violencedetection.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "macOS-13.4.1-arm64-arm-64bit\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "import platform\n",
    "from IPython.display import clear_output\n",
    "print(platform.platform())\n",
    "\n",
    "def resolve_dir(Dir):\n",
    "    if not os.path.exists(Dir):\n",
    "        os.mkdir(Dir)\n",
    "\n",
    "def reset_path(Dir):\n",
    "    if not os.path.exists(Dir):\n",
    "        os.mkdir(Dir)\n",
    "    else:\n",
    "        os.system('rm -f {}/*'.format( Dir))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "zsh:1: command not found: nvidia-smi\n",
      "Tensorflow version 2.13.0-rc2\n"
     ]
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "tf.random.set_seed(73)\n",
    "TPU_INIT = False\n",
    "\n",
    "if TPU_INIT:\n",
    "    try:\n",
    "        tpu = tf.distribute.cluster_resolver.TPUClusterResolver.connect()\n",
    "        tpu_strategy = tf.distribute.experimental.TPUStrategy(tpu)\n",
    "    \n",
    "    except ValueError:\n",
    "        raise BaseException('ERROR: Not connected to a TPU runtime!')\n",
    "else:\n",
    "    !nvidia-smi\n",
    ";    \n",
    "print(\"Tensorflow version \" + tf.__version__)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import cv2\n",
    "import os\n",
    "import imageio\n",
    "import imgaug.augmenters as iaa\n",
    "import imgaug as ia\n",
    "\n",
    "IMG_SIZE = 128\n",
    "ColorChannels = 3\n",
    "\n",
    "def video_to_frames(video):\n",
    "    vidcap = cv2.VideoCapture(video)\n",
    "    \n",
    "    import math\n",
    "    rate = math.floor(vidcap.get(3))\n",
    "    count = 0\n",
    "    \n",
    "    ImageFrames = []\n",
    "    while vidcap.isOpened():\n",
    "        ID = vidcap.get(1)\n",
    "        success, image = vidcap.read()\n",
    "        \n",
    "        if success:\n",
    "            if (ID % 7 == 0):\n",
    "                flip = iaa.Fliplr(1.0)\n",
    "                zoom = iaa.Affine(scale=1.3)\n",
    "                random_brightness = iaa.Multiply((1, 1.3))\n",
    "                rotate = iaa.Affine(rotate=(-25, 25))\n",
    "                \n",
    "                image_aug = flip(image = image)\n",
    "                image_aug = random_brightness(image = image_aug)\n",
    "                image_aug = zoom(image = image_aug)\n",
    "                image_aug = rotate(image = image_aug)\n",
    "                \n",
    "                rgb_img = cv2.cvtColor(image_aug, cv2.COLOR_BGR2RGB)\n",
    "                resized = cv2.resize(rgb_img, (IMG_SIZE, IMG_SIZE))\n",
    "                ImageFrames.append(resized)\n",
    "            count+=1\n",
    "        else:\n",
    "            break\n",
    "    \n",
    "    vidcap.release()\n",
    "    \n",
    "    return ImageFrames"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "we have \n",
      "1000 Violence videos \n",
      "1000 NonViolence videos\n"
     ]
    }
   ],
   "source": [
    "print('we have \\n{} Violence videos \\n{} NonViolence videos'.format(\n",
    "              len(os.listdir('/Users/faizahkureshi/Desktop/GAIP Project/violencedataset/NonViolence')), \n",
    "              len(os.listdir('/Users/faizahkureshi/Desktop/GAIP Project/violencedataset/NonViolence'))))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "we have \n",
      "1000 Violence videos \n",
      "1000 NonViolence videos\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 350/350 [00:21<00:00, 16.53it/s]\n",
      "100%|██████████| 350/350 [00:59<00:00,  5.88it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 3min 28s, sys: 5.94 s, total: 3min 34s\n",
      "Wall time: 1min 20s\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "#tqdm is used here to visulaize the progress of dataset creation\n",
    "from tqdm import tqdm\n",
    "\n",
    "VideoDataDir ='/Users/faizahkureshi/Desktop/GAIP Project/violencedataset'\n",
    "print('we have \\n{} Violence videos \\n{} NonViolence videos'.format(\n",
    "              len(os.listdir('/Users/faizahkureshi/Desktop/GAIP Project/violencedataset/NonViolence')), \n",
    "              len(os.listdir('/Users/faizahkureshi/Desktop/GAIP Project/violencedataset/NonViolence'))))\n",
    "\n",
    "X_original = []\n",
    "y_original = []\n",
    "CLASSES = [\"NonViolence\", \"Violence\"]\n",
    "\n",
    "for category in os.listdir(VideoDataDir):\n",
    "    category_path = os.path.join(VideoDataDir, category)\n",
    "    if os.path.isdir(category_path):\n",
    "        class_num = CLASSES.index(category)\n",
    "        try:\n",
    "            video_list = [video for video in os.listdir(category_path)[:350] if video != '.DS_Store']\n",
    "            for i, video in enumerate(tqdm(video_list)):\n",
    "                video_path = os.path.join(category_path, video)\n",
    "                frames = video_to_frames(video_path)\n",
    "                for j, frame in enumerate(frames):\n",
    "                    X_original.append(frame)\n",
    "                    y_original.append(class_num)\n",
    "        except OSError as e:\n",
    "            print(f\"Error processing directory: {category_path}. Skipping...\")\n",
    "            continue\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "13487"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy as np\n",
    "X_original = np.array(X_original).reshape(-1 , IMG_SIZE * IMG_SIZE * 3)\n",
    "y_original = np.array(y_original)\n",
    "len(X_original)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import StratifiedShuffleSplit\n",
    "\n",
    "stratified_sample = StratifiedShuffleSplit(n_splits=2, test_size=0.3, random_state=73)\n",
    "\n",
    "for train_index, test_index in stratified_sample.split(X_original, y_original):\n",
    "    X_train, X_test = X_original[train_index], X_original[test_index]\n",
    "    y_train, y_test = y_original[train_index], y_original[test_index]\n",
    "\n",
    "X_train_nn = X_train.reshape(-1, IMG_SIZE, IMG_SIZE, 3) / 255\n",
    "X_test_nn = X_test.reshape(-1, IMG_SIZE, IMG_SIZE, 3) / 255"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Model Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.layers import Input, Dropout, Flatten, Dense\n",
    "from tensorflow.keras.models import Model\n",
    "import matplotlib.pyplot as plt\n",
    "import cv2\n",
    "import os\n",
    "import numpy as np\n",
    "import pickle\n",
    "import matplotlib\n",
    "matplotlib.use(\"Agg\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "import ssl\n",
    "ssl._create_default_https_context = ssl._create_unverified_context"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:`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",
      "Compiling model...\n",
      "Model: \"model_1\"\n",
      "__________________________________________________________________________________________________\n",
      " Layer (type)                Output Shape                 Param #   Connected to                  \n",
      "==================================================================================================\n",
      " input_2 (InputLayer)        [(None, 128, 128, 3)]        0         []                            \n",
      "                                                                                                  \n",
      " Conv1 (Conv2D)              (None, 64, 64, 32)           864       ['input_2[0][0]']             \n",
      "                                                                                                  \n",
      " bn_Conv1 (BatchNormalizati  (None, 64, 64, 32)           128       ['Conv1[0][0]']               \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " Conv1_relu (ReLU)           (None, 64, 64, 32)           0         ['bn_Conv1[0][0]']            \n",
      "                                                                                                  \n",
      " expanded_conv_depthwise (D  (None, 64, 64, 32)           288       ['Conv1_relu[0][0]']          \n",
      " epthwiseConv2D)                                                                                  \n",
      "                                                                                                  \n",
      " expanded_conv_depthwise_BN  (None, 64, 64, 32)           128       ['expanded_conv_depthwise[0][0\n",
      "  (BatchNormalization)                                              ]']                           \n",
      "                                                                                                  \n",
      " expanded_conv_depthwise_re  (None, 64, 64, 32)           0         ['expanded_conv_depthwise_BN[0\n",
      " lu (ReLU)                                                          ][0]']                        \n",
      "                                                                                                  \n",
      " expanded_conv_project (Con  (None, 64, 64, 16)           512       ['expanded_conv_depthwise_relu\n",
      " v2D)                                                               [0][0]']                      \n",
      "                                                                                                  \n",
      " expanded_conv_project_BN (  (None, 64, 64, 16)           64        ['expanded_conv_project[0][0]'\n",
      " BatchNormalization)                                                ]                             \n",
      "                                                                                                  \n",
      " block_1_expand (Conv2D)     (None, 64, 64, 96)           1536      ['expanded_conv_project_BN[0][\n",
      "                                                                    0]']                          \n",
      "                                                                                                  \n",
      " block_1_expand_BN (BatchNo  (None, 64, 64, 96)           384       ['block_1_expand[0][0]']      \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " block_1_expand_relu (ReLU)  (None, 64, 64, 96)           0         ['block_1_expand_BN[0][0]']   \n",
      "                                                                                                  \n",
      " block_1_pad (ZeroPadding2D  (None, 65, 65, 96)           0         ['block_1_expand_relu[0][0]'] \n",
      " )                                                                                                \n",
      "                                                                                                  \n",
      " block_1_depthwise (Depthwi  (None, 32, 32, 96)           864       ['block_1_pad[0][0]']         \n",
      " seConv2D)                                                                                        \n",
      "                                                                                                  \n",
      " block_1_depthwise_BN (Batc  (None, 32, 32, 96)           384       ['block_1_depthwise[0][0]']   \n",
      " hNormalization)                                                                                  \n",
      "                                                                                                  \n",
      " block_1_depthwise_relu (Re  (None, 32, 32, 96)           0         ['block_1_depthwise_BN[0][0]']\n",
      " LU)                                                                                              \n",
      "                                                                                                  \n",
      " block_1_project (Conv2D)    (None, 32, 32, 24)           2304      ['block_1_depthwise_relu[0][0]\n",
      "                                                                    ']                            \n",
      "                                                                                                  \n",
      " block_1_project_BN (BatchN  (None, 32, 32, 24)           96        ['block_1_project[0][0]']     \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " block_2_expand (Conv2D)     (None, 32, 32, 144)          3456      ['block_1_project_BN[0][0]']  \n",
      "                                                                                                  \n",
      " block_2_expand_BN (BatchNo  (None, 32, 32, 144)          576       ['block_2_expand[0][0]']      \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " block_2_expand_relu (ReLU)  (None, 32, 32, 144)          0         ['block_2_expand_BN[0][0]']   \n",
      "                                                                                                  \n",
      " block_2_depthwise (Depthwi  (None, 32, 32, 144)          1296      ['block_2_expand_relu[0][0]'] \n",
      " seConv2D)                                                                                        \n",
      "                                                                                                  \n",
      " block_2_depthwise_BN (Batc  (None, 32, 32, 144)          576       ['block_2_depthwise[0][0]']   \n",
      " hNormalization)                                                                                  \n",
      "                                                                                                  \n",
      " block_2_depthwise_relu (Re  (None, 32, 32, 144)          0         ['block_2_depthwise_BN[0][0]']\n",
      " LU)                                                                                              \n",
      "                                                                                                  \n",
      " block_2_project (Conv2D)    (None, 32, 32, 24)           3456      ['block_2_depthwise_relu[0][0]\n",
      "                                                                    ']                            \n",
      "                                                                                                  \n",
      " block_2_project_BN (BatchN  (None, 32, 32, 24)           96        ['block_2_project[0][0]']     \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " block_2_add (Add)           (None, 32, 32, 24)           0         ['block_1_project_BN[0][0]',  \n",
      "                                                                     'block_2_project_BN[0][0]']  \n",
      "                                                                                                  \n",
      " block_3_expand (Conv2D)     (None, 32, 32, 144)          3456      ['block_2_add[0][0]']         \n",
      "                                                                                                  \n",
      " block_3_expand_BN (BatchNo  (None, 32, 32, 144)          576       ['block_3_expand[0][0]']      \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " block_3_expand_relu (ReLU)  (None, 32, 32, 144)          0         ['block_3_expand_BN[0][0]']   \n",
      "                                                                                                  \n",
      " block_3_pad (ZeroPadding2D  (None, 33, 33, 144)          0         ['block_3_expand_relu[0][0]'] \n",
      " )                                                                                                \n",
      "                                                                                                  \n",
      " block_3_depthwise (Depthwi  (None, 16, 16, 144)          1296      ['block_3_pad[0][0]']         \n",
      " seConv2D)                                                                                        \n",
      "                                                                                                  \n",
      " block_3_depthwise_BN (Batc  (None, 16, 16, 144)          576       ['block_3_depthwise[0][0]']   \n",
      " hNormalization)                                                                                  \n",
      "                                                                                                  \n",
      " block_3_depthwise_relu (Re  (None, 16, 16, 144)          0         ['block_3_depthwise_BN[0][0]']\n",
      " LU)                                                                                              \n",
      "                                                                                                  \n",
      " block_3_project (Conv2D)    (None, 16, 16, 32)           4608      ['block_3_depthwise_relu[0][0]\n",
      "                                                                    ']                            \n",
      "                                                                                                  \n",
      " block_3_project_BN (BatchN  (None, 16, 16, 32)           128       ['block_3_project[0][0]']     \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " block_4_expand (Conv2D)     (None, 16, 16, 192)          6144      ['block_3_project_BN[0][0]']  \n",
      "                                                                                                  \n",
      " block_4_expand_BN (BatchNo  (None, 16, 16, 192)          768       ['block_4_expand[0][0]']      \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " block_4_expand_relu (ReLU)  (None, 16, 16, 192)          0         ['block_4_expand_BN[0][0]']   \n",
      "                                                                                                  \n",
      " block_4_depthwise (Depthwi  (None, 16, 16, 192)          1728      ['block_4_expand_relu[0][0]'] \n",
      " seConv2D)                                                                                        \n",
      "                                                                                                  \n",
      " block_4_depthwise_BN (Batc  (None, 16, 16, 192)          768       ['block_4_depthwise[0][0]']   \n",
      " hNormalization)                                                                                  \n",
      "                                                                                                  \n",
      " block_4_depthwise_relu (Re  (None, 16, 16, 192)          0         ['block_4_depthwise_BN[0][0]']\n",
      " LU)                                                                                              \n",
      "                                                                                                  \n",
      " block_4_project (Conv2D)    (None, 16, 16, 32)           6144      ['block_4_depthwise_relu[0][0]\n",
      "                                                                    ']                            \n",
      "                                                                                                  \n",
      " block_4_project_BN (BatchN  (None, 16, 16, 32)           128       ['block_4_project[0][0]']     \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " block_4_add (Add)           (None, 16, 16, 32)           0         ['block_3_project_BN[0][0]',  \n",
      "                                                                     'block_4_project_BN[0][0]']  \n",
      "                                                                                                  \n",
      " block_5_expand (Conv2D)     (None, 16, 16, 192)          6144      ['block_4_add[0][0]']         \n",
      "                                                                                                  \n",
      " block_5_expand_BN (BatchNo  (None, 16, 16, 192)          768       ['block_5_expand[0][0]']      \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " block_5_expand_relu (ReLU)  (None, 16, 16, 192)          0         ['block_5_expand_BN[0][0]']   \n",
      "                                                                                                  \n",
      " block_5_depthwise (Depthwi  (None, 16, 16, 192)          1728      ['block_5_expand_relu[0][0]'] \n",
      " seConv2D)                                                                                        \n",
      "                                                                                                  \n",
      " block_5_depthwise_BN (Batc  (None, 16, 16, 192)          768       ['block_5_depthwise[0][0]']   \n",
      " hNormalization)                                                                                  \n",
      "                                                                                                  \n",
      " block_5_depthwise_relu (Re  (None, 16, 16, 192)          0         ['block_5_depthwise_BN[0][0]']\n",
      " LU)                                                                                              \n",
      "                                                                                                  \n",
      " block_5_project (Conv2D)    (None, 16, 16, 32)           6144      ['block_5_depthwise_relu[0][0]\n",
      "                                                                    ']                            \n",
      "                                                                                                  \n",
      " block_5_project_BN (BatchN  (None, 16, 16, 32)           128       ['block_5_project[0][0]']     \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " block_5_add (Add)           (None, 16, 16, 32)           0         ['block_4_add[0][0]',         \n",
      "                                                                     'block_5_project_BN[0][0]']  \n",
      "                                                                                                  \n",
      " block_6_expand (Conv2D)     (None, 16, 16, 192)          6144      ['block_5_add[0][0]']         \n",
      "                                                                                                  \n",
      " block_6_expand_BN (BatchNo  (None, 16, 16, 192)          768       ['block_6_expand[0][0]']      \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " block_6_expand_relu (ReLU)  (None, 16, 16, 192)          0         ['block_6_expand_BN[0][0]']   \n",
      "                                                                                                  \n",
      " block_6_pad (ZeroPadding2D  (None, 17, 17, 192)          0         ['block_6_expand_relu[0][0]'] \n",
      " )                                                                                                \n",
      "                                                                                                  \n",
      " block_6_depthwise (Depthwi  (None, 8, 8, 192)            1728      ['block_6_pad[0][0]']         \n",
      " seConv2D)                                                                                        \n",
      "                                                                                                  \n",
      " block_6_depthwise_BN (Batc  (None, 8, 8, 192)            768       ['block_6_depthwise[0][0]']   \n",
      " hNormalization)                                                                                  \n",
      "                                                                                                  \n",
      " block_6_depthwise_relu (Re  (None, 8, 8, 192)            0         ['block_6_depthwise_BN[0][0]']\n",
      " LU)                                                                                              \n",
      "                                                                                                  \n",
      " block_6_project (Conv2D)    (None, 8, 8, 64)             12288     ['block_6_depthwise_relu[0][0]\n",
      "                                                                    ']                            \n",
      "                                                                                                  \n",
      " block_6_project_BN (BatchN  (None, 8, 8, 64)             256       ['block_6_project[0][0]']     \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " block_7_expand (Conv2D)     (None, 8, 8, 384)            24576     ['block_6_project_BN[0][0]']  \n",
      "                                                                                                  \n",
      " block_7_expand_BN (BatchNo  (None, 8, 8, 384)            1536      ['block_7_expand[0][0]']      \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " block_7_expand_relu (ReLU)  (None, 8, 8, 384)            0         ['block_7_expand_BN[0][0]']   \n",
      "                                                                                                  \n",
      " block_7_depthwise (Depthwi  (None, 8, 8, 384)            3456      ['block_7_expand_relu[0][0]'] \n",
      " seConv2D)                                                                                        \n",
      "                                                                                                  \n",
      " block_7_depthwise_BN (Batc  (None, 8, 8, 384)            1536      ['block_7_depthwise[0][0]']   \n",
      " hNormalization)                                                                                  \n",
      "                                                                                                  \n",
      " block_7_depthwise_relu (Re  (None, 8, 8, 384)            0         ['block_7_depthwise_BN[0][0]']\n",
      " LU)                                                                                              \n",
      "                                                                                                  \n",
      " block_7_project (Conv2D)    (None, 8, 8, 64)             24576     ['block_7_depthwise_relu[0][0]\n",
      "                                                                    ']                            \n",
      "                                                                                                  \n",
      " block_7_project_BN (BatchN  (None, 8, 8, 64)             256       ['block_7_project[0][0]']     \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " block_7_add (Add)           (None, 8, 8, 64)             0         ['block_6_project_BN[0][0]',  \n",
      "                                                                     'block_7_project_BN[0][0]']  \n",
      "                                                                                                  \n",
      " block_8_expand (Conv2D)     (None, 8, 8, 384)            24576     ['block_7_add[0][0]']         \n",
      "                                                                                                  \n",
      " block_8_expand_BN (BatchNo  (None, 8, 8, 384)            1536      ['block_8_expand[0][0]']      \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " block_8_expand_relu (ReLU)  (None, 8, 8, 384)            0         ['block_8_expand_BN[0][0]']   \n",
      "                                                                                                  \n",
      " block_8_depthwise (Depthwi  (None, 8, 8, 384)            3456      ['block_8_expand_relu[0][0]'] \n",
      " seConv2D)                                                                                        \n",
      "                                                                                                  \n",
      " block_8_depthwise_BN (Batc  (None, 8, 8, 384)            1536      ['block_8_depthwise[0][0]']   \n",
      " hNormalization)                                                                                  \n",
      "                                                                                                  \n",
      " block_8_depthwise_relu (Re  (None, 8, 8, 384)            0         ['block_8_depthwise_BN[0][0]']\n",
      " LU)                                                                                              \n",
      "                                                                                                  \n",
      " block_8_project (Conv2D)    (None, 8, 8, 64)             24576     ['block_8_depthwise_relu[0][0]\n",
      "                                                                    ']                            \n",
      "                                                                                                  \n",
      " block_8_project_BN (BatchN  (None, 8, 8, 64)             256       ['block_8_project[0][0]']     \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " block_8_add (Add)           (None, 8, 8, 64)             0         ['block_7_add[0][0]',         \n",
      "                                                                     'block_8_project_BN[0][0]']  \n",
      "                                                                                                  \n",
      " block_9_expand (Conv2D)     (None, 8, 8, 384)            24576     ['block_8_add[0][0]']         \n",
      "                                                                                                  \n",
      " block_9_expand_BN (BatchNo  (None, 8, 8, 384)            1536      ['block_9_expand[0][0]']      \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " block_9_expand_relu (ReLU)  (None, 8, 8, 384)            0         ['block_9_expand_BN[0][0]']   \n",
      "                                                                                                  \n",
      " block_9_depthwise (Depthwi  (None, 8, 8, 384)            3456      ['block_9_expand_relu[0][0]'] \n",
      " seConv2D)                                                                                        \n",
      "                                                                                                  \n",
      " block_9_depthwise_BN (Batc  (None, 8, 8, 384)            1536      ['block_9_depthwise[0][0]']   \n",
      " hNormalization)                                                                                  \n",
      "                                                                                                  \n",
      " block_9_depthwise_relu (Re  (None, 8, 8, 384)            0         ['block_9_depthwise_BN[0][0]']\n",
      " LU)                                                                                              \n",
      "                                                                                                  \n",
      " block_9_project (Conv2D)    (None, 8, 8, 64)             24576     ['block_9_depthwise_relu[0][0]\n",
      "                                                                    ']                            \n",
      "                                                                                                  \n",
      " block_9_project_BN (BatchN  (None, 8, 8, 64)             256       ['block_9_project[0][0]']     \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " block_9_add (Add)           (None, 8, 8, 64)             0         ['block_8_add[0][0]',         \n",
      "                                                                     'block_9_project_BN[0][0]']  \n",
      "                                                                                                  \n",
      " block_10_expand (Conv2D)    (None, 8, 8, 384)            24576     ['block_9_add[0][0]']         \n",
      "                                                                                                  \n",
      " block_10_expand_BN (BatchN  (None, 8, 8, 384)            1536      ['block_10_expand[0][0]']     \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " block_10_expand_relu (ReLU  (None, 8, 8, 384)            0         ['block_10_expand_BN[0][0]']  \n",
      " )                                                                                                \n",
      "                                                                                                  \n",
      " block_10_depthwise (Depthw  (None, 8, 8, 384)            3456      ['block_10_expand_relu[0][0]']\n",
      " iseConv2D)                                                                                       \n",
      "                                                                                                  \n",
      " block_10_depthwise_BN (Bat  (None, 8, 8, 384)            1536      ['block_10_depthwise[0][0]']  \n",
      " chNormalization)                                                                                 \n",
      "                                                                                                  \n",
      " block_10_depthwise_relu (R  (None, 8, 8, 384)            0         ['block_10_depthwise_BN[0][0]'\n",
      " eLU)                                                               ]                             \n",
      "                                                                                                  \n",
      " block_10_project (Conv2D)   (None, 8, 8, 96)             36864     ['block_10_depthwise_relu[0][0\n",
      "                                                                    ]']                           \n",
      "                                                                                                  \n",
      " block_10_project_BN (Batch  (None, 8, 8, 96)             384       ['block_10_project[0][0]']    \n",
      " Normalization)                                                                                   \n",
      "                                                                                                  \n",
      " block_11_expand (Conv2D)    (None, 8, 8, 576)            55296     ['block_10_project_BN[0][0]'] \n",
      "                                                                                                  \n",
      " block_11_expand_BN (BatchN  (None, 8, 8, 576)            2304      ['block_11_expand[0][0]']     \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " block_11_expand_relu (ReLU  (None, 8, 8, 576)            0         ['block_11_expand_BN[0][0]']  \n",
      " )                                                                                                \n",
      "                                                                                                  \n",
      " block_11_depthwise (Depthw  (None, 8, 8, 576)            5184      ['block_11_expand_relu[0][0]']\n",
      " iseConv2D)                                                                                       \n",
      "                                                                                                  \n",
      " block_11_depthwise_BN (Bat  (None, 8, 8, 576)            2304      ['block_11_depthwise[0][0]']  \n",
      " chNormalization)                                                                                 \n",
      "                                                                                                  \n",
      " block_11_depthwise_relu (R  (None, 8, 8, 576)            0         ['block_11_depthwise_BN[0][0]'\n",
      " eLU)                                                               ]                             \n",
      "                                                                                                  \n",
      " block_11_project (Conv2D)   (None, 8, 8, 96)             55296     ['block_11_depthwise_relu[0][0\n",
      "                                                                    ]']                           \n",
      "                                                                                                  \n",
      " block_11_project_BN (Batch  (None, 8, 8, 96)             384       ['block_11_project[0][0]']    \n",
      " Normalization)                                                                                   \n",
      "                                                                                                  \n",
      " block_11_add (Add)          (None, 8, 8, 96)             0         ['block_10_project_BN[0][0]', \n",
      "                                                                     'block_11_project_BN[0][0]'] \n",
      "                                                                                                  \n",
      " block_12_expand (Conv2D)    (None, 8, 8, 576)            55296     ['block_11_add[0][0]']        \n",
      "                                                                                                  \n",
      " block_12_expand_BN (BatchN  (None, 8, 8, 576)            2304      ['block_12_expand[0][0]']     \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " block_12_expand_relu (ReLU  (None, 8, 8, 576)            0         ['block_12_expand_BN[0][0]']  \n",
      " )                                                                                                \n",
      "                                                                                                  \n",
      " block_12_depthwise (Depthw  (None, 8, 8, 576)            5184      ['block_12_expand_relu[0][0]']\n",
      " iseConv2D)                                                                                       \n",
      "                                                                                                  \n",
      " block_12_depthwise_BN (Bat  (None, 8, 8, 576)            2304      ['block_12_depthwise[0][0]']  \n",
      " chNormalization)                                                                                 \n",
      "                                                                                                  \n",
      " block_12_depthwise_relu (R  (None, 8, 8, 576)            0         ['block_12_depthwise_BN[0][0]'\n",
      " eLU)                                                               ]                             \n",
      "                                                                                                  \n",
      " block_12_project (Conv2D)   (None, 8, 8, 96)             55296     ['block_12_depthwise_relu[0][0\n",
      "                                                                    ]']                           \n",
      "                                                                                                  \n",
      " block_12_project_BN (Batch  (None, 8, 8, 96)             384       ['block_12_project[0][0]']    \n",
      " Normalization)                                                                                   \n",
      "                                                                                                  \n",
      " block_12_add (Add)          (None, 8, 8, 96)             0         ['block_11_add[0][0]',        \n",
      "                                                                     'block_12_project_BN[0][0]'] \n",
      "                                                                                                  \n",
      " block_13_expand (Conv2D)    (None, 8, 8, 576)            55296     ['block_12_add[0][0]']        \n",
      "                                                                                                  \n",
      " block_13_expand_BN (BatchN  (None, 8, 8, 576)            2304      ['block_13_expand[0][0]']     \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " block_13_expand_relu (ReLU  (None, 8, 8, 576)            0         ['block_13_expand_BN[0][0]']  \n",
      " )                                                                                                \n",
      "                                                                                                  \n",
      " block_13_pad (ZeroPadding2  (None, 9, 9, 576)            0         ['block_13_expand_relu[0][0]']\n",
      " D)                                                                                               \n",
      "                                                                                                  \n",
      " block_13_depthwise (Depthw  (None, 4, 4, 576)            5184      ['block_13_pad[0][0]']        \n",
      " iseConv2D)                                                                                       \n",
      "                                                                                                  \n",
      " block_13_depthwise_BN (Bat  (None, 4, 4, 576)            2304      ['block_13_depthwise[0][0]']  \n",
      " chNormalization)                                                                                 \n",
      "                                                                                                  \n",
      " block_13_depthwise_relu (R  (None, 4, 4, 576)            0         ['block_13_depthwise_BN[0][0]'\n",
      " eLU)                                                               ]                             \n",
      "                                                                                                  \n",
      " block_13_project (Conv2D)   (None, 4, 4, 160)            92160     ['block_13_depthwise_relu[0][0\n",
      "                                                                    ]']                           \n",
      "                                                                                                  \n",
      " block_13_project_BN (Batch  (None, 4, 4, 160)            640       ['block_13_project[0][0]']    \n",
      " Normalization)                                                                                   \n",
      "                                                                                                  \n",
      " block_14_expand (Conv2D)    (None, 4, 4, 960)            153600    ['block_13_project_BN[0][0]'] \n",
      "                                                                                                  \n",
      " block_14_expand_BN (BatchN  (None, 4, 4, 960)            3840      ['block_14_expand[0][0]']     \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " block_14_expand_relu (ReLU  (None, 4, 4, 960)            0         ['block_14_expand_BN[0][0]']  \n",
      " )                                                                                                \n",
      "                                                                                                  \n",
      " block_14_depthwise (Depthw  (None, 4, 4, 960)            8640      ['block_14_expand_relu[0][0]']\n",
      " iseConv2D)                                                                                       \n",
      "                                                                                                  \n",
      " block_14_depthwise_BN (Bat  (None, 4, 4, 960)            3840      ['block_14_depthwise[0][0]']  \n",
      " chNormalization)                                                                                 \n",
      "                                                                                                  \n",
      " block_14_depthwise_relu (R  (None, 4, 4, 960)            0         ['block_14_depthwise_BN[0][0]'\n",
      " eLU)                                                               ]                             \n",
      "                                                                                                  \n",
      " block_14_project (Conv2D)   (None, 4, 4, 160)            153600    ['block_14_depthwise_relu[0][0\n",
      "                                                                    ]']                           \n",
      "                                                                                                  \n",
      " block_14_project_BN (Batch  (None, 4, 4, 160)            640       ['block_14_project[0][0]']    \n",
      " Normalization)                                                                                   \n",
      "                                                                                                  \n",
      " block_14_add (Add)          (None, 4, 4, 160)            0         ['block_13_project_BN[0][0]', \n",
      "                                                                     'block_14_project_BN[0][0]'] \n",
      "                                                                                                  \n",
      " block_15_expand (Conv2D)    (None, 4, 4, 960)            153600    ['block_14_add[0][0]']        \n",
      "                                                                                                  \n",
      " block_15_expand_BN (BatchN  (None, 4, 4, 960)            3840      ['block_15_expand[0][0]']     \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " block_15_expand_relu (ReLU  (None, 4, 4, 960)            0         ['block_15_expand_BN[0][0]']  \n",
      " )                                                                                                \n",
      "                                                                                                  \n",
      " block_15_depthwise (Depthw  (None, 4, 4, 960)            8640      ['block_15_expand_relu[0][0]']\n",
      " iseConv2D)                                                                                       \n",
      "                                                                                                  \n",
      " block_15_depthwise_BN (Bat  (None, 4, 4, 960)            3840      ['block_15_depthwise[0][0]']  \n",
      " chNormalization)                                                                                 \n",
      "                                                                                                  \n",
      " block_15_depthwise_relu (R  (None, 4, 4, 960)            0         ['block_15_depthwise_BN[0][0]'\n",
      " eLU)                                                               ]                             \n",
      "                                                                                                  \n",
      " block_15_project (Conv2D)   (None, 4, 4, 160)            153600    ['block_15_depthwise_relu[0][0\n",
      "                                                                    ]']                           \n",
      "                                                                                                  \n",
      " block_15_project_BN (Batch  (None, 4, 4, 160)            640       ['block_15_project[0][0]']    \n",
      " Normalization)                                                                                   \n",
      "                                                                                                  \n",
      " block_15_add (Add)          (None, 4, 4, 160)            0         ['block_14_add[0][0]',        \n",
      "                                                                     'block_15_project_BN[0][0]'] \n",
      "                                                                                                  \n",
      " block_16_expand (Conv2D)    (None, 4, 4, 960)            153600    ['block_15_add[0][0]']        \n",
      "                                                                                                  \n",
      " block_16_expand_BN (BatchN  (None, 4, 4, 960)            3840      ['block_16_expand[0][0]']     \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " block_16_expand_relu (ReLU  (None, 4, 4, 960)            0         ['block_16_expand_BN[0][0]']  \n",
      " )                                                                                                \n",
      "                                                                                                  \n",
      " block_16_depthwise (Depthw  (None, 4, 4, 960)            8640      ['block_16_expand_relu[0][0]']\n",
      " iseConv2D)                                                                                       \n",
      "                                                                                                  \n",
      " block_16_depthwise_BN (Bat  (None, 4, 4, 960)            3840      ['block_16_depthwise[0][0]']  \n",
      " chNormalization)                                                                                 \n",
      "                                                                                                  \n",
      " block_16_depthwise_relu (R  (None, 4, 4, 960)            0         ['block_16_depthwise_BN[0][0]'\n",
      " eLU)                                                               ]                             \n",
      "                                                                                                  \n",
      " block_16_project (Conv2D)   (None, 4, 4, 320)            307200    ['block_16_depthwise_relu[0][0\n",
      "                                                                    ]']                           \n",
      "                                                                                                  \n",
      " block_16_project_BN (Batch  (None, 4, 4, 320)            1280      ['block_16_project[0][0]']    \n",
      " Normalization)                                                                                   \n",
      "                                                                                                  \n",
      " Conv_1 (Conv2D)             (None, 4, 4, 1280)           409600    ['block_16_project_BN[0][0]'] \n",
      "                                                                                                  \n",
      " Conv_1_bn (BatchNormalizat  (None, 4, 4, 1280)           5120      ['Conv_1[0][0]']              \n",
      " ion)                                                                                             \n",
      "                                                                                                  \n",
      " out_relu (ReLU)             (None, 4, 4, 1280)           0         ['Conv_1_bn[0][0]']           \n",
      "                                                                                                  \n",
      " global_average_pooling2d_1  (None, 1280)                 0         ['out_relu[0][0]']            \n",
      "  (GlobalAveragePooling2D)                                                                        \n",
      "                                                                                                  \n",
      " dense_1 (Dense)             (None, 1)                    1281      ['global_average_pooling2d_1[0\n",
      "                                                                    ][0]']                        \n",
      "                                                                                                  \n",
      "==================================================================================================\n",
      "Total params: 2259265 (8.62 MB)\n",
      "Trainable params: 1281 (5.00 KB)\n",
      "Non-trainable params: 2257984 (8.61 MB)\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "epochs = 50\n",
    "\n",
    "from keras import regularizers\n",
    "kernel_regularizer = regularizers.l2(0.0001)\n",
    "\n",
    "from keras.applications.mobilenet_v2 import MobileNetV2\n",
    "\n",
    "def load_layers():\n",
    "    input_tensor = Input(shape=(IMG_SIZE, IMG_SIZE, ColorChannels))\n",
    "    baseModel = MobileNetV2(pooling='avg',\n",
    "                            include_top=False, \n",
    "                            input_tensor=input_tensor)\n",
    "    \n",
    "    headModel = baseModel.output   \n",
    "    headModel = Dense(1, activation=\"sigmoid\")(headModel)\n",
    "    model = Model(inputs=baseModel.input, outputs=headModel)\n",
    "\n",
    "    for layer in baseModel.layers:\n",
    "        layer.trainable = False\n",
    "\n",
    "    print(\"Compiling model...\")\n",
    "    model.compile(loss=\"binary_crossentropy\",\n",
    "                    optimizer='adam',\n",
    "                    metrics=[\"accuracy\"])\n",
    "\n",
    "    return model\n",
    "\n",
    "if TPU_INIT:\n",
    "    with tpu_strategy.scope():\n",
    "        model = load_layers()\n",
    "else:\n",
    "    model = load_layers()\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.callbacks import Callback, ModelCheckpoint, LearningRateScheduler, TensorBoard, EarlyStopping, ReduceLROnPlateau\n",
    "\n",
    "patience = 3\n",
    "\n",
    "start_lr = 0.00001\n",
    "min_lr = 0.00001\n",
    "max_lr = 0.00005\n",
    "\n",
    "batch_size = 4\n",
    "\n",
    "if TPU_INIT:\n",
    "    max_lr = max_lr * tpu_strategy.num_replicas_in_sync\n",
    "    batch_size = batch_size * tpu_strategy.num_replicas_in_sync\n",
    "\n",
    "rampup_epochs = 5\n",
    "sustain_epochs = 0\n",
    "exp_decay = .8\n",
    "\n",
    "def lrfn(epoch):\n",
    "    if epoch < rampup_epochs:\n",
    "        return (max_lr - start_lr)/rampup_epochs * epoch + start_lr\n",
    "    elif epoch < rampup_epochs + sustain_epochs:\n",
    "        return max_lr\n",
    "    else:\n",
    "        return (max_lr - min_lr) * exp_decay**(epoch-rampup_epochs-sustain_epochs) + min_lr\n",
    "\n",
    "\n",
    "class myCallback(Callback):\n",
    "    def on_epoch_end(self, epoch, logs={}):\n",
    "        if ((logs.get('accuracy')>=0.999)):\n",
    "            print(\"\\nLimits Reached cancelling training!\")\n",
    "            self.model.stop_training = True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "end_callback = myCallback()\n",
    "\n",
    "lr_callback = LearningRateScheduler(lambda epoch: lrfn(epoch), verbose=False)\n",
    "\n",
    "early_stopping = EarlyStopping(patience = patience, monitor='val_loss',\n",
    "                                 mode='min', restore_best_weights=True, \n",
    "                                 verbose = 1, min_delta = .00075)\n",
    "\n",
    "PROJECT_DIR ='/Users/faizahkureshi/Desktop/GAIP Project/RiskDetection'\n",
    "\n",
    "lr_plat = ReduceLROnPlateau(patience = 2, mode = 'min')\n",
    "\n",
    "os.system('rm -rf ./logs/')\n",
    "\n",
    "import datetime\n",
    "log_dir=\"logs/fit/\" + datetime.datetime.now().strftime(\"%Y%m%d-%H%M%S\")\n",
    "tensorboard_callback = TensorBoard(log_dir = log_dir, write_graph=True, histogram_freq=1)\n",
    "\n",
    "checkpoint_filepath = 'ModelWeights.h5'\n",
    "\n",
    "model_checkpoints = ModelCheckpoint(filepath=checkpoint_filepath,\n",
    "                                        save_weights_only=True,\n",
    "                                        monitor='val_loss',\n",
    "                                        mode='min',\n",
    "                                        verbose = 1,\n",
    "                                        save_best_only=True)\n",
    "\n",
    "\n",
    "callbacks = [end_callback, lr_callback, model_checkpoints, tensorboard_callback, early_stopping, lr_plat]\n",
    "\n",
    "if TPU_INIT:\n",
    "    callbacks = [end_callback, lr_callback, model_checkpoints, early_stopping, lr_plat]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training head...\n",
      "Epoch 1/50\n",
      "2360/2360 [==============================] - ETA: 0s - loss: 0.8058 - accuracy: 0.5502\n",
      "Epoch 1: val_loss improved from inf to 0.70456, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 65s 27ms/step - loss: 0.8058 - accuracy: 0.5502 - val_loss: 0.7046 - val_accuracy: 0.6140 - lr: 1.0000e-05\n",
      "Epoch 2/50\n",
      "2358/2360 [============================>.] - ETA: 0s - loss: 0.6042 - accuracy: 0.6792\n",
      "Epoch 2: val_loss improved from 0.70456 to 0.52811, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 62s 26ms/step - loss: 0.6042 - accuracy: 0.6791 - val_loss: 0.5281 - val_accuracy: 0.7401 - lr: 1.8000e-05\n",
      "Epoch 3/50\n",
      "2359/2360 [============================>.] - ETA: 0s - loss: 0.4746 - accuracy: 0.7738\n",
      "Epoch 3: val_loss improved from 0.52811 to 0.43506, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 60s 25ms/step - loss: 0.4746 - accuracy: 0.7738 - val_loss: 0.4351 - val_accuracy: 0.7994 - lr: 2.6000e-05\n",
      "Epoch 4/50\n",
      "2360/2360 [==============================] - ETA: 0s - loss: 0.4053 - accuracy: 0.8130\n",
      "Epoch 4: val_loss improved from 0.43506 to 0.38379, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 60s 25ms/step - loss: 0.4053 - accuracy: 0.8130 - val_loss: 0.3838 - val_accuracy: 0.8283 - lr: 3.4000e-05\n",
      "Epoch 5/50\n",
      "2359/2360 [============================>.] - ETA: 0s - loss: 0.3628 - accuracy: 0.8384\n",
      "Epoch 5: val_loss improved from 0.38379 to 0.34884, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 66s 28ms/step - loss: 0.3628 - accuracy: 0.8385 - val_loss: 0.3488 - val_accuracy: 0.8463 - lr: 4.2000e-05\n",
      "Epoch 6/50\n",
      "2359/2360 [============================>.] - ETA: 0s - loss: 0.3312 - accuracy: 0.8566\n",
      "Epoch 6: val_loss improved from 0.34884 to 0.32267, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 65s 27ms/step - loss: 0.3311 - accuracy: 0.8567 - val_loss: 0.3227 - val_accuracy: 0.8604 - lr: 5.0000e-05\n",
      "Epoch 7/50\n",
      "2360/2360 [==============================] - ETA: 0s - loss: 0.3078 - accuracy: 0.8702\n",
      "Epoch 7: val_loss improved from 0.32267 to 0.30704, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 65s 28ms/step - loss: 0.3078 - accuracy: 0.8702 - val_loss: 0.3070 - val_accuracy: 0.8685 - lr: 4.2000e-05\n",
      "Epoch 8/50\n",
      "2360/2360 [==============================] - ETA: 0s - loss: 0.2930 - accuracy: 0.8762\n",
      "Epoch 8: val_loss improved from 0.30704 to 0.29620, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 64s 27ms/step - loss: 0.2930 - accuracy: 0.8762 - val_loss: 0.2962 - val_accuracy: 0.8747 - lr: 3.5600e-05\n",
      "Epoch 9/50\n",
      "2360/2360 [==============================] - ETA: 0s - loss: 0.2822 - accuracy: 0.8865\n",
      "Epoch 9: val_loss improved from 0.29620 to 0.28832, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 60s 25ms/step - loss: 0.2822 - accuracy: 0.8865 - val_loss: 0.2883 - val_accuracy: 0.8774 - lr: 3.0480e-05\n",
      "Epoch 10/50\n",
      "2360/2360 [==============================] - ETA: 0s - loss: 0.2744 - accuracy: 0.8892\n",
      "Epoch 10: val_loss improved from 0.28832 to 0.28239, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 54s 23ms/step - loss: 0.2744 - accuracy: 0.8892 - val_loss: 0.2824 - val_accuracy: 0.8809 - lr: 2.6384e-05\n",
      "Epoch 11/50\n",
      "2357/2360 [============================>.] - ETA: 0s - loss: 0.2684 - accuracy: 0.8922\n",
      "Epoch 11: val_loss improved from 0.28239 to 0.27766, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 52s 22ms/step - loss: 0.2682 - accuracy: 0.8924 - val_loss: 0.2777 - val_accuracy: 0.8831 - lr: 2.3107e-05\n",
      "Epoch 12/50\n",
      "2360/2360 [==============================] - ETA: 0s - loss: 0.2631 - accuracy: 0.8942\n",
      "Epoch 12: val_loss improved from 0.27766 to 0.27422, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 54s 23ms/step - loss: 0.2631 - accuracy: 0.8942 - val_loss: 0.2742 - val_accuracy: 0.8849 - lr: 2.0486e-05\n",
      "Epoch 13/50\n",
      "2358/2360 [============================>.] - ETA: 0s - loss: 0.2589 - accuracy: 0.8961\n",
      "Epoch 13: val_loss improved from 0.27422 to 0.27108, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 54s 23ms/step - loss: 0.2588 - accuracy: 0.8961 - val_loss: 0.2711 - val_accuracy: 0.8861 - lr: 1.8389e-05\n",
      "Epoch 14/50\n",
      "2360/2360 [==============================] - ETA: 0s - loss: 0.2556 - accuracy: 0.8976\n",
      "Epoch 14: val_loss improved from 0.27108 to 0.26800, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 54s 23ms/step - loss: 0.2556 - accuracy: 0.8976 - val_loss: 0.2680 - val_accuracy: 0.8883 - lr: 1.6711e-05\n",
      "Epoch 15/50\n",
      "2357/2360 [============================>.] - ETA: 0s - loss: 0.2524 - accuracy: 0.8989\n",
      "Epoch 15: val_loss improved from 0.26800 to 0.26562, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 53s 22ms/step - loss: 0.2524 - accuracy: 0.8989 - val_loss: 0.2656 - val_accuracy: 0.8903 - lr: 1.5369e-05\n",
      "Epoch 16/50\n",
      "2357/2360 [============================>.] - ETA: 0s - loss: 0.2498 - accuracy: 0.8997\n",
      "Epoch 16: val_loss improved from 0.26562 to 0.26358, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 52s 22ms/step - loss: 0.2498 - accuracy: 0.8997 - val_loss: 0.2636 - val_accuracy: 0.8910 - lr: 1.4295e-05\n",
      "Epoch 17/50\n",
      "2357/2360 [============================>.] - ETA: 0s - loss: 0.2474 - accuracy: 0.9005\n",
      "Epoch 17: val_loss improved from 0.26358 to 0.26179, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 52s 22ms/step - loss: 0.2473 - accuracy: 0.9005 - val_loss: 0.2618 - val_accuracy: 0.8910 - lr: 1.3436e-05\n",
      "Epoch 18/50\n",
      "2357/2360 [============================>.] - ETA: 0s - loss: 0.2450 - accuracy: 0.9019\n",
      "Epoch 18: val_loss improved from 0.26179 to 0.26017, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 51s 22ms/step - loss: 0.2453 - accuracy: 0.9018 - val_loss: 0.2602 - val_accuracy: 0.8915 - lr: 1.2749e-05\n",
      "Epoch 19/50\n",
      "2359/2360 [============================>.] - ETA: 0s - loss: 0.2434 - accuracy: 0.9022\n",
      "Epoch 19: val_loss improved from 0.26017 to 0.25873, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 51s 21ms/step - loss: 0.2433 - accuracy: 0.9022 - val_loss: 0.2587 - val_accuracy: 0.8933 - lr: 1.2199e-05\n",
      "Epoch 20/50\n",
      "2357/2360 [============================>.] - ETA: 0s - loss: 0.2416 - accuracy: 0.9031\n",
      "Epoch 20: val_loss improved from 0.25873 to 0.25732, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 51s 22ms/step - loss: 0.2414 - accuracy: 0.9031 - val_loss: 0.2573 - val_accuracy: 0.8937 - lr: 1.1759e-05\n",
      "Epoch 21/50\n",
      "2360/2360 [==============================] - ETA: 0s - loss: 0.2398 - accuracy: 0.9041\n",
      "Epoch 21: val_loss improved from 0.25732 to 0.25599, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 53s 23ms/step - loss: 0.2398 - accuracy: 0.9041 - val_loss: 0.2560 - val_accuracy: 0.8955 - lr: 1.1407e-05\n",
      "Epoch 22/50\n",
      "2359/2360 [============================>.] - ETA: 0s - loss: 0.2382 - accuracy: 0.9049\n",
      "Epoch 22: val_loss improved from 0.25599 to 0.25476, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 53s 22ms/step - loss: 0.2381 - accuracy: 0.9050 - val_loss: 0.2548 - val_accuracy: 0.8952 - lr: 1.1126e-05\n",
      "Epoch 23/50\n",
      "2358/2360 [============================>.] - ETA: 0s - loss: 0.2369 - accuracy: 0.9050\n",
      "Epoch 23: val_loss improved from 0.25476 to 0.25338, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 51s 21ms/step - loss: 0.2368 - accuracy: 0.9050 - val_loss: 0.2534 - val_accuracy: 0.8962 - lr: 1.0901e-05\n",
      "Epoch 24/50\n",
      "2360/2360 [==============================] - ETA: 0s - loss: 0.2353 - accuracy: 0.9061\n",
      "Epoch 24: val_loss improved from 0.25338 to 0.25238, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 52s 22ms/step - loss: 0.2353 - accuracy: 0.9061 - val_loss: 0.2524 - val_accuracy: 0.8982 - lr: 1.0721e-05\n",
      "Epoch 25/50\n",
      "2357/2360 [============================>.] - ETA: 0s - loss: 0.2341 - accuracy: 0.9053\n",
      "Epoch 25: val_loss improved from 0.25238 to 0.25127, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 56s 24ms/step - loss: 0.2340 - accuracy: 0.9053 - val_loss: 0.2513 - val_accuracy: 0.8987 - lr: 1.0576e-05\n",
      "Epoch 26/50\n",
      "2360/2360 [==============================] - ETA: 0s - loss: 0.2327 - accuracy: 0.9062\n",
      "Epoch 26: val_loss improved from 0.25127 to 0.25011, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 53s 22ms/step - loss: 0.2327 - accuracy: 0.9062 - val_loss: 0.2501 - val_accuracy: 0.8982 - lr: 1.0461e-05\n",
      "Epoch 27/50\n",
      "2359/2360 [============================>.] - ETA: 0s - loss: 0.2315 - accuracy: 0.9058\n",
      "Epoch 27: val_loss improved from 0.25011 to 0.24907, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 53s 22ms/step - loss: 0.2314 - accuracy: 0.9058 - val_loss: 0.2491 - val_accuracy: 0.8997 - lr: 1.0369e-05\n",
      "Epoch 28/50\n",
      "2357/2360 [============================>.] - ETA: 0s - loss: 0.2303 - accuracy: 0.9071\n",
      "Epoch 28: val_loss improved from 0.24907 to 0.24833, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 50s 21ms/step - loss: 0.2302 - accuracy: 0.9072 - val_loss: 0.2483 - val_accuracy: 0.8997 - lr: 1.0295e-05\n",
      "Epoch 29/50\n",
      "2359/2360 [============================>.] - ETA: 0s - loss: 0.2289 - accuracy: 0.9074\n",
      "Epoch 29: val_loss improved from 0.24833 to 0.24743, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 50s 21ms/step - loss: 0.2289 - accuracy: 0.9074 - val_loss: 0.2474 - val_accuracy: 0.9002 - lr: 1.0236e-05\n",
      "Epoch 30/50\n",
      "2357/2360 [============================>.] - ETA: 0s - loss: 0.2277 - accuracy: 0.9078\n",
      "Epoch 30: val_loss improved from 0.24743 to 0.24628, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 58s 24ms/step - loss: 0.2278 - accuracy: 0.9078 - val_loss: 0.2463 - val_accuracy: 0.9017 - lr: 1.0189e-05\n",
      "Epoch 31/50\n",
      "2358/2360 [============================>.] - ETA: 0s - loss: 0.2266 - accuracy: 0.9086\n",
      "Epoch 31: val_loss improved from 0.24628 to 0.24540, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 59s 25ms/step - loss: 0.2266 - accuracy: 0.9086 - val_loss: 0.2454 - val_accuracy: 0.9017 - lr: 1.0151e-05\n",
      "Epoch 32/50\n",
      "2360/2360 [==============================] - ETA: 0s - loss: 0.2256 - accuracy: 0.9100\n",
      "Epoch 32: val_loss improved from 0.24540 to 0.24451, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 55s 23ms/step - loss: 0.2256 - accuracy: 0.9100 - val_loss: 0.2445 - val_accuracy: 0.9019 - lr: 1.0121e-05\n",
      "Epoch 33/50\n",
      "2358/2360 [============================>.] - ETA: 0s - loss: 0.2246 - accuracy: 0.9100\n",
      "Epoch 33: val_loss improved from 0.24451 to 0.24382, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 54s 23ms/step - loss: 0.2245 - accuracy: 0.9100 - val_loss: 0.2438 - val_accuracy: 0.9024 - lr: 1.0097e-05\n",
      "Epoch 34/50\n",
      "2358/2360 [============================>.] - ETA: 0s - loss: 0.2235 - accuracy: 0.9086\n",
      "Epoch 34: val_loss improved from 0.24382 to 0.24295, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 55s 23ms/step - loss: 0.2234 - accuracy: 0.9087 - val_loss: 0.2430 - val_accuracy: 0.9026 - lr: 1.0077e-05\n",
      "Epoch 35/50\n",
      "2360/2360 [==============================] - ETA: 0s - loss: 0.2225 - accuracy: 0.9103\n",
      "Epoch 35: val_loss improved from 0.24295 to 0.24216, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 55s 23ms/step - loss: 0.2225 - accuracy: 0.9103 - val_loss: 0.2422 - val_accuracy: 0.9039 - lr: 1.0062e-05\n",
      "Epoch 36/50\n",
      "2360/2360 [==============================] - ETA: 0s - loss: 0.2214 - accuracy: 0.9112\n",
      "Epoch 36: val_loss improved from 0.24216 to 0.24146, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 53s 23ms/step - loss: 0.2214 - accuracy: 0.9112 - val_loss: 0.2415 - val_accuracy: 0.9041 - lr: 1.0050e-05\n",
      "Epoch 37/50\n",
      "2360/2360 [==============================] - ETA: 0s - loss: 0.2206 - accuracy: 0.9108\n",
      "Epoch 37: val_loss improved from 0.24146 to 0.24058, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 54s 23ms/step - loss: 0.2206 - accuracy: 0.9108 - val_loss: 0.2406 - val_accuracy: 0.9034 - lr: 1.0040e-05\n",
      "Epoch 38/50\n",
      "2358/2360 [============================>.] - ETA: 0s - loss: 0.2197 - accuracy: 0.9112\n",
      "Epoch 38: val_loss improved from 0.24058 to 0.23983, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 56s 24ms/step - loss: 0.2196 - accuracy: 0.9112 - val_loss: 0.2398 - val_accuracy: 0.9039 - lr: 1.0032e-05\n",
      "Epoch 39/50\n",
      "2359/2360 [============================>.] - ETA: 0s - loss: 0.2184 - accuracy: 0.9108\n",
      "Epoch 39: val_loss improved from 0.23983 to 0.23912, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 55s 23ms/step - loss: 0.2185 - accuracy: 0.9107 - val_loss: 0.2391 - val_accuracy: 0.9044 - lr: 1.0025e-05\n",
      "Epoch 40/50\n",
      "2359/2360 [============================>.] - ETA: 0s - loss: 0.2178 - accuracy: 0.9126\n",
      "Epoch 40: val_loss improved from 0.23912 to 0.23848, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 57s 24ms/step - loss: 0.2177 - accuracy: 0.9126 - val_loss: 0.2385 - val_accuracy: 0.9049 - lr: 1.0020e-05\n",
      "Epoch 41/50\n",
      "2360/2360 [==============================] - ETA: 0s - loss: 0.2169 - accuracy: 0.9127\n",
      "Epoch 41: val_loss improved from 0.23848 to 0.23778, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 53s 22ms/step - loss: 0.2169 - accuracy: 0.9127 - val_loss: 0.2378 - val_accuracy: 0.9054 - lr: 1.0016e-05\n",
      "Epoch 42/50\n",
      "2360/2360 [==============================] - ETA: 0s - loss: 0.2155 - accuracy: 0.9124\n",
      "Epoch 42: val_loss improved from 0.23778 to 0.23773, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 52s 22ms/step - loss: 0.2155 - accuracy: 0.9124 - val_loss: 0.2377 - val_accuracy: 0.9036 - lr: 1.0013e-05\n",
      "Epoch 43/50\n",
      "2359/2360 [============================>.] - ETA: 0s - loss: 0.2149 - accuracy: 0.9130\n",
      "Epoch 43: val_loss improved from 0.23773 to 0.23638, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 53s 22ms/step - loss: 0.2150 - accuracy: 0.9129 - val_loss: 0.2364 - val_accuracy: 0.9046 - lr: 1.0010e-05\n",
      "Epoch 44/50\n",
      "2357/2360 [============================>.] - ETA: 0s - loss: 0.2144 - accuracy: 0.9122\n",
      "Epoch 44: val_loss improved from 0.23638 to 0.23571, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 52s 22ms/step - loss: 0.2142 - accuracy: 0.9123 - val_loss: 0.2357 - val_accuracy: 0.9039 - lr: 1.0008e-05\n",
      "Epoch 45/50\n",
      "2358/2360 [============================>.] - ETA: 0s - loss: 0.2134 - accuracy: 0.9151\n",
      "Epoch 45: val_loss improved from 0.23571 to 0.23515, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 52s 22ms/step - loss: 0.2134 - accuracy: 0.9150 - val_loss: 0.2351 - val_accuracy: 0.9049 - lr: 1.0007e-05\n",
      "Epoch 46/50\n",
      "2357/2360 [============================>.] - ETA: 0s - loss: 0.2125 - accuracy: 0.9138\n",
      "Epoch 46: val_loss improved from 0.23515 to 0.23446, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 52s 22ms/step - loss: 0.2127 - accuracy: 0.9137 - val_loss: 0.2345 - val_accuracy: 0.9044 - lr: 1.0005e-05\n",
      "Epoch 47/50\n",
      "2357/2360 [============================>.] - ETA: 0s - loss: 0.2112 - accuracy: 0.9143\n",
      "Epoch 47: val_loss improved from 0.23446 to 0.23406, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 52s 22ms/step - loss: 0.2118 - accuracy: 0.9141 - val_loss: 0.2341 - val_accuracy: 0.9064 - lr: 1.0004e-05\n",
      "Epoch 48/50\n",
      "2359/2360 [============================>.] - ETA: 0s - loss: 0.2111 - accuracy: 0.9151\n",
      "Epoch 48: val_loss improved from 0.23406 to 0.23343, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 54s 23ms/step - loss: 0.2111 - accuracy: 0.9151 - val_loss: 0.2334 - val_accuracy: 0.9059 - lr: 1.0003e-05\n",
      "Epoch 49/50\n",
      "2357/2360 [============================>.] - ETA: 0s - loss: 0.2103 - accuracy: 0.9166\n",
      "Epoch 49: val_loss improved from 0.23343 to 0.23274, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 56s 24ms/step - loss: 0.2103 - accuracy: 0.9165 - val_loss: 0.2327 - val_accuracy: 0.9059 - lr: 1.0003e-05\n",
      "Epoch 50/50\n",
      "2357/2360 [============================>.] - ETA: 0s - loss: 0.2095 - accuracy: 0.9159\n",
      "Epoch 50: val_loss improved from 0.23274 to 0.23236, saving model to ModelWeights.h5\n",
      "2360/2360 [==============================] - 55s 23ms/step - loss: 0.2095 - accuracy: 0.9159 - val_loss: 0.2324 - val_accuracy: 0.9054 - lr: 1.0002e-05\n",
      "\n",
      "Restoring best Weights for MobileNetV2\n"
     ]
    }
   ],
   "source": [
    "print('Training head...')\n",
    "#model.load_weights('./Model_Weights.h5')\n",
    "\n",
    "history = model.fit(X_train_nn ,y_train, epochs=epochs,\n",
    "                        callbacks=callbacks,\n",
    "                        validation_data = (X_test_nn, y_test),\n",
    "                        batch_size=batch_size)\n",
    "\n",
    "print('\\nRestoring best Weights for MobileNetV2')\n",
    "model.load_weights(checkpoint_filepath)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "def print_graph(item, index, history):\n",
    "    plt.figure()\n",
    "    train_values = history.history[item][0:index]\n",
    "    plt.plot(train_values)\n",
    "    test_values = history.history['val_' + item][0:index]\n",
    "    plt.plot(test_values)\n",
    "    plt.legend(['training','validation'])\n",
    "    plt.title('Training and validation '+ item)\n",
    "    plt.xlabel('epoch')\n",
    "    plt.show()\n",
    "    plot = '{}.png'.format(item)\n",
    "    plt.savefig(plot)\n",
    "\n",
    "\n",
    "def get_best_epoch(test_loss, history):\n",
    "    for key, item in enumerate(history.history.items()):\n",
    "        (name, arr) = item\n",
    "        if name == 'val_loss':\n",
    "            for i in range(len(arr)):\n",
    "                if round(test_loss, 2) == round(arr[i], 2):\n",
    "                    return i\n",
    "                \n",
    "def model_summary(model, history):\n",
    "    print('---'*30)\n",
    "    test_loss, test_accuracy = model.evaluate(X_test_nn, y_test, verbose=0)\n",
    "\n",
    "    if history:\n",
    "        index = get_best_epoch(test_loss, history)\n",
    "        print('Best Epochs: ', index)\n",
    "\n",
    "        train_accuracy = history.history['accuracy'][index]\n",
    "        train_loss = history.history['loss'][index]\n",
    "\n",
    "        print('Accuracy on train:',train_accuracy,'\\tLoss on train:',train_loss)\n",
    "        print('Accuracy on test:',test_accuracy,'\\tLoss on test:',test_loss)\n",
    "        print_graph('loss', index, history)\n",
    "        print_graph('accuracy', index, history)\n",
    "        print('---'*30)   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "------------------------------------------------------------------------------------------\n",
      "Best Epochs:  12\n",
      "Accuracy on train: 0.9171609878540039 \tLoss on train: 0.21271847188472748\n",
      "Accuracy on test: 0.9103039503097534 \tLoss on test: 0.22530776262283325\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 640x480 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "------------------------------------------------------------------------------------------\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 640x480 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model_summary(model, history)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Evaluation on test dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Evaluating network...\n",
      "127/127 [==============================] - 10s 79ms/step\n"
     ]
    }
   ],
   "source": [
    "# evaluate the network\n",
    "print(\"Evaluating network...\")\n",
    "predictions = model.predict(X_test_nn)\n",
    "preds = predictions > 0.5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from scikitplot.metrics import plot_roc_curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.metrics import roc_curve, roc_auc_score\n",
    "\n",
    "# Example predicted probabilities and true labels\n",
    "predicted_probabilities = predictions\n",
    "true_labels =y_test\n",
    "\n",
    "# Compute the false positive rate (FPR), true positive rate (TPR), and thresholds\n",
    "fpr, tpr, thresholds = roc_curve(true_labels, predicted_probabilities)\n",
    "\n",
    "# Calculate the Area Under the Curve (AUC)\n",
    "auc = roc_auc_score(true_labels, predicted_probabilities)\n",
    "\n",
    "# Plot the ROC curve\n",
    "plt.plot(fpr, tpr, label='ROC curve (AUC = {:.2f})'.format(auc))\n",
    "plt.plot([0, 1], [0, 1], linestyle='--', color='r', label='Random Classifier')\n",
    "plt.plot([0, 0, 1], [0, 1, 1], linestyle=':', color='g', label='Perfect Classifier')\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('Receiver Operating Characteristic (ROC) Curve')\n",
    "plt.legend()\n",
    "plt.show()\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "args_model = \"modelnew.h5\"\n",
    "model.save(args_model)"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "aee8b7b246df8f9039afb4144a1f6fd8d2ca17a180786b69acc140d282b71a49"
  },
  "kernelspec": {
   "display_name": "Python 3.10.1 64-bit",
   "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": "3.10.1"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}