CSE 465 V2.1.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "6efb43fb",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Convolution2D, MaxPooling2D, Flatten, Dense, Dropout, GlobalAveragePooling2D"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "8852fd60",
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "import keras\n",
    "import os \n",
    "import time\n",
    "#import cv2\n",
    "import numpy as np\n",
    "#import matplotlib.pyplot as plt\n",
    "from keras import optimizers\n",
    "import keras\n",
    "import tensorflow as tf\n",
    "import keras.backend as K\n",
    "from sklearn.metrics import confusion_matrix, classification_report\n",
    "from keras.models import load_model\n",
    "from keras.models import Sequential\n",
    "from keras.regularizers import l2\n",
    "from keras.applications.vgg16 import VGG16\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from keras.callbacks import ModelCheckpoint, CSVLogger, EarlyStopping, ReduceLROnPlateau\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.metrics import confusion_matrix, classification_report\n",
    "from sklearn.metrics import roc_curve, auc, roc_auc_score\n",
    "#import matplotlib.pyplot as plt\n",
    "#from tqdm import tqdm\n",
    "from keras.utils import np_utils\n",
    "#from imgaug import augmenters as iaa    \n",
    "import itertools\n",
    "%matplotlib inline\n",
    "#import matplotlib.pyplot as plt\n",
    "import matplotlib.image as mpimg\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from keras.models import Sequential\n",
    "from tensorflow.keras.optimizers import RMSprop\n",
    "from keras.layers import Conv2D, MaxPooling2D\n",
    "from keras.layers import Activation, Dropout, Flatten, Dense\n",
    "from keras.callbacks import CSVLogger\n",
    "#from livelossplot import PlotLossesKeras\n",
    "import os\n",
    "#from classification_models.keras import Classifiers\n",
    "# GPU test\n",
    "from tensorflow.python.client import device_lib\n",
    "np.random.seed(42)\n",
    "from keras.models import load_model\n",
    "# Print version\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
    "from keras.layers import Input, Activation, merge, Dense, Flatten, Dropout, BatchNormalization, GlobalAveragePooling2D\n",
    "from keras.models import Model\n",
    "from keras.layers import Input, Activation, merge, Dense, Flatten, Dropout, BatchNormalization, GlobalAveragePooling2D\n",
    "from keras.layers.convolutional import Convolution2D, MaxPooling2D\n",
    "from keras.applications.vgg16 import VGG16\n",
    "from keras.regularizers import l2\n",
    "import os, sys\n",
    "import scipy.misc\n",
    "from glob import glob\n",
    "import numpy as np\n",
    "import random \n",
    "import shutil \n",
    "import keras\n",
    "import time\n",
    "import matplotlib.pyplot as plt\n",
    "from keras.layers import Dense, Input, Conv2D, Flatten, MaxPool2D, Activation,Dropout, GlobalAveragePooling2D\n",
    "from keras.models import Model\n",
    "from keras.callbacks import ModelCheckpoint\n",
    "from keras import backend as K\n",
    "from keras.models import Sequential\n",
    "from keras.applications.densenet import DenseNet121\n",
    "from keras.models import load_model\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "from keras import optimizers\n",
    "import pickle\n",
    "import sys\n",
    "import cv2\n",
    "\n",
    "import tensorflow as tf\n",
    "\n",
    "\n",
    "np.random.seed(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "fe9d31eb",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_dir = \"C:/Users/Havoc/OneDrive/Documents/Corona Dataset Final/train\"\n",
    "valid_dir = \"C:/Users/Havoc/OneDrive/Documents/Corona Dataset Final/test\"\n",
    "\n",
    "img_width, img_height = 224, 224 "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dd3bf7f5",
   "metadata": {},
   "source": [
    "# Using DenseNet201"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "715c6cfe",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/densenet/densenet201_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
      "74842112/74836368 [==============================] - 31s 0us/step\n",
      "74850304/74836368 [==============================] - 31s 0us/step\n",
      "Model: \"densenet201\"\n",
      "__________________________________________________________________________________________________\n",
      " Layer (type)                   Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      " input_1 (InputLayer)           [(None, 224, 224, 3  0           []                               \n",
      "                                )]                                                                \n",
      "                                                                                                  \n",
      " zero_padding2d (ZeroPadding2D)  (None, 230, 230, 3)  0          ['input_1[0][0]']                \n",
      "                                                                                                  \n",
      " conv1/conv (Conv2D)            (None, 112, 112, 64  9408        ['zero_padding2d[0][0]']         \n",
      "                                )                                                                 \n",
      "                                                                                                  \n",
      " conv1/bn (BatchNormalization)  (None, 112, 112, 64  256         ['conv1/conv[0][0]']             \n",
      "                                )                                                                 \n",
      "                                                                                                  \n",
      " conv1/relu (Activation)        (None, 112, 112, 64  0           ['conv1/bn[0][0]']               \n",
      "                                )                                                                 \n",
      "                                                                                                  \n",
      " zero_padding2d_1 (ZeroPadding2  (None, 114, 114, 64  0          ['conv1/relu[0][0]']             \n",
      " D)                             )                                                                 \n",
      "                                                                                                  \n",
      " pool1 (MaxPooling2D)           (None, 56, 56, 64)   0           ['zero_padding2d_1[0][0]']       \n",
      "                                                                                                  \n",
      " conv2_block1_0_bn (BatchNormal  (None, 56, 56, 64)  256         ['pool1[0][0]']                  \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv2_block1_0_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block1_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv2_block1_1_conv (Conv2D)   (None, 56, 56, 128)  8192        ['conv2_block1_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv2_block1_1_bn (BatchNormal  (None, 56, 56, 128)  512        ['conv2_block1_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv2_block1_1_relu (Activatio  (None, 56, 56, 128)  0          ['conv2_block1_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv2_block1_2_conv (Conv2D)   (None, 56, 56, 32)   36864       ['conv2_block1_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv2_block1_concat (Concatena  (None, 56, 56, 96)  0           ['pool1[0][0]',                  \n",
      " te)                                                              'conv2_block1_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv2_block2_0_bn (BatchNormal  (None, 56, 56, 96)  384         ['conv2_block1_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv2_block2_0_relu (Activatio  (None, 56, 56, 96)  0           ['conv2_block2_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv2_block2_1_conv (Conv2D)   (None, 56, 56, 128)  12288       ['conv2_block2_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv2_block2_1_bn (BatchNormal  (None, 56, 56, 128)  512        ['conv2_block2_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv2_block2_1_relu (Activatio  (None, 56, 56, 128)  0          ['conv2_block2_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv2_block2_2_conv (Conv2D)   (None, 56, 56, 32)   36864       ['conv2_block2_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv2_block2_concat (Concatena  (None, 56, 56, 128)  0          ['conv2_block1_concat[0][0]',    \n",
      " te)                                                              'conv2_block2_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv2_block3_0_bn (BatchNormal  (None, 56, 56, 128)  512        ['conv2_block2_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv2_block3_0_relu (Activatio  (None, 56, 56, 128)  0          ['conv2_block3_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv2_block3_1_conv (Conv2D)   (None, 56, 56, 128)  16384       ['conv2_block3_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv2_block3_1_bn (BatchNormal  (None, 56, 56, 128)  512        ['conv2_block3_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv2_block3_1_relu (Activatio  (None, 56, 56, 128)  0          ['conv2_block3_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv2_block3_2_conv (Conv2D)   (None, 56, 56, 32)   36864       ['conv2_block3_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv2_block3_concat (Concatena  (None, 56, 56, 160)  0          ['conv2_block2_concat[0][0]',    \n",
      " te)                                                              'conv2_block3_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv2_block4_0_bn (BatchNormal  (None, 56, 56, 160)  640        ['conv2_block3_concat[0][0]']    \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv2_block4_0_relu (Activatio  (None, 56, 56, 160)  0          ['conv2_block4_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv2_block4_1_conv (Conv2D)   (None, 56, 56, 128)  20480       ['conv2_block4_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv2_block4_1_bn (BatchNormal  (None, 56, 56, 128)  512        ['conv2_block4_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv2_block4_1_relu (Activatio  (None, 56, 56, 128)  0          ['conv2_block4_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv2_block4_2_conv (Conv2D)   (None, 56, 56, 32)   36864       ['conv2_block4_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv2_block4_concat (Concatena  (None, 56, 56, 192)  0          ['conv2_block3_concat[0][0]',    \n",
      " te)                                                              'conv2_block4_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv2_block5_0_bn (BatchNormal  (None, 56, 56, 192)  768        ['conv2_block4_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv2_block5_0_relu (Activatio  (None, 56, 56, 192)  0          ['conv2_block5_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv2_block5_1_conv (Conv2D)   (None, 56, 56, 128)  24576       ['conv2_block5_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv2_block5_1_bn (BatchNormal  (None, 56, 56, 128)  512        ['conv2_block5_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv2_block5_1_relu (Activatio  (None, 56, 56, 128)  0          ['conv2_block5_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv2_block5_2_conv (Conv2D)   (None, 56, 56, 32)   36864       ['conv2_block5_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv2_block5_concat (Concatena  (None, 56, 56, 224)  0          ['conv2_block4_concat[0][0]',    \n",
      " te)                                                              'conv2_block5_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv2_block6_0_bn (BatchNormal  (None, 56, 56, 224)  896        ['conv2_block5_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv2_block6_0_relu (Activatio  (None, 56, 56, 224)  0          ['conv2_block6_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv2_block6_1_conv (Conv2D)   (None, 56, 56, 128)  28672       ['conv2_block6_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv2_block6_1_bn (BatchNormal  (None, 56, 56, 128)  512        ['conv2_block6_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv2_block6_1_relu (Activatio  (None, 56, 56, 128)  0          ['conv2_block6_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv2_block6_2_conv (Conv2D)   (None, 56, 56, 32)   36864       ['conv2_block6_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv2_block6_concat (Concatena  (None, 56, 56, 256)  0          ['conv2_block5_concat[0][0]',    \n",
      " te)                                                              'conv2_block6_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " pool2_bn (BatchNormalization)  (None, 56, 56, 256)  1024        ['conv2_block6_concat[0][0]']    \n",
      "                                                                                                  \n",
      " pool2_relu (Activation)        (None, 56, 56, 256)  0           ['pool2_bn[0][0]']               \n",
      "                                                                                                  \n",
      " pool2_conv (Conv2D)            (None, 56, 56, 128)  32768       ['pool2_relu[0][0]']             \n",
      "                                                                                                  \n",
      " pool2_pool (AveragePooling2D)  (None, 28, 28, 128)  0           ['pool2_conv[0][0]']             \n",
      "                                                                                                  \n",
      " conv3_block1_0_bn (BatchNormal  (None, 28, 28, 128)  512        ['pool2_pool[0][0]']             \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv3_block1_0_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block1_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv3_block1_1_conv (Conv2D)   (None, 28, 28, 128)  16384       ['conv3_block1_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block1_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block1_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv3_block1_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block1_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv3_block1_2_conv (Conv2D)   (None, 28, 28, 32)   36864       ['conv3_block1_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block1_concat (Concatena  (None, 28, 28, 160)  0          ['pool2_pool[0][0]',             \n",
      " te)                                                              'conv3_block1_2_conv[0][0]']    \n",
      "                                                                                                  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " conv3_block2_0_bn (BatchNormal  (None, 28, 28, 160)  640        ['conv3_block1_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv3_block2_0_relu (Activatio  (None, 28, 28, 160)  0          ['conv3_block2_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv3_block2_1_conv (Conv2D)   (None, 28, 28, 128)  20480       ['conv3_block2_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block2_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block2_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv3_block2_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block2_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv3_block2_2_conv (Conv2D)   (None, 28, 28, 32)   36864       ['conv3_block2_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block2_concat (Concatena  (None, 28, 28, 192)  0          ['conv3_block1_concat[0][0]',    \n",
      " te)                                                              'conv3_block2_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block3_0_bn (BatchNormal  (None, 28, 28, 192)  768        ['conv3_block2_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv3_block3_0_relu (Activatio  (None, 28, 28, 192)  0          ['conv3_block3_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv3_block3_1_conv (Conv2D)   (None, 28, 28, 128)  24576       ['conv3_block3_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block3_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block3_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv3_block3_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block3_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv3_block3_2_conv (Conv2D)   (None, 28, 28, 32)   36864       ['conv3_block3_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block3_concat (Concatena  (None, 28, 28, 224)  0          ['conv3_block2_concat[0][0]',    \n",
      " te)                                                              'conv3_block3_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block4_0_bn (BatchNormal  (None, 28, 28, 224)  896        ['conv3_block3_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv3_block4_0_relu (Activatio  (None, 28, 28, 224)  0          ['conv3_block4_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv3_block4_1_conv (Conv2D)   (None, 28, 28, 128)  28672       ['conv3_block4_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block4_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block4_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv3_block4_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block4_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv3_block4_2_conv (Conv2D)   (None, 28, 28, 32)   36864       ['conv3_block4_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block4_concat (Concatena  (None, 28, 28, 256)  0          ['conv3_block3_concat[0][0]',    \n",
      " te)                                                              'conv3_block4_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block5_0_bn (BatchNormal  (None, 28, 28, 256)  1024       ['conv3_block4_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv3_block5_0_relu (Activatio  (None, 28, 28, 256)  0          ['conv3_block5_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv3_block5_1_conv (Conv2D)   (None, 28, 28, 128)  32768       ['conv3_block5_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block5_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block5_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv3_block5_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block5_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv3_block5_2_conv (Conv2D)   (None, 28, 28, 32)   36864       ['conv3_block5_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block5_concat (Concatena  (None, 28, 28, 288)  0          ['conv3_block4_concat[0][0]',    \n",
      " te)                                                              'conv3_block5_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block6_0_bn (BatchNormal  (None, 28, 28, 288)  1152       ['conv3_block5_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv3_block6_0_relu (Activatio  (None, 28, 28, 288)  0          ['conv3_block6_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv3_block6_1_conv (Conv2D)   (None, 28, 28, 128)  36864       ['conv3_block6_0_relu[0][0]']    \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                                                                  \n",
      " conv3_block6_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block6_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv3_block6_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block6_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv3_block6_2_conv (Conv2D)   (None, 28, 28, 32)   36864       ['conv3_block6_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block6_concat (Concatena  (None, 28, 28, 320)  0          ['conv3_block5_concat[0][0]',    \n",
      " te)                                                              'conv3_block6_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block7_0_bn (BatchNormal  (None, 28, 28, 320)  1280       ['conv3_block6_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv3_block7_0_relu (Activatio  (None, 28, 28, 320)  0          ['conv3_block7_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv3_block7_1_conv (Conv2D)   (None, 28, 28, 128)  40960       ['conv3_block7_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block7_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block7_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv3_block7_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block7_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv3_block7_2_conv (Conv2D)   (None, 28, 28, 32)   36864       ['conv3_block7_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block7_concat (Concatena  (None, 28, 28, 352)  0          ['conv3_block6_concat[0][0]',    \n",
      " te)                                                              'conv3_block7_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block8_0_bn (BatchNormal  (None, 28, 28, 352)  1408       ['conv3_block7_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv3_block8_0_relu (Activatio  (None, 28, 28, 352)  0          ['conv3_block8_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv3_block8_1_conv (Conv2D)   (None, 28, 28, 128)  45056       ['conv3_block8_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block8_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block8_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv3_block8_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block8_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv3_block8_2_conv (Conv2D)   (None, 28, 28, 32)   36864       ['conv3_block8_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block8_concat (Concatena  (None, 28, 28, 384)  0          ['conv3_block7_concat[0][0]',    \n",
      " te)                                                              'conv3_block8_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block9_0_bn (BatchNormal  (None, 28, 28, 384)  1536       ['conv3_block8_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv3_block9_0_relu (Activatio  (None, 28, 28, 384)  0          ['conv3_block9_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv3_block9_1_conv (Conv2D)   (None, 28, 28, 128)  49152       ['conv3_block9_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block9_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block9_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv3_block9_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block9_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv3_block9_2_conv (Conv2D)   (None, 28, 28, 32)   36864       ['conv3_block9_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block9_concat (Concatena  (None, 28, 28, 416)  0          ['conv3_block8_concat[0][0]',    \n",
      " te)                                                              'conv3_block9_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv3_block10_0_bn (BatchNorma  (None, 28, 28, 416)  1664       ['conv3_block9_concat[0][0]']    \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv3_block10_0_relu (Activati  (None, 28, 28, 416)  0          ['conv3_block10_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv3_block10_1_conv (Conv2D)  (None, 28, 28, 128)  53248       ['conv3_block10_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv3_block10_1_bn (BatchNorma  (None, 28, 28, 128)  512        ['conv3_block10_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv3_block10_1_relu (Activati  (None, 28, 28, 128)  0          ['conv3_block10_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " conv3_block10_2_conv (Conv2D)  (None, 28, 28, 32)   36864       ['conv3_block10_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv3_block10_concat (Concaten  (None, 28, 28, 448)  0          ['conv3_block9_concat[0][0]',    \n",
      " ate)                                                             'conv3_block10_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv3_block11_0_bn (BatchNorma  (None, 28, 28, 448)  1792       ['conv3_block10_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv3_block11_0_relu (Activati  (None, 28, 28, 448)  0          ['conv3_block11_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv3_block11_1_conv (Conv2D)  (None, 28, 28, 128)  57344       ['conv3_block11_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv3_block11_1_bn (BatchNorma  (None, 28, 28, 128)  512        ['conv3_block11_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv3_block11_1_relu (Activati  (None, 28, 28, 128)  0          ['conv3_block11_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv3_block11_2_conv (Conv2D)  (None, 28, 28, 32)   36864       ['conv3_block11_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv3_block11_concat (Concaten  (None, 28, 28, 480)  0          ['conv3_block10_concat[0][0]',   \n",
      " ate)                                                             'conv3_block11_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv3_block12_0_bn (BatchNorma  (None, 28, 28, 480)  1920       ['conv3_block11_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv3_block12_0_relu (Activati  (None, 28, 28, 480)  0          ['conv3_block12_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv3_block12_1_conv (Conv2D)  (None, 28, 28, 128)  61440       ['conv3_block12_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv3_block12_1_bn (BatchNorma  (None, 28, 28, 128)  512        ['conv3_block12_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv3_block12_1_relu (Activati  (None, 28, 28, 128)  0          ['conv3_block12_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv3_block12_2_conv (Conv2D)  (None, 28, 28, 32)   36864       ['conv3_block12_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv3_block12_concat (Concaten  (None, 28, 28, 512)  0          ['conv3_block11_concat[0][0]',   \n",
      " ate)                                                             'conv3_block12_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " pool3_bn (BatchNormalization)  (None, 28, 28, 512)  2048        ['conv3_block12_concat[0][0]']   \n",
      "                                                                                                  \n",
      " pool3_relu (Activation)        (None, 28, 28, 512)  0           ['pool3_bn[0][0]']               \n",
      "                                                                                                  \n",
      " pool3_conv (Conv2D)            (None, 28, 28, 256)  131072      ['pool3_relu[0][0]']             \n",
      "                                                                                                  \n",
      " pool3_pool (AveragePooling2D)  (None, 14, 14, 256)  0           ['pool3_conv[0][0]']             \n",
      "                                                                                                  \n",
      " conv4_block1_0_bn (BatchNormal  (None, 14, 14, 256)  1024       ['pool3_pool[0][0]']             \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv4_block1_0_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block1_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv4_block1_1_conv (Conv2D)   (None, 14, 14, 128)  32768       ['conv4_block1_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block1_1_bn (BatchNormal  (None, 14, 14, 128)  512        ['conv4_block1_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv4_block1_1_relu (Activatio  (None, 14, 14, 128)  0          ['conv4_block1_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv4_block1_2_conv (Conv2D)   (None, 14, 14, 32)   36864       ['conv4_block1_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block1_concat (Concatena  (None, 14, 14, 288)  0          ['pool3_pool[0][0]',             \n",
      " te)                                                              'conv4_block1_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block2_0_bn (BatchNormal  (None, 14, 14, 288)  1152       ['conv4_block1_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv4_block2_0_relu (Activatio  (None, 14, 14, 288)  0          ['conv4_block2_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv4_block2_1_conv (Conv2D)   (None, 14, 14, 128)  36864       ['conv4_block2_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block2_1_bn (BatchNormal  (None, 14, 14, 128)  512        ['conv4_block2_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv4_block2_1_relu (Activatio  (None, 14, 14, 128)  0          ['conv4_block2_1_bn[0][0]']      \n",
      " n)                                                                                               \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                                                                  \n",
      " conv4_block2_2_conv (Conv2D)   (None, 14, 14, 32)   36864       ['conv4_block2_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block2_concat (Concatena  (None, 14, 14, 320)  0          ['conv4_block1_concat[0][0]',    \n",
      " te)                                                              'conv4_block2_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block3_0_bn (BatchNormal  (None, 14, 14, 320)  1280       ['conv4_block2_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv4_block3_0_relu (Activatio  (None, 14, 14, 320)  0          ['conv4_block3_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv4_block3_1_conv (Conv2D)   (None, 14, 14, 128)  40960       ['conv4_block3_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block3_1_bn (BatchNormal  (None, 14, 14, 128)  512        ['conv4_block3_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv4_block3_1_relu (Activatio  (None, 14, 14, 128)  0          ['conv4_block3_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv4_block3_2_conv (Conv2D)   (None, 14, 14, 32)   36864       ['conv4_block3_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block3_concat (Concatena  (None, 14, 14, 352)  0          ['conv4_block2_concat[0][0]',    \n",
      " te)                                                              'conv4_block3_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block4_0_bn (BatchNormal  (None, 14, 14, 352)  1408       ['conv4_block3_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv4_block4_0_relu (Activatio  (None, 14, 14, 352)  0          ['conv4_block4_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv4_block4_1_conv (Conv2D)   (None, 14, 14, 128)  45056       ['conv4_block4_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block4_1_bn (BatchNormal  (None, 14, 14, 128)  512        ['conv4_block4_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv4_block4_1_relu (Activatio  (None, 14, 14, 128)  0          ['conv4_block4_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv4_block4_2_conv (Conv2D)   (None, 14, 14, 32)   36864       ['conv4_block4_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block4_concat (Concatena  (None, 14, 14, 384)  0          ['conv4_block3_concat[0][0]',    \n",
      " te)                                                              'conv4_block4_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block5_0_bn (BatchNormal  (None, 14, 14, 384)  1536       ['conv4_block4_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv4_block5_0_relu (Activatio  (None, 14, 14, 384)  0          ['conv4_block5_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv4_block5_1_conv (Conv2D)   (None, 14, 14, 128)  49152       ['conv4_block5_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block5_1_bn (BatchNormal  (None, 14, 14, 128)  512        ['conv4_block5_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv4_block5_1_relu (Activatio  (None, 14, 14, 128)  0          ['conv4_block5_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv4_block5_2_conv (Conv2D)   (None, 14, 14, 32)   36864       ['conv4_block5_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block5_concat (Concatena  (None, 14, 14, 416)  0          ['conv4_block4_concat[0][0]',    \n",
      " te)                                                              'conv4_block5_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block6_0_bn (BatchNormal  (None, 14, 14, 416)  1664       ['conv4_block5_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv4_block6_0_relu (Activatio  (None, 14, 14, 416)  0          ['conv4_block6_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv4_block6_1_conv (Conv2D)   (None, 14, 14, 128)  53248       ['conv4_block6_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block6_1_bn (BatchNormal  (None, 14, 14, 128)  512        ['conv4_block6_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv4_block6_1_relu (Activatio  (None, 14, 14, 128)  0          ['conv4_block6_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv4_block6_2_conv (Conv2D)   (None, 14, 14, 32)   36864       ['conv4_block6_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block6_concat (Concatena  (None, 14, 14, 448)  0          ['conv4_block5_concat[0][0]',    \n",
      " te)                                                              'conv4_block6_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block7_0_bn (BatchNormal  (None, 14, 14, 448)  1792       ['conv4_block6_concat[0][0]']    \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv4_block7_0_relu (Activatio  (None, 14, 14, 448)  0          ['conv4_block7_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv4_block7_1_conv (Conv2D)   (None, 14, 14, 128)  57344       ['conv4_block7_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block7_1_bn (BatchNormal  (None, 14, 14, 128)  512        ['conv4_block7_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv4_block7_1_relu (Activatio  (None, 14, 14, 128)  0          ['conv4_block7_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv4_block7_2_conv (Conv2D)   (None, 14, 14, 32)   36864       ['conv4_block7_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block7_concat (Concatena  (None, 14, 14, 480)  0          ['conv4_block6_concat[0][0]',    \n",
      " te)                                                              'conv4_block7_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block8_0_bn (BatchNormal  (None, 14, 14, 480)  1920       ['conv4_block7_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv4_block8_0_relu (Activatio  (None, 14, 14, 480)  0          ['conv4_block8_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv4_block8_1_conv (Conv2D)   (None, 14, 14, 128)  61440       ['conv4_block8_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block8_1_bn (BatchNormal  (None, 14, 14, 128)  512        ['conv4_block8_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv4_block8_1_relu (Activatio  (None, 14, 14, 128)  0          ['conv4_block8_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv4_block8_2_conv (Conv2D)   (None, 14, 14, 32)   36864       ['conv4_block8_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block8_concat (Concatena  (None, 14, 14, 512)  0          ['conv4_block7_concat[0][0]',    \n",
      " te)                                                              'conv4_block8_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block9_0_bn (BatchNormal  (None, 14, 14, 512)  2048       ['conv4_block8_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv4_block9_0_relu (Activatio  (None, 14, 14, 512)  0          ['conv4_block9_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv4_block9_1_conv (Conv2D)   (None, 14, 14, 128)  65536       ['conv4_block9_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block9_1_bn (BatchNormal  (None, 14, 14, 128)  512        ['conv4_block9_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv4_block9_1_relu (Activatio  (None, 14, 14, 128)  0          ['conv4_block9_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv4_block9_2_conv (Conv2D)   (None, 14, 14, 32)   36864       ['conv4_block9_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block9_concat (Concatena  (None, 14, 14, 544)  0          ['conv4_block8_concat[0][0]',    \n",
      " te)                                                              'conv4_block9_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv4_block10_0_bn (BatchNorma  (None, 14, 14, 544)  2176       ['conv4_block9_concat[0][0]']    \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block10_0_relu (Activati  (None, 14, 14, 544)  0          ['conv4_block10_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block10_1_conv (Conv2D)  (None, 14, 14, 128)  69632       ['conv4_block10_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block10_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block10_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block10_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block10_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block10_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block10_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block10_concat (Concaten  (None, 14, 14, 576)  0          ['conv4_block9_concat[0][0]',    \n",
      " ate)                                                             'conv4_block10_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block11_0_bn (BatchNorma  (None, 14, 14, 576)  2304       ['conv4_block10_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block11_0_relu (Activati  (None, 14, 14, 576)  0          ['conv4_block11_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block11_1_conv (Conv2D)  (None, 14, 14, 128)  73728       ['conv4_block11_0_relu[0][0]']   \n",
      "                                                                                                  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " conv4_block11_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block11_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block11_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block11_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block11_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block11_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block11_concat (Concaten  (None, 14, 14, 608)  0          ['conv4_block10_concat[0][0]',   \n",
      " ate)                                                             'conv4_block11_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block12_0_bn (BatchNorma  (None, 14, 14, 608)  2432       ['conv4_block11_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block12_0_relu (Activati  (None, 14, 14, 608)  0          ['conv4_block12_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block12_1_conv (Conv2D)  (None, 14, 14, 128)  77824       ['conv4_block12_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block12_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block12_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block12_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block12_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block12_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block12_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block12_concat (Concaten  (None, 14, 14, 640)  0          ['conv4_block11_concat[0][0]',   \n",
      " ate)                                                             'conv4_block12_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block13_0_bn (BatchNorma  (None, 14, 14, 640)  2560       ['conv4_block12_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block13_0_relu (Activati  (None, 14, 14, 640)  0          ['conv4_block13_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block13_1_conv (Conv2D)  (None, 14, 14, 128)  81920       ['conv4_block13_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block13_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block13_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block13_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block13_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block13_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block13_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block13_concat (Concaten  (None, 14, 14, 672)  0          ['conv4_block12_concat[0][0]',   \n",
      " ate)                                                             'conv4_block13_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block14_0_bn (BatchNorma  (None, 14, 14, 672)  2688       ['conv4_block13_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block14_0_relu (Activati  (None, 14, 14, 672)  0          ['conv4_block14_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block14_1_conv (Conv2D)  (None, 14, 14, 128)  86016       ['conv4_block14_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block14_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block14_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block14_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block14_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block14_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block14_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block14_concat (Concaten  (None, 14, 14, 704)  0          ['conv4_block13_concat[0][0]',   \n",
      " ate)                                                             'conv4_block14_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block15_0_bn (BatchNorma  (None, 14, 14, 704)  2816       ['conv4_block14_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block15_0_relu (Activati  (None, 14, 14, 704)  0          ['conv4_block15_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block15_1_conv (Conv2D)  (None, 14, 14, 128)  90112       ['conv4_block15_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block15_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block15_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block15_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block15_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block15_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block15_1_relu[0][0]']   \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                                                                  \n",
      " conv4_block15_concat (Concaten  (None, 14, 14, 736)  0          ['conv4_block14_concat[0][0]',   \n",
      " ate)                                                             'conv4_block15_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block16_0_bn (BatchNorma  (None, 14, 14, 736)  2944       ['conv4_block15_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block16_0_relu (Activati  (None, 14, 14, 736)  0          ['conv4_block16_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block16_1_conv (Conv2D)  (None, 14, 14, 128)  94208       ['conv4_block16_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block16_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block16_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block16_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block16_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block16_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block16_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block16_concat (Concaten  (None, 14, 14, 768)  0          ['conv4_block15_concat[0][0]',   \n",
      " ate)                                                             'conv4_block16_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block17_0_bn (BatchNorma  (None, 14, 14, 768)  3072       ['conv4_block16_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block17_0_relu (Activati  (None, 14, 14, 768)  0          ['conv4_block17_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block17_1_conv (Conv2D)  (None, 14, 14, 128)  98304       ['conv4_block17_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block17_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block17_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block17_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block17_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block17_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block17_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block17_concat (Concaten  (None, 14, 14, 800)  0          ['conv4_block16_concat[0][0]',   \n",
      " ate)                                                             'conv4_block17_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block18_0_bn (BatchNorma  (None, 14, 14, 800)  3200       ['conv4_block17_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block18_0_relu (Activati  (None, 14, 14, 800)  0          ['conv4_block18_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block18_1_conv (Conv2D)  (None, 14, 14, 128)  102400      ['conv4_block18_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block18_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block18_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block18_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block18_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block18_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block18_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block18_concat (Concaten  (None, 14, 14, 832)  0          ['conv4_block17_concat[0][0]',   \n",
      " ate)                                                             'conv4_block18_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block19_0_bn (BatchNorma  (None, 14, 14, 832)  3328       ['conv4_block18_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block19_0_relu (Activati  (None, 14, 14, 832)  0          ['conv4_block19_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block19_1_conv (Conv2D)  (None, 14, 14, 128)  106496      ['conv4_block19_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block19_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block19_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block19_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block19_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block19_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block19_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block19_concat (Concaten  (None, 14, 14, 864)  0          ['conv4_block18_concat[0][0]',   \n",
      " ate)                                                             'conv4_block19_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block20_0_bn (BatchNorma  (None, 14, 14, 864)  3456       ['conv4_block19_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " conv4_block20_0_relu (Activati  (None, 14, 14, 864)  0          ['conv4_block20_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block20_1_conv (Conv2D)  (None, 14, 14, 128)  110592      ['conv4_block20_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block20_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block20_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block20_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block20_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block20_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block20_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block20_concat (Concaten  (None, 14, 14, 896)  0          ['conv4_block19_concat[0][0]',   \n",
      " ate)                                                             'conv4_block20_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block21_0_bn (BatchNorma  (None, 14, 14, 896)  3584       ['conv4_block20_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block21_0_relu (Activati  (None, 14, 14, 896)  0          ['conv4_block21_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block21_1_conv (Conv2D)  (None, 14, 14, 128)  114688      ['conv4_block21_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block21_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block21_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block21_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block21_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block21_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block21_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block21_concat (Concaten  (None, 14, 14, 928)  0          ['conv4_block20_concat[0][0]',   \n",
      " ate)                                                             'conv4_block21_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block22_0_bn (BatchNorma  (None, 14, 14, 928)  3712       ['conv4_block21_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block22_0_relu (Activati  (None, 14, 14, 928)  0          ['conv4_block22_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block22_1_conv (Conv2D)  (None, 14, 14, 128)  118784      ['conv4_block22_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block22_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block22_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block22_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block22_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block22_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block22_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block22_concat (Concaten  (None, 14, 14, 960)  0          ['conv4_block21_concat[0][0]',   \n",
      " ate)                                                             'conv4_block22_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block23_0_bn (BatchNorma  (None, 14, 14, 960)  3840       ['conv4_block22_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block23_0_relu (Activati  (None, 14, 14, 960)  0          ['conv4_block23_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block23_1_conv (Conv2D)  (None, 14, 14, 128)  122880      ['conv4_block23_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block23_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block23_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block23_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block23_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block23_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block23_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block23_concat (Concaten  (None, 14, 14, 992)  0          ['conv4_block22_concat[0][0]',   \n",
      " ate)                                                             'conv4_block23_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block24_0_bn (BatchNorma  (None, 14, 14, 992)  3968       ['conv4_block23_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block24_0_relu (Activati  (None, 14, 14, 992)  0          ['conv4_block24_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block24_1_conv (Conv2D)  (None, 14, 14, 128)  126976      ['conv4_block24_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block24_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block24_1_conv[0][0]']   \n",
      " lization)                                                                                        \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                                                                  \n",
      " conv4_block24_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block24_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block24_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block24_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block24_concat (Concaten  (None, 14, 14, 1024  0          ['conv4_block23_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block24_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block25_0_bn (BatchNorma  (None, 14, 14, 1024  4096       ['conv4_block24_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block25_0_relu (Activati  (None, 14, 14, 1024  0          ['conv4_block25_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block25_1_conv (Conv2D)  (None, 14, 14, 128)  131072      ['conv4_block25_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block25_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block25_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block25_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block25_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block25_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block25_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block25_concat (Concaten  (None, 14, 14, 1056  0          ['conv4_block24_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block25_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block26_0_bn (BatchNorma  (None, 14, 14, 1056  4224       ['conv4_block25_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block26_0_relu (Activati  (None, 14, 14, 1056  0          ['conv4_block26_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block26_1_conv (Conv2D)  (None, 14, 14, 128)  135168      ['conv4_block26_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block26_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block26_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block26_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block26_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block26_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block26_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block26_concat (Concaten  (None, 14, 14, 1088  0          ['conv4_block25_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block26_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block27_0_bn (BatchNorma  (None, 14, 14, 1088  4352       ['conv4_block26_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block27_0_relu (Activati  (None, 14, 14, 1088  0          ['conv4_block27_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block27_1_conv (Conv2D)  (None, 14, 14, 128)  139264      ['conv4_block27_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block27_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block27_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block27_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block27_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block27_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block27_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block27_concat (Concaten  (None, 14, 14, 1120  0          ['conv4_block26_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block27_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block28_0_bn (BatchNorma  (None, 14, 14, 1120  4480       ['conv4_block27_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block28_0_relu (Activati  (None, 14, 14, 1120  0          ['conv4_block28_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block28_1_conv (Conv2D)  (None, 14, 14, 128)  143360      ['conv4_block28_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block28_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block28_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block28_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block28_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block28_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block28_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block28_concat (Concaten  (None, 14, 14, 1152  0          ['conv4_block27_concat[0][0]',   \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " ate)                           )                                 'conv4_block28_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block29_0_bn (BatchNorma  (None, 14, 14, 1152  4608       ['conv4_block28_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block29_0_relu (Activati  (None, 14, 14, 1152  0          ['conv4_block29_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block29_1_conv (Conv2D)  (None, 14, 14, 128)  147456      ['conv4_block29_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block29_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block29_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block29_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block29_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block29_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block29_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block29_concat (Concaten  (None, 14, 14, 1184  0          ['conv4_block28_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block29_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block30_0_bn (BatchNorma  (None, 14, 14, 1184  4736       ['conv4_block29_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block30_0_relu (Activati  (None, 14, 14, 1184  0          ['conv4_block30_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block30_1_conv (Conv2D)  (None, 14, 14, 128)  151552      ['conv4_block30_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block30_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block30_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block30_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block30_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block30_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block30_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block30_concat (Concaten  (None, 14, 14, 1216  0          ['conv4_block29_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block30_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block31_0_bn (BatchNorma  (None, 14, 14, 1216  4864       ['conv4_block30_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block31_0_relu (Activati  (None, 14, 14, 1216  0          ['conv4_block31_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block31_1_conv (Conv2D)  (None, 14, 14, 128)  155648      ['conv4_block31_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block31_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block31_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block31_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block31_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block31_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block31_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block31_concat (Concaten  (None, 14, 14, 1248  0          ['conv4_block30_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block31_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block32_0_bn (BatchNorma  (None, 14, 14, 1248  4992       ['conv4_block31_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block32_0_relu (Activati  (None, 14, 14, 1248  0          ['conv4_block32_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block32_1_conv (Conv2D)  (None, 14, 14, 128)  159744      ['conv4_block32_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block32_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block32_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block32_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block32_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block32_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block32_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block32_concat (Concaten  (None, 14, 14, 1280  0          ['conv4_block31_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block32_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block33_0_bn (BatchNorma  (None, 14, 14, 1280  5120       ['conv4_block32_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block33_0_relu (Activati  (None, 14, 14, 1280  0          ['conv4_block33_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                                                                  \n",
      " conv4_block33_1_conv (Conv2D)  (None, 14, 14, 128)  163840      ['conv4_block33_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block33_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block33_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block33_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block33_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block33_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block33_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block33_concat (Concaten  (None, 14, 14, 1312  0          ['conv4_block32_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block33_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block34_0_bn (BatchNorma  (None, 14, 14, 1312  5248       ['conv4_block33_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block34_0_relu (Activati  (None, 14, 14, 1312  0          ['conv4_block34_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block34_1_conv (Conv2D)  (None, 14, 14, 128)  167936      ['conv4_block34_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block34_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block34_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block34_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block34_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block34_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block34_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block34_concat (Concaten  (None, 14, 14, 1344  0          ['conv4_block33_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block34_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block35_0_bn (BatchNorma  (None, 14, 14, 1344  5376       ['conv4_block34_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block35_0_relu (Activati  (None, 14, 14, 1344  0          ['conv4_block35_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block35_1_conv (Conv2D)  (None, 14, 14, 128)  172032      ['conv4_block35_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block35_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block35_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block35_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block35_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block35_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block35_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block35_concat (Concaten  (None, 14, 14, 1376  0          ['conv4_block34_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block35_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block36_0_bn (BatchNorma  (None, 14, 14, 1376  5504       ['conv4_block35_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block36_0_relu (Activati  (None, 14, 14, 1376  0          ['conv4_block36_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block36_1_conv (Conv2D)  (None, 14, 14, 128)  176128      ['conv4_block36_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block36_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block36_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block36_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block36_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block36_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block36_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block36_concat (Concaten  (None, 14, 14, 1408  0          ['conv4_block35_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block36_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block37_0_bn (BatchNorma  (None, 14, 14, 1408  5632       ['conv4_block36_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block37_0_relu (Activati  (None, 14, 14, 1408  0          ['conv4_block37_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block37_1_conv (Conv2D)  (None, 14, 14, 128)  180224      ['conv4_block37_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block37_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block37_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block37_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block37_1_bn[0][0]']     \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block37_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block37_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block37_concat (Concaten  (None, 14, 14, 1440  0          ['conv4_block36_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block37_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block38_0_bn (BatchNorma  (None, 14, 14, 1440  5760       ['conv4_block37_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block38_0_relu (Activati  (None, 14, 14, 1440  0          ['conv4_block38_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block38_1_conv (Conv2D)  (None, 14, 14, 128)  184320      ['conv4_block38_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block38_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block38_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block38_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block38_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block38_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block38_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block38_concat (Concaten  (None, 14, 14, 1472  0          ['conv4_block37_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block38_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block39_0_bn (BatchNorma  (None, 14, 14, 1472  5888       ['conv4_block38_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block39_0_relu (Activati  (None, 14, 14, 1472  0          ['conv4_block39_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block39_1_conv (Conv2D)  (None, 14, 14, 128)  188416      ['conv4_block39_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block39_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block39_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block39_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block39_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block39_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block39_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block39_concat (Concaten  (None, 14, 14, 1504  0          ['conv4_block38_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block39_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block40_0_bn (BatchNorma  (None, 14, 14, 1504  6016       ['conv4_block39_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block40_0_relu (Activati  (None, 14, 14, 1504  0          ['conv4_block40_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block40_1_conv (Conv2D)  (None, 14, 14, 128)  192512      ['conv4_block40_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block40_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block40_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block40_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block40_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block40_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block40_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block40_concat (Concaten  (None, 14, 14, 1536  0          ['conv4_block39_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block40_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block41_0_bn (BatchNorma  (None, 14, 14, 1536  6144       ['conv4_block40_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block41_0_relu (Activati  (None, 14, 14, 1536  0          ['conv4_block41_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block41_1_conv (Conv2D)  (None, 14, 14, 128)  196608      ['conv4_block41_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block41_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block41_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block41_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block41_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block41_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block41_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block41_concat (Concaten  (None, 14, 14, 1568  0          ['conv4_block40_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block41_2_conv[0][0]']   \n",
      "                                                                                                  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " conv4_block42_0_bn (BatchNorma  (None, 14, 14, 1568  6272       ['conv4_block41_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block42_0_relu (Activati  (None, 14, 14, 1568  0          ['conv4_block42_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block42_1_conv (Conv2D)  (None, 14, 14, 128)  200704      ['conv4_block42_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block42_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block42_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block42_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block42_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block42_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block42_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block42_concat (Concaten  (None, 14, 14, 1600  0          ['conv4_block41_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block42_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block43_0_bn (BatchNorma  (None, 14, 14, 1600  6400       ['conv4_block42_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block43_0_relu (Activati  (None, 14, 14, 1600  0          ['conv4_block43_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block43_1_conv (Conv2D)  (None, 14, 14, 128)  204800      ['conv4_block43_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block43_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block43_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block43_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block43_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block43_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block43_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block43_concat (Concaten  (None, 14, 14, 1632  0          ['conv4_block42_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block43_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block44_0_bn (BatchNorma  (None, 14, 14, 1632  6528       ['conv4_block43_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block44_0_relu (Activati  (None, 14, 14, 1632  0          ['conv4_block44_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block44_1_conv (Conv2D)  (None, 14, 14, 128)  208896      ['conv4_block44_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block44_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block44_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block44_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block44_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block44_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block44_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block44_concat (Concaten  (None, 14, 14, 1664  0          ['conv4_block43_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block44_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block45_0_bn (BatchNorma  (None, 14, 14, 1664  6656       ['conv4_block44_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block45_0_relu (Activati  (None, 14, 14, 1664  0          ['conv4_block45_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block45_1_conv (Conv2D)  (None, 14, 14, 128)  212992      ['conv4_block45_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block45_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block45_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block45_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block45_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block45_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block45_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block45_concat (Concaten  (None, 14, 14, 1696  0          ['conv4_block44_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block45_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block46_0_bn (BatchNorma  (None, 14, 14, 1696  6784       ['conv4_block45_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block46_0_relu (Activati  (None, 14, 14, 1696  0          ['conv4_block46_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block46_1_conv (Conv2D)  (None, 14, 14, 128)  217088      ['conv4_block46_0_relu[0][0]']   \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                                                                  \n",
      " conv4_block46_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block46_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block46_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block46_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block46_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block46_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block46_concat (Concaten  (None, 14, 14, 1728  0          ['conv4_block45_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block46_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block47_0_bn (BatchNorma  (None, 14, 14, 1728  6912       ['conv4_block46_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block47_0_relu (Activati  (None, 14, 14, 1728  0          ['conv4_block47_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block47_1_conv (Conv2D)  (None, 14, 14, 128)  221184      ['conv4_block47_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block47_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block47_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block47_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block47_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block47_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block47_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block47_concat (Concaten  (None, 14, 14, 1760  0          ['conv4_block46_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block47_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block48_0_bn (BatchNorma  (None, 14, 14, 1760  7040       ['conv4_block47_concat[0][0]']   \n",
      " lization)                      )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block48_0_relu (Activati  (None, 14, 14, 1760  0          ['conv4_block48_0_bn[0][0]']     \n",
      " on)                            )                                                                 \n",
      "                                                                                                  \n",
      " conv4_block48_1_conv (Conv2D)  (None, 14, 14, 128)  225280      ['conv4_block48_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block48_1_bn (BatchNorma  (None, 14, 14, 128)  512        ['conv4_block48_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv4_block48_1_relu (Activati  (None, 14, 14, 128)  0          ['conv4_block48_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv4_block48_2_conv (Conv2D)  (None, 14, 14, 32)   36864       ['conv4_block48_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv4_block48_concat (Concaten  (None, 14, 14, 1792  0          ['conv4_block47_concat[0][0]',   \n",
      " ate)                           )                                 'conv4_block48_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " pool4_bn (BatchNormalization)  (None, 14, 14, 1792  7168        ['conv4_block48_concat[0][0]']   \n",
      "                                )                                                                 \n",
      "                                                                                                  \n",
      " pool4_relu (Activation)        (None, 14, 14, 1792  0           ['pool4_bn[0][0]']               \n",
      "                                )                                                                 \n",
      "                                                                                                  \n",
      " pool4_conv (Conv2D)            (None, 14, 14, 896)  1605632     ['pool4_relu[0][0]']             \n",
      "                                                                                                  \n",
      " pool4_pool (AveragePooling2D)  (None, 7, 7, 896)    0           ['pool4_conv[0][0]']             \n",
      "                                                                                                  \n",
      " conv5_block1_0_bn (BatchNormal  (None, 7, 7, 896)   3584        ['pool4_pool[0][0]']             \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv5_block1_0_relu (Activatio  (None, 7, 7, 896)   0           ['conv5_block1_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv5_block1_1_conv (Conv2D)   (None, 7, 7, 128)    114688      ['conv5_block1_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block1_1_bn (BatchNormal  (None, 7, 7, 128)   512         ['conv5_block1_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv5_block1_1_relu (Activatio  (None, 7, 7, 128)   0           ['conv5_block1_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv5_block1_2_conv (Conv2D)   (None, 7, 7, 32)     36864       ['conv5_block1_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block1_concat (Concatena  (None, 7, 7, 928)   0           ['pool4_pool[0][0]',             \n",
      " te)                                                              'conv5_block1_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block2_0_bn (BatchNormal  (None, 7, 7, 928)   3712        ['conv5_block1_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv5_block2_0_relu (Activatio  (None, 7, 7, 928)   0           ['conv5_block2_0_bn[0][0]']      \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv5_block2_1_conv (Conv2D)   (None, 7, 7, 128)    118784      ['conv5_block2_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block2_1_bn (BatchNormal  (None, 7, 7, 128)   512         ['conv5_block2_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv5_block2_1_relu (Activatio  (None, 7, 7, 128)   0           ['conv5_block2_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv5_block2_2_conv (Conv2D)   (None, 7, 7, 32)     36864       ['conv5_block2_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block2_concat (Concatena  (None, 7, 7, 960)   0           ['conv5_block1_concat[0][0]',    \n",
      " te)                                                              'conv5_block2_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block3_0_bn (BatchNormal  (None, 7, 7, 960)   3840        ['conv5_block2_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv5_block3_0_relu (Activatio  (None, 7, 7, 960)   0           ['conv5_block3_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv5_block3_1_conv (Conv2D)   (None, 7, 7, 128)    122880      ['conv5_block3_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block3_1_bn (BatchNormal  (None, 7, 7, 128)   512         ['conv5_block3_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv5_block3_1_relu (Activatio  (None, 7, 7, 128)   0           ['conv5_block3_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv5_block3_2_conv (Conv2D)   (None, 7, 7, 32)     36864       ['conv5_block3_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block3_concat (Concatena  (None, 7, 7, 992)   0           ['conv5_block2_concat[0][0]',    \n",
      " te)                                                              'conv5_block3_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block4_0_bn (BatchNormal  (None, 7, 7, 992)   3968        ['conv5_block3_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv5_block4_0_relu (Activatio  (None, 7, 7, 992)   0           ['conv5_block4_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv5_block4_1_conv (Conv2D)   (None, 7, 7, 128)    126976      ['conv5_block4_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block4_1_bn (BatchNormal  (None, 7, 7, 128)   512         ['conv5_block4_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv5_block4_1_relu (Activatio  (None, 7, 7, 128)   0           ['conv5_block4_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv5_block4_2_conv (Conv2D)   (None, 7, 7, 32)     36864       ['conv5_block4_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block4_concat (Concatena  (None, 7, 7, 1024)  0           ['conv5_block3_concat[0][0]',    \n",
      " te)                                                              'conv5_block4_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block5_0_bn (BatchNormal  (None, 7, 7, 1024)  4096        ['conv5_block4_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv5_block5_0_relu (Activatio  (None, 7, 7, 1024)  0           ['conv5_block5_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv5_block5_1_conv (Conv2D)   (None, 7, 7, 128)    131072      ['conv5_block5_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block5_1_bn (BatchNormal  (None, 7, 7, 128)   512         ['conv5_block5_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv5_block5_1_relu (Activatio  (None, 7, 7, 128)   0           ['conv5_block5_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv5_block5_2_conv (Conv2D)   (None, 7, 7, 32)     36864       ['conv5_block5_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block5_concat (Concatena  (None, 7, 7, 1056)  0           ['conv5_block4_concat[0][0]',    \n",
      " te)                                                              'conv5_block5_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block6_0_bn (BatchNormal  (None, 7, 7, 1056)  4224        ['conv5_block5_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv5_block6_0_relu (Activatio  (None, 7, 7, 1056)  0           ['conv5_block6_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv5_block6_1_conv (Conv2D)   (None, 7, 7, 128)    135168      ['conv5_block6_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block6_1_bn (BatchNormal  (None, 7, 7, 128)   512         ['conv5_block6_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " conv5_block6_1_relu (Activatio  (None, 7, 7, 128)   0           ['conv5_block6_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv5_block6_2_conv (Conv2D)   (None, 7, 7, 32)     36864       ['conv5_block6_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block6_concat (Concatena  (None, 7, 7, 1088)  0           ['conv5_block5_concat[0][0]',    \n",
      " te)                                                              'conv5_block6_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block7_0_bn (BatchNormal  (None, 7, 7, 1088)  4352        ['conv5_block6_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv5_block7_0_relu (Activatio  (None, 7, 7, 1088)  0           ['conv5_block7_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv5_block7_1_conv (Conv2D)   (None, 7, 7, 128)    139264      ['conv5_block7_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block7_1_bn (BatchNormal  (None, 7, 7, 128)   512         ['conv5_block7_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv5_block7_1_relu (Activatio  (None, 7, 7, 128)   0           ['conv5_block7_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv5_block7_2_conv (Conv2D)   (None, 7, 7, 32)     36864       ['conv5_block7_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block7_concat (Concatena  (None, 7, 7, 1120)  0           ['conv5_block6_concat[0][0]',    \n",
      " te)                                                              'conv5_block7_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block8_0_bn (BatchNormal  (None, 7, 7, 1120)  4480        ['conv5_block7_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv5_block8_0_relu (Activatio  (None, 7, 7, 1120)  0           ['conv5_block8_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv5_block8_1_conv (Conv2D)   (None, 7, 7, 128)    143360      ['conv5_block8_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block8_1_bn (BatchNormal  (None, 7, 7, 128)   512         ['conv5_block8_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv5_block8_1_relu (Activatio  (None, 7, 7, 128)   0           ['conv5_block8_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv5_block8_2_conv (Conv2D)   (None, 7, 7, 32)     36864       ['conv5_block8_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block8_concat (Concatena  (None, 7, 7, 1152)  0           ['conv5_block7_concat[0][0]',    \n",
      " te)                                                              'conv5_block8_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block9_0_bn (BatchNormal  (None, 7, 7, 1152)  4608        ['conv5_block8_concat[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv5_block9_0_relu (Activatio  (None, 7, 7, 1152)  0           ['conv5_block9_0_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv5_block9_1_conv (Conv2D)   (None, 7, 7, 128)    147456      ['conv5_block9_0_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block9_1_bn (BatchNormal  (None, 7, 7, 128)   512         ['conv5_block9_1_conv[0][0]']    \n",
      " ization)                                                                                         \n",
      "                                                                                                  \n",
      " conv5_block9_1_relu (Activatio  (None, 7, 7, 128)   0           ['conv5_block9_1_bn[0][0]']      \n",
      " n)                                                                                               \n",
      "                                                                                                  \n",
      " conv5_block9_2_conv (Conv2D)   (None, 7, 7, 32)     36864       ['conv5_block9_1_relu[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block9_concat (Concatena  (None, 7, 7, 1184)  0           ['conv5_block8_concat[0][0]',    \n",
      " te)                                                              'conv5_block9_2_conv[0][0]']    \n",
      "                                                                                                  \n",
      " conv5_block10_0_bn (BatchNorma  (None, 7, 7, 1184)  4736        ['conv5_block9_concat[0][0]']    \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block10_0_relu (Activati  (None, 7, 7, 1184)  0           ['conv5_block10_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block10_1_conv (Conv2D)  (None, 7, 7, 128)    151552      ['conv5_block10_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block10_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block10_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block10_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block10_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block10_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block10_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block10_concat (Concaten  (None, 7, 7, 1216)  0           ['conv5_block9_concat[0][0]',    \n",
      " ate)                                                             'conv5_block10_2_conv[0][0]']   \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                                                                  \n",
      " conv5_block11_0_bn (BatchNorma  (None, 7, 7, 1216)  4864        ['conv5_block10_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block11_0_relu (Activati  (None, 7, 7, 1216)  0           ['conv5_block11_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block11_1_conv (Conv2D)  (None, 7, 7, 128)    155648      ['conv5_block11_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block11_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block11_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block11_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block11_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block11_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block11_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block11_concat (Concaten  (None, 7, 7, 1248)  0           ['conv5_block10_concat[0][0]',   \n",
      " ate)                                                             'conv5_block11_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block12_0_bn (BatchNorma  (None, 7, 7, 1248)  4992        ['conv5_block11_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block12_0_relu (Activati  (None, 7, 7, 1248)  0           ['conv5_block12_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block12_1_conv (Conv2D)  (None, 7, 7, 128)    159744      ['conv5_block12_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block12_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block12_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block12_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block12_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block12_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block12_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block12_concat (Concaten  (None, 7, 7, 1280)  0           ['conv5_block11_concat[0][0]',   \n",
      " ate)                                                             'conv5_block12_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block13_0_bn (BatchNorma  (None, 7, 7, 1280)  5120        ['conv5_block12_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block13_0_relu (Activati  (None, 7, 7, 1280)  0           ['conv5_block13_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block13_1_conv (Conv2D)  (None, 7, 7, 128)    163840      ['conv5_block13_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block13_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block13_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block13_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block13_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block13_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block13_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block13_concat (Concaten  (None, 7, 7, 1312)  0           ['conv5_block12_concat[0][0]',   \n",
      " ate)                                                             'conv5_block13_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block14_0_bn (BatchNorma  (None, 7, 7, 1312)  5248        ['conv5_block13_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block14_0_relu (Activati  (None, 7, 7, 1312)  0           ['conv5_block14_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block14_1_conv (Conv2D)  (None, 7, 7, 128)    167936      ['conv5_block14_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block14_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block14_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block14_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block14_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block14_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block14_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block14_concat (Concaten  (None, 7, 7, 1344)  0           ['conv5_block13_concat[0][0]',   \n",
      " ate)                                                             'conv5_block14_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block15_0_bn (BatchNorma  (None, 7, 7, 1344)  5376        ['conv5_block14_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block15_0_relu (Activati  (None, 7, 7, 1344)  0           ['conv5_block15_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " conv5_block15_1_conv (Conv2D)  (None, 7, 7, 128)    172032      ['conv5_block15_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block15_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block15_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block15_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block15_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block15_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block15_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block15_concat (Concaten  (None, 7, 7, 1376)  0           ['conv5_block14_concat[0][0]',   \n",
      " ate)                                                             'conv5_block15_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block16_0_bn (BatchNorma  (None, 7, 7, 1376)  5504        ['conv5_block15_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block16_0_relu (Activati  (None, 7, 7, 1376)  0           ['conv5_block16_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block16_1_conv (Conv2D)  (None, 7, 7, 128)    176128      ['conv5_block16_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block16_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block16_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block16_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block16_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block16_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block16_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block16_concat (Concaten  (None, 7, 7, 1408)  0           ['conv5_block15_concat[0][0]',   \n",
      " ate)                                                             'conv5_block16_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block17_0_bn (BatchNorma  (None, 7, 7, 1408)  5632        ['conv5_block16_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block17_0_relu (Activati  (None, 7, 7, 1408)  0           ['conv5_block17_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block17_1_conv (Conv2D)  (None, 7, 7, 128)    180224      ['conv5_block17_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block17_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block17_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block17_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block17_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block17_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block17_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block17_concat (Concaten  (None, 7, 7, 1440)  0           ['conv5_block16_concat[0][0]',   \n",
      " ate)                                                             'conv5_block17_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block18_0_bn (BatchNorma  (None, 7, 7, 1440)  5760        ['conv5_block17_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block18_0_relu (Activati  (None, 7, 7, 1440)  0           ['conv5_block18_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block18_1_conv (Conv2D)  (None, 7, 7, 128)    184320      ['conv5_block18_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block18_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block18_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block18_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block18_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block18_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block18_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block18_concat (Concaten  (None, 7, 7, 1472)  0           ['conv5_block17_concat[0][0]',   \n",
      " ate)                                                             'conv5_block18_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block19_0_bn (BatchNorma  (None, 7, 7, 1472)  5888        ['conv5_block18_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block19_0_relu (Activati  (None, 7, 7, 1472)  0           ['conv5_block19_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block19_1_conv (Conv2D)  (None, 7, 7, 128)    188416      ['conv5_block19_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block19_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block19_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block19_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block19_1_bn[0][0]']     \n",
      " on)                                                                                              \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                                                                  \n",
      " conv5_block19_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block19_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block19_concat (Concaten  (None, 7, 7, 1504)  0           ['conv5_block18_concat[0][0]',   \n",
      " ate)                                                             'conv5_block19_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block20_0_bn (BatchNorma  (None, 7, 7, 1504)  6016        ['conv5_block19_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block20_0_relu (Activati  (None, 7, 7, 1504)  0           ['conv5_block20_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block20_1_conv (Conv2D)  (None, 7, 7, 128)    192512      ['conv5_block20_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block20_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block20_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block20_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block20_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block20_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block20_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block20_concat (Concaten  (None, 7, 7, 1536)  0           ['conv5_block19_concat[0][0]',   \n",
      " ate)                                                             'conv5_block20_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block21_0_bn (BatchNorma  (None, 7, 7, 1536)  6144        ['conv5_block20_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block21_0_relu (Activati  (None, 7, 7, 1536)  0           ['conv5_block21_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block21_1_conv (Conv2D)  (None, 7, 7, 128)    196608      ['conv5_block21_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block21_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block21_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block21_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block21_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block21_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block21_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block21_concat (Concaten  (None, 7, 7, 1568)  0           ['conv5_block20_concat[0][0]',   \n",
      " ate)                                                             'conv5_block21_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block22_0_bn (BatchNorma  (None, 7, 7, 1568)  6272        ['conv5_block21_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block22_0_relu (Activati  (None, 7, 7, 1568)  0           ['conv5_block22_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block22_1_conv (Conv2D)  (None, 7, 7, 128)    200704      ['conv5_block22_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block22_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block22_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block22_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block22_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block22_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block22_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block22_concat (Concaten  (None, 7, 7, 1600)  0           ['conv5_block21_concat[0][0]',   \n",
      " ate)                                                             'conv5_block22_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block23_0_bn (BatchNorma  (None, 7, 7, 1600)  6400        ['conv5_block22_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block23_0_relu (Activati  (None, 7, 7, 1600)  0           ['conv5_block23_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block23_1_conv (Conv2D)  (None, 7, 7, 128)    204800      ['conv5_block23_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block23_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block23_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block23_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block23_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block23_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block23_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block23_concat (Concaten  (None, 7, 7, 1632)  0           ['conv5_block22_concat[0][0]',   \n",
      " ate)                                                             'conv5_block23_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block24_0_bn (BatchNorma  (None, 7, 7, 1632)  6528        ['conv5_block23_concat[0][0]']   \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block24_0_relu (Activati  (None, 7, 7, 1632)  0           ['conv5_block24_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block24_1_conv (Conv2D)  (None, 7, 7, 128)    208896      ['conv5_block24_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block24_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block24_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block24_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block24_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block24_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block24_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block24_concat (Concaten  (None, 7, 7, 1664)  0           ['conv5_block23_concat[0][0]',   \n",
      " ate)                                                             'conv5_block24_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block25_0_bn (BatchNorma  (None, 7, 7, 1664)  6656        ['conv5_block24_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block25_0_relu (Activati  (None, 7, 7, 1664)  0           ['conv5_block25_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block25_1_conv (Conv2D)  (None, 7, 7, 128)    212992      ['conv5_block25_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block25_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block25_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block25_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block25_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block25_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block25_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block25_concat (Concaten  (None, 7, 7, 1696)  0           ['conv5_block24_concat[0][0]',   \n",
      " ate)                                                             'conv5_block25_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block26_0_bn (BatchNorma  (None, 7, 7, 1696)  6784        ['conv5_block25_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block26_0_relu (Activati  (None, 7, 7, 1696)  0           ['conv5_block26_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block26_1_conv (Conv2D)  (None, 7, 7, 128)    217088      ['conv5_block26_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block26_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block26_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block26_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block26_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block26_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block26_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block26_concat (Concaten  (None, 7, 7, 1728)  0           ['conv5_block25_concat[0][0]',   \n",
      " ate)                                                             'conv5_block26_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block27_0_bn (BatchNorma  (None, 7, 7, 1728)  6912        ['conv5_block26_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block27_0_relu (Activati  (None, 7, 7, 1728)  0           ['conv5_block27_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block27_1_conv (Conv2D)  (None, 7, 7, 128)    221184      ['conv5_block27_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block27_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block27_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block27_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block27_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block27_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block27_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block27_concat (Concaten  (None, 7, 7, 1760)  0           ['conv5_block26_concat[0][0]',   \n",
      " ate)                                                             'conv5_block27_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block28_0_bn (BatchNorma  (None, 7, 7, 1760)  7040        ['conv5_block27_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block28_0_relu (Activati  (None, 7, 7, 1760)  0           ['conv5_block28_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block28_1_conv (Conv2D)  (None, 7, 7, 128)    225280      ['conv5_block28_0_relu[0][0]']   \n",
      "                                                                                                  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " conv5_block28_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block28_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block28_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block28_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block28_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block28_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block28_concat (Concaten  (None, 7, 7, 1792)  0           ['conv5_block27_concat[0][0]',   \n",
      " ate)                                                             'conv5_block28_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block29_0_bn (BatchNorma  (None, 7, 7, 1792)  7168        ['conv5_block28_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block29_0_relu (Activati  (None, 7, 7, 1792)  0           ['conv5_block29_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block29_1_conv (Conv2D)  (None, 7, 7, 128)    229376      ['conv5_block29_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block29_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block29_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block29_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block29_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block29_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block29_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block29_concat (Concaten  (None, 7, 7, 1824)  0           ['conv5_block28_concat[0][0]',   \n",
      " ate)                                                             'conv5_block29_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block30_0_bn (BatchNorma  (None, 7, 7, 1824)  7296        ['conv5_block29_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block30_0_relu (Activati  (None, 7, 7, 1824)  0           ['conv5_block30_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block30_1_conv (Conv2D)  (None, 7, 7, 128)    233472      ['conv5_block30_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block30_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block30_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block30_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block30_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block30_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block30_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block30_concat (Concaten  (None, 7, 7, 1856)  0           ['conv5_block29_concat[0][0]',   \n",
      " ate)                                                             'conv5_block30_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block31_0_bn (BatchNorma  (None, 7, 7, 1856)  7424        ['conv5_block30_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block31_0_relu (Activati  (None, 7, 7, 1856)  0           ['conv5_block31_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block31_1_conv (Conv2D)  (None, 7, 7, 128)    237568      ['conv5_block31_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block31_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block31_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block31_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block31_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block31_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block31_1_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block31_concat (Concaten  (None, 7, 7, 1888)  0           ['conv5_block30_concat[0][0]',   \n",
      " ate)                                                             'conv5_block31_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block32_0_bn (BatchNorma  (None, 7, 7, 1888)  7552        ['conv5_block31_concat[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block32_0_relu (Activati  (None, 7, 7, 1888)  0           ['conv5_block32_0_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block32_1_conv (Conv2D)  (None, 7, 7, 128)    241664      ['conv5_block32_0_relu[0][0]']   \n",
      "                                                                                                  \n",
      " conv5_block32_1_bn (BatchNorma  (None, 7, 7, 128)   512         ['conv5_block32_1_conv[0][0]']   \n",
      " lization)                                                                                        \n",
      "                                                                                                  \n",
      " conv5_block32_1_relu (Activati  (None, 7, 7, 128)   0           ['conv5_block32_1_bn[0][0]']     \n",
      " on)                                                                                              \n",
      "                                                                                                  \n",
      " conv5_block32_2_conv (Conv2D)  (None, 7, 7, 32)     36864       ['conv5_block32_1_relu[0][0]']   \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                                                                  \n",
      " conv5_block32_concat (Concaten  (None, 7, 7, 1920)  0           ['conv5_block31_concat[0][0]',   \n",
      " ate)                                                             'conv5_block32_2_conv[0][0]']   \n",
      "                                                                                                  \n",
      " bn (BatchNormalization)        (None, 7, 7, 1920)   7680        ['conv5_block32_concat[0][0]']   \n",
      "                                                                                                  \n",
      " relu (Activation)              (None, 7, 7, 1920)   0           ['bn[0][0]']                     \n",
      "                                                                                                  \n",
      "==================================================================================================\n",
      "Total params: 18,321,984\n",
      "Trainable params: 18,092,928\n",
      "Non-trainable params: 229,056\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "conv_base = tf.keras.applications.DenseNet201(\n",
    "    include_top=False,\n",
    "    weights=\"imagenet\",\n",
    "    input_shape=(img_width, img_height, 3),\n",
    ")\n",
    "\n",
    "\n",
    "conv_base.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "f8e58b55",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 2623 images belonging to 3 classes.\n",
      "Found 660 images belonging to 3 classes.\n"
     ]
    }
   ],
   "source": [
    "import os, shutil\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "\n",
    "\n",
    "\n",
    "datagen = ImageDataGenerator(rescale=1./255)\n",
    "batch_size = 32\n",
    "\n",
    "def extract_features(directory, sample_count):\n",
    "    features = np.zeros(shape=(sample_count, 7, 7, 1920)) # Must be equal to the output of the convolutional base\n",
    "    labels = np.zeros(shape=(sample_count,3))\n",
    "    # Preprocess data\n",
    "    generator = datagen.flow_from_directory(directory,\n",
    "                                            target_size=(img_width,img_height),\n",
    "                                            batch_size = batch_size,\n",
    "                                            class_mode='categorical')\n",
    " \n",
    "    i = 0\n",
    "    for inputs_batch, labels_batch in generator:\n",
    "        features_batch = conv_base.predict(inputs_batch)\n",
    "        features[i * batch_size: (i + 1) * batch_size] = features_batch\n",
    "        labels[i * batch_size: (i + 1) * batch_size] = labels_batch\n",
    "        i += 1\n",
    "        if i * batch_size >= sample_count:\n",
    "            break\n",
    "    return features, labels\n",
    "    \n",
    "train_features, train_labels = extract_features(train_dir, 2623) \n",
    "validation_features, validation_labels = extract_features(valid_dir, 660)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "dc690b6c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " conv2d (Conv2D)             (None, 5, 5, 32)          552992    \n",
      "                                                                 \n",
      " max_pooling2d (MaxPooling2D  (None, 2, 2, 32)         0         \n",
      " )                                                               \n",
      "                                                                 \n",
      " dropout (Dropout)           (None, 2, 2, 32)          0         \n",
      "                                                                 \n",
      " flatten (Flatten)           (None, 128)               0         \n",
      "                                                                 \n",
      " dense (Dense)               (None, 32)                4128      \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 3)                 99        \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 557,219\n",
      "Trainable params: 557,219\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "epochs = 40\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "# Create the model\n",
    "model = Sequential()\n",
    "model.add(Conv2D(32, kernel_size=(3, 3), activation='relu', input_shape=(7,7,1920)))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "model.add(Dropout(0.25))\n",
    "\n",
    "model.add(Flatten())\n",
    "model.add(Dense(32, activation='relu'))\n",
    "model.add(Dense(3, activation='softmax'))\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "f60886d6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.3364 - acc: 0.8883\n",
      "Epoch 1: val_acc improved from -inf to 0.92424, saving model to model-001-0.888296-0.924242.h5\n",
      "82/82 [==============================] - 2s 24ms/step - loss: 0.3364 - acc: 0.8883 - val_loss: 0.2340 - val_acc: 0.9242\n",
      "Epoch 2/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.1103 - acc: 0.9604\n",
      "Epoch 2: val_acc improved from 0.92424 to 0.93030, saving model to model-002-0.960351-0.930303.h5\n",
      "82/82 [==============================] - 2s 22ms/step - loss: 0.1103 - acc: 0.9604 - val_loss: 0.1988 - val_acc: 0.9303\n",
      "Epoch 3/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0683 - acc: 0.9783\n",
      "Epoch 3: val_acc improved from 0.93030 to 0.93788, saving model to model-003-0.978269-0.937879.h5\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0683 - acc: 0.9783 - val_loss: 0.2283 - val_acc: 0.9379\n",
      "Epoch 4/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0614 - acc: 0.9787\n",
      "Epoch 4: val_acc improved from 0.93788 to 0.94242, saving model to model-004-0.978650-0.942424.h5\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0614 - acc: 0.9787 - val_loss: 0.1404 - val_acc: 0.9424\n",
      "Epoch 5/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0487 - acc: 0.9836\n",
      "Epoch 5: val_acc did not improve from 0.94242\n",
      "82/82 [==============================] - 2s 20ms/step - loss: 0.0487 - acc: 0.9836 - val_loss: 0.2741 - val_acc: 0.9333\n",
      "Epoch 6/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0570 - acc: 0.9802\n",
      "Epoch 6: val_acc improved from 0.94242 to 0.94545, saving model to model-006-0.980175-0.945455.h5\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0570 - acc: 0.9802 - val_loss: 0.1653 - val_acc: 0.9455\n",
      "Epoch 7/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0332 - acc: 0.9863\n",
      "Epoch 7: val_acc did not improve from 0.94545\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0332 - acc: 0.9863 - val_loss: 0.2180 - val_acc: 0.9379\n",
      "Epoch 8/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0207 - acc: 0.9939\n",
      "Epoch 8: val_acc did not improve from 0.94545\n",
      "82/82 [==============================] - 2s 22ms/step - loss: 0.0207 - acc: 0.9939 - val_loss: 0.2100 - val_acc: 0.9455\n",
      "Epoch 9/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0245 - acc: 0.9905\n",
      "Epoch 9: val_acc improved from 0.94545 to 0.95000, saving model to model-009-0.990469-0.950000.h5\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0245 - acc: 0.9905 - val_loss: 0.1602 - val_acc: 0.9500\n",
      "Epoch 10/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0181 - acc: 0.9939\n",
      "Epoch 10: val_acc did not improve from 0.95000\n",
      "82/82 [==============================] - 2s 22ms/step - loss: 0.0181 - acc: 0.9939 - val_loss: 0.1831 - val_acc: 0.9485\n",
      "Epoch 11/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0238 - acc: 0.9905\n",
      "Epoch 11: val_acc improved from 0.95000 to 0.95606, saving model to model-011-0.990469-0.956061.h5\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0238 - acc: 0.9905 - val_loss: 0.1476 - val_acc: 0.9561\n",
      "Epoch 12/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0293 - acc: 0.9901\n",
      "Epoch 12: val_acc did not improve from 0.95606\n",
      "82/82 [==============================] - 2s 22ms/step - loss: 0.0293 - acc: 0.9901 - val_loss: 0.1385 - val_acc: 0.9515\n",
      "Epoch 13/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0216 - acc: 0.9928\n",
      "Epoch 13: val_acc did not improve from 0.95606\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0216 - acc: 0.9928 - val_loss: 0.1613 - val_acc: 0.9545\n",
      "Epoch 14/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0147 - acc: 0.9947\n",
      "Epoch 14: val_acc did not improve from 0.95606\n",
      "82/82 [==============================] - 2s 23ms/step - loss: 0.0147 - acc: 0.9947 - val_loss: 0.1307 - val_acc: 0.9545\n",
      "Epoch 15/40\n",
      "81/82 [============================>.] - ETA: 0s - loss: 0.0101 - acc: 0.9973\n",
      "Epoch 15: val_acc did not improve from 0.95606\n",
      "82/82 [==============================] - 2s 25ms/step - loss: 0.0100 - acc: 0.9973 - val_loss: 0.1854 - val_acc: 0.9485\n",
      "Epoch 16/40\n",
      "81/82 [============================>.] - ETA: 0s - loss: 0.0105 - acc: 0.9969\n",
      "Epoch 16: val_acc did not improve from 0.95606\n",
      "82/82 [==============================] - 2s 22ms/step - loss: 0.0103 - acc: 0.9970 - val_loss: 0.2460 - val_acc: 0.9439\n",
      "Epoch 17/40\n",
      "81/82 [============================>.] - ETA: 0s - loss: 0.0093 - acc: 0.9965\n",
      "Epoch 17: val_acc did not improve from 0.95606\n",
      "82/82 [==============================] - 2s 23ms/step - loss: 0.0092 - acc: 0.9966 - val_loss: 0.2174 - val_acc: 0.9500\n",
      "Epoch 18/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0195 - acc: 0.9939\n",
      "Epoch 18: val_acc did not improve from 0.95606\n",
      "82/82 [==============================] - 2s 23ms/step - loss: 0.0195 - acc: 0.9939 - val_loss: 0.3503 - val_acc: 0.9227\n",
      "Epoch 19/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0166 - acc: 0.9954\n",
      "Epoch 19: val_acc improved from 0.95606 to 0.96061, saving model to model-019-0.995425-0.960606.h5\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0166 - acc: 0.9954 - val_loss: 0.1000 - val_acc: 0.9606\n",
      "Epoch 20/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0197 - acc: 0.9935\n",
      "Epoch 20: val_acc did not improve from 0.96061\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0197 - acc: 0.9935 - val_loss: 0.2810 - val_acc: 0.9409\n",
      "Epoch 21/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0362 - acc: 0.9867\n",
      "Epoch 21: val_acc did not improve from 0.96061\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0362 - acc: 0.9867 - val_loss: 0.1676 - val_acc: 0.9470\n",
      "Epoch 22/40\n",
      "81/82 [============================>.] - ETA: 0s - loss: 0.0401 - acc: 0.9861\n",
      "Epoch 22: val_acc did not improve from 0.96061\n",
      "82/82 [==============================] - 2s 22ms/step - loss: 0.0405 - acc: 0.9859 - val_loss: 0.2658 - val_acc: 0.9333\n",
      "Epoch 23/40\n",
      "81/82 [============================>.] - ETA: 0s - loss: 0.0483 - acc: 0.9884\n",
      "Epoch 23: val_acc did not improve from 0.96061\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0477 - acc: 0.9886 - val_loss: 0.1738 - val_acc: 0.9545\n",
      "Epoch 24/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0195 - acc: 0.9935\n",
      "Epoch 24: val_acc did not improve from 0.96061\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0195 - acc: 0.9935 - val_loss: 0.1757 - val_acc: 0.9500\n",
      "Epoch 25/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0161 - acc: 0.9935\n",
      "Epoch 25: val_acc did not improve from 0.96061\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0161 - acc: 0.9935 - val_loss: 0.3145 - val_acc: 0.9424\n",
      "Epoch 26/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0152 - acc: 0.9962\n",
      "Epoch 26: val_acc did not improve from 0.96061\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0152 - acc: 0.9962 - val_loss: 0.1899 - val_acc: 0.9515\n",
      "Epoch 27/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0107 - acc: 0.9958\n",
      "Epoch 27: val_acc did not improve from 0.96061\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0107 - acc: 0.9958 - val_loss: 0.1718 - val_acc: 0.9591\n",
      "Epoch 28/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0075 - acc: 0.9985\n",
      "Epoch 28: val_acc did not improve from 0.96061\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0075 - acc: 0.9985 - val_loss: 0.2222 - val_acc: 0.9561\n",
      "Epoch 29/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0065 - acc: 0.9981\n",
      "Epoch 29: val_acc did not improve from 0.96061\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0065 - acc: 0.9981 - val_loss: 0.1649 - val_acc: 0.9576\n",
      "Epoch 30/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0080 - acc: 0.9962\n",
      "Epoch 30: val_acc did not improve from 0.96061\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0080 - acc: 0.9962 - val_loss: 0.3379 - val_acc: 0.9439\n",
      "Epoch 31/40\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "82/82 [==============================] - ETA: 0s - loss: 0.0131 - acc: 0.9958\n",
      "Epoch 31: val_acc did not improve from 0.96061\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0131 - acc: 0.9958 - val_loss: 0.1486 - val_acc: 0.9530\n",
      "Epoch 32/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0230 - acc: 0.9924\n",
      "Epoch 32: val_acc did not improve from 0.96061\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0230 - acc: 0.9924 - val_loss: 0.1906 - val_acc: 0.9424\n",
      "Epoch 33/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0163 - acc: 0.9939\n",
      "Epoch 33: val_acc improved from 0.96061 to 0.96364, saving model to model-033-0.993900-0.963636.h5\n",
      "82/82 [==============================] - 2s 22ms/step - loss: 0.0163 - acc: 0.9939 - val_loss: 0.1554 - val_acc: 0.9636\n",
      "Epoch 34/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0152 - acc: 0.9943\n",
      "Epoch 34: val_acc did not improve from 0.96364\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0152 - acc: 0.9943 - val_loss: 0.1477 - val_acc: 0.9636\n",
      "Epoch 35/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0112 - acc: 0.9970\n",
      "Epoch 35: val_acc did not improve from 0.96364\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0112 - acc: 0.9970 - val_loss: 0.2249 - val_acc: 0.9530\n",
      "Epoch 36/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0081 - acc: 0.9977\n",
      "Epoch 36: val_acc did not improve from 0.96364\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0081 - acc: 0.9977 - val_loss: 0.1771 - val_acc: 0.9530\n",
      "Epoch 37/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0061 - acc: 0.9981\n",
      "Epoch 37: val_acc did not improve from 0.96364\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0061 - acc: 0.9981 - val_loss: 0.2277 - val_acc: 0.9545\n",
      "Epoch 38/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0172 - acc: 0.9935\n",
      "Epoch 38: val_acc did not improve from 0.96364\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0172 - acc: 0.9935 - val_loss: 0.2020 - val_acc: 0.9591\n",
      "Epoch 39/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.0114 - acc: 0.9970\n",
      "Epoch 39: val_acc did not improve from 0.96364\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0114 - acc: 0.9970 - val_loss: 0.2240 - val_acc: 0.9485\n",
      "Epoch 40/40\n",
      "79/82 [===========================>..] - ETA: 0s - loss: 0.0179 - acc: 0.9941\n",
      "Epoch 40: val_acc did not improve from 0.96364\n",
      "82/82 [==============================] - 2s 21ms/step - loss: 0.0177 - acc: 0.9943 - val_loss: 0.2409 - val_acc: 0.9530\n"
     ]
    }
   ],
   "source": [
    "from keras.callbacks import ModelCheckpoint\n",
    "checkpoint = ModelCheckpoint('model-{epoch:03d}-{acc:03f}-{val_acc:03f}.h5', verbose=1, monitor='val_acc',save_best_only=True, mode='auto')  \n",
    "\n",
    "\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "model.compile(optimizer='adam',\n",
    "              loss='categorical_crossentropy',\n",
    "              metrics=['acc'])\n",
    "\n",
    "\n",
    "history = model.fit(train_features, train_labels,\n",
    "                    epochs=epochs,\n",
    "                    batch_size=batch_size, \n",
    "                    callbacks=[checkpoint],\n",
    "                    validation_data=(validation_features, validation_labels))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "18f697c9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot results\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "acc = history.history['acc']\n",
    "\n",
    "val_acc = history.history['val_acc']\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "epochs = range(1, len(acc)+1)\n",
    "\n",
    "plt.plot(epochs, acc, 'g', label='Training accuracy')\n",
    "plt.plot(epochs, val_acc, 'r', label='Validation accuracy')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\n",
    "\n",
    "plt.figure()\n",
    "\n",
    "plt.plot(epochs, loss, 'g', label='Training loss')\n",
    "plt.plot(epochs, val_loss, 'r', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "\n",
    "\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "a5f04165",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import confusion_matrix, classification_report"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "05bf01cc",
   "metadata": {},
   "outputs": [],
   "source": [
    "prediction = model.predict(validation_features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "fc77587f",
   "metadata": {},
   "outputs": [],
   "source": [
    "pred = np.argmax(prediction, axis=1)\n",
    "# label\n",
    "validation_labels = np.argmax(validation_labels, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "b9fdc8bd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "       Covid       1.00      1.00      1.00       220\n",
      "    NonCovid       0.91      0.97      0.94       220\n",
      "   Pneumonia       0.96      0.90      0.93       220\n",
      "\n",
      "    accuracy                           0.95       660\n",
      "   macro avg       0.95      0.95      0.95       660\n",
      "weighted avg       0.95      0.95      0.95       660\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(validation_labels, pred, target_names = ['Covid','NonCovid','Pneumonia']))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "af0e43d3",
   "metadata": {},
   "source": [
    "# Using VGG 16"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "29a704f1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/vgg16/vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
      "58892288/58889256 [==============================] - 24s 0us/step\n",
      "58900480/58889256 [==============================] - 24s 0us/step\n",
      "Model: \"vgg16\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " input_2 (InputLayer)        [(None, 224, 224, 3)]     0         \n",
      "                                                                 \n",
      " block1_conv1 (Conv2D)       (None, 224, 224, 64)      1792      \n",
      "                                                                 \n",
      " block1_conv2 (Conv2D)       (None, 224, 224, 64)      36928     \n",
      "                                                                 \n",
      " block1_pool (MaxPooling2D)  (None, 112, 112, 64)      0         \n",
      "                                                                 \n",
      " block2_conv1 (Conv2D)       (None, 112, 112, 128)     73856     \n",
      "                                                                 \n",
      " block2_conv2 (Conv2D)       (None, 112, 112, 128)     147584    \n",
      "                                                                 \n",
      " block2_pool (MaxPooling2D)  (None, 56, 56, 128)       0         \n",
      "                                                                 \n",
      " block3_conv1 (Conv2D)       (None, 56, 56, 256)       295168    \n",
      "                                                                 \n",
      " block3_conv2 (Conv2D)       (None, 56, 56, 256)       590080    \n",
      "                                                                 \n",
      " block3_conv3 (Conv2D)       (None, 56, 56, 256)       590080    \n",
      "                                                                 \n",
      " block3_pool (MaxPooling2D)  (None, 28, 28, 256)       0         \n",
      "                                                                 \n",
      " block4_conv1 (Conv2D)       (None, 28, 28, 512)       1180160   \n",
      "                                                                 \n",
      " block4_conv2 (Conv2D)       (None, 28, 28, 512)       2359808   \n",
      "                                                                 \n",
      " block4_conv3 (Conv2D)       (None, 28, 28, 512)       2359808   \n",
      "                                                                 \n",
      " block4_pool (MaxPooling2D)  (None, 14, 14, 512)       0         \n",
      "                                                                 \n",
      " block5_conv1 (Conv2D)       (None, 14, 14, 512)       2359808   \n",
      "                                                                 \n",
      " block5_conv2 (Conv2D)       (None, 14, 14, 512)       2359808   \n",
      "                                                                 \n",
      " block5_conv3 (Conv2D)       (None, 14, 14, 512)       2359808   \n",
      "                                                                 \n",
      " block5_pool (MaxPooling2D)  (None, 7, 7, 512)         0         \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 14,714,688\n",
      "Trainable params: 14,714,688\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from keras.applications.vgg16 import VGG16\n",
    "convolutional1_VGG16_base = tf.keras.applications.VGG16(\n",
    "    include_top=False,\n",
    "    weights=\"imagenet\",\n",
    "    input_shape=(img_width, img_height,3))\n",
    "\n",
    "\n",
    "convolutional1_VGG16_base.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "65d60521",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 2623 images belonging to 3 classes.\n",
      "Found 660 images belonging to 3 classes.\n"
     ]
    }
   ],
   "source": [
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "datagen = ImageDataGenerator(rescale=1./255)\n",
    "\n",
    "batch_size = 32\n",
    "\n",
    "def extract_features(directory, sample_count):\n",
    "  features = np.zeros(shape = (sample_count,7, 7, 512)) # equals to convolutional base\n",
    "  labels = np.zeros(shape = (sample_count,3)) # as we have 3 class for labels \n",
    "\n",
    "  #process data\n",
    "  generator = datagen.flow_from_directory(directory, target_size = (img_width, img_height),batch_size = batch_size, class_mode = 'categorical')\n",
    "  \n",
    "  #pass data into to convolutional base\n",
    "  i = 0;\n",
    "  for inputs_batch, labels_batch in generator:\n",
    "    features_batch = convolutional1_VGG16_base.predict(inputs_batch)\n",
    "    features[i * batch_size: (i + 1) * batch_size] = features_batch\n",
    "    labels[i * batch_size: (i + 1) * batch_size] = labels_batch\n",
    "    i += 1\n",
    "    if i * batch_size >= sample_count:\n",
    "      break\n",
    "  return features, labels\n",
    "    \n",
    "train_features, train_labels = extract_features(train_dir,2623)\n",
    "test_features, test_labels = extract_features(valid_dir,660 )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "430f84c7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_2\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " global_average_pooling2d (G  (None, 512)              0         \n",
      " lobalAveragePooling2D)                                          \n",
      "                                                                 \n",
      " dense_2 (Dense)             (None, 3)                 1539      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 1,539\n",
      "Trainable params: 1,539\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "epochs = 40\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "model.add(GlobalAveragePooling2D(input_shape=(7, 7, 512)))\n",
    "model.add(Dense(3, activation = 'softmax'))\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "acb9c995",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/40\n",
      "80/82 [============================>.] - ETA: 0s - loss: 0.9665 - acc: 0.6387\n",
      "Epoch 1: val_acc improved from -inf to 0.83788, saving model to model-001-0.643919-0.837879.h5\n",
      "82/82 [==============================] - 0s 3ms/step - loss: 0.9626 - acc: 0.6439 - val_loss: 0.8109 - val_acc: 0.8379\n",
      "Epoch 2/40\n",
      "45/82 [===============>..............] - ETA: 0s - loss: 0.7338 - acc: 0.8722\n",
      "Epoch 2: val_acc did not improve from 0.83788\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.6985 - acc: 0.8662 - val_loss: 0.6558 - val_acc: 0.8348\n",
      "Epoch 3/40\n",
      "45/82 [===============>..............] - ETA: 0s - loss: 0.6001 - acc: 0.8764\n",
      "Epoch 3: val_acc improved from 0.83788 to 0.85000, saving model to model-003-0.878002-0.850000.h5\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.5733 - acc: 0.8780 - val_loss: 0.5625 - val_acc: 0.8500\n",
      "Epoch 4/40\n",
      "71/82 [========================>.....] - ETA: 0s - loss: 0.5046 - acc: 0.8790\n",
      "Epoch 4: val_acc improved from 0.85000 to 0.86061, saving model to model-004-0.878384-0.860606.h5\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.5010 - acc: 0.8784 - val_loss: 0.5026 - val_acc: 0.8606\n",
      "Epoch 5/40\n",
      "40/82 [=============>................] - ETA: 0s - loss: 0.4583 - acc: 0.8875\n",
      "Epoch 5: val_acc improved from 0.86061 to 0.86212, saving model to model-005-0.887915-0.862121.h5\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.4518 - acc: 0.8879 - val_loss: 0.4646 - val_acc: 0.8621\n",
      "Epoch 6/40\n",
      "44/82 [===============>..............] - ETA: 0s - loss: 0.4241 - acc: 0.8963\n",
      "Epoch 6: val_acc did not improve from 0.86212\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.4162 - acc: 0.8933 - val_loss: 0.4358 - val_acc: 0.8606\n",
      "Epoch 7/40\n",
      "76/82 [==========================>...] - ETA: 0s - loss: 0.3899 - acc: 0.8976\n",
      "Epoch 7: val_acc improved from 0.86212 to 0.88182, saving model to model-007-0.898589-0.881818.h5\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.3884 - acc: 0.8986 - val_loss: 0.4076 - val_acc: 0.8818\n",
      "Epoch 8/40\n",
      "45/82 [===============>..............] - ETA: 0s - loss: 0.3727 - acc: 0.9083\n",
      "Epoch 8: val_acc did not improve from 0.88182\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.3654 - acc: 0.9032 - val_loss: 0.3889 - val_acc: 0.8803\n",
      "Epoch 9/40\n",
      "40/82 [=============>................] - ETA: 0s - loss: 0.3582 - acc: 0.9062\n",
      "Epoch 9: val_acc did not improve from 0.88182\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.3475 - acc: 0.9081 - val_loss: 0.3763 - val_acc: 0.8803\n",
      "Epoch 10/40\n",
      "81/82 [============================>.] - ETA: 0s - loss: 0.3354 - acc: 0.9059\n",
      "Epoch 10: val_acc improved from 0.88182 to 0.88939, saving model to model-010-0.906977-0.889394.h5\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.3335 - acc: 0.9070 - val_loss: 0.3587 - val_acc: 0.8894\n",
      "Epoch 11/40\n",
      "77/82 [===========================>..] - ETA: 0s - loss: 0.3188 - acc: 0.9103\n",
      "Epoch 11: val_acc improved from 0.88939 to 0.89091, saving model to model-011-0.910408-0.890909.h5\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.3192 - acc: 0.9104 - val_loss: 0.3492 - val_acc: 0.8909\n",
      "Epoch 12/40\n",
      "42/82 [==============>...............] - ETA: 0s - loss: 0.3029 - acc: 0.9182\n",
      "Epoch 12: val_acc improved from 0.89091 to 0.89545, saving model to model-012-0.912695-0.895455.h5\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.3067 - acc: 0.9127 - val_loss: 0.3333 - val_acc: 0.8955\n",
      "Epoch 13/40\n",
      "78/82 [===========================>..] - ETA: 0s - loss: 0.2985 - acc: 0.9171\n",
      "Epoch 13: val_acc did not improve from 0.89545\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.2971 - acc: 0.9173 - val_loss: 0.3316 - val_acc: 0.8955\n",
      "Epoch 14/40\n",
      "39/82 [=============>................] - ETA: 0s - loss: 0.2917 - acc: 0.9111\n",
      "Epoch 14: val_acc improved from 0.89545 to 0.89848, saving model to model-014-0.918414-0.898485.h5\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.2891 - acc: 0.9184 - val_loss: 0.3228 - val_acc: 0.8985\n",
      "Epoch 15/40\n",
      "79/82 [===========================>..] - ETA: 0s - loss: 0.2798 - acc: 0.9169\n",
      "Epoch 15: val_acc did not improve from 0.89848\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.2794 - acc: 0.9177 - val_loss: 0.3117 - val_acc: 0.8970\n",
      "Epoch 16/40\n",
      "79/82 [===========================>..] - ETA: 0s - loss: 0.2717 - acc: 0.9197\n",
      "Epoch 16: val_acc did not improve from 0.89848\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.2716 - acc: 0.9207 - val_loss: 0.3100 - val_acc: 0.8955\n",
      "Epoch 17/40\n",
      "38/82 [============>.................] - ETA: 0s - loss: 0.2655 - acc: 0.9153\n",
      "Epoch 17: val_acc improved from 0.89848 to 0.90303, saving model to model-017-0.923751-0.903030.h5\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.2656 - acc: 0.9238 - val_loss: 0.3039 - val_acc: 0.9030\n",
      "Epoch 18/40\n",
      "45/82 [===============>..............] - ETA: 0s - loss: 0.2498 - acc: 0.9243\n",
      "Epoch 18: val_acc did not improve from 0.90303\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.2589 - acc: 0.9241 - val_loss: 0.2963 - val_acc: 0.9015\n",
      "Epoch 19/40\n",
      "43/82 [==============>...............] - ETA: 0s - loss: 0.2462 - acc: 0.9266\n",
      "Epoch 19: val_acc improved from 0.90303 to 0.90455, saving model to model-019-0.924895-0.904545.h5\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.2524 - acc: 0.9249 - val_loss: 0.2913 - val_acc: 0.9045\n",
      "Epoch 20/40\n",
      "75/82 [==========================>...] - ETA: 0s - loss: 0.2456 - acc: 0.9279\n",
      "Epoch 20: val_acc did not improve from 0.90455\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.2472 - acc: 0.9276 - val_loss: 0.2873 - val_acc: 0.9030\n",
      "Epoch 21/40\n",
      "47/82 [================>.............] - ETA: 0s - loss: 0.2445 - acc: 0.9255\n",
      "Epoch 21: val_acc improved from 0.90455 to 0.90909, saving model to model-021-0.926801-0.909091.h5\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.2422 - acc: 0.9268 - val_loss: 0.2816 - val_acc: 0.9091\n",
      "Epoch 22/40\n",
      "44/82 [===============>..............] - ETA: 0s - loss: 0.2295 - acc: 0.9268\n",
      "Epoch 22: val_acc did not improve from 0.90909\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.2354 - acc: 0.9287 - val_loss: 0.2741 - val_acc: 0.9076\n",
      "Epoch 23/40\n",
      "81/82 [============================>.] - ETA: 0s - loss: 0.2326 - acc: 0.9317\n",
      "Epoch 23: val_acc did not improve from 0.90909\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.2323 - acc: 0.9321 - val_loss: 0.2695 - val_acc: 0.9091\n",
      "Epoch 24/40\n",
      "75/82 [==========================>...] - ETA: 0s - loss: 0.2264 - acc: 0.9308\n",
      "Epoch 24: val_acc improved from 0.90909 to 0.91061, saving model to model-024-0.930995-0.910606.h5\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.2276 - acc: 0.9310 - val_loss: 0.2683 - val_acc: 0.9106\n",
      "Epoch 25/40\n",
      "47/82 [================>.............] - ETA: 0s - loss: 0.2251 - acc: 0.9328\n",
      "Epoch 25: val_acc improved from 0.91061 to 0.91212, saving model to model-025-0.934045-0.912121.h5\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.2236 - acc: 0.9340 - val_loss: 0.2610 - val_acc: 0.9121\n",
      "Epoch 26/40\n",
      "80/82 [============================>.] - ETA: 0s - loss: 0.2186 - acc: 0.9328\n",
      "Epoch 26: val_acc did not improve from 0.91212\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.2202 - acc: 0.9318 - val_loss: 0.2574 - val_acc: 0.9121\n",
      "Epoch 27/40\n",
      "78/82 [===========================>..] - ETA: 0s - loss: 0.2188 - acc: 0.9343\n",
      "Epoch 27: val_acc improved from 0.91212 to 0.91364, saving model to model-027-0.935189-0.913636.h5\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.2168 - acc: 0.9352 - val_loss: 0.2545 - val_acc: 0.9136\n",
      "Epoch 28/40\n",
      "43/82 [==============>...............] - ETA: 0s - loss: 0.2175 - acc: 0.9346\n",
      "Epoch 28: val_acc did not improve from 0.91364\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.2131 - acc: 0.9360 - val_loss: 0.2511 - val_acc: 0.9121\n",
      "Epoch 29/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.2083 - acc: 0.9363\n",
      "Epoch 29: val_acc improved from 0.91364 to 0.91515, saving model to model-029-0.936332-0.915152.h5\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.2083 - acc: 0.9363 - val_loss: 0.2554 - val_acc: 0.9152\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 30/40\n",
      "80/82 [============================>.] - ETA: 0s - loss: 0.2055 - acc: 0.9363\n",
      "Epoch 30: val_acc improved from 0.91515 to 0.91667, saving model to model-030-0.935570-0.916667.h5\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.2061 - acc: 0.9356 - val_loss: 0.2492 - val_acc: 0.9167\n",
      "Epoch 31/40\n",
      "42/82 [==============>...............] - ETA: 0s - loss: 0.2190 - acc: 0.9286\n",
      "Epoch 31: val_acc did not improve from 0.91667\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.2031 - acc: 0.9379 - val_loss: 0.2470 - val_acc: 0.9152\n",
      "Epoch 32/40\n",
      "43/82 [==============>...............] - ETA: 0s - loss: 0.1936 - acc: 0.9404\n",
      "Epoch 32: val_acc improved from 0.91667 to 0.91818, saving model to model-032-0.936714-0.918182.h5\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.1997 - acc: 0.9367 - val_loss: 0.2472 - val_acc: 0.9182\n",
      "Epoch 33/40\n",
      "77/82 [===========================>..] - ETA: 0s - loss: 0.1960 - acc: 0.9399\n",
      "Epoch 33: val_acc did not improve from 0.91818\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.1966 - acc: 0.9394 - val_loss: 0.2362 - val_acc: 0.9182\n",
      "Epoch 34/40\n",
      "47/82 [================>.............] - ETA: 0s - loss: 0.2079 - acc: 0.9342\n",
      "Epoch 34: val_acc improved from 0.91818 to 0.91970, saving model to model-034-0.940907-0.919697.h5\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.1936 - acc: 0.9409 - val_loss: 0.2356 - val_acc: 0.9197\n",
      "Epoch 35/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.1915 - acc: 0.9398\n",
      "Epoch 35: val_acc did not improve from 0.91970\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.1915 - acc: 0.9398 - val_loss: 0.2335 - val_acc: 0.9167\n",
      "Epoch 36/40\n",
      "82/82 [==============================] - ETA: 0s - loss: 0.1887 - acc: 0.9394\n",
      "Epoch 36: val_acc improved from 0.91970 to 0.92424, saving model to model-036-0.939382-0.924242.h5\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.1887 - acc: 0.9394 - val_loss: 0.2360 - val_acc: 0.9242\n",
      "Epoch 37/40\n",
      "71/82 [========================>.....] - ETA: 0s - loss: 0.1866 - acc: 0.9419\n",
      "Epoch 37: val_acc did not improve from 0.92424\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.1866 - acc: 0.9417 - val_loss: 0.2267 - val_acc: 0.9182\n",
      "Epoch 38/40\n",
      "60/82 [====================>.........] - ETA: 0s - loss: 0.1846 - acc: 0.9432\n",
      "Epoch 38: val_acc did not improve from 0.92424\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.1841 - acc: 0.9428 - val_loss: 0.2241 - val_acc: 0.9167\n",
      "Epoch 39/40\n",
      "80/82 [============================>.] - ETA: 0s - loss: 0.1803 - acc: 0.9422\n",
      "Epoch 39: val_acc did not improve from 0.92424\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.1820 - acc: 0.9413 - val_loss: 0.2212 - val_acc: 0.9197\n",
      "Epoch 40/40\n",
      "73/82 [=========================>....] - ETA: 0s - loss: 0.1843 - acc: 0.9418\n",
      "Epoch 40: val_acc did not improve from 0.92424\n",
      "82/82 [==============================] - 0s 2ms/step - loss: 0.1799 - acc: 0.9436 - val_loss: 0.2196 - val_acc: 0.9197\n"
     ]
    }
   ],
   "source": [
    "# training the model\n",
    "#collect the best weights\n",
    "from keras.callbacks import ModelCheckpoint\n",
    "checkpoint = ModelCheckpoint('model-{epoch:03d}-{acc:03f}-{val_acc:03f}.h5', verbose=1, monitor='val_acc',save_best_only=True, mode='auto')  \n",
    "\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "model.compile(optimizer='adam',\n",
    "              loss='categorical_crossentropy',\n",
    "              metrics=['acc'])\n",
    "\n",
    "history = model.fit(train_features, train_labels,\n",
    "                    epochs=epochs,\n",
    "                    batch_size=batch_size, \n",
    "                    callbacks=[checkpoint],\n",
    "                    validation_data=(test_features, test_labels))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "4379fbc6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAEICAYAAABRSj9aAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8/fFQqAAAACXBIWXMAAAsTAAALEwEAmpwYAAA340lEQVR4nO3deXxU1fn48c+TPZBAAmEJBEhQZFEkQMQCrQJqRfAnYi2CS0VtVVxR+Vq01lLUb+1XW22/pSpWxWIt7hQVtbKp3+KSAAEhLGIIENYAWSF7zu+Pc5PMZJ2EJJPMPO/Xa165c+feO89cwpMz55z7XDHGoJRSyncFeDsApZRSrUsTvVJK+ThN9Eop5eM00SullI/TRK+UUj5OE71SSvk4TfR+SEQ+EpEbW3pbbxKRDBG5uBWOa0TkTGf5eRH5tSfbNuN9rhORfzc3TqUaIjqPvmMQkQKXp52AYqDceX6bMeYfbR9V+yEiGcDPjTGrWvi4BhhkjNndUtuKSDywBwg2xpS1SKBKNSDI2wEozxhjIiqXG0pqIhKkyUO1F/r72D5o100HJyITRCRTRH4pIoeBV0QkWkQ+EJEsEcl2luNc9lknIj93lmeLyP+JyNPOtntE5LJmbpsgIp+LSL6IrBKRRSLyWj1xexLjYyLyH+d4/xaRGJfXbxCRvSJyXER+1cD5OV9EDotIoMu66SKyxVkeIyJfikiOiBwSkb+ISEg9x1oiIo+7PP8vZ5+DInJzjW2nisgmEckTkf0issDl5c+dnzkiUiAiYyvPrcv+40QkWURynZ/jPD03TTzP3UTkFeczZIvIcpfXpolIqvMZvheRyc56t24yEVlQ+e8sIvFOF9YtIrIPWOOsf8v5d8h1fkfOdtk/XET+4Px75jq/Y+Ei8qGI3F3j82wRkel1fVZVP030vqE30A0YANyK/Xd9xXneHygE/tLA/ucDO4EY4H+Al0REmrHt68A3QHdgAXBDA+/pSYzXAjcBPYEQYB6AiAwDnnOO38d5vzjqYIz5GjgJTKpx3Ned5XLgPufzjAUuAu5oIG6cGCY78VwCDAJqjg+cBH4GRAFTgTkicqXz2gXOzyhjTIQx5ssax+4GfAj82flsfwQ+FJHuNT5DrXNTh8bO81JsV+DZzrGecWIYA/wd+C/nM1wAZNTzHnW5EBgKXOo8/wh7nnoCGwHXrsangdHAOOzv8YNABfAqcH3lRiIyAuiLPTeqKYwx+uhgD+x/uIud5QlACRDWwPaJQLbL83XYrh+A2cBul9c6AQbo3ZRtsUmkDOjk8vprwGsefqa6YnzE5fkdwMfO8qPAMpfXOjvn4OJ6jv048LKzHIlNwgPq2XYu8J7LcwOc6SwvAR53ll8GnnTZ7izXbes47rPAM85yvLNtkMvrs4H/c5ZvAL6psf+XwOzGzk1TzjMQi02o0XVs90JlvA39/jnPF1T+O7t8toENxBDlbNMV+4eoEBhRx3ZhQDZ23APsH4S/tsb/KV9/aIveN2QZY4oqn4hIJxF5wfkqnIftKohy7b6o4XDlgjHmlLMY0cRt+wAnXNYB7K8vYA9jPOyyfMolpj6uxzbGnASO1/de2Nb7VSISClwFbDTG7HXiOMvpzjjsxPHf2NZ9Y9xiAPbW+Hzni8hap8skF7jdw+NWHntvjXV7sa3ZSvWdGzeNnOd+2H+z7Dp27Qd872G8dak6NyISKCJPOt0/eVR/M4hxHmF1vZfzO/0GcL2IBACzsN9AVBNpovcNNadOPQAMBs43xnShuqugvu6YlnAI6CYinVzW9Wtg+9OJ8ZDrsZ337F7fxsaYNGyivAz3bhuwXUA7sK3GLsDDzYkB+43G1evACqCfMaYr8LzLcRub6nYQ29Xiqj9wwIO4amroPO/H/ptF1bHffuCMeo55EvttrlLvOrZx/YzXAtOw3Vtdsa3+yhiOAUUNvNerwHXYLrVTpkY3l/KMJnrfFIn9Opzj9Pf+prXf0GkhpwALRCRERMYC/6+VYnwbuFxEfugMnC6k8d/l14F7sYnurRpx5AEFIjIEmONhDG8Cs0VkmPOHpmb8kdjWcpHT332ty2tZ2C6TgfUceyVwlohcKyJBInINMAz4wMPYasZR53k2xhzC9p3/1Rm0DRaRyj8ELwE3ichFIhIgIn2d8wOQCsx0tk8CrvYghmLst65O2G9NlTFUYLvB/igifZzW/1jn2xdOYq8A/oC25ptNE71vehYIx7aWvgI+bqP3vQ47oHkc2y/+BvY/eF2epZkxGmO2AXdik/chbD9uZiO7/RM7QLjGGHPMZf08bBLOB150YvYkho+cz7AG2O38dHUHsFBE8rFjCm+67HsKeAL4j9jZPj+ocezjwOXY1vhx7ODk5TXi9tSzNHyebwBKsd9qjmLHKDDGfIMd7H0GyAU+o/pbxq+xLfBs4Le4f0Oqy9+x36gOAGlOHK7mAd8CycAJ4Pe456a/A8OxYz6qGfSCKdVqROQNYIcxptW/USjfJSI/A241xvzQ27F0VNqiVy1GRM4TkTOcr/qTsf2yy70clurAnG6xO4DF3o6lI9NEr1pSb+zUvwLsHPA5xphNXo1IdVgicil2POMIjXcPqQZo141SSvk4bdErpZSPa3dFzWJiYkx8fLy3w1BKqQ5lw4YNx4wxPep6rd0l+vj4eFJSUrwdhlJKdSgiUvNq6iradaOUUj7Oo0QvIpNFZKeI7BaR+XW8PkBEVjslRNeJexnUcqfUaaqIrGjJ4JVSSjWu0a4bp/jRImw51kwgWURWOPVDKj0N/N0Y86qITAJ+R3WJ2kJjTGLLhq2UUspTnrTox2BL06YbY0qAZdgLYVwNo/oS8LV1vK6UUspLPEn0fXEvx5qJe7lUgM3Y8q8A04FIl5skhIlIioh85XLjBTcicquzTUpWVpbn0SullGpUSw3GzgMuFJFN2MJRB6i+cfUAY0wStnDUsyJSqxypMWaxMSbJGJPUo0eds4OUUko1kyfTKw/gXnc7jhp1sY0xB3Fa9CISAfzEGJPjvHbA+ZkuIuuAkZzeDQ2UUko1gSeJPhkYJCIJ2AQ/E/fa2oi9MfEJp7b0Q9j60ohINPZmAcXONuOx9xlVSikFFJQUsPnwZjYd3kRwQDC3Jd3W4u/RaKI3xpSJyF3AJ0Ag9t6b20RkIZBijFmBvW/p70TEYG9Vdqez+1DgBRGpwHYTPVljto5SSnUIecV5fHvkWzLzMjlUcIhD+Yc4VHCIwwWHq57nFufSN7IvCdEJJEQ5j+jqn6GBoaQeTmXT4U1sPLSRjYc2suv4LoxzQ66xcWNbJdG3u6JmSUlJRq+MVUo1prismL25e9mTvYc9OXuqfuYV59ErohexEbHERsTSO6I3sZHVy5GhkY0e++jJo2w6tKkqIW86vIndJ3a7bRMSGGKPHRFLbGQsvTv3pmtYVzLzMqviOVRwqN736NelH6NiRzEqdhQje49kVOwo+kT2QaR5d/wUkQ3OeGgt7a4EglJK1ZSZl8mX+79k/f71bDi0gfTsdA7mH6xqCYNNvAO6DqBrWFfSstI4XHCY0orSWscKCggiLCiM8KBwwoLC7HKwXQ4JDGFP9h4O5FcPQyZEJTAydiSzR8wmsXciA6IGEBsRS7fwbo0m5cLSQrc/RqdKTzGi1whGxo4kppOn94o/fZrolVKtyhhDZl5mVct485HNBEpgna3t2MhYosKi+PbIt3yZaRP7+v3r2Z9nZ3iHBYUxKnYUFw+8uFa3SJ/IPgRIgNv7nig84dbNUtm9UlRWRFFZEYVlhVXLlY8J8ROqWtmJvROJDo9u9mcPDw5nSMwQhsQMaXzjVqSJXik/ZIyhsKyQnKIcsguz7c+i7KrnrsmwqKyIwtJCisqrnxtjiAqLIjos2v4Mj3Z7fuzUsarEvvHQRo4XHgcgQAI4q/tZCMLqPavJKcppMM5+Xfoxrt84xsaNZVy/cYzoPYKQwBCPPqOI0L1Td7p36s45Pc853VPWoWmiV6oD2puzl3JTTkJUgsd9uvty97Hyu5Ws/G4lazPWUlBS0OD2ARLg1r3h2sVhjGHX8V3kFOWQU5RDuSmvtX9wQDDDew3nyiFXVrWQz+11Lp1DOldtU1hayJGTR9wGNo+dOsaQmCGMjRtLv679ah1XNZ0meqU6gEP5h1ibsZY1e9awZs8a9uTsAaBraFdGxo6sGswb2Xskg2MGExQQRGl5Kev3r7fJffdKth7dCkB8VDzXD7+eAVED6m2Rdw3rSnBAsEd/RIwxFJQUuH0j6BLahbN7nt1o6zs8OJz4qHjio+JP+xyp+umsG6XaiDGGtKy0qsSbejiVmE4x1TM3KmdvOM9Plp6sSuzbj20HICosignxE5gYP5GwoDA2HdrExsMb2XJkC0VlRQCEB4UzrMcwdp/YTW5xLkEBQVww4AKmnDmFqWdNZXD3wc2e2aHar4Zm3WiiV6oVnSo9xZo9a6q6TPbm2ntDnNvrXMbFjSO3ONdtsDCvOM9t/87BnfnRgB8xKX4SkxImkdg7kcCAwFrvU1ZRxo5jO2ziP7SRLUe3kBCVwJRBU7h44MV0Ce3SJp9XeY8meqXaQG5RLtuPbWd71na2H9vO5iOb+SzjM4rLi+kc3JmLB17M1EFTuWzQZcR1iavzGKdKT9kLcPIPESABjO4z2uPBR+XfdB69Uh4oqyhj69GtVVP6vsr8ivySfKLCotxnmLj0Yx/IO8D2Y9tJy0pzuzgmJDCEITFDmJM0hymDpnDBgAsIDQptNIZOwZ0YGD2QgdEDW/OjKj+jiV75ndLy0qrphLtP7LYX4mSu5+vMrzlZehKA3hG9GddvHD069ajaNrsom/Ts9KrnZRVlRIREMDRmKD8+48cMjRnK0B5DGRozlIToBIIC9L+Xah/0N1F1SBWmgs2HN/NV5lfkFee5z/l2uQjmVOkpt9kgOUU5Vcm8UqAEMqL3CG5KvImx/ex87QFdBzQ4YFk5Dz08KFwHNn1FZiY88QTExsK990LXrt6OqMVoolcdgjGGHcd22FkoGWtYl7GOE4Un3LYJCQxxm/NdeZl7dHg0Z3Y7s1bXS3R4NHFd4kjqk0RESEST4hEROgV3asmPqLylogKefx7mz4fiYigpgWefhf/6L7j7boho2u9Ge6SJXrVbe7L3VCX2NXvWcLjgMAD9u/Zn2uBpTEqYxAUDLiCmUwxhQWFul78r5ZHt2+EXv4D//AcuuQReeAGys+HRR+Hhh+GZZ+CXv4Q77oDwcG9H22w660a1GwfzD7J2z9qq5J6RkwFAr869mJQwqerRlKtBlQ8rLISMDNizx/1RVgYXXAATJ8KIERBYezoqJSXw5JO2qyYiwib0G24A19+rr76CX/8aVq2y3TkPP2z/KIQ2Pqju5uhRWLvWPrZvb3jbYcPgueeadnyHTq9Uraa8opztx7ZzMP+gW+Eo1zrdRWVFteuihNqfXUO78n3297UuCpoYP5FJCZOYGD+RYT2GaWL3dcXFNrGuXQtffglFRfVvW1QEe/fCkSPu68PCID4eysvhu+/suuhomDDBJv1Jk2wi/eorm7C3bYNrr7VJvmfP+t/v889twv/8c4iLgx/9yL5PQkL1o39/CA6222dnw2efwZo19vNstVck06ULnHsuBDXQkTJ0KPz1r42crLppolctbk/2Hl5JfYUlqUuqKgtWigyJdKtIGB4cTm5Rbq1B0dziXMBeFHTBgAuqWuwjeo2o86Ig1Q7l5FS3pF1b1yK1k2FCQvUAZ2kppKTYRLhmje06KSqCgACbDKOi6n/P4GCbWCuPWfk+vXrZ/QEOHIB16+yx16yxsQH06AHHjtmE/fzzMGWKZ5/TGFi92vbdb98O+/bZbw6VAgKgb1+bzNPS7Pbh4faPQuUfmVGjGk7yp0kTvWoRRWVFvLf9PV7a9BKr96xGEC4981JmnTOLgdEDqxK7a9GqhpRXlJNbnEtkSCTBgcGtHL06bYcOVXdBbNhgE3pOjvs2XbrYpGuMfT0/3/31qCibpNPTocApqjZ8uE2EkybZLpeGknxz7dlj4163Dvr0gV/9CiIbvwFJvcrK7B+Tml1HJ07AmDH2s4wZAyFtd7GbJnrVbBWmgtTDqbyy6RX+8e0/yC7KJj4qnpsTb2Z24mytLtjelZfbhFSzxX3woG3dura24+PduyCOHbOJsbLVvWOHXR8VBT/4AZxxRu1We1RUdT+3MbYbo2Yf+t699n0mTbLdKj16tPFJ8U16ZazySGl5KWlZaW63T0s9nEpBSQGhgaFcNfQqbhl5CxMTJuoMl7aUnV275ZiRAXl59e9TUQGHD9suhlKXuyyJ2C6Gvn3h++/hzTftH4NKAQG2W6Nz5+qBw4gI2wVxyy02Odc3wFmTCHTrZh+jRzfnk6sWoonejxWUFLA6fTUf7/6Y5IPJfHv0W0rKSwDbb57YO5HZI2Yzus9orhh8Bd3Cu3k5Yh918qR7Iq+Z1HNz3bfv2tW2nqMbufNRUhL89Kfu/dj9+7vPGikrsxcK1XzfnBy4/nrbv5yUVN3KVx2SJnofUHlJf/dO3RtsaVfeLKKyTO7nez+npLyEyJBIzo87n3vPv7eqrvmZ3c7UAdGGfP01/PGPtjVcs/sjPt59znVxsW1Z15XEMzLs9DtX4eHViXn8+NqDjo0l+KYICqqOWfksTfQdWGZeJi+kvMDijYs5evIoQQFB9Orcq9Z9OHtH9GbHsR2s/G4l32d/D8CwHsO4Z8w9TD1rKuP6jdMKiZ7auNFeTPPhh9C9O8TEwAcf2GTuqndvO+h35IjtD3cdCwsOhgEDbHKdNq12Iu/Z030+t1KnSRN9B2OMYW3GWhYlL+JfO/5Fhalg6llTuSjhIo6ePFo1d31/7n6SDyRz9ORRDIbwoHAmJUzigbEPcNmgy/SOPk317bfwm9/Ae+/ZFvV//3f15fEVFTah12ypHzhgZ5TUTOR9+njWx61UC9FE30HkFeexdPNSFiUvYvux7XQP784DYx/g9qTbSYhOqHe/sooysk5mERUWRXhwx72Eu0VlZlbPJDl0yLaua873jomxreqdO2HBAnjjDTsdb8ECmDvXveBVQIC9cjI2FsaN89KHUqp+HiV6EZkM/AkIBP5mjHmyxusDgJeBHsAJ4HpjTKbz2o3AI86mjxtjXm2h2H2eMYbkg8m8tPElXt/6OgUlBST1SWLJtCXMOHuGR4k7KCCI2MjYNojWCz77DJ5+2s7drjnNr1+/6gHEI0eqL55Zu7b6qslu3ex+KSlw/Lj7sTt3tsfYtcv2mc+fD/Pm2X2U6mAanUcvIoHALuASIBNIBmYZY9JctnkL+MAY86qITAJuMsbcICLdgBQgCTDABmC0MSa7vvfTefRw7NQxXtvyGi9teomtR7cSHhTOjLNncMd5dzCm7xhvh+d9OTnw4IPw4ou2FR0aCvv31z1NMCzMJmuwfxAuvLD6SsXhw6uvpMzPr3ugdPBgeOCBhi+RV6odON159GOA3caYdOdgy4BpQJrLNsOA+53ltcByZ/lS4FNjzAln30+BycA/m/gZfF55RTmr0lfx0qaXWL5jOaUVpYzpO4YXLn+BmefM1Ht+VnrvPbjzTttKnzcPfvtb6NTJfZpgzbnmN99sE/vIkfVfgh4ZaRP/8OFt+nGUagueJPq+gGsxk0zg/BrbbAauwnbvTAciRaR7Pfv2rfkGInIrcCtA//79PY3dJ+QU5bB4w2IWJS9iX+4+uod3587z7uTmkTczvFcHTjr79lX3gx89WrsfPD6+uh/cEwcP2sHPd9+FxER4/333i3BcpwlOnNjyn0epDqylBmPnAX8RkdnA58ABoLzBPVwYYxYDi8F23bRQTO3a3py9/OnrP/HixhcpKClgYvxEnrrkKaYNnubRvUXbncOHqxP72rX2qkuwybx/f0hOrt0PHhHh3rdes5+9Sxc7o+Vvf7NdNcXFtrTs/ffrBTxKNYEnif4A4FrQJM5ZV8UYcxDbokdEIoCfGGNyROQAMKHGvutOI94OL+VgCk+vf5q3095GRLjm7Gt4YOwDjIwd6e3Qajt+HF5/Hf71r9rzxF1lZdnZKWBno0yYYFvfkybB2WfX3w/uurxuXe0CWN262WSfkWGPuXgxDBrU0p9SKZ/nyWBsEHYw9iJsgk8GrjXGbHPZJgY4YYypEJEngHJjzKPOYOwGYJSz6UbsYKz7PeBc+OJgbIWp4MNdH/KHL//AZ3s/o0toF24ddSv3nH9P+ysKVl5ub7Tw8suwfLm9QcPw4Q0Xnurc2dZCmTTJdqs0Z454fQWwDhyAK66w/ex6EZFS9TqtwVhjTJmI3AV8gp1e+bIxZpuILARSjDErsK3234mIwXbd3Onse0JEHsP+cQBY2FCS9zWFpYUs3bKUP375R3Ye30n/rv35w4//wM9H/bz9Da6mp8OSJfaxf79tTc+ZAzfdZItYtTYtgKVUq9Eyxa0g62QWf03+K4uSF5F1KotRsaOYN3YeVw+72rt11ysq7AVCNWem7Nhh7+ojApdealvPV1zR9FumKaW8RssUt5Gdx3byzFfP8OrmVykqK+Lysy7ngbEPcOGAC713K7yCAns3+9WrbR3wkhL312Nj7cDnwoUwe7a9SEgp5VM00beAvOI85nw4h39++09CAkP42Yifcd8P7mNoj6HeDWz7dvjJT+xA6bRpMH26+8yWAQPsBUVKKZ+mif407Tq+iyuXXcmu47t46IcPcc/599Arope3w4J//tPeALlzZ/j0UztQqpTyS5roT8PK71Zy7TvXEhwYzKqfrWJC/ARvh2SnQT7wACxaBD/8ISxbZu8mpJTyW3o/uGYwxvC7L37H5a9fzsDogaT8IqV9JPl9++zNlRctssl+zRpN8kopbdE31cmSk9z0r5t4K+0tZp0zi79d8Tc6BXc6vYPu2we33WbrnNcsl+tahbEhn3wC111n7w/6zjtw1VWnF5NSymdoom+C9Ox0rlx2JduytvHUJU/xwNgHWmY2zYMP2rIBffrUf7Pmhm5WUV5ub203fDi8/bZePaqUcqOJ3lFSXsL6/es5WXKSorKiqkdhWSFFZUUUlBTwv9/8L8YYPrruI358xo9b5o2/+sre1OLXv7ZTHOu7WfPhw+63o6tp7lx4/HFbyVEppVxoosd2x0x9fSqf7f2swe0Seyfy9k/f5oxuZ7TMGxtjC3T17m1b9aA3a1ZKtTi/T/SVSf6LfV/wl8v+wnl9zyMsKIywoDDCg8KrlsOCwlr+qtZ33rFXpL74oq3kqJRSrcCvE71rkl86fSnXDr+27d68uBh++Us45xxbT0YppVqJ3yZ6ryZ5sFMg09Ph44+bV+1RKaU85JeJ/mTJSaa8PoX/2/d/3knyx4/DY4/ZAmKXXtq2762U8jt+l+hdk/xr019j1vBZbR/EY4/Ze5k+/XTbv7dSyu/41ZWx7SLJf/ed7ba55RbbP6+UUq3MbxJ9UVmR95M8wPz5tmLkwoXeeX+llN/xm0S/LmMdn+/9nOemPue9JP/FF/Duu3a2Te/e3olBKeV3/CbRZxdmA/Cj/j/yTgAVFbbQWN++9iIppZRqI34zGJtfkg9w+vdqNQaOHKldoiAjw956r7IYmesNPqKibLng5GR7T1YtU6CUakN+k+jzivMAiAyNbPrOxsCTT8Lf/24TelGR++s9e9q7NZWU2O6ZvDz316OibFXJkSPhhhuaFb9SSjWX3yT6/GLboo8IaWKpgbIyW0L45Zdh4kSYMsW9xR4fb+/iVMkYyMlxb+nv2QMHDsDDD9tqlEop1Yb8JtHnFecRGRJJgDQh0Z48CddcAx9+CL/5jX00VpZYxNaVj46GUaNOL2illGoB/pXom9Jtc+wYTJ0KKSnw/PO2Va+UUh2Q3yT6/JJ8zwdi9+yByZPtnZ/eeQeuvLJVY1NKqdbkUT+GiEwWkZ0isltE5tfxen8RWSsim0Rki4hMcdbHi0ihiKQ6j+db+gN4Kq84z7NEv2kTjBsHWVmwapUmeaVUh9doi15EAoFFwCVAJpAsIiuMMWkumz0CvGmMeU5EhgErgXjnte+NMYktGnUzVPbRN2jVKnuv1agoWL0ahg1rk9iUUqo1edKiHwPsNsakG2NKgGXAtBrbGKCyudwVONhyIbaMRrtu3n/fzqgZMMDeDESTvFLKR3iS6PsC+12eZzrrXC0ArheRTGxr/m6X1xKcLp3PRKTOy1JF5FYRSRGRlKysLM+jb4IGB2N37YLrroNzz7Xz4PvW/HhKKdVxtdSk7lnAEmNMHDAFWCoiAcAhoL8xZiRwP/C6iNRqVhtjFhtjkowxST169GihkNzlF+fTJaSOFn1Bge2uCQmxdWiiolrl/ZVSyls8SfQHgH4uz+Ocda5uAd4EMMZ8CYQBMcaYYmPMcWf9BuB74KzTDbqpjDF1D8YaAz//OWzfbksU9O/f1qEppVSr8yTRJwODRCRBREKAmcCKGtvsAy4CEJGh2ESfJSI9nMFcRGQgMAhIb6ngPVVYVki5Ka/ddfPnP8Mbb8Djj8PFF7d1WEop1SYanXVjjCkTkbuAT4BA4GVjzDYRWQikGGNWAA8AL4rIfdiB2dnGGCMiFwALRaQUqABuN8acaLVPU4/K8gduLfovvoB582DaNFs2WCmlfJRHF0wZY1ZiB1ld1z3qspwGjK9jv3eAd04zxtNWWdCsKtEfOgQzZtg6Na++qvVnlFI+zS+ujK2qXBkSaatIXnONrTD5739D165ejk4ppVqX7yT64mI47zwYOxYmTYIJE6BXL6BGLfpf/tJ227z+Ogwf7sWAlVKqbfhOoj9+3F7stGwZLF5s1519NkycSKfBEUSfgv4ffwnPPAP33AOzvHQ7QaWUamNijPF2DG6SkpJMSkpK8w9QVgYbN8KaNbB2rW29FxZSAUhQEHL++fa1kJAWi1kppbxNRDYYY5Lqes33RiGDgmDMGJg/Hz75BLKzWf7C/SyYAEXT/x+8+aYmeaWUX/G9RF9TaCg7h/XksQlQ8dpS6NPH2xEppVSb8v1Ejx2MDZAAOgXrTbmVUv7HLxJ9ZfkDaew2gEop5YP8JtE3WoteKaV8lF8k+ibdRlAppXyMXyT6Jt8YXCmlfIjfJHpt0Sul/JVfJPr8Yu26UUr5L79I9DoYq5TyZ36R6HUwVinlz3w+0VeYCu26UUr5NZ9P9CdLTmIw2nWjlPJbPp/o3WrRK6WUH/L5RF91dymdR6+U8lN+k+i1Ra+U8lc+n+jzi7XrRinl33w+0bvdGFwppfyQzyd6HYxVSvk7n0/02kevlPJ3HiV6EZksIjtFZLeIzK/j9f4islZENonIFhGZ4vLaQ85+O0Xk0pYM3hM660Yp5e+CGttARAKBRcAlQCaQLCIrjDFpLps9ArxpjHlORIYBK4F4Z3kmcDbQB1glImcZY8pb+oPUJ784n+CAYEIDQ9vqLZVSql3xpEU/BthtjEk3xpQAy4BpNbYxQGXfSFfgoLM8DVhmjCk2xuwBdjvHazOVtej1NoJKKX/lSaLvC+x3eZ7prHO1ALheRDKxrfm7m7AvInKriKSISEpWVpaHoXsmr0Rr0Sul/FtLDcbOApYYY+KAKcBSEfH42MaYxcaYJGNMUo8ePVooJEsLmiml/F2jffTAAaCfy/M4Z52rW4DJAMaYL0UkDIjxcN9WpbXolVL+zpNWdzIwSEQSRCQEO7i6osY2+4CLAERkKBAGZDnbzRSRUBFJAAYB37RU8J7QWvRKKX/XaIveGFMmIncBnwCBwMvGmG0ishBIMcasAB4AXhSR+7ADs7ONMQbYJiJvAmlAGXBnW864AduiPyP6jLZ8S6WUalc86brBGLMSO8jquu5Rl+U0YHw9+z4BPHEaMZ4W7bpRSvk7n78yVgdjlVL+zqcTfXlFOSdLT+pVsUopv+bTiV4LmimllK8neq1Fr5RSvp3otRa9Ukr5eKLXrhullPLxRK+16JVSyk8Svc66UUr5M59O9DoYq5RSPp7odTBWKaX8JdFr141Syo/5dKLPL8knLCiMkMAQb4eilFJe49OJXguaKaWUjyd6rUWvlFI+nujzivV+sUop5fOJXgdilVL+zqcTvdaiV0opH0/0OhirlFJ+kOi1Ra+U8nc+neh11o1SSvlwoi8tL6WorEi7bpRSfs9nE73WoldKKctnE73WoldKKcvnE73Oo1dK+TuPEr2ITBaRnSKyW0Tm1/H6MyKS6jx2iUiOy2vlLq+taMHYG6S16JVSygpqbAMRCQQWAZcAmUCyiKwwxqRVbmOMuc9l+7uBkS6HKDTGJLZYxB7SWvRKKWV50qIfA+w2xqQbY0qAZcC0BrafBfyzJYI7HdpHr5RSlieJvi+w3+V5prOuFhEZACQAa1xWh4lIioh8JSJX1rPfrc42KVlZWZ5F3giddaOUUlZLD8bOBN42xpS7rBtgjEkCrgWeFZEzau5kjFlsjEkyxiT16NGjRQLRwVillLI8SfQHgH4uz+OcdXWZSY1uG2PMAednOrAO9/77VlM5GKt99Eopf+dJok8GBolIgoiEYJN5rdkzIjIEiAa+dFkXLSKhznIMMB5Iq7lva8grzqNzcGcCAwLb4u2UUqrdanTWjTGmTETuAj4BAoGXjTHbRGQhkGKMqUz6M4FlxhjjsvtQ4AURqcD+UXnSdbZOa9Ja9EopZTWa6AGMMSuBlTXWPVrj+YI69lsPDD+N+JpNC5oppZTl01fGav+8Ukr5eKLXFr1SSvlwoteuG6WUsnw20etgrFJKWT6b6POL8+kSoi16pZTyyURvjNE+eqWUcvhkoi8uL6a0olS7bpRSCh9N9FqLXimlqvlkotda9EopVc2nE7226JVSykcTvdaiV0qpaj6Z6LUWvVJKVfPJRK+DsUopVc0nE7320SulVDWfTvQ660YppXw00eeX5CMInUM6ezsUpZTyOp9M9HnFeUSERBAgPvnxlFKqSXwyE2qdG6WUquaTiV5r0SulVDWfTPRai14ppar5ZKLPL9YWvVJKVfLJRK999EopVc1nE73OoVdKKcsnE70OxiqlVDWPEr2ITBaRnSKyW0Tm1/H6MyKS6jx2iUiOy2s3ish3zuPGFoy9TpW3EdQWvVJKWUGNbSAigcAi4BIgE0gWkRXGmLTKbYwx97lsfzcw0lnuBvwGSAIMsMHZN7tFP4WLU6WnqDAV2qJXSimHJy36McBuY0y6MaYEWAZMa2D7WcA/neVLgU+NMSec5P4pMPl0Am6M1qJXSil3niT6vsB+l+eZzrpaRGQAkACsacq+InKriKSISEpWVpYncddLa9ErpZS7lh6MnQm8bYwpb8pOxpjFxpgkY0xSjx49TisArUWvlFLuPEn0B4B+Ls/jnHV1mUl1t01T920RWoteKaXceZLok4FBIpIgIiHYZL6i5kYiMgSIBr50Wf0J8GMRiRaRaODHzrpWo7XolVLKXaOzbowxZSJyFzZBBwIvG2O2ichCIMUYU5n0ZwLLjDHGZd8TIvIY9o8FwEJjzImW/QjudDBWKaXcNZroAYwxK4GVNdY9WuP5gnr2fRl4uZnxNZkOxiqllDufuzJW++iVUsqdzyX6/OJ8AiWQ8KBwb4eilFLtgs8l+spa9CLi7VCUUqpd8KiPviPRgmbKl5SWlpKZmUlRUZG3Q1HtRFhYGHFxcQQHB3u8j88leq1Fr3xJZmYmkZGRxMfH67dUhTGG48ePk5mZSUJCgsf7+WbXjc6hVz6iqKiI7t27a5JXAIgI3bt3b/I3PJ9L9Np1o3yNJnnlqjm/Dz6X6PXG4Eop5c4nE32XEG3RK9USjh8/TmJiIomJifTu3Zu+fftWPS8pKWlw35SUFO65555G32PcuHEtFa6qh88NxuYXa9eNUi2le/fupKamArBgwQIiIiKYN29e1etlZWUEBdWdRpKSkkhKSmr0PdavX98isbal8vJyAgMDvR2Gx3wq0VeYCvJL8rXrRvmkuR/PJfVwaoseM7F3Is9OfrZJ+8yePZuwsDA2bdrE+PHjmTlzJvfeey9FRUWEh4fzyiuvMHjwYNatW8fTTz/NBx98wIIFC9i3bx/p6ens27ePuXPnVrX2IyIiKCgoYN26dSxYsICYmBi2bt3K6NGjee211xARVq5cyf3330/nzp0ZP3486enpfPDBB25xZWRkcMMNN3Dy5EkA/vKXv1R9W/j973/Pa6+9RkBAAJdddhlPPvkku3fv5vbbbycrK4vAwEDeeust9u/fXxUzwF133UVSUhKzZ88mPj6ea665hk8//ZQHH3yQ/Px8Fi9eTElJCWeeeSZLly6lU6dOHDlyhNtvv5309HQAnnvuOT7++GO6devG3LlzAfjVr35Fz549uffee5v7T9ckPpXoT5bYf2Bt0SvVujIzM1m/fj2BgYHk5eXxxRdfEBQUxKpVq3j44Yd55513au2zY8cO1q5dS35+PoMHD2bOnDm15oJv2rSJbdu20adPH8aPH89//vMfkpKSuO222/j8889JSEhg1qxZdcbUs2dPPv30U8LCwvjuu++YNWsWKSkpfPTRR/zrX//i66+/plOnTpw4YesqXnfddcyfP5/p06dTVFRERUUF+/fvr/PYlbp3787GjRsB2631i1/8AoBHHnmEl156ibvvvpt77rmHCy+8kPfee4/y8nIKCgro06cPV111FXPnzqWiooJly5bxzTffNPm8N5dPJXqtc6N8WVNb3q3ppz/9aVXXRW5uLjfeeCPfffcdIkJpaWmd+0ydOpXQ0FBCQ0Pp2bMnR44cIS4uzm2bMWPGVK1LTEwkIyODiIgIBg4cWDVvfNasWSxevLjW8UtLS7nrrrtITU0lMDCQXbt2AbBq1SpuuukmOnXqBEC3bt3Iz8/nwIEDTJ8+HbAXIXnimmuuqVreunUrjzzyCDk5ORQUFHDppZcCsGbNGv7+978DEBgYSNeuXenatSvdu3dn06ZNHDlyhJEjR9K9e3eP3rMl+GSi13n0SrWuzp07Vy3/+te/ZuLEibz33ntkZGQwYcKEOvcJDQ2tWg4MDKSsrKxZ29TnmWeeoVevXmzevJmKigqPk7eroKAgKioqqp7XnK/u+rlnz57N8uXLGTFiBEuWLGHdunUNHvvnP/85S5Ys4fDhw9x8881Nju10+NSsG61Fr1Tby83NpW9feyvoJUuWtPjxBw8eTHp6OhkZGQC88cYb9cYRGxtLQEAAS5cupbzc3tH0kksu4ZVXXuHUqVMAnDhxgsjISOLi4li+fDkAxcXFnDp1igEDBpCWlkZxcTE5OTmsXr263rjy8/OJjY2ltLSUf/zjH1XrL7roIp577jnADtrm5uYCMH36dD7++GOSk5OrWv9txacSvdaiV6rtPfjggzz00EOMHDmySS1wT4WHh/PXv/6VyZMnM3r0aCIjI+natWut7e644w5effVVRowYwY4dO6pa35MnT+aKK64gKSmJxMREnn76aQCWLl3Kn//8Z84991zGjRvH4cOH6devHzNmzOCcc85hxowZjBw5st64HnvsMc4//3zGjx/PkCFDqtb/6U9/Yu3atQwfPpzRo0eTlpYGQEhICBMnTmTGjBltPmNHXG4I1S4kJSWZlJSUZu377vZ3+cmbP2HTbZtI7J3YsoEp5QXbt29n6NCh3g7D6woKCoiIiMAYw5133smgQYO47777vB1Wk1RUVDBq1CjeeustBg0adFrHquv3QkQ2GGPqnM/qUy36/GLtulHKF7344oskJiZy9tlnk5uby2233ebtkJokLS2NM888k4suuui0k3xz6GCsUqrdu++++zpcC97VsGHDqubVe4Nvteh1MFYppWrxqUSfV5xHSGAIoUGhjW+slFJ+wucSvXbbKKWUO59K9FqLXimlavOpRK+16JVqWRMnTuSTTz5xW/fss88yZ86ceveZMGEClVOkp0yZQk5OTq1tFixYUDWfvT7Lly+vmoMO8Oijj7Jq1aomRK8qeZToRWSyiOwUkd0iMr+ebWaISJqIbBOR113Wl4tIqvNY0VKB10XvF6tUy5o1axbLli1zW7ds2bJ6C4vVtHLlSqKiopr13jUT/cKFC7n44oubdSxvqbw619sanV4pIoHAIuASIBNIFpEVxpg0l20GAQ8B440x2SLS0+UQhcaYxJYNu275xfn0iujVFm+lVNubOxec2vAtJjERnn223pevvvpqHnnkEUpKSggJCSEjI4ODBw/yox/9iDlz5pCcnExhYSFXX301v/3tb2vtHx8fT0pKCjExMTzxxBO8+uqr9OzZk379+jF69GjAzpGvWe43NTWVFStW8Nlnn/H444/zzjvv8Nhjj3H55Zdz9dVXs3r1aubNm0dZWRnnnXcezz33HKGhocTHx3PjjTfy/vvvU1payltvveV21Sr4ZzljT1r0Y4Ddxph0Y0wJsAyYVmObXwCLjDHZAMaYo6cVVTPpYKxSLatbt26MGTOGjz76CLCt+RkzZiAiPPHEE6SkpLBlyxY+++wztmzZUu9xNmzYwLJly0hNTWXlypUkJydXvXbVVVeRnJzM5s2bGTp0KC+99BLjxo3jiiuu4KmnniI1NZUzzjijavuioiJmz57NG2+8wbfffktZWVlVbRmAmJgYNm7cyJw5c+rsHqosZ7xx40beeOONqrr4ruWMN2/ezIMPPgjYcsZ33nknmzdvZv369cTGxjZ63irLGc+cObPOzwdUlTPevHkzGzdu5Oyzz+bmm2+uqnxZWc74+uuvb/T9GuPJBVN9AdcizZnA+TW2OQtARP4DBAILjDEfO6+FiUgKUAY8aYxZXvMNRORW4FaA/v37NyV+NzoYq3xaAy3v1lTZfTNt2jSWLVtWlajefPNNFi9eTFlZGYcOHSItLY1zzz23zmN88cUXTJ8+vapU8BVXXFH1Wn3lfuuzc+dOEhISOOusswC48cYbWbRoUVUr+KqrrgJg9OjRvPvuu7X298dyxi11ZWwQMAiYAMQBn4vIcGNMDjDAGHNARAYCa0TkW2PM9647G2MWA4vB1rppbhDaR69Uy5s2bRr33XcfGzdu5NSpU4wePZo9e/bw9NNPk5ycTHR0NLNnz65V0tdTTS3325jKUsf1lTn2x3LGnnTdHAD6uTyPc9a5ygRWGGNKjTF7gF3YxI8x5oDzMx1YB9RfDu40lFWUcar0lHbdKNXCIiIimDhxIjfffHPVIGxeXh6dO3ema9euHDlypKprpz4XXHABy5cvp7CwkPz8fN5///2q1+or9xsZGUl+fn6tYw0ePJiMjAx2794N2CqUF154ocefxx/LGXuS6JOBQSKSICIhwEyg5uyZ5djWPCISg+3KSReRaBEJdVk/HkijFRSUFABa/kCp1jBr1iw2b95clehHjBjByJEjGTJkCNdeey3jx49vcP9Ro0ZxzTXXMGLECC677DLOO++8qtfqK/c7c+ZMnnrqKUaOHMn331d3AoSFhfHKK6/w05/+lOHDhxMQEMDtt9/u8Wfxx3LGHpUpFpEpwLPY/veXjTFPiMhCIMUYs0JEBPgDMBkoB54wxiwTkXHAC0AF9o/Ks8aYlxp6r+aWKc4uzGbOh3O4KfEmLj2zbYv6K9VatEyx//GknHFTyxR71EdvjFkJrKyx7lGXZQPc7zxct1kPDPfkPU5XdHg0y65e1viGSinVTqWlpXH55Zczffr0Fi1n7FNlipVSqiNrrXLGPlUCQSlf1N7uAqe8qzm/D5rolWrHwsLCOH78uCZ7Bdgkf/z48SZPCdWuG6Xasbi4ODIzM8nKyvJ2KKqdCAsLIy4urkn7aKJXqh0LDg4mISHB22GoDk67bpRSysdpoldKKR+niV4ppXycR1fGtiURyQL2NrBJDHCsjcJpKo2teTS25tHYmsdXYxtgjOlR1wvtLtE3RkRS6rvM19s0tubR2JpHY2sef4xNu26UUsrHaaJXSikf1xET/WJvB9AAja15NLbm0diax+9i63B99EoppZqmI7bolVJKNYEmeqWU8nEdJtGLyGQR2Skiu0VkvrfjqUlEMkTkWxFJFZGm3yKrZWN5WUSOishWl3XdRORTEfnO+RndjmJbICIHnHOX6tzRrK3j6icia0UkTUS2ici9znqvn7cGYmsP5y1MRL4Rkc1ObL911ieIyNfO/9c3nNuQtpfYlojIHpfzltjWsbnEGCgim0TkA+d565w3Y0y7f2BvYfg9MBAIATYDw7wdV40YM4AYb8fhxHIBMArY6rLuf4D5zvJ84PftKLYFwDwvn7NYYJSzHIm9wf2w9nDeGoitPZw3ASKc5WDga+AHwJvATGf988CcdhTbEuBqb543lxjvB14HPnCet8p56ygt+jHAbmNMujGmBFgGTPNyTO2WMeZz4ESN1dOAV53lV4Er2zKmSvXE5nXGmEPGmI3Ocj6wHehLOzhvDcTmdcYqcJ4GOw8DTALedtZ767zVF1u7ICJxwFTgb85zoZXOW0dJ9H2B/S7PM2knv+guDPBvEdkgIrd6O5g69DLGHHKWDwO9vBlMHe4SkS1O145XupUqiUg8MBLbAmxX561GbNAOzpvT/ZAKHAU+xX77zjHGlDmbeO3/a83YjDGV5+0J57w9IyKh3ogNeBZ4EKhwnnenlc5bR0n0HcEPjTGjgMuAO0XkAm8HVB9jvxe2m5YN8BxwBpAIHAL+4K1ARCQCeAeYa4zJc33N2+etjtjaxXkzxpQbYxKBOOy37yHeiKMuNWMTkXOAh7Axngd0A37Z1nGJyOXAUWPMhrZ4v46S6A8A/Vyexznr2g1jzAHn51HgPewvfHtyRERiAZyfR70cTxVjzBHnP2QF8CJeOnciEoxNpP8wxrzrrG4X562u2NrLeatkjMkB1gJjgSgRqbyxkdf/v7rENtnpCjPGmGLgFbxz3sYDV4hIBrYrehLwJ1rpvHWURJ8MDHJGpEOAmcAKL8dURUQ6i0hk5TLwY2Brw3u1uRXAjc7yjcC/vBiLm8pE6piOF86d0z/6ErDdGPNHl5e8ft7qi62dnLceIhLlLIcDl2DHENYCVzubeeu81RXbDpc/3ILtA2/z82aMecgYE2eMicfmszXGmOtorfPm7VHnJoxOT8HONvge+JW346kR20DsTKDNwDZvxwf8E/tVvhTbz3cLtv9vNfAdsAro1o5iWwp8C2zBJtZYL8T1Q2y3zBYg1XlMaQ/nrYHY2sN5OxfY5MSwFXjUWT8Q+AbYDbwFhLaj2NY4520r8BrOzBxvPYAJVM+6aZXzpiUQlFLKx3WUrhullFLNpIleKaV8nCZ6pZTycZrolVLKx2miV0opH6eJXimlfJwmeqWU8nH/H2CBSv5J9WbpAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plotting results\n",
    "acc = history.history['acc']\n",
    "\n",
    "val_acc = history.history['val_acc']\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "epochs = range(1, len(acc)+1)\n",
    "\n",
    "plt.plot(epochs, acc, 'g', label='Training accuracy')\n",
    "plt.plot(epochs, val_acc, 'r', label='Validation accuracy')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\n",
    "\n",
    "plt.figure()\n",
    "\n",
    "plt.plot(epochs, loss, 'g', label='Training loss')\n",
    "plt.plot(epochs, val_loss, 'r', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "\n",
    "\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "a7db0553",
   "metadata": {},
   "outputs": [],
   "source": [
    "prediction = model.predict(test_features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "325b23ae",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "pred = np.argmax(prediction, axis=1)\n",
    "# label\n",
    "test_labels = np.argmax(test_labels, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "5dd65c90",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "       Covid       0.95      0.97      0.96       220\n",
      "    NonCovid       0.90      0.90      0.90       220\n",
      "   Pneumonia       0.91      0.89      0.90       220\n",
      "\n",
      "    accuracy                           0.92       660\n",
      "   macro avg       0.92      0.92      0.92       660\n",
      "weighted avg       0.92      0.92      0.92       660\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(test_labels, pred, target_names = ['Covid','NonCovid','Pneumonia']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6ef55d48",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}