098042  0.64705884  0.6156863 ]\n",
      "   [ 0.43529415  0.64705884  0.60784316]\n",
      "   [ 0.4039216   0.6313726   0.5921569 ]\n",
      "   ...\n",
      "   [-0.23137254  0.39607847  0.37254906]\n",
      "   [-0.00392157  0.58431375  0.5529412 ]\n",
      "   [ 0.2313726   0.7411765   0.70980394]]]\n",
      "\n",
      "\n",
      " ...\n",
      "\n",
      "\n",
      " [[[-1.          0.09803927  0.7176471 ]\n",
      "   [-1.          0.09803927  0.7176471 ]\n",
      "   [-1.          0.09803927  0.7176471 ]\n",
      "   ...\n",
      "   [ 0.4901961   0.4666667   0.12941182]\n",
      "   [ 0.5137255   0.4901961   0.1686275 ]\n",
      "   [ 0.54509807  0.5294118   0.22352946]]\n",
      "\n",
      "  [[-1.          0.09803927  0.7176471 ]\n",
      "   [-1.          0.09803927  0.7176471 ]\n",
      "   [-0.99215686  0.10588241  0.7254902 ]\n",
      "   ...\n",
      "   [ 0.4431373   0.41960788  0.09803927]\n",
      "   [ 0.45882356  0.427451    0.12941182]\n",
      "   [ 0.5294118   0.49803925  0.20784318]]\n",
      "\n",
      "  [[-1.          0.09803927  0.7019608 ]\n",
      "   [-1.          0.09803927  0.7019608 ]\n",
      "   [-0.99215686  0.10588241  0.70980394]\n",
      "   ...\n",
      "   [ 0.41176474  0.3803922   0.082353  ]\n",
      "   [ 0.427451    0.39607847  0.09803927]\n",
      "   [ 0.5294118   0.49803925  0.20784318]]\n",
      "\n",
      "  ...\n",
      "\n",
      "  [[-0.8901961  -0.8745098  -0.9137255 ]\n",
      "   [-0.8901961  -0.8745098  -0.9137255 ]\n",
      "   [-0.8901961  -0.8745098  -0.9137255 ]\n",
      "   ...\n",
      "   [-0.92941177 -0.92941177 -0.94509804]\n",
      "   [-0.94509804 -0.94509804 -0.9607843 ]\n",
      "   [-0.94509804 -0.94509804 -0.94509804]]\n",
      "\n",
      "  [[-0.8980392  -0.88235295 -0.92156863]\n",
      "   [-0.8901961  -0.8745098  -0.9137255 ]\n",
      "   [-0.8901961  -0.8745098  -0.9137255 ]\n",
      "   ...\n",
      "   [-0.92156863 -0.92156863 -0.9372549 ]\n",
      "   [-0.92941177 -0.92941177 -0.94509804]\n",
      "   [-0.9372549  -0.9372549  -0.9372549 ]]\n",
      "\n",
      "  [[-0.8980392  -0.88235295 -0.92156863]\n",
      "   [-0.8980392  -0.88235295 -0.92156863]\n",
      "   [-0.8980392  -0.88235295 -0.92156863]\n",
      "   ...\n",
      "   [-0.92156863 -0.92156863 -0.9372549 ]\n",
      "   [-0.92156863 -0.92156863 -0.9372549 ]\n",
      "   [-0.92156863 -0.92156863 -0.92156863]]]\n",
      "\n",
      "\n",
      " [[[ 0.9843137   0.96862745  1.        ]\n",
      "   [ 1.          0.99215686  1.        ]\n",
      "   [ 1.          0.99215686  1.        ]\n",
      "   ...\n",
      "   [ 0.9607843   1.          1.        ]\n",
      "   [ 0.9607843   1.          1.        ]\n",
      "   [ 0.9764706   1.          1.        ]]\n",
      "\n",
      "  [[ 1.          1.          1.        ]\n",
      "   [ 0.99215686  0.99215686  1.        ]\n",
      "   [ 0.99215686  0.99215686  1.        ]\n",
      "   ...\n",
      "   [ 0.99215686  0.99215686  1.        ]\n",
      "   [ 0.9843137   0.9843137   1.        ]\n",
      "   [ 0.99215686  0.9764706   1.        ]]\n",
      "\n",
      "  [[ 0.99215686  1.          0.9843137 ]\n",
      "   [ 0.9764706   0.99215686  0.96862745]\n",
      "   [ 0.9843137   1.          0.9764706 ]\n",
      "   ...\n",
      "   [ 0.8980392   0.8117647   0.7647059 ]\n",
      "   [ 0.8980392   0.8117647   0.7647059 ]\n",
      "   [ 0.90588236  0.8039216   0.75686276]]\n",
      "\n",
      "  ...\n",
      "\n",
      "  [[ 1.          1.          1.        ]\n",
      "   [ 1.          1.          1.        ]\n",
      "   [ 1.          1.          1.        ]\n",
      "   ...\n",
      "   [ 0.7882353   0.6862745   0.62352943]\n",
      "   [ 0.8509804   0.78039217  0.7411765 ]\n",
      "   [ 0.9529412   0.9137255   0.8901961 ]]\n",
      "\n",
      "  [[ 1.          1.          1.        ]\n",
      "   [ 1.          1.          1.        ]\n",
      "   [ 1.          1.          1.        ]\n",
      "   ...\n",
      "   [ 0.69411767  0.5764706   0.52156866]\n",
      "   [ 0.7490196   0.6627451   0.6313726 ]\n",
      "   [ 0.8352941   0.77254903  0.75686276]]\n",
      "\n",
      "  [[ 1.          1.          1.        ]\n",
      "   [ 1.          1.          1.        ]\n",
      "   [ 1.          1.          1.        ]\n",
      "   ...\n",
      "   [ 0.6313726   0.52156866  0.45098042]\n",
      "   [ 0.67058825  0.58431375  0.5372549 ]\n",
      "   [ 0.73333335  0.67058825  0.654902  ]]]\n",
      "\n",
      "\n",
      " [[[ 0.84313726  0.8980392   0.5764706 ]\n",
      "   [ 0.9764706   1.          0.73333335]\n",
      "   [ 0.8666667   0.8352941   0.60784316]\n",
      "   ...\n",
      "   [ 1.          1.          0.94509804]\n",
      "   [ 0.90588236  0.8980392   0.8666667 ]\n",
      "   [ 0.84313726  0.84313726  0.85882354]]\n",
      "\n",
      "  [[ 0.8901961   0.92156863  0.6313726 ]\n",
      "   [ 0.85882354  0.8666667   0.6       ]\n",
      "   [ 0.94509804  0.9137255   0.7019608 ]\n",
      "   ...\n",
      "   [ 0.9843137   0.99215686  0.9529412 ]\n",
      "   [ 0.94509804  0.94509804  0.92941177]\n",
      "   [ 0.9372549   0.92941177  0.96862745]]\n",
      "\n",
      "  [[ 0.90588236  0.9137255   0.64705884]\n",
      "   [ 0.78039217  0.77254903  0.5294118 ]\n",
      "   [ 0.92156863  0.90588236  0.70980394]\n",
      "   ...\n",
      "   [ 0.85882354  0.8745098   0.8666667 ]\n",
      "   [ 0.92941177  0.92941177  0.94509804]\n",
      "   [ 0.99215686  0.9843137   1.        ]]\n",
      "\n",
      "  ...\n",
      "\n",
      "  [[ 0.06666672 -0.35686272 -0.56078434]\n",
      "   [-0.00392157 -0.44313723 -0.62352943]\n",
      "   [ 0.17647064 -0.31764704 -0.4352941 ]\n",
      "   ...\n",
      "   [ 0.94509804 -0.62352943 -0.54509807]\n",
      "   [ 0.9372549  -0.64705884 -0.6627451 ]\n",
      "   [ 0.8980392  -0.7411765  -0.79607844]]\n",
      "\n",
      "  [[-0.09803921 -0.7411765  -0.70980394]\n",
      "   [ 0.04313731 -0.5686275  -0.7411765 ]\n",
      "   [ 0.19215691 -0.29411763 -0.6156863 ]\n",
      "   ...\n",
      "   [ 0.7882353  -0.60784316 -0.60784316]\n",
      "   [ 1.         -0.60784316 -0.84313726]\n",
      "   [ 0.96862745 -0.49019605 -0.654902  ]]\n",
      "\n",
      "  [[ 0.15294123 -0.49019605 -0.4588235 ]\n",
      "   [-0.17647058 -0.7411765  -0.92941177]\n",
      "   [ 0.01176476 -0.40392154 -0.7490196 ]\n",
      "   ...\n",
      "   [ 0.8901961  -0.54509807 -0.5058824 ]\n",
      "   [ 1.         -0.6392157  -0.81960785]\n",
      "   [ 0.8666667  -0.6392157  -0.7254902 ]]]]\n"
     ]
    }
   ],
   "source": [
    "print(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "06ca1056",
   "metadata": {},
   "outputs": [],
   "source": [
    "# train_test_split() performs the split and returns four sequences\n",
    "(trainX, testX, trainY, testY) = train_test_split(data, targets,\n",
    "\ttest_size=0.20, stratify=targets, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "40541e10",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1100, 224, 224, 3) (276, 224, 224, 3) (1100, 2) (276, 2) 276\n"
     ]
    }
   ],
   "source": [
    "print( trainX.shape, testX.shape, trainY.shape, testY.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "99ecc9c7",
   "metadata": {},
   "outputs": [],
   "source": [
    "aug = ImageDataGenerator(\n",
    "\trotation_range=20,\n",
    "\tzoom_range=0.15,\n",
    "\twidth_shift_range=0.2,\n",
    "\theight_shift_range=0.2,\n",
    "\tshear_range=0.15,\n",
    "\thorizontal_flip=True,\n",
    "\tfill_mode=\"nearest\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "ac0d1648",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/roshitab/opt/anaconda3/envs/CWPProject1/lib/python3.7/site-packages/keras_applications/mobilenet_v2.py:294: UserWarning: `input_shape` is undefined or non-square, or `rows` is not in [96, 128, 160, 192, 224]. Weights for input shape (224, 224) will be loaded as the default.\n",
      "  warnings.warn('`input_shape` is undefined or non-square, '\n"
     ]
    }
   ],
   "source": [
    "# load the MobileNetV2 network, ensuring the head FC layer sets are\n",
    "# left off\n",
    "baseModel = MobileNetV2(weights=\"imagenet\", include_top=False,\n",
    "\tinput_tensor=Input(shape=(224, 224, 3)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "55300942",
   "metadata": {},
   "outputs": [],
   "source": [
    "headModel = baseModel.output\n",
    "headModel = AveragePooling2D(pool_size=(7, 7))(headModel)\n",
    "headModel = Flatten(name=\"flatten\")(headModel)\n",
    "\n",
    "# 128 refers to the neurons in the hidden layer\n",
    "headModel = Dense(128, activation=\"relu\")(headModel)\n",
    "\n",
    "# 0.5 refers to the fraction of the input that is dropped\n",
    "headModel = Dropout(0.5)(headModel)\n",
    "\n",
    "# 2 refers to the neurons in the output layer\n",
    "headModel = Dense(2, activation=\"softmax\")(headModel)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "456880f2",
   "metadata": {},
   "outputs": [],
   "source": [
    "# place the head FC model on top of the base model (this will become\n",
    "# the actual model we will train)\n",
    "model = Model(inputs=baseModel.input, outputs=headModel)\n",
    "\n",
    "# loop over all layers in the base model and freeze them so they will\n",
    "# *not* be updated during the first training process\n",
    "for layer in baseModel.layers:\n",
    "\tlayer.trainable = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "2fe73e1b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[INFO] compiling model...\n"
     ]
    }
   ],
   "source": [
    "# compile our model\n",
    "print(\"[INFO] compiling model...\")\n",
    "opt = Adam(lr=INIT_LR, decay=INIT_LR / EPOCHS)\n",
    "model.compile(loss=\"binary_crossentropy\", optimizer=opt,\n",
    "\tmetrics=[\"accuracy\"])\n",
    "#accuracy computers the accuracy rate across all our predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "73cead80",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[INFO] training head...\n",
      "Train for 34 steps, validate on 276 samples\n",
      "Epoch 1/20\n",
      "34/34 [==============================] - 98s 3s/step - loss: 0.7153 - accuracy: 0.6283 - val_loss: 0.2656 - val_accuracy: 0.9219\n",
      "Epoch 2/20\n",
      "34/34 [==============================] - 94s 3s/step - loss: 0.4371 - accuracy: 0.8015 - val_loss: 0.1654 - val_accuracy: 0.9609\n",
      "Epoch 3/20\n",
      "34/34 [==============================] - 92s 3s/step - loss: 0.3288 - accuracy: 0.8605 - val_loss: 0.1208 - val_accuracy: 0.9766\n",
      "Epoch 4/20\n",
      "34/34 [==============================] - 95s 3s/step - loss: 0.2690 - accuracy: 0.8933 - val_loss: 0.0925 - val_accuracy: 0.9844\n",
      "Epoch 5/20\n",
      "34/34 [==============================] - 94s 3s/step - loss: 0.2152 - accuracy: 0.9213 - val_loss: 0.0831 - val_accuracy: 0.9844\n",
      "Epoch 6/20\n",
      "34/34 [==============================] - 96s 3s/step - loss: 0.2003 - accuracy: 0.9298 - val_loss: 0.0688 - val_accuracy: 0.9883\n",
      "Epoch 7/20\n",
      "34/34 [==============================] - 96s 3s/step - loss: 0.1717 - accuracy: 0.9260 - val_loss: 0.0551 - val_accuracy: 0.9883\n",
      "Epoch 8/20\n",
      "34/34 [==============================] - 96s 3s/step - loss: 0.1678 - accuracy: 0.9354 - val_loss: 0.0475 - val_accuracy: 0.9922\n",
      "Epoch 9/20\n",
      "34/34 [==============================] - 96s 3s/step - loss: 0.1325 - accuracy: 0.9569 - val_loss: 0.0640 - val_accuracy: 0.9844\n",
      "Epoch 10/20\n",
      "34/34 [==============================] - 95s 3s/step - loss: 0.1483 - accuracy: 0.9494 - val_loss: 0.0408 - val_accuracy: 0.9883\n",
      "Epoch 11/20\n",
      "34/34 [==============================] - 94s 3s/step - loss: 0.1247 - accuracy: 0.9560 - val_loss: 0.0483 - val_accuracy: 0.9883\n",
      "Epoch 12/20\n",
      "34/34 [==============================] - 94s 3s/step - loss: 0.1292 - accuracy: 0.9485 - val_loss: 0.0504 - val_accuracy: 0.9922\n",
      "Epoch 13/20\n",
      "34/34 [==============================] - 94s 3s/step - loss: 0.1278 - accuracy: 0.9494 - val_loss: 0.0478 - val_accuracy: 0.9883\n",
      "Epoch 14/20\n",
      "34/34 [==============================] - 97s 3s/step - loss: 0.1259 - accuracy: 0.9466 - val_loss: 0.0394 - val_accuracy: 0.9922\n",
      "Epoch 15/20\n",
      "34/34 [==============================] - 94s 3s/step - loss: 0.1034 - accuracy: 0.9616 - val_loss: 0.0327 - val_accuracy: 0.9922\n",
      "Epoch 16/20\n",
      "34/34 [==============================] - 95s 3s/step - loss: 0.1011 - accuracy: 0.9616 - val_loss: 0.0409 - val_accuracy: 0.9922\n",
      "Epoch 17/20\n",
      "34/34 [==============================] - 97s 3s/step - loss: 0.0993 - accuracy: 0.9616 - val_loss: 0.0264 - val_accuracy: 0.9922\n",
      "Epoch 18/20\n",
      "34/34 [==============================] - 94s 3s/step - loss: 0.1018 - accuracy: 0.9597 - val_loss: 0.0311 - val_accuracy: 0.9922\n",
      "Epoch 19/20\n",
      "34/34 [==============================] - 95s 3s/step - loss: 0.0853 - accuracy: 0.9700 - val_loss: 0.0346 - val_accuracy: 0.9922\n",
      "Epoch 20/20\n",
      "34/34 [==============================] - 96s 3s/step - loss: 0.0898 - accuracy: 0.9672 - val_loss: 0.0350 - val_accuracy: 0.9922\n"
     ]
    }
   ],
   "source": [
    "print(\"[INFO] training head...\")\n",
    "H = model.fit(\n",
    "\taug.flow(trainX, trainY, batch_size=BS),\n",
    "\tsteps_per_epoch=len(trainX) // BS,\n",
    "\tvalidation_data=(testX, testY),\n",
    "\tvalidation_steps=len(testX) // BS,\n",
    "\tepochs=EPOCHS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "b6ec5be1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[INFO] evaluating network...\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "   with mask       0.99      1.00      0.99       138\n",
      "without mask       1.00      0.99      0.99       138\n",
      "\n",
      "    accuracy                           0.99       276\n",
      "   macro avg       0.99      0.99      0.99       276\n",
      "weighted avg       0.99      0.99      0.99       276\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(\"[INFO] evaluating network...\")\n",
    "predIdxs = model.predict(testX, batch_size=BS)\n",
    "\n",
    "# for each image in the testing set we need to find the index of the\n",
    "# label with corresponding largest predicted probability\n",
    "predIdxs = np.argmax(predIdxs, axis=1)\n",
    "\n",
    "# show a nicely formatted classification report\n",
    "print(classification_report(testY.argmax(axis=1), predIdxs,\n",
    "\ttarget_names=lb.classes_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "3661b28d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[INFO] saving mask detector model...\n"
     ]
    }
   ],
   "source": [
    "print(\"[INFO] saving mask detector model...\")\n",
    "model.save(\"mask_detector.model\", save_format=\"h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "ad67b6cf",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "N = EPOCHS\n",
    "plt.style.use(\"ggplot\")\n",
    "plt.figure()\n",
    "plt.plot(np.arange(0, N), H.history[\"loss\"], label=\"train_loss\")\n",
    "plt.plot(np.arange(0, N), H.history[\"val_loss\"], label=\"val_loss\")\n",
    "plt.plot(np.arange(0, N), H.history[\"accuracy\"], label=\"train_acc\")\n",
    "plt.plot(np.arange(0, N), H.history[\"val_accuracy\"], label=\"val_acc\")\n",
    "plt.title(\"Training Loss and Accuracy\")\n",
    "plt.xlabel(\"Epoch #\")\n",
    "plt.ylabel(\"Loss/Accuracy\")\n",
    "plt.legend(loc=\"lower left\")\n",
    "plt.savefig(\"plot.png\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}