nc-severstal03-classification-densenet-3.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2021-03-19T05:29:11.479513Z",
     "iopub.status.busy": "2021-03-19T05:29:11.478888Z",
     "iopub.status.idle": "2021-03-19T05:29:17.680661Z",
     "shell.execute_reply": "2021-03-19T05:29:17.681768Z"
    },
    "papermill": {
     "duration": 6.225143,
     "end_time": "2021-03-19T05:29:17.682176",
     "exception": false,
     "start_time": "2021-03-19T05:29:11.457033",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "import os\n",
    "import cv2\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras import layers\n",
    "from tensorflow.keras.callbacks import Callback, ModelCheckpoint, ReduceLROnPlateau,EarlyStopping\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
    "from tensorflow.keras.models import Sequential, Model, load_model\n",
    "from tensorflow.keras.layers import Input, Dense, Conv2D, MaxPooling2D\n",
    "from tensorflow.keras.layers import GlobalAveragePooling2D, AveragePooling2D\n",
    "from tensorflow.keras.layers import ZeroPadding2D, Dropout, Flatten, add, Reshape\n",
    "from tensorflow.keras.layers import Activation, BatchNormalization\n",
    "from tensorflow.keras.optimizers import Adam, SGD, Nadam\n",
    "from tensorflow.keras.utils import plot_model,to_categorical\n",
    "from tensorflow.keras import backend as K\n",
    "from tensorflow.keras import initializers\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
    "from tensorflow.keras.applications import DenseNet121\n",
    "from tqdm import tqdm\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import confusion_matrix, classification_report\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.021236,
     "end_time": "2021-03-19T05:29:17.726902",
     "exception": false,
     "start_time": "2021-03-19T05:29:17.705666",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## Read train and test files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2021-03-19T05:29:17.782904Z",
     "iopub.status.busy": "2021-03-19T05:29:17.778402Z",
     "iopub.status.idle": "2021-03-19T05:29:18.338652Z",
     "shell.execute_reply": "2021-03-19T05:29:18.339640Z"
    },
    "papermill": {
     "duration": 0.590672,
     "end_time": "2021-03-19T05:29:18.339837",
     "exception": false,
     "start_time": "2021-03-19T05:29:17.749165",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(7095, 3)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ImageId</th>\n",
       "      <th>ClassId</th>\n",
       "      <th>EncodedPixels</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0002cc93b.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>29102 12 29346 24 29602 24 29858 24 30114 24 3...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0007a71bf.jpg</td>\n",
       "      <td>3</td>\n",
       "      <td>18661 28 18863 82 19091 110 19347 110 19603 11...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>000a4bcdd.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>37607 3 37858 8 38108 14 38359 20 38610 25 388...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>000f6bf48.jpg</td>\n",
       "      <td>4</td>\n",
       "      <td>131973 1 132228 4 132483 6 132738 8 132993 11 ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0014fce06.jpg</td>\n",
       "      <td>3</td>\n",
       "      <td>229501 11 229741 33 229981 55 230221 77 230468...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         ImageId  ClassId                                      EncodedPixels\n",
       "0  0002cc93b.jpg        1  29102 12 29346 24 29602 24 29858 24 30114 24 3...\n",
       "1  0007a71bf.jpg        3  18661 28 18863 82 19091 110 19347 110 19603 11...\n",
       "2  000a4bcdd.jpg        1  37607 3 37858 8 38108 14 38359 20 38610 25 388...\n",
       "3  000f6bf48.jpg        4  131973 1 132228 4 132483 6 132738 8 132993 11 ...\n",
       "4  0014fce06.jpg        3  229501 11 229741 33 229981 55 230221 77 230468..."
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df = pd.read_csv('../input/severstal-steel-defect-detection/train.csv')\n",
    "\n",
    "print(train_df.shape)\n",
    "train_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2021-03-19T05:29:18.388998Z",
     "iopub.status.busy": "2021-03-19T05:29:18.388236Z",
     "iopub.status.idle": "2021-03-19T05:29:18.416578Z",
     "shell.execute_reply": "2021-03-19T05:29:18.417558Z"
    },
    "papermill": {
     "duration": 0.056333,
     "end_time": "2021-03-19T05:29:18.417744",
     "exception": false,
     "start_time": "2021-03-19T05:29:18.361411",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(5506, 3)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ImageId</th>\n",
       "      <th>EncodedPixels</th>\n",
       "      <th>ClassId</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0000f269f.jpg</td>\n",
       "      <td>1 409600</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>000ccc2ac.jpg</td>\n",
       "      <td>1 409600</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>002451917.jpg</td>\n",
       "      <td>1 409600</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>003c5da97.jpg</td>\n",
       "      <td>1 409600</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0042e163f.jpg</td>\n",
       "      <td>1 409600</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         ImageId EncodedPixels  ClassId\n",
       "0  0000f269f.jpg      1 409600        0\n",
       "1  000ccc2ac.jpg      1 409600        0\n",
       "2  002451917.jpg      1 409600        0\n",
       "3  003c5da97.jpg      1 409600        0\n",
       "4  0042e163f.jpg      1 409600        0"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "submission_df = pd.read_csv('../input/severstal-steel-defect-detection/sample_submission.csv')\n",
    "print(submission_df.shape)\n",
    "submission_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2021-03-19T05:29:18.472375Z",
     "iopub.status.busy": "2021-03-19T05:29:18.471634Z",
     "iopub.status.idle": "2021-03-19T05:29:18.487791Z",
     "shell.execute_reply": "2021-03-19T05:29:18.488729Z"
    },
    "papermill": {
     "duration": 0.048676,
     "end_time": "2021-03-19T05:29:18.488913",
     "exception": false,
     "start_time": "2021-03-19T05:29:18.440237",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(5506, 3)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ImageId</th>\n",
       "      <th>EncodedPixels</th>\n",
       "      <th>ClassId</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0000f269f.jpg</td>\n",
       "      <td>1 409600</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>000ccc2ac.jpg</td>\n",
       "      <td>1 409600</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>002451917.jpg</td>\n",
       "      <td>1 409600</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>003c5da97.jpg</td>\n",
       "      <td>1 409600</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0042e163f.jpg</td>\n",
       "      <td>1 409600</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         ImageId EncodedPixels  ClassId\n",
       "0  0000f269f.jpg      1 409600        0\n",
       "1  000ccc2ac.jpg      1 409600        0\n",
       "2  002451917.jpg      1 409600        0\n",
       "3  003c5da97.jpg      1 409600        0\n",
       "4  0042e163f.jpg      1 409600        0"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "submission_df = pd.read_csv('../input/severstal-steel-defect-detection/sample_submission.csv')\n",
    "print(submission_df.shape)\n",
    "submission_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.023111,
     "end_time": "2021-03-19T05:29:18.535495",
     "exception": false,
     "start_time": "2021-03-19T05:29:18.512384",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## Combine defected and no-defect images"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2021-03-19T05:29:18.590395Z",
     "iopub.status.busy": "2021-03-19T05:29:18.589660Z",
     "iopub.status.idle": "2021-03-19T05:29:20.673645Z",
     "shell.execute_reply": "2021-03-19T05:29:20.674056Z"
    },
    "papermill": {
     "duration": 2.115413,
     "end_time": "2021-03-19T05:29:20.674199",
     "exception": false,
     "start_time": "2021-03-19T05:29:18.558786",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(12568, 2)\n",
      "1    6666\n",
      "0    5902\n",
      "Name: defect_label, dtype: int64\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ImageId</th>\n",
       "      <th>defect_label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0002cc93b.jpg</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         ImageId  defect_label\n",
       "0  0002cc93b.jpg             1"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Data Pre-Processing\n",
    "TRAIN_PATH = '../input/severstal-steel-defect-detection/train_images/'\n",
    "TEST_PATH = '../input/severstal-steel-defect-detection/test_images/'\n",
    "\n",
    "train_flist = os.listdir(TRAIN_PATH) \n",
    "\n",
    "image_defects = train_df['ImageId'].unique().tolist()\n",
    "image_no_defects=[item for item in train_flist if item not in image_defects]\n",
    "\n",
    "#df_bin is the dataframe with label defect based on train_df \n",
    "df_bin = train_df.drop_duplicates(subset=['ImageId'], keep='last').copy()\n",
    "df_bin.loc[:, 'defect_label'] = 1\n",
    "\n",
    "df_noDefect_bin = pd.DataFrame(image_no_defects,columns=['ImageId'])\n",
    "df_noDefect_bin.loc[:, 'defect_label'] = 0\n",
    "\n",
    "frames = [df_bin[['ImageId', \"defect_label\"]], df_noDefect_bin]\n",
    "df = pd.concat(frames).reset_index(drop=True)\n",
    "print(df.shape)\n",
    "print(df['defect_label'].value_counts())\n",
    "df.head(1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.016067,
     "end_time": "2021-03-19T05:29:20.705831",
     "exception": false,
     "start_time": "2021-03-19T05:29:20.689764",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## Split train data, load and resave images"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2021-03-19T05:29:20.741385Z",
     "iopub.status.busy": "2021-03-19T05:29:20.740884Z",
     "iopub.status.idle": "2021-03-19T05:29:20.763608Z",
     "shell.execute_reply": "2021-03-19T05:29:20.763188Z"
    },
    "papermill": {
     "duration": 0.042768,
     "end_time": "2021-03-19T05:29:20.763733",
     "exception": false,
     "start_time": "2021-03-19T05:29:20.720965",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "nb_images = df.shape[0]\n",
    "df_train, df_test = train_test_split(df, test_size=0.2, \n",
    "                                                    stratify = df['defect_label'], \n",
    "                                                    random_state=2021)\n",
    "\n",
    "df_train, df_val = train_test_split(df_train, test_size=0.25, \n",
    "                                                  stratify = df_train['defect_label'],\n",
    "                                                  random_state=2021) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2021-03-19T05:29:20.802670Z",
     "iopub.status.busy": "2021-03-19T05:29:20.801858Z",
     "iopub.status.idle": "2021-03-19T05:29:20.806467Z",
     "shell.execute_reply": "2021-03-19T05:29:20.805984Z"
    },
    "papermill": {
     "duration": 0.027182,
     "end_time": "2021-03-19T05:29:20.806579",
     "exception": false,
     "start_time": "2021-03-19T05:29:20.779397",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# 0.25 x 0.8 = 0.2\n",
    "\n",
    "IMAGE_SIZE = 256\n",
    "\n",
    "def load_img(code, base, resize=True):\n",
    "    path = f'{base}/{code}'\n",
    "    img = cv2.imread(path)\n",
    "    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)\n",
    "    if resize:\n",
    "        img = cv2.resize(img, (IMAGE_SIZE, IMAGE_SIZE))\n",
    "    return img\n",
    "\n",
    "def validate_path(path):\n",
    "    if not os.path.exists(path):\n",
    "        os.makedirs(path)\n",
    "\n",
    "train_path = '../tmp/train'\n",
    "validate_path(train_path)\n",
    "val_path = '../tmp/val'\n",
    "validate_path(val_path)\n",
    "test_path = '../tmp/test'\n",
    "validate_path(test_path)\n",
    "\n",
    "def process_img(df, train_path):\n",
    "    #read image, resize and save\n",
    "    for code in tqdm(df['ImageId']):\n",
    "        img = load_img(\n",
    "            code,\n",
    "            base='../input/severstal-steel-defect-detection/train_images'\n",
    "        )\n",
    "        path = code.replace('.jpg', '')\n",
    "        cv2.imwrite(f'{train_path}/{path}.png', img)\n",
    "\n",
    "    df['ImageId'] = df['ImageId'].apply(lambda x: x.replace('.jpg', '.png'))\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2021-03-19T05:29:20.842598Z",
     "iopub.status.busy": "2021-03-19T05:29:20.841893Z",
     "iopub.status.idle": "2021-03-19T05:34:34.713727Z",
     "shell.execute_reply": "2021-03-19T05:34:34.714210Z"
    },
    "papermill": {
     "duration": 313.89222,
     "end_time": "2021-03-19T05:34:34.714388",
     "exception": false,
     "start_time": "2021-03-19T05:29:20.822168",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 7540/7540 [03:11<00:00, 39.28it/s]\n",
      "100%|██████████| 2514/2514 [01:01<00:00, 41.08it/s]\n",
      "100%|██████████| 2514/2514 [01:00<00:00, 41.42it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(7540, 2) (2514, 2) (2514, 2)\n",
      "1    3999\n",
      "0    3541\n",
      "Name: defect_label, dtype: int64\n",
      "1    1334\n",
      "0    1180\n",
      "Name: defect_label, dtype: int64\n",
      "1    1333\n",
      "0    1181\n",
      "Name: defect_label, dtype: int64\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "/opt/conda/lib/python3.7/site-packages/ipykernel_launcher.py:34: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    }
   ],
   "source": [
    "df_train = process_img(df_train,train_path)\n",
    "df_val = process_img(df_val,val_path)\n",
    "df_test = process_img(df_test,test_path)\n",
    "\n",
    "print(df_train.shape,df_val.shape,df_test.shape)\n",
    "print(df_train['defect_label'].value_counts())\n",
    "print(df_val['defect_label'].value_counts())\n",
    "print(df_test['defect_label'].value_counts())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.981318,
     "end_time": "2021-03-19T05:34:36.592512",
     "exception": false,
     "start_time": "2021-03-19T05:34:35.611194",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## Define ImageDataGenerator and data augmentation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2021-03-19T05:34:37.992858Z",
     "iopub.status.busy": "2021-03-19T05:34:37.991680Z",
     "iopub.status.idle": "2021-03-19T05:34:50.329235Z",
     "shell.execute_reply": "2021-03-19T05:34:50.329835Z"
    },
    "papermill": {
     "duration": 13.04498,
     "end_time": "2021-03-19T05:34:50.330077",
     "exception": false,
     "start_time": "2021-03-19T05:34:37.285097",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 7540 validated image filenames.\n",
      "Found 2514 validated image filenames.\n",
      "Found 2514 validated image filenames.\n",
      "Found 5506 validated image filenames.\n"
     ]
    }
   ],
   "source": [
    "BATCH_SIZE = 32\n",
    "\n",
    "# Using original generator\n",
    "data_generator = ImageDataGenerator(\n",
    "        zoom_range=0.1,  # set range for random zoom\n",
    "        # set mode for filling points outside the input boundaries\n",
    "        fill_mode='constant',\n",
    "        cval=0.,\n",
    "        rotation_range=10,\n",
    "        height_shift_range=0.05,\n",
    "        width_shift_range=0.05,\n",
    "        horizontal_flip=True,\n",
    "        vertical_flip=True,\n",
    "        rescale=1/255.,\n",
    "    )\n",
    "val_generator = ImageDataGenerator(\n",
    "        rescale=1/255.\n",
    "    )\n",
    "\n",
    "train_gen = data_generator.flow_from_dataframe(\n",
    "        df_train, \n",
    "        directory='../tmp/train',\n",
    "        x_col='ImageId', \n",
    "        y_col='defect_label', \n",
    "        class_mode='raw',\n",
    "        target_size=(IMAGE_SIZE, IMAGE_SIZE),\n",
    "        batch_size=BATCH_SIZE,\n",
    "        )\n",
    "\n",
    "val_gen =  val_generator.flow_from_dataframe(\n",
    "        df_val, \n",
    "        directory='../tmp/val',\n",
    "        x_col='ImageId', \n",
    "        y_col='defect_label', \n",
    "        class_mode='raw',\n",
    "        target_size=(IMAGE_SIZE, IMAGE_SIZE),\n",
    "        batch_size=BATCH_SIZE,\n",
    "        )\n",
    "test_gen =  val_generator.flow_from_dataframe(\n",
    "        df_test, \n",
    "        directory='../tmp/test',\n",
    "        x_col='ImageId', \n",
    "        y_col='defect_label', \n",
    "        class_mode='raw',\n",
    "        target_size=(IMAGE_SIZE, IMAGE_SIZE),\n",
    "        batch_size=BATCH_SIZE,\n",
    "        shuffle=False\n",
    "        )\n",
    "\n",
    "submission_gen = ImageDataGenerator(rescale=1/255.).flow_from_dataframe(\n",
    "        submission_df,\n",
    "        directory='../input/severstal-steel-defect-detection/test_images/',\n",
    "        x_col='ImageId',\n",
    "        class_mode=None,\n",
    "        target_size=(64, 64),\n",
    "        batch_size=BATCH_SIZE,\n",
    "        shuffle=False\n",
    "    )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.676991,
     "end_time": "2021-03-19T05:34:51.725114",
     "exception": false,
     "start_time": "2021-03-19T05:34:51.048123",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## Define Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2021-03-19T05:34:53.095649Z",
     "iopub.status.busy": "2021-03-19T05:34:53.095119Z",
     "iopub.status.idle": "2021-03-19T05:35:00.989743Z",
     "shell.execute_reply": "2021-03-19T05:35:00.989225Z"
    },
    "papermill": {
     "duration": 8.583678,
     "end_time": "2021-03-19T05:35:00.989908",
     "exception": false,
     "start_time": "2021-03-19T05:34:52.406230",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "densenet121 (Functional)     (None, 8, 8, 1024)        7037504   \n",
      "_________________________________________________________________\n",
      "global_average_pooling2d (Gl (None, 1024)              0         \n",
      "_________________________________________________________________\n",
      "batch_normalization (BatchNo (None, 1024)              4096      \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 1024)              0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 512)               524800    \n",
      "_________________________________________________________________\n",
      "batch_normalization_1 (Batch (None, 512)               2048      \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 1)                 513       \n",
      "=================================================================\n",
      "Total params: 7,568,961\n",
      "Trainable params: 7,482,241\n",
      "Non-trainable params: 86,720\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "input_shape = (IMAGE_SIZE,IMAGE_SIZE,3)\n",
    "\n",
    "def build_model():\n",
    "    \n",
    "    densenet = DenseNet121(\n",
    "        include_top=False,\n",
    "        input_shape=(256,256,3),\n",
    "        weights='../input/keras-densenet/DenseNet-BC-121-32-no-top.h5'\n",
    "    )\n",
    "    \n",
    "    model = Sequential()\n",
    "    model.add(densenet)\n",
    "    model.add(layers.GlobalAveragePooling2D())\n",
    "    model.add(layers.BatchNormalization())\n",
    "    model.add(layers.Dropout(0.5))\n",
    "    model.add(layers.Dense(512, activation='relu'))\n",
    "    model.add(layers.BatchNormalization())\n",
    "    model.add(layers.Dropout(0.5))\n",
    "    model.add(layers.Dense(1, activation='sigmoid'))\n",
    "    \n",
    "    model.compile(\n",
    "        loss='binary_crossentropy',\n",
    "        optimizer=Nadam(),\n",
    "        metrics=['accuracy']    )\n",
    "    \n",
    "    return model\n",
    "model = build_model()\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.669196,
     "end_time": "2021-03-19T05:35:02.387405",
     "exception": false,
     "start_time": "2021-03-19T05:35:01.718209",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## Model training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2021-03-19T05:35:03.769609Z",
     "iopub.status.busy": "2021-03-19T05:35:03.767876Z",
     "iopub.status.idle": "2021-03-19T05:35:03.770200Z",
     "shell.execute_reply": "2021-03-19T05:35:03.770588Z"
    },
    "papermill": {
     "duration": 0.682451,
     "end_time": "2021-03-19T05:35:03.770717",
     "exception": false,
     "start_time": "2021-03-19T05:35:03.088266",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "epochs = 100\n",
    "\n",
    "import math\n",
    "# We take the ceiling because we do not drop the remainder of the batch\n",
    "compute_steps_per_epoch = lambda x: int(math.ceil(1. * x / BATCH_SIZE))\n",
    "\n",
    "steps_per_epoch = compute_steps_per_epoch(df_train.shape[0])\n",
    "validation_steps = compute_steps_per_epoch(df_val.shape[0])\n",
    "\n",
    "es = EarlyStopping(monitor='val_loss', mode='min', verbose=2, patience=36)\n",
    "reduce_lr = ReduceLROnPlateau(monitor='val_loss', patience=15, verbose=1, min_lr=1e-6)\n",
    "\n",
    "mc = ModelCheckpoint(\n",
    "    'model.h5', \n",
    "    monitor='val_accuracy', \n",
    "    verbose=1, \n",
    "    save_best_only=True, \n",
    "    save_weights_only=False,\n",
    "    mode='auto'\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2021-03-19T05:35:05.162215Z",
     "iopub.status.busy": "2021-03-19T05:35:05.161541Z",
     "iopub.status.idle": "2021-03-19T09:34:43.894414Z",
     "shell.execute_reply": "2021-03-19T09:34:43.893716Z"
    },
    "papermill": {
     "duration": 14379.411275,
     "end_time": "2021-03-19T09:34:43.894599",
     "exception": false,
     "start_time": "2021-03-19T05:35:04.483324",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "236/236 [==============================] - 181s 630ms/step - loss: 0.7063 - accuracy: 0.6956 - val_loss: 0.9273 - val_accuracy: 0.6874\n",
      "\n",
      "Epoch 00001: val_accuracy improved from -inf to 0.68735, saving model to model.h5\n",
      "Epoch 2/100\n",
      "236/236 [==============================] - 146s 617ms/step - loss: 0.3773 - accuracy: 0.8229 - val_loss: 0.3437 - val_accuracy: 0.8604\n",
      "\n",
      "Epoch 00002: val_accuracy improved from 0.68735 to 0.86038, saving model to model.h5\n",
      "Epoch 3/100\n",
      "236/236 [==============================] - 145s 615ms/step - loss: 0.3294 - accuracy: 0.8564 - val_loss: 0.3078 - val_accuracy: 0.8747\n",
      "\n",
      "Epoch 00003: val_accuracy improved from 0.86038 to 0.87470, saving model to model.h5\n",
      "Epoch 4/100\n",
      "236/236 [==============================] - 146s 617ms/step - loss: 0.2649 - accuracy: 0.8911 - val_loss: 2.2885 - val_accuracy: 0.5084\n",
      "\n",
      "Epoch 00004: val_accuracy did not improve from 0.87470\n",
      "Epoch 5/100\n",
      "236/236 [==============================] - 147s 621ms/step - loss: 0.2410 - accuracy: 0.8955 - val_loss: 0.9822 - val_accuracy: 0.5959\n",
      "\n",
      "Epoch 00005: val_accuracy did not improve from 0.87470\n",
      "Epoch 6/100\n",
      "236/236 [==============================] - 146s 618ms/step - loss: 0.2375 - accuracy: 0.9015 - val_loss: 0.2389 - val_accuracy: 0.9037\n",
      "\n",
      "Epoch 00006: val_accuracy improved from 0.87470 to 0.90374, saving model to model.h5\n",
      "Epoch 7/100\n",
      "236/236 [==============================] - 146s 617ms/step - loss: 0.2369 - accuracy: 0.8994 - val_loss: 0.5154 - val_accuracy: 0.8099\n",
      "\n",
      "Epoch 00007: val_accuracy did not improve from 0.90374\n",
      "Epoch 8/100\n",
      "236/236 [==============================] - 146s 617ms/step - loss: 0.2238 - accuracy: 0.9129 - val_loss: 0.2545 - val_accuracy: 0.8962\n",
      "\n",
      "Epoch 00008: val_accuracy did not improve from 0.90374\n",
      "Epoch 9/100\n",
      "236/236 [==============================] - 147s 619ms/step - loss: 0.2268 - accuracy: 0.9124 - val_loss: 0.2427 - val_accuracy: 0.9037\n",
      "\n",
      "Epoch 00009: val_accuracy did not improve from 0.90374\n",
      "Epoch 10/100\n",
      "236/236 [==============================] - 147s 620ms/step - loss: 0.2249 - accuracy: 0.9121 - val_loss: 1.4382 - val_accuracy: 0.7554\n",
      "\n",
      "Epoch 00010: val_accuracy did not improve from 0.90374\n",
      "Epoch 11/100\n",
      "236/236 [==============================] - 147s 621ms/step - loss: 0.2242 - accuracy: 0.9135 - val_loss: 2.0215 - val_accuracy: 0.5426\n",
      "\n",
      "Epoch 00011: val_accuracy did not improve from 0.90374\n",
      "Epoch 12/100\n",
      "236/236 [==============================] - 146s 620ms/step - loss: 0.2082 - accuracy: 0.9140 - val_loss: 2.2203 - val_accuracy: 0.5008\n",
      "\n",
      "Epoch 00012: val_accuracy did not improve from 0.90374\n",
      "Epoch 13/100\n",
      "236/236 [==============================] - 147s 624ms/step - loss: 0.1953 - accuracy: 0.9201 - val_loss: 0.4637 - val_accuracy: 0.8198\n",
      "\n",
      "Epoch 00013: val_accuracy did not improve from 0.90374\n",
      "Epoch 14/100\n",
      "236/236 [==============================] - 146s 620ms/step - loss: 0.1879 - accuracy: 0.9220 - val_loss: 0.2241 - val_accuracy: 0.9097\n",
      "\n",
      "Epoch 00014: val_accuracy improved from 0.90374 to 0.90971, saving model to model.h5\n",
      "Epoch 15/100\n",
      "236/236 [==============================] - 148s 628ms/step - loss: 0.2004 - accuracy: 0.9175 - val_loss: 2.0206 - val_accuracy: 0.5843\n",
      "\n",
      "Epoch 00015: val_accuracy did not improve from 0.90971\n",
      "Epoch 16/100\n",
      "236/236 [==============================] - 147s 622ms/step - loss: 0.1802 - accuracy: 0.9267 - val_loss: 0.2701 - val_accuracy: 0.8982\n",
      "\n",
      "Epoch 00016: val_accuracy did not improve from 0.90971\n",
      "Epoch 17/100\n",
      "236/236 [==============================] - 147s 621ms/step - loss: 0.1873 - accuracy: 0.9249 - val_loss: 0.5918 - val_accuracy: 0.8111\n",
      "\n",
      "Epoch 00017: val_accuracy did not improve from 0.90971\n",
      "Epoch 18/100\n",
      "236/236 [==============================] - 149s 630ms/step - loss: 0.1864 - accuracy: 0.9260 - val_loss: 0.3991 - val_accuracy: 0.8150\n",
      "\n",
      "Epoch 00018: val_accuracy did not improve from 0.90971\n",
      "Epoch 19/100\n",
      "236/236 [==============================] - 147s 624ms/step - loss: 0.1798 - accuracy: 0.9266 - val_loss: 0.2815 - val_accuracy: 0.8850\n",
      "\n",
      "Epoch 00019: val_accuracy did not improve from 0.90971\n",
      "Epoch 20/100\n",
      "236/236 [==============================] - 147s 621ms/step - loss: 0.1861 - accuracy: 0.9301 - val_loss: 0.3849 - val_accuracy: 0.8457\n",
      "\n",
      "Epoch 00020: val_accuracy did not improve from 0.90971\n",
      "Epoch 21/100\n",
      "236/236 [==============================] - 149s 629ms/step - loss: 0.1810 - accuracy: 0.9302 - val_loss: 3.4619 - val_accuracy: 0.5362\n",
      "\n",
      "Epoch 00021: val_accuracy did not improve from 0.90971\n",
      "Epoch 22/100\n",
      "236/236 [==============================] - 149s 630ms/step - loss: 0.1701 - accuracy: 0.9351 - val_loss: 1.2171 - val_accuracy: 0.7096\n",
      "\n",
      "Epoch 00022: val_accuracy did not improve from 0.90971\n",
      "Epoch 23/100\n",
      "236/236 [==============================] - 147s 623ms/step - loss: 0.1781 - accuracy: 0.9304 - val_loss: 0.2740 - val_accuracy: 0.8886\n",
      "\n",
      "Epoch 00023: val_accuracy did not improve from 0.90971\n",
      "Epoch 24/100\n",
      "236/236 [==============================] - 149s 630ms/step - loss: 0.1678 - accuracy: 0.9340 - val_loss: 0.7427 - val_accuracy: 0.7936\n",
      "\n",
      "Epoch 00024: val_accuracy did not improve from 0.90971\n",
      "Epoch 25/100\n",
      "236/236 [==============================] - 148s 625ms/step - loss: 0.1620 - accuracy: 0.9363 - val_loss: 0.2582 - val_accuracy: 0.9041\n",
      "\n",
      "Epoch 00025: val_accuracy did not improve from 0.90971\n",
      "Epoch 26/100\n",
      "236/236 [==============================] - 149s 629ms/step - loss: 0.1732 - accuracy: 0.9308 - val_loss: 0.2120 - val_accuracy: 0.9125\n",
      "\n",
      "Epoch 00026: val_accuracy improved from 0.90971 to 0.91249, saving model to model.h5\n",
      "Epoch 27/100\n",
      "236/236 [==============================] - 148s 626ms/step - loss: 0.1709 - accuracy: 0.9327 - val_loss: 0.3373 - val_accuracy: 0.8771\n",
      "\n",
      "Epoch 00027: val_accuracy did not improve from 0.91249\n",
      "Epoch 28/100\n",
      "236/236 [==============================] - 148s 625ms/step - loss: 0.1515 - accuracy: 0.9395 - val_loss: 0.4272 - val_accuracy: 0.8039\n",
      "\n",
      "Epoch 00028: val_accuracy did not improve from 0.91249\n",
      "Epoch 29/100\n",
      "236/236 [==============================] - 148s 626ms/step - loss: 0.1436 - accuracy: 0.9464 - val_loss: 3.9107 - val_accuracy: 0.5322\n",
      "\n",
      "Epoch 00029: val_accuracy did not improve from 0.91249\n",
      "Epoch 30/100\n",
      "236/236 [==============================] - 148s 626ms/step - loss: 0.1550 - accuracy: 0.9387 - val_loss: 3.4257 - val_accuracy: 0.6054\n",
      "\n",
      "Epoch 00030: val_accuracy did not improve from 0.91249\n",
      "Epoch 31/100\n",
      "236/236 [==============================] - 148s 624ms/step - loss: 0.1608 - accuracy: 0.9379 - val_loss: 0.2192 - val_accuracy: 0.9185\n",
      "\n",
      "Epoch 00031: val_accuracy improved from 0.91249 to 0.91846, saving model to model.h5\n",
      "Epoch 32/100\n",
      "236/236 [==============================] - 148s 627ms/step - loss: 0.1368 - accuracy: 0.9493 - val_loss: 0.3820 - val_accuracy: 0.8496\n",
      "\n",
      "Epoch 00032: val_accuracy did not improve from 0.91846\n",
      "Epoch 33/100\n",
      "236/236 [==============================] - 149s 631ms/step - loss: 0.1484 - accuracy: 0.9414 - val_loss: 0.4927 - val_accuracy: 0.8083\n",
      "\n",
      "Epoch 00033: val_accuracy did not improve from 0.91846\n",
      "Epoch 34/100\n",
      "236/236 [==============================] - 150s 634ms/step - loss: 0.1490 - accuracy: 0.9459 - val_loss: 0.2527 - val_accuracy: 0.9105\n",
      "\n",
      "Epoch 00034: val_accuracy did not improve from 0.91846\n",
      "Epoch 35/100\n",
      "236/236 [==============================] - 149s 630ms/step - loss: 0.1447 - accuracy: 0.9497 - val_loss: 0.7239 - val_accuracy: 0.7896\n",
      "\n",
      "Epoch 00035: val_accuracy did not improve from 0.91846\n",
      "Epoch 36/100\n",
      "236/236 [==============================] - 150s 634ms/step - loss: 0.1464 - accuracy: 0.9426 - val_loss: 0.4454 - val_accuracy: 0.8107\n",
      "\n",
      "Epoch 00036: val_accuracy did not improve from 0.91846\n",
      "Epoch 37/100\n",
      "236/236 [==============================] - 150s 635ms/step - loss: 0.1407 - accuracy: 0.9473 - val_loss: 0.5215 - val_accuracy: 0.8051\n",
      "\n",
      "Epoch 00037: val_accuracy did not improve from 0.91846\n",
      "Epoch 38/100\n",
      "236/236 [==============================] - 150s 632ms/step - loss: 0.1257 - accuracy: 0.9512 - val_loss: 0.9186 - val_accuracy: 0.7359\n",
      "\n",
      "Epoch 00038: val_accuracy did not improve from 0.91846\n",
      "Epoch 39/100\n",
      "236/236 [==============================] - 150s 636ms/step - loss: 0.1314 - accuracy: 0.9477 - val_loss: 0.2610 - val_accuracy: 0.9065\n",
      "\n",
      "Epoch 00039: val_accuracy did not improve from 0.91846\n",
      "Epoch 40/100\n",
      "236/236 [==============================] - 151s 636ms/step - loss: 0.1266 - accuracy: 0.9475 - val_loss: 0.2888 - val_accuracy: 0.9014\n",
      "\n",
      "Epoch 00040: val_accuracy did not improve from 0.91846\n",
      "Epoch 41/100\n",
      "236/236 [==============================] - 150s 633ms/step - loss: 0.1333 - accuracy: 0.9446 - val_loss: 0.6286 - val_accuracy: 0.8369\n",
      "\n",
      "Epoch 00041: val_accuracy did not improve from 0.91846\n",
      "\n",
      "Epoch 00041: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
      "Epoch 42/100\n",
      "236/236 [==============================] - 152s 640ms/step - loss: 0.1029 - accuracy: 0.9637 - val_loss: 0.4293 - val_accuracy: 0.8644\n",
      "\n",
      "Epoch 00042: val_accuracy did not improve from 0.91846\n",
      "Epoch 43/100\n",
      "236/236 [==============================] - 151s 639ms/step - loss: 0.0824 - accuracy: 0.9675 - val_loss: 1.6453 - val_accuracy: 0.6675\n",
      "\n",
      "Epoch 00043: val_accuracy did not improve from 0.91846\n",
      "Epoch 44/100\n",
      "236/236 [==============================] - 151s 638ms/step - loss: 0.0888 - accuracy: 0.9710 - val_loss: 0.9662 - val_accuracy: 0.7621\n",
      "\n",
      "Epoch 00044: val_accuracy did not improve from 0.91846\n",
      "Epoch 45/100\n",
      "236/236 [==============================] - 151s 639ms/step - loss: 0.0778 - accuracy: 0.9726 - val_loss: 0.3404 - val_accuracy: 0.9021\n",
      "\n",
      "Epoch 00045: val_accuracy did not improve from 0.91846\n",
      "Epoch 46/100\n",
      "236/236 [==============================] - 152s 643ms/step - loss: 0.0708 - accuracy: 0.9738 - val_loss: 1.0539 - val_accuracy: 0.7705\n",
      "\n",
      "Epoch 00046: val_accuracy did not improve from 0.91846\n",
      "Epoch 47/100\n",
      "236/236 [==============================] - 151s 639ms/step - loss: 0.0731 - accuracy: 0.9717 - val_loss: 0.2322 - val_accuracy: 0.9328\n",
      "\n",
      "Epoch 00047: val_accuracy improved from 0.91846 to 0.93278, saving model to model.h5\n",
      "Epoch 48/100\n",
      "236/236 [==============================] - 151s 639ms/step - loss: 0.0791 - accuracy: 0.9707 - val_loss: 0.2621 - val_accuracy: 0.9280\n",
      "\n",
      "Epoch 00048: val_accuracy did not improve from 0.93278\n",
      "Epoch 49/100\n",
      "236/236 [==============================] - 150s 633ms/step - loss: 0.0584 - accuracy: 0.9794 - val_loss: 0.8662 - val_accuracy: 0.8186\n",
      "\n",
      "Epoch 00049: val_accuracy did not improve from 0.93278\n",
      "Epoch 50/100\n",
      "236/236 [==============================] - 147s 622ms/step - loss: 0.0548 - accuracy: 0.9786 - val_loss: 0.2041 - val_accuracy: 0.9491\n",
      "\n",
      "Epoch 00050: val_accuracy improved from 0.93278 to 0.94909, saving model to model.h5\n",
      "Epoch 51/100\n",
      "236/236 [==============================] - 146s 619ms/step - loss: 0.0633 - accuracy: 0.9779 - val_loss: 0.2614 - val_accuracy: 0.9264\n",
      "\n",
      "Epoch 00051: val_accuracy did not improve from 0.94909\n",
      "Epoch 52/100\n",
      "236/236 [==============================] - 150s 636ms/step - loss: 0.0605 - accuracy: 0.9779 - val_loss: 0.2762 - val_accuracy: 0.9296\n",
      "\n",
      "Epoch 00052: val_accuracy did not improve from 0.94909\n",
      "Epoch 53/100\n",
      "236/236 [==============================] - 147s 622ms/step - loss: 0.0581 - accuracy: 0.9769 - val_loss: 0.4764 - val_accuracy: 0.8763\n",
      "\n",
      "Epoch 00053: val_accuracy did not improve from 0.94909\n",
      "Epoch 54/100\n",
      "236/236 [==============================] - 148s 628ms/step - loss: 0.0595 - accuracy: 0.9762 - val_loss: 0.3750 - val_accuracy: 0.9045\n",
      "\n",
      "Epoch 00054: val_accuracy did not improve from 0.94909\n",
      "Epoch 55/100\n",
      "236/236 [==============================] - 148s 625ms/step - loss: 0.0586 - accuracy: 0.9771 - val_loss: 0.8054 - val_accuracy: 0.8262\n",
      "\n",
      "Epoch 00055: val_accuracy did not improve from 0.94909\n",
      "Epoch 56/100\n",
      "236/236 [==============================] - 148s 628ms/step - loss: 0.0576 - accuracy: 0.9784 - val_loss: 0.2780 - val_accuracy: 0.9288\n",
      "\n",
      "Epoch 00056: val_accuracy did not improve from 0.94909\n",
      "Epoch 57/100\n",
      "236/236 [==============================] - 147s 624ms/step - loss: 0.0562 - accuracy: 0.9799 - val_loss: 0.2559 - val_accuracy: 0.9304\n",
      "\n",
      "Epoch 00057: val_accuracy did not improve from 0.94909\n",
      "Epoch 58/100\n",
      "236/236 [==============================] - 150s 633ms/step - loss: 0.0541 - accuracy: 0.9816 - val_loss: 0.2861 - val_accuracy: 0.9232\n",
      "\n",
      "Epoch 00058: val_accuracy did not improve from 0.94909\n",
      "Epoch 59/100\n",
      "236/236 [==============================] - 147s 624ms/step - loss: 0.0528 - accuracy: 0.9784 - val_loss: 0.4022 - val_accuracy: 0.9014\n",
      "\n",
      "Epoch 00059: val_accuracy did not improve from 0.94909\n",
      "Epoch 60/100\n",
      "236/236 [==============================] - 148s 628ms/step - loss: 0.0451 - accuracy: 0.9831 - val_loss: 0.1949 - val_accuracy: 0.9491\n",
      "\n",
      "Epoch 00060: val_accuracy did not improve from 0.94909\n",
      "Epoch 61/100\n",
      "236/236 [==============================] - 148s 628ms/step - loss: 0.0477 - accuracy: 0.9807 - val_loss: 0.3494 - val_accuracy: 0.9169\n",
      "\n",
      "Epoch 00061: val_accuracy did not improve from 0.94909\n",
      "Epoch 62/100\n",
      "236/236 [==============================] - 148s 629ms/step - loss: 0.0404 - accuracy: 0.9841 - val_loss: 0.9139 - val_accuracy: 0.8274\n",
      "\n",
      "Epoch 00062: val_accuracy did not improve from 0.94909\n",
      "Epoch 63/100\n",
      "236/236 [==============================] - 148s 625ms/step - loss: 0.0436 - accuracy: 0.9834 - val_loss: 0.4410 - val_accuracy: 0.8918\n",
      "\n",
      "Epoch 00063: val_accuracy did not improve from 0.94909\n",
      "Epoch 64/100\n",
      "236/236 [==============================] - 149s 627ms/step - loss: 0.0422 - accuracy: 0.9849 - val_loss: 0.5470 - val_accuracy: 0.8743\n",
      "\n",
      "Epoch 00064: val_accuracy did not improve from 0.94909\n",
      "Epoch 65/100\n",
      "236/236 [==============================] - 151s 636ms/step - loss: 0.0404 - accuracy: 0.9824 - val_loss: 0.2051 - val_accuracy: 0.9491\n",
      "\n",
      "Epoch 00065: val_accuracy did not improve from 0.94909\n",
      "Epoch 66/100\n",
      "236/236 [==============================] - 149s 630ms/step - loss: 0.0417 - accuracy: 0.9844 - val_loss: 1.2745 - val_accuracy: 0.7681\n",
      "\n",
      "Epoch 00066: val_accuracy did not improve from 0.94909\n",
      "Epoch 67/100\n",
      "236/236 [==============================] - 149s 631ms/step - loss: 0.0368 - accuracy: 0.9867 - val_loss: 0.4669 - val_accuracy: 0.8906\n",
      "\n",
      "Epoch 00067: val_accuracy did not improve from 0.94909\n",
      "Epoch 68/100\n",
      "236/236 [==============================] - 149s 630ms/step - loss: 0.0368 - accuracy: 0.9863 - val_loss: 0.5305 - val_accuracy: 0.8783\n",
      "\n",
      "Epoch 00068: val_accuracy did not improve from 0.94909\n",
      "Epoch 69/100\n",
      "236/236 [==============================] - 149s 631ms/step - loss: 0.0358 - accuracy: 0.9865 - val_loss: 0.9639 - val_accuracy: 0.8270\n",
      "\n",
      "Epoch 00069: val_accuracy did not improve from 0.94909\n",
      "Epoch 70/100\n",
      "236/236 [==============================] - 148s 629ms/step - loss: 0.0319 - accuracy: 0.9891 - val_loss: 0.2490 - val_accuracy: 0.9387\n",
      "\n",
      "Epoch 00070: val_accuracy did not improve from 0.94909\n",
      "Epoch 71/100\n",
      "236/236 [==============================] - 150s 634ms/step - loss: 0.0307 - accuracy: 0.9877 - val_loss: 0.5124 - val_accuracy: 0.8839\n",
      "\n",
      "Epoch 00071: val_accuracy did not improve from 0.94909\n",
      "Epoch 72/100\n",
      "236/236 [==============================] - 148s 626ms/step - loss: 0.0369 - accuracy: 0.9872 - val_loss: 0.3357 - val_accuracy: 0.9065\n",
      "\n",
      "Epoch 00072: val_accuracy did not improve from 0.94909\n",
      "Epoch 73/100\n",
      "236/236 [==============================] - 150s 634ms/step - loss: 0.0403 - accuracy: 0.9855 - val_loss: 0.2337 - val_accuracy: 0.9387\n",
      "\n",
      "Epoch 00073: val_accuracy did not improve from 0.94909\n",
      "Epoch 74/100\n",
      "236/236 [==============================] - 155s 655ms/step - loss: 0.0337 - accuracy: 0.9864 - val_loss: 0.2342 - val_accuracy: 0.9459\n",
      "\n",
      "Epoch 00074: val_accuracy did not improve from 0.94909\n",
      "Epoch 75/100\n",
      "236/236 [==============================] - 152s 642ms/step - loss: 0.0421 - accuracy: 0.9863 - val_loss: 1.7245 - val_accuracy: 0.7478\n",
      "\n",
      "Epoch 00075: val_accuracy did not improve from 0.94909\n",
      "\n",
      "Epoch 00075: ReduceLROnPlateau reducing learning rate to 1.0000000474974514e-05.\n",
      "Epoch 76/100\n",
      "236/236 [==============================] - 150s 634ms/step - loss: 0.0307 - accuracy: 0.9899 - val_loss: 0.4099 - val_accuracy: 0.9081\n",
      "\n",
      "Epoch 00076: val_accuracy did not improve from 0.94909\n",
      "Epoch 77/100\n",
      "236/236 [==============================] - 150s 635ms/step - loss: 0.0244 - accuracy: 0.9906 - val_loss: 0.3416 - val_accuracy: 0.9208\n",
      "\n",
      "Epoch 00077: val_accuracy did not improve from 0.94909\n",
      "Epoch 78/100\n",
      "236/236 [==============================] - 151s 639ms/step - loss: 0.0266 - accuracy: 0.9919 - val_loss: 0.5159 - val_accuracy: 0.8914\n",
      "\n",
      "Epoch 00078: val_accuracy did not improve from 0.94909\n",
      "Epoch 79/100\n",
      "236/236 [==============================] - 152s 642ms/step - loss: 0.0248 - accuracy: 0.9911 - val_loss: 0.4261 - val_accuracy: 0.9061\n",
      "\n",
      "Epoch 00079: val_accuracy did not improve from 0.94909\n",
      "Epoch 80/100\n",
      "236/236 [==============================] - 151s 638ms/step - loss: 0.0260 - accuracy: 0.9905 - val_loss: 0.3804 - val_accuracy: 0.9161\n",
      "\n",
      "Epoch 00080: val_accuracy did not improve from 0.94909\n",
      "Epoch 81/100\n",
      "236/236 [==============================] - 152s 640ms/step - loss: 0.0285 - accuracy: 0.9905 - val_loss: 0.4970 - val_accuracy: 0.8934\n",
      "\n",
      "Epoch 00081: val_accuracy did not improve from 0.94909\n",
      "Epoch 82/100\n",
      "236/236 [==============================] - 151s 636ms/step - loss: 0.0256 - accuracy: 0.9928 - val_loss: 0.4238 - val_accuracy: 0.9093\n",
      "\n",
      "Epoch 00082: val_accuracy did not improve from 0.94909\n",
      "Epoch 83/100\n",
      "236/236 [==============================] - 151s 638ms/step - loss: 0.0196 - accuracy: 0.9941 - val_loss: 0.4122 - val_accuracy: 0.9101\n",
      "\n",
      "Epoch 00083: val_accuracy did not improve from 0.94909\n",
      "Epoch 84/100\n",
      "236/236 [==============================] - 152s 641ms/step - loss: 0.0234 - accuracy: 0.9923 - val_loss: 0.4533 - val_accuracy: 0.9025\n",
      "\n",
      "Epoch 00084: val_accuracy did not improve from 0.94909\n",
      "Epoch 85/100\n",
      "236/236 [==============================] - 152s 643ms/step - loss: 0.0224 - accuracy: 0.9915 - val_loss: 0.3627 - val_accuracy: 0.9185\n",
      "\n",
      "Epoch 00085: val_accuracy did not improve from 0.94909\n",
      "Epoch 86/100\n",
      "236/236 [==============================] - 151s 639ms/step - loss: 0.0298 - accuracy: 0.9895 - val_loss: 0.5034 - val_accuracy: 0.8942\n",
      "\n",
      "Epoch 00086: val_accuracy did not improve from 0.94909\n",
      "Epoch 87/100\n",
      "236/236 [==============================] - 152s 641ms/step - loss: 0.0301 - accuracy: 0.9892 - val_loss: 0.4090 - val_accuracy: 0.9105\n",
      "\n",
      "Epoch 00087: val_accuracy did not improve from 0.94909\n",
      "Epoch 88/100\n",
      "236/236 [==============================] - 152s 643ms/step - loss: 0.0267 - accuracy: 0.9917 - val_loss: 0.4173 - val_accuracy: 0.9097\n",
      "\n",
      "Epoch 00088: val_accuracy did not improve from 0.94909\n",
      "Epoch 89/100\n",
      "236/236 [==============================] - 152s 643ms/step - loss: 0.0252 - accuracy: 0.9913 - val_loss: 0.5063 - val_accuracy: 0.8914\n",
      "\n",
      "Epoch 00089: val_accuracy did not improve from 0.94909\n",
      "Epoch 90/100\n",
      "236/236 [==============================] - 152s 642ms/step - loss: 0.0295 - accuracy: 0.9908 - val_loss: 0.4686 - val_accuracy: 0.9025\n",
      "\n",
      "Epoch 00090: val_accuracy did not improve from 0.94909\n",
      "\n",
      "Epoch 00090: ReduceLROnPlateau reducing learning rate to 1.0000000656873453e-06.\n",
      "Epoch 91/100\n",
      "236/236 [==============================] - 152s 644ms/step - loss: 0.0305 - accuracy: 0.9907 - val_loss: 0.4596 - val_accuracy: 0.9045\n",
      "\n",
      "Epoch 00091: val_accuracy did not improve from 0.94909\n",
      "Epoch 92/100\n",
      "236/236 [==============================] - 151s 636ms/step - loss: 0.0233 - accuracy: 0.9907 - val_loss: 0.4493 - val_accuracy: 0.9049\n",
      "\n",
      "Epoch 00092: val_accuracy did not improve from 0.94909\n",
      "Epoch 93/100\n",
      "236/236 [==============================] - 152s 640ms/step - loss: 0.0187 - accuracy: 0.9933 - val_loss: 0.4557 - val_accuracy: 0.9049\n",
      "\n",
      "Epoch 00093: val_accuracy did not improve from 0.94909\n",
      "Epoch 94/100\n",
      "236/236 [==============================] - 152s 640ms/step - loss: 0.0222 - accuracy: 0.9917 - val_loss: 0.4579 - val_accuracy: 0.9045\n",
      "\n",
      "Epoch 00094: val_accuracy did not improve from 0.94909\n",
      "Epoch 95/100\n",
      "236/236 [==============================] - 153s 645ms/step - loss: 0.0319 - accuracy: 0.9886 - val_loss: 0.4446 - val_accuracy: 0.9057\n",
      "\n",
      "Epoch 00095: val_accuracy did not improve from 0.94909\n",
      "Epoch 96/100\n",
      "236/236 [==============================] - 153s 646ms/step - loss: 0.0252 - accuracy: 0.9904 - val_loss: 0.4560 - val_accuracy: 0.9053\n",
      "\n",
      "Epoch 00096: val_accuracy did not improve from 0.94909\n",
      "Epoch 00096: early stopping\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(\n",
    "    train_gen,\n",
    "    steps_per_epoch= steps_per_epoch,\n",
    "    validation_data=val_gen,\n",
    "    validation_steps= validation_steps,\n",
    "    epochs=epochs,\n",
    "    callbacks=[mc,es,reduce_lr]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2021-03-19T09:34:57.790434Z",
     "iopub.status.busy": "2021-03-19T09:34:57.789590Z",
     "iopub.status.idle": "2021-03-19T09:34:58.116229Z",
     "shell.execute_reply": "2021-03-19T09:34:58.116640Z"
    },
    "papermill": {
     "duration": 7.161715,
     "end_time": "2021-03-19T09:34:58.116789",
     "exception": false,
     "start_time": "2021-03-19T09:34:50.955074",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x1152 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def history_plot(model_id, history1):\n",
    "\tplt.figure(figsize=(14,16)) \n",
    "\t# plot loss\n",
    "\tplt.subplot(2, 1, 1)\n",
    "\tplt.title('Cross Entropy Loss - ' + model_id, fontsize = 12)\n",
    "\tplt.plot(history1.history['loss'], color='blue', label='train base')\n",
    "\tplt.plot(history1.history['val_loss'], color='cyan', label='val base')\n",
    "\tplt.xlabel(\"Epochs\",fontsize = 12)\n",
    "\tplt.ylabel(\"Loss\", fontsize = 12)\n",
    "\tplt.legend(loc='upper right')\n",
    "\n",
    "\t# plot accuracy\n",
    "\tplt.subplot(2, 1, 2)\n",
    "\tplt.title('Classification Accuracy ' + model_id, fontsize = 10)\n",
    "\tplt.plot(history1.history['accuracy'], color='blue', label='train base')\n",
    "\tplt.plot(history1.history['val_accuracy'], color='cyan', label='val base')\n",
    "\tplt.xlabel(\"Epochs\", fontsize = 12)\n",
    "\tplt.ylabel(\"Accuracy\", fontsize = 12)\n",
    "\tplt.legend(loc='lower right')\n",
    "\n",
    "\tplt.show()\n",
    "history_plot('CNN Model', history)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2021-03-19T09:35:12.074949Z",
     "iopub.status.busy": "2021-03-19T09:35:12.069975Z",
     "iopub.status.idle": "2021-03-19T09:35:24.593821Z",
     "shell.execute_reply": "2021-03-19T09:35:24.592999Z"
    },
    "papermill": {
     "duration": 19.88706,
     "end_time": "2021-03-19T09:35:24.593991",
     "exception": false,
     "start_time": "2021-03-19T09:35:04.706931",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "predicting...\n",
      "79/79 [==============================] - 12s 133ms/step\n"
     ]
    }
   ],
   "source": [
    "# Save results as CSV files\n",
    "model.load_weights('model.h5')\n",
    "#saved_model = load_model(checkpoint_best_model)\n",
    "\n",
    "category_names = [\"No Defect\", \"Defect\"]\n",
    "\n",
    "y_test = df_test['defect_label']\n",
    "\n",
    "STEP_SIZE_TEST = compute_steps_per_epoch(y_test.shape[0])\n",
    "print('predicting...')\n",
    "\n",
    "predictions = model.predict(\n",
    "    test_gen,\n",
    "    steps = STEP_SIZE_TEST,\n",
    "    verbose=1)\n",
    "\n",
    "y_pred = (predictions > 0.5).astype(\"int32\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2021-03-19T09:35:38.095027Z",
     "iopub.status.busy": "2021-03-19T09:35:38.094177Z",
     "iopub.status.idle": "2021-03-19T09:35:38.105614Z",
     "shell.execute_reply": "2021-03-19T09:35:38.106311Z"
    },
    "papermill": {
     "duration": 6.667446,
     "end_time": "2021-03-19T09:35:38.106495",
     "exception": false,
     "start_time": "2021-03-19T09:35:31.439049",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "   No Defect       0.94      0.95      0.95      1181\n",
      "      Defect       0.96      0.95      0.95      1333\n",
      "\n",
      "    accuracy                           0.95      2514\n",
      "   macro avg       0.95      0.95      0.95      2514\n",
      "weighted avg       0.95      0.95      0.95      2514\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# classification report\n",
    "report_string = classification_report(y_test, y_pred, target_names=category_names)\n",
    "print(report_string)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2021-03-19T09:35:52.287273Z",
     "iopub.status.busy": "2021-03-19T09:35:52.286441Z",
     "iopub.status.idle": "2021-03-19T09:35:52.461202Z",
     "shell.execute_reply": "2021-03-19T09:35:52.461614Z"
    },
    "papermill": {
     "duration": 7.008489,
     "end_time": "2021-03-19T09:35:52.461758",
     "exception": false,
     "start_time": "2021-03-19T09:35:45.453269",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot_cm(y_true, y_pred, figsize=(6,5)):\n",
    "    cm = confusion_matrix(y_true, y_pred, labels=np.unique(y_true))\n",
    "    cm_sum = np.sum(cm, axis=1, keepdims=True)\n",
    "    cm_perc = cm / cm_sum.astype(float) * 100\n",
    "    annot = np.empty_like(cm).astype(str)\n",
    "    nrows, ncols = cm.shape\n",
    "    for i in range(nrows):\n",
    "        for j in range(ncols):\n",
    "            c = cm[i, j]\n",
    "            p = cm_perc[i, j]\n",
    "            if i == j:\n",
    "                s = cm_sum[i]\n",
    "                annot[i, j] = '%.1f%%\\n%d/%d' % (p, c, s)\n",
    "            elif c == 0:\n",
    "                annot[i, j] = ''\n",
    "            elif i<j:\n",
    "                annot[i, j] = '%.1f%%\\n%d\\nFP' % (p, c)\n",
    "            else:\n",
    "                annot[i, j] = '%.1f%%\\n%d\\nFN' % (p, c)\n",
    "    #cm = pd.DataFrame(cm, index=np.unique(y_true), columns=np.unique(y_true))\n",
    "    cm = pd.DataFrame(cm, index=category_names, columns=category_names)\n",
    "    cm.index.name = 'Actual'\n",
    "    cm.columns.name = 'Predicted'\n",
    "    fig, ax = plt.subplots(figsize=figsize)\n",
    "    sns.set(font_scale=1.4)\n",
    "    sns.heatmap(cm, cmap= \"YlGnBu\", annot=annot, fmt='', ax=ax, annot_kws={\"fontsize\":12})\n",
    "plot_cm(y_test, y_pred, figsize=(6,5))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2021-03-19T09:36:05.815313Z",
     "iopub.status.busy": "2021-03-19T09:36:05.814439Z",
     "iopub.status.idle": "2021-03-19T09:36:06.008659Z",
     "shell.execute_reply": "2021-03-19T09:36:06.009077Z"
    },
    "papermill": {
     "duration": 7.052635,
     "end_time": "2021-03-19T09:36:06.009237",
     "exception": false,
     "start_time": "2021-03-19T09:35:58.956602",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROC AUC=0.990\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_curve, auc, roc_auc_score\n",
    "\n",
    "# calculate scores\n",
    "auc = roc_auc_score(y_test, predictions)\n",
    "# summarize scores\n",
    "print('ROC AUC=%.3f' % (auc))\n",
    "# calculate roc curves\n",
    "\n",
    "ns_fpr, ns_tpr, _ = roc_curve(y_test, predictions)\n",
    "\n",
    "# plot the roc curve for the model\n",
    "plt.plot(ns_fpr, ns_tpr, linestyle='--', label=\"AUC = \"+ str(auc))\n",
    "plt.plot([0, 1], [0, 1], color='navy', lw=2, linestyle='--')\n",
    "# axis labels\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "# show the legend\n",
    "plt.legend()\n",
    "# show the plot\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 7.330812,
     "end_time": "2021-03-19T09:36:19.927652",
     "exception": false,
     "start_time": "2021-03-19T09:36:12.596840",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## Save Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2021-03-19T09:36:33.230986Z",
     "iopub.status.busy": "2021-03-19T09:36:33.215686Z",
     "iopub.status.idle": "2021-03-19T09:36:34.169413Z",
     "shell.execute_reply": "2021-03-19T09:36:34.173879Z"
    },
    "papermill": {
     "duration": 7.438631,
     "end_time": "2021-03-19T09:36:34.174134",
     "exception": false,
     "start_time": "2021-03-19T09:36:26.735503",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "model.save('severstal-classification-DenseNet.h5', save_format=\"h5\")\n"
   ]
  }
 ],
 "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": "3.8.3"
  },
  "papermill": {
   "default_parameters": {},
   "duration": 14858.957012,
   "end_time": "2021-03-19T09:36:45.475728",
   "environment_variables": {},
   "exception": null,
   "input_path": "__notebook__.ipynb",
   "output_path": "__notebook__.ipynb",
   "parameters": {},
   "start_time": "2021-03-19T05:29:06.518716",
   "version": "2.2.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}