Torax Deep CNN Especifico 9c 450p.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU",
    "gpuClass": "standard"
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "cwfdyx0gI1OW",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 17
        },
        "outputId": "af43d05a-9800-494f-e674-a3da8c365d94"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ],
            "application/javascript": [
              "\n",
              "function ClickConnect(){\n",
              "console.log(\"Working\");\n",
              "document.querySelector(\"colab-toolbar-button#connect\").click()\n",
              "}\n",
              "setInterval(ClickConnect,60000)\n"
            ]
          },
          "metadata": {}
        }
      ],
      "source": [
        "import IPython\n",
        "js_code = '''\n",
        "function ClickConnect(){\n",
        "console.log(\"Working\");\n",
        "document.querySelector(\"colab-toolbar-button#connect\").click()\n",
        "}\n",
        "setInterval(ClickConnect,60000)\n",
        "'''\n",
        "display(IPython.display.Javascript(js_code))\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "mHJ9nWauqtiJ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "2e3c9503-cfd7-4551-8bf1-9d1c3a8e8ca0"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found GPU at: /device:GPU:0\n"
          ]
        }
      ],
      "source": [
        "%tensorflow_version 2.x\n",
        "\n",
        "import tensorflow as tf\n",
        "\n",
        "device_name = tf.test.gpu_device_name()\n",
        "\n",
        "if device_name != '/device:GPU:0':\n",
        "  raise SystemError('GPU device not found')\n",
        "\n",
        "print('Found GPU at: {}'.format(device_name))\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "nRtH1u2CLRg2",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "be9fabdb-7bcf-4b91-aee0-b901082d1b74"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ]
        }
      ],
      "source": [
        "from google.colab import drive\n",
        "\n",
        "drive.mount('/content/drive')\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "51RSMhae_FBg"
      },
      "outputs": [],
      "source": [
        "!pip install -q -U tensorflow-addons\n",
        "!pip install -q -U watermark\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install -q -U keras_sequential_ascii\n",
        "!pip install -q -U wandb\n"
      ],
      "metadata": {
        "id": "7G8fVMV3Oz-v"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "qCAFn3SRLgxZ"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "import keras\n",
        "import numpy as np\n",
        "import cv2\n",
        "import PIL\n",
        "import glob\n",
        "import matplotlib.pyplot as plt\n",
        "import tensorflow as tf\n",
        "import keras.layers.convolutional as conv\n",
        "from math import sqrt\n",
        "from PIL import Image\n",
        "from tensorflow import keras\n",
        "from numpy import mean\n",
        "from keras.models import Sequential, load_model\n",
        "from tensorflow.keras import regularizers, layers, Model\n",
        "from keras.preprocessing.image import ImageDataGenerator, load_img\n",
        "from keras.preprocessing import image\n",
        "from tensorflow.keras.preprocessing import image_dataset_from_directory\n",
        "from keras.callbacks import LearningRateScheduler, ReduceLROnPlateau, EarlyStopping\n",
        "from keras.utils.vis_utils import plot_model\n",
        "from keras import backend as B\n",
        "from tensorflow.keras.layers import ReLU\n",
        "from tensorflow.keras.layers import BatchNormalization,MaxPool2D\n",
        "from tensorflow.keras import Model\n",
        "from keras.layers import Dense, Add, Conv2D, GlobalAveragePooling2D, MaxPool2D, Flatten, BatchNormalization, Dropout, ZeroPadding2D, MaxPooling2D, Activation, Input, UpSampling2D, AveragePooling2D, Reshape, InputLayer, SeparableConv2D\n",
        "from tensorflow.keras.optimizers import Adam, Nadam, RMSprop, SGD\n",
        "from keras.regularizers import l2\n",
        "from keras.initializers import glorot_uniform\n",
        "from skimage import measure, feature, morphology, segmentation, color\n",
        "from skimage.feature import peak_local_max\n",
        "from skimage.morphology import disk\n",
        "from skimage.segmentation import watershed\n",
        "from skimage.exposure import equalize_adapthist\n",
        "from skimage.io import imread\n",
        "from skimage.color import rgb2gray\n",
        "from skimage.filters import sobel, rank, median\n",
        "from skimage.util import img_as_ubyte\n",
        "from scipy import ndimage as ndi\n",
        "from scipy.ndimage import distance_transform_edt\n",
        "import tensorflow_addons as tfa\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.utils import class_weight\n",
        "from sklearn.metrics import confusion_matrix, accuracy_score\n",
        "from tensorflow.keras.metrics import Metric, Recall, Precision, CategoricalAccuracy\n",
        "from keras import backend as K\n",
        "from sklearn.metrics import precision_recall_fscore_support, f1_score\n",
        "from tensorflow.keras.models import Model\n",
        "from keras_sequential_ascii import keras2ascii\n",
        "from wandb.keras import WandbCallback\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "WY04CdGsvo0d",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e627e133-b4d2-4f82-b434-5c9f54c0b370"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Bibliotecas utilizadas e suas respectivas versões:\n",
            "cv2              : 4.6.0\n",
            "IPython          : 5.5.0\n",
            "keras            : 2.8.0\n",
            "matplotlib       : 3.2.2\n",
            "tensorflow       : 2.8.2+zzzcolab20220719082949\n",
            "google           : 2.0.3\n",
            "sklearn          : 0.0\n",
            "PIL              : 7.1.2\n",
            "scipy            : 1.7.3\n",
            "numpy            : 1.21.6\n",
            "tensorflow_addons: 0.17.1\n",
            "skimage          : 0.18.3\n",
            "\n"
          ]
        }
      ],
      "source": [
        "%reload_ext watermark\n",
        "\n",
        "print('Bibliotecas utilizadas e suas respectivas versões:')\n",
        "%watermark --iversions\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "LJ0RefpJLg5y"
      },
      "outputs": [],
      "source": [
        "import warnings\n",
        "\n",
        "warnings.simplefilter(action = 'ignore', category = FutureWarning)\n",
        "warnings.simplefilter(action = 'ignore', category = UserWarning)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tx-CD5L0B5nT"
      },
      "source": [
        "## Modelo"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "-M1pg67ilBUH"
      },
      "outputs": [],
      "source": [
        "N_classes = 9\n",
        "N_epocas = [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,\n",
        "            21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,\n",
        "            41,42,43,44,45,46,47,48,49,50]\n",
        "            "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Anu2RjhLK7MU",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ed8004ef-a0b6-4be5-adb0-45fe20a8176d"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 2802 images belonging to 9 classes.\n",
            "Found 693 images belonging to 9 classes.\n"
          ]
        }
      ],
      "source": [
        "data = ImageDataGenerator(rescale = 1./255,\n",
        "                          rotation_range=5,\n",
        "                          horizontal_flip = True,\n",
        "                          zoom_range=0.15,\n",
        "                          width_shift_range=0.2,\n",
        "                          height_shift_range=0.2,\n",
        "                          shear_range=0.5,\n",
        "                          validation_split = 0.20)\n",
        "\n",
        "traindata = data.flow_from_directory(directory='/content/drive/MyDrive/Colab Notebooks/Datasets/RXTorax/1 - Especifico 9 Classes/',\n",
        "                                     target_size = (384,384),\n",
        "                                     class_mode = 'categorical',\n",
        "                                     batch_size = 16,\n",
        "                                     shuffle = True,\n",
        "                                     subset = 'training',\n",
        "                                     interpolation = 'nearest')\n",
        "\n",
        "testdata = data.flow_from_directory(directory='/content/drive/MyDrive/Colab Notebooks/Datasets/RXTorax/1 - Especifico 9 Classes/',\n",
        "                                     target_size = (384,384),\n",
        "                                     class_mode = 'categorical',\n",
        "                                     batch_size = 8,\n",
        "                                     shuffle = True,\n",
        "                                     subset = 'validation',\n",
        "                                     interpolation = 'nearest')\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(traindata[0][0]) # Retornará as arrays referentes as amostras\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WP9AIvwrc-7c",
        "outputId": "b83b582a-44c8-4572-866d-bacf91a7dbd1"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[[[[6.5238804e-01 6.5238804e-01 6.5238804e-01]\n",
            "   [6.5259498e-01 6.5259498e-01 6.5259498e-01]\n",
            "   [6.5280205e-01 6.5280205e-01 6.5280205e-01]\n",
            "   ...\n",
            "   [6.9275063e-01 6.9275063e-01 6.9275063e-01]\n",
            "   [6.9347519e-01 6.9347519e-01 6.9347519e-01]\n",
            "   [6.9419974e-01 6.9419974e-01 6.9419974e-01]]\n",
            "\n",
            "  [[6.4015883e-01 6.4015883e-01 6.4015883e-01]\n",
            "   [6.4057285e-01 6.4057285e-01 6.4057285e-01]\n",
            "   [6.4098692e-01 6.4098692e-01 6.4098692e-01]\n",
            "   ...\n",
            "   [6.8130648e-01 6.8130648e-01 6.8130648e-01]\n",
            "   [6.8151349e-01 6.8151349e-01 6.8151349e-01]\n",
            "   [6.8172055e-01 6.8172055e-01 6.8172055e-01]]\n",
            "\n",
            "  [[6.3090813e-01 6.3090813e-01 6.3090813e-01]\n",
            "   [6.3111514e-01 6.3111514e-01 6.3111514e-01]\n",
            "   [6.3132215e-01 6.3132215e-01 6.3132215e-01]\n",
            "   ...\n",
            "   [6.7843139e-01 6.7843139e-01 6.7843139e-01]\n",
            "   [6.7843139e-01 6.7843139e-01 6.7843139e-01]\n",
            "   [6.7843139e-01 6.7843139e-01 6.7843139e-01]]\n",
            "\n",
            "  ...\n",
            "\n",
            "  [[4.6290064e-01 4.6290064e-01 4.6290064e-01]\n",
            "   [4.6279714e-01 4.6279714e-01 4.6279714e-01]\n",
            "   [4.6264210e-01 4.6264210e-01 4.6264210e-01]\n",
            "   ...\n",
            "   [6.8241888e-01 6.8241888e-01 6.8241888e-01]\n",
            "   [6.8252230e-01 6.8252230e-01 6.8252230e-01]\n",
            "   [6.8262583e-01 6.8262583e-01 6.8262583e-01]]\n",
            "\n",
            "  [[4.6630982e-01 4.6630982e-01 4.6630982e-01]\n",
            "   [4.6620631e-01 4.6620631e-01 4.6620631e-01]\n",
            "   [4.6610281e-01 4.6610281e-01 4.6610281e-01]\n",
            "   ...\n",
            "   [6.8569630e-01 6.8569630e-01 6.8569630e-01]\n",
            "   [6.8559283e-01 6.8559283e-01 6.8559283e-01]\n",
            "   [6.8548930e-01 6.8548930e-01 6.8548930e-01]]\n",
            "\n",
            "  [[4.5140514e-01 4.5140514e-01 4.5140514e-01]\n",
            "   [4.5192268e-01 4.5192268e-01 4.5192268e-01]\n",
            "   [4.5244023e-01 4.5244023e-01 4.5244023e-01]\n",
            "   ...\n",
            "   [6.7778176e-01 6.7778176e-01 6.7778176e-01]\n",
            "   [6.7809224e-01 6.7809224e-01 6.7809224e-01]\n",
            "   [6.7840278e-01 6.7840278e-01 6.7840278e-01]]]\n",
            "\n",
            "\n",
            " [[[7.8431377e-03 7.8431377e-03 7.8431377e-03]\n",
            "   [7.8431377e-03 7.8431377e-03 7.8431377e-03]\n",
            "   [7.8431377e-03 7.8431377e-03 7.8431377e-03]\n",
            "   ...\n",
            "   [2.3390682e-01 2.3390682e-01 2.3390682e-01]\n",
            "   [2.3529413e-01 2.3529413e-01 2.3529413e-01]\n",
            "   [2.3529413e-01 2.3529413e-01 2.3529413e-01]]\n",
            "\n",
            "  [[7.8431377e-03 7.8431377e-03 7.8431377e-03]\n",
            "   [7.8431377e-03 7.8431377e-03 7.8431377e-03]\n",
            "   [7.8431377e-03 7.8431377e-03 7.8431377e-03]\n",
            "   ...\n",
            "   [2.3422889e-01 2.3422889e-01 2.3422889e-01]\n",
            "   [2.3529413e-01 2.3529413e-01 2.3529413e-01]\n",
            "   [2.3529413e-01 2.3529413e-01 2.3529413e-01]]\n",
            "\n",
            "  [[7.8431377e-03 7.8431377e-03 7.8431377e-03]\n",
            "   [7.8431377e-03 7.8431377e-03 7.8431377e-03]\n",
            "   [7.8431377e-03 7.8431377e-03 7.8431377e-03]\n",
            "   ...\n",
            "   [2.3455095e-01 2.3455095e-01 2.3455095e-01]\n",
            "   [2.3529413e-01 2.3529413e-01 2.3529413e-01]\n",
            "   [2.3552386e-01 2.3552386e-01 2.3552386e-01]]\n",
            "\n",
            "  ...\n",
            "\n",
            "  [[6.6274512e-01 6.6274512e-01 6.6274512e-01]\n",
            "   [6.6274512e-01 6.6274512e-01 6.6274512e-01]\n",
            "   [6.6278392e-01 6.6278392e-01 6.6278392e-01]\n",
            "   ...\n",
            "   [2.1176472e-01 2.1176472e-01 2.1176472e-01]\n",
            "   [2.1176472e-01 2.1176472e-01 2.1176472e-01]\n",
            "   [2.1176472e-01 2.1176472e-01 2.1176472e-01]]\n",
            "\n",
            "  [[6.6042900e-01 6.6042900e-01 6.6042900e-01]\n",
            "   [6.6161287e-01 6.6161287e-01 6.6161287e-01]\n",
            "   [6.6274512e-01 6.6274512e-01 6.6274512e-01]\n",
            "   ...\n",
            "   [2.0840922e-01 2.0840922e-01 2.0840922e-01]\n",
            "   [2.0880386e-01 2.0880386e-01 2.0880386e-01]\n",
            "   [2.0919850e-01 2.0919850e-01 2.0919850e-01]]\n",
            "\n",
            "  [[6.4859962e-01 6.4859962e-01 6.4859962e-01]\n",
            "   [6.4978361e-01 6.4978361e-01 6.4978361e-01]\n",
            "   [6.5096748e-01 6.5096748e-01 6.5096748e-01]\n",
            "   ...\n",
            "   [2.1459723e-01 2.1459723e-01 2.1459723e-01]\n",
            "   [2.1380794e-01 2.1380794e-01 2.1380794e-01]\n",
            "   [2.1301866e-01 2.1301866e-01 2.1301866e-01]]]\n",
            "\n",
            "\n",
            " [[[1.5499569e-01 1.5499569e-01 1.5499569e-01]\n",
            "   [1.6197631e-01 1.6197631e-01 1.6197631e-01]\n",
            "   [1.7133652e-01 1.7133652e-01 1.7133652e-01]\n",
            "   ...\n",
            "   [5.0617874e-01 5.0617874e-01 5.0617874e-01]\n",
            "   [4.6328661e-01 4.6328661e-01 4.6328661e-01]\n",
            "   [4.4676307e-01 4.4676307e-01 4.4676307e-01]]\n",
            "\n",
            "  [[1.5512490e-01 1.5512490e-01 1.5512490e-01]\n",
            "   [1.6236392e-01 1.6236392e-01 1.6236392e-01]\n",
            "   [1.7159492e-01 1.7159492e-01 1.7159492e-01]\n",
            "   ...\n",
            "   [5.0346547e-01 5.0346547e-01 5.0346547e-01]\n",
            "   [4.6302819e-01 4.6302819e-01 4.6302819e-01]\n",
            "   [4.4560024e-01 4.4560024e-01 4.4560024e-01]]\n",
            "\n",
            "  [[1.5525411e-01 1.5525411e-01 1.5525411e-01]\n",
            "   [1.6275154e-01 1.6275154e-01 1.6275154e-01]\n",
            "   [1.7185333e-01 1.7185333e-01 1.7185333e-01]\n",
            "   ...\n",
            "   [5.0075209e-01 5.0075209e-01 5.0075209e-01]\n",
            "   [4.6276978e-01 4.6276978e-01 4.6276978e-01]\n",
            "   [4.4443735e-01 4.4443735e-01 4.4443735e-01]]\n",
            "\n",
            "  ...\n",
            "\n",
            "  [[7.2770524e-01 7.2770524e-01 7.2770524e-01]\n",
            "   [7.2596467e-01 7.2596467e-01 7.2596467e-01]\n",
            "   [7.2916734e-01 7.2916734e-01 7.2916734e-01]\n",
            "   ...\n",
            "   [3.6343660e-03 3.6343660e-03 3.6343660e-03]\n",
            "   [6.2791572e-05 6.2791572e-05 6.2791572e-05]\n",
            "   [0.0000000e+00 0.0000000e+00 0.0000000e+00]]\n",
            "\n",
            "  [[7.3154879e-01 7.3154879e-01 7.3154879e-01]\n",
            "   [7.3148000e-01 7.3148000e-01 7.3148000e-01]\n",
            "   [7.3490202e-01 7.3490202e-01 7.3490202e-01]\n",
            "   ...\n",
            "   [3.5051601e-03 3.5051601e-03 3.5051601e-03]\n",
            "   [0.0000000e+00 0.0000000e+00 0.0000000e+00]\n",
            "   [0.0000000e+00 0.0000000e+00 0.0000000e+00]]\n",
            "\n",
            "  [[7.3139644e-01 7.3139644e-01 7.3139644e-01]\n",
            "   [7.3175609e-01 7.3175609e-01 7.3175609e-01]\n",
            "   [7.3547727e-01 7.3547727e-01 7.3547727e-01]\n",
            "   ...\n",
            "   [3.3759538e-03 3.3759538e-03 3.3759538e-03]\n",
            "   [0.0000000e+00 0.0000000e+00 0.0000000e+00]\n",
            "   [0.0000000e+00 0.0000000e+00 0.0000000e+00]]]\n",
            "\n",
            "\n",
            " ...\n",
            "\n",
            "\n",
            " [[[5.4117650e-01 5.4117650e-01 5.4117650e-01]\n",
            "   [5.4117650e-01 5.4117650e-01 5.4117650e-01]\n",
            "   [5.4174685e-01 5.4174685e-01 5.4174685e-01]\n",
            "   ...\n",
            "   [3.9215688e-02 3.9215688e-02 3.9215688e-02]\n",
            "   [3.9215688e-02 3.9215688e-02 3.9215688e-02]\n",
            "   [3.9215688e-02 3.9215688e-02 3.9215688e-02]]\n",
            "\n",
            "  [[5.4117650e-01 5.4117650e-01 5.4117650e-01]\n",
            "   [5.4117650e-01 5.4117650e-01 5.4117650e-01]\n",
            "   [5.4182166e-01 5.4182166e-01 5.4182166e-01]\n",
            "   ...\n",
            "   [3.9215688e-02 3.9215688e-02 3.9215688e-02]\n",
            "   [3.9215688e-02 3.9215688e-02 3.9215688e-02]\n",
            "   [3.9215688e-02 3.9215688e-02 3.9215688e-02]]\n",
            "\n",
            "  [[5.4117650e-01 5.4117650e-01 5.4117650e-01]\n",
            "   [5.4117650e-01 5.4117650e-01 5.4117650e-01]\n",
            "   [5.4189646e-01 5.4189646e-01 5.4189646e-01]\n",
            "   ...\n",
            "   [3.9215688e-02 3.9215688e-02 3.9215688e-02]\n",
            "   [3.9215688e-02 3.9215688e-02 3.9215688e-02]\n",
            "   [3.9215688e-02 3.9215688e-02 3.9215688e-02]]\n",
            "\n",
            "  ...\n",
            "\n",
            "  [[6.2135911e-01 6.2135911e-01 6.2135911e-01]\n",
            "   [6.2857842e-01 6.2857842e-01 6.2857842e-01]\n",
            "   [6.3831800e-01 6.3831800e-01 6.3831800e-01]\n",
            "   ...\n",
            "   [6.2600994e-01 6.2600994e-01 6.2600994e-01]\n",
            "   [6.1617714e-01 6.1617714e-01 6.1617714e-01]\n",
            "   [6.0923111e-01 6.0923111e-01 6.0923111e-01]]\n",
            "\n",
            "  [[6.2352943e-01 6.2352943e-01 6.2352943e-01]\n",
            "   [6.2528366e-01 6.2528366e-01 6.2528366e-01]\n",
            "   [6.3247699e-01 6.3247699e-01 6.3247699e-01]\n",
            "   ...\n",
            "   [6.2745100e-01 6.2745100e-01 6.2745100e-01]\n",
            "   [6.1967897e-01 6.1967897e-01 6.1967897e-01]\n",
            "   [6.1658818e-01 6.1658818e-01 6.1658818e-01]]\n",
            "\n",
            "  [[6.2437558e-01 6.2437558e-01 6.2437558e-01]\n",
            "   [6.2623823e-01 6.2623823e-01 6.2623823e-01]\n",
            "   [6.3184375e-01 6.3184375e-01 6.3184375e-01]\n",
            "   ...\n",
            "   [6.2485403e-01 6.2485403e-01 6.2485403e-01]\n",
            "   [6.2044966e-01 6.2044966e-01 6.2044966e-01]\n",
            "   [6.1757696e-01 6.1757696e-01 6.1757696e-01]]]\n",
            "\n",
            "\n",
            " [[[5.5780017e-01 5.5780017e-01 5.5780017e-01]\n",
            "   [5.6873941e-01 5.6873941e-01 5.6873941e-01]\n",
            "   [5.7801872e-01 5.7801872e-01 5.7801872e-01]\n",
            "   ...\n",
            "   [4.7058827e-01 4.7058827e-01 4.7058827e-01]\n",
            "   [4.7058827e-01 4.7058827e-01 4.7058827e-01]\n",
            "   [4.7058827e-01 4.7058827e-01 4.7058827e-01]]\n",
            "\n",
            "  [[5.5773664e-01 5.5773664e-01 5.5773664e-01]\n",
            "   [5.6861246e-01 5.6861246e-01 5.6861246e-01]\n",
            "   [5.7798702e-01 5.7798702e-01 5.7798702e-01]\n",
            "   ...\n",
            "   [4.7058827e-01 4.7058827e-01 4.7058827e-01]\n",
            "   [4.7058827e-01 4.7058827e-01 4.7058827e-01]\n",
            "   [4.7058827e-01 4.7058827e-01 4.7058827e-01]]\n",
            "\n",
            "  [[5.5767316e-01 5.5767316e-01 5.5767316e-01]\n",
            "   [5.6848556e-01 5.6848556e-01 5.6848556e-01]\n",
            "   [5.7795519e-01 5.7795519e-01 5.7795519e-01]\n",
            "   ...\n",
            "   [4.7058827e-01 4.7058827e-01 4.7058827e-01]\n",
            "   [4.7058827e-01 4.7058827e-01 4.7058827e-01]\n",
            "   [4.7058827e-01 4.7058827e-01 4.7058827e-01]]\n",
            "\n",
            "  ...\n",
            "\n",
            "  [[7.1334863e-01 7.1334863e-01 7.1334863e-01]\n",
            "   [7.0996898e-01 7.0996898e-01 7.0996898e-01]\n",
            "   [7.0491374e-01 7.0491374e-01 7.0491374e-01]\n",
            "   ...\n",
            "   [5.9571487e-01 5.9571487e-01 5.9571487e-01]\n",
            "   [5.9592271e-01 5.9592271e-01 5.9592271e-01]\n",
            "   [5.9613055e-01 5.9613055e-01 5.9613055e-01]]\n",
            "\n",
            "  [[7.1119851e-01 7.1119851e-01 7.1119851e-01]\n",
            "   [7.0796311e-01 7.0796311e-01 7.0796311e-01]\n",
            "   [7.0588237e-01 7.0588237e-01 7.0588237e-01]\n",
            "   ...\n",
            "   [6.0834002e-01 6.0834002e-01 6.0834002e-01]\n",
            "   [6.0839194e-01 6.0839194e-01 6.0839194e-01]\n",
            "   [6.0844392e-01 6.0844392e-01 6.0844392e-01]]\n",
            "\n",
            "  [[7.0600361e-01 7.0600361e-01 7.0600361e-01]\n",
            "   [7.0590019e-01 7.0590019e-01 7.0590019e-01]\n",
            "   [7.0588887e-01 7.0588887e-01 7.0588887e-01]\n",
            "   ...\n",
            "   [6.1197305e-01 6.1197305e-01 6.1197305e-01]\n",
            "   [6.1207694e-01 6.1207694e-01 6.1207694e-01]\n",
            "   [6.1218083e-01 6.1218083e-01 6.1218083e-01]]]\n",
            "\n",
            "\n",
            " [[[2.9019609e-01 2.9019609e-01 2.9019609e-01]\n",
            "   [2.9019609e-01 2.9019609e-01 2.9019609e-01]\n",
            "   [2.9019609e-01 2.9019609e-01 2.9019609e-01]\n",
            "   ...\n",
            "   [5.5526823e-01 5.5526823e-01 5.5526823e-01]\n",
            "   [5.2861714e-01 5.2861714e-01 5.2861714e-01]\n",
            "   [4.9827233e-01 4.9827233e-01 4.9827233e-01]]\n",
            "\n",
            "  [[2.9019609e-01 2.9019609e-01 2.9019609e-01]\n",
            "   [2.9019609e-01 2.9019609e-01 2.9019609e-01]\n",
            "   [2.9019609e-01 2.9019609e-01 2.9019609e-01]\n",
            "   ...\n",
            "   [5.5377287e-01 5.5377287e-01 5.5377287e-01]\n",
            "   [5.2690816e-01 5.2690816e-01 5.2690816e-01]\n",
            "   [4.9711680e-01 4.9711680e-01 4.9711680e-01]]\n",
            "\n",
            "  [[2.9019609e-01 2.9019609e-01 2.9019609e-01]\n",
            "   [2.9019609e-01 2.9019609e-01 2.9019609e-01]\n",
            "   [2.9019609e-01 2.9019609e-01 2.9019609e-01]\n",
            "   ...\n",
            "   [5.5227751e-01 5.5227751e-01 5.5227751e-01]\n",
            "   [5.2519917e-01 5.2519917e-01 5.2519917e-01]\n",
            "   [4.9604869e-01 4.9604869e-01 4.9604869e-01]]\n",
            "\n",
            "  ...\n",
            "\n",
            "  [[0.0000000e+00 0.0000000e+00 0.0000000e+00]\n",
            "   [0.0000000e+00 0.0000000e+00 0.0000000e+00]\n",
            "   [0.0000000e+00 0.0000000e+00 0.0000000e+00]\n",
            "   ...\n",
            "   [1.0955009e-02 1.0955009e-02 1.0955009e-02]\n",
            "   [1.1764707e-02 1.1764707e-02 1.1764707e-02]\n",
            "   [2.9749444e-03 2.9749444e-03 2.9749444e-03]]\n",
            "\n",
            "  [[0.0000000e+00 0.0000000e+00 0.0000000e+00]\n",
            "   [0.0000000e+00 0.0000000e+00 0.0000000e+00]\n",
            "   [0.0000000e+00 0.0000000e+00 0.0000000e+00]\n",
            "   ...\n",
            "   [4.4841040e-03 4.4841040e-03 4.4841040e-03]\n",
            "   [1.3522735e-03 1.3522735e-03 1.3522735e-03]\n",
            "   [8.5765673e-03 8.5765673e-03 8.5765673e-03]]\n",
            "\n",
            "  [[0.0000000e+00 0.0000000e+00 0.0000000e+00]\n",
            "   [0.0000000e+00 0.0000000e+00 0.0000000e+00]\n",
            "   [0.0000000e+00 0.0000000e+00 0.0000000e+00]\n",
            "   ...\n",
            "   [4.2290371e-03 4.2290371e-03 4.2290371e-03]\n",
            "   [4.3593510e-04 4.3593510e-04 4.3593510e-04]\n",
            "   [1.2964684e-02 1.2964684e-02 1.2964684e-02]]]]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(traindata[0][1]) # Retornará uma array referentes as labels das 16 amostras geradas para cada batch\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kRscR7fokxGe",
        "outputId": "9eb47fea-2868-458d-f9bc-61030fb89774"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[[0. 0. 0. 0. 1. 0. 0. 0. 0.]\n",
            " [0. 1. 0. 0. 0. 0. 0. 0. 0.]\n",
            " [1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            " [0. 0. 0. 0. 0. 1. 0. 0. 0.]\n",
            " [0. 0. 0. 0. 0. 1. 0. 0. 0.]\n",
            " [1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            " [0. 0. 0. 0. 0. 1. 0. 0. 0.]\n",
            " [1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            " [0. 1. 0. 0. 0. 0. 0. 0. 0.]\n",
            " [0. 1. 0. 0. 0. 0. 0. 0. 0.]\n",
            " [1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            " [0. 1. 0. 0. 0. 0. 0. 0. 0.]\n",
            " [0. 1. 0. 0. 0. 0. 0. 0. 0.]\n",
            " [0. 0. 0. 0. 0. 0. 0. 1. 0.]\n",
            " [0. 1. 0. 0. 0. 0. 0. 0. 0.]\n",
            " [0. 0. 0. 0. 0. 0. 1. 0. 0.]]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(traindata[0][0][0]) # Array referente a primeira amostra\n",
        "print(traindata[0][1][0]) # Retorna as labels referentes a primeira amostra\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JqZKdsbTnrVN",
        "outputId": "ec949858-5d05-4079-955d-f0e510f0bcdf"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[[[0.41294286 0.41294286 0.41294286]\n",
            "  [0.4156863  0.4156863  0.4156863 ]\n",
            "  [0.41043422 0.41043422 0.41043422]\n",
            "  ...\n",
            "  [0.15686275 0.15686275 0.15686275]\n",
            "  [0.15686275 0.15686275 0.15686275]\n",
            "  [0.15686275 0.15686275 0.15686275]]\n",
            "\n",
            " [[0.4132689  0.4132689  0.4132689 ]\n",
            "  [0.4156863  0.4156863  0.4156863 ]\n",
            "  [0.4091302  0.4091302  0.4091302 ]\n",
            "  ...\n",
            "  [0.15686275 0.15686275 0.15686275]\n",
            "  [0.15686275 0.15686275 0.15686275]\n",
            "  [0.15686275 0.15686275 0.15686275]]\n",
            "\n",
            " [[0.4135949  0.4135949  0.4135949 ]\n",
            "  [0.4156863  0.4156863  0.4156863 ]\n",
            "  [0.40782616 0.40782616 0.40782616]\n",
            "  ...\n",
            "  [0.15686275 0.15686275 0.15686275]\n",
            "  [0.15686275 0.15686275 0.15686275]\n",
            "  [0.15686275 0.15686275 0.15686275]]\n",
            "\n",
            " ...\n",
            "\n",
            " [[0.77238536 0.77238536 0.77238536]\n",
            "  [0.7815108  0.7815108  0.7815108 ]\n",
            "  [0.77647066 0.77647066 0.77647066]\n",
            "  ...\n",
            "  [0.6862745  0.6862745  0.6862745 ]\n",
            "  [0.6862745  0.6862745  0.6862745 ]\n",
            "  [0.6862745  0.6862745  0.6862745 ]]\n",
            "\n",
            " [[0.7733173  0.7733173  0.7733173 ]\n",
            "  [0.7850643  0.7850643  0.7850643 ]\n",
            "  [0.7774501  0.7774501  0.7774501 ]\n",
            "  ...\n",
            "  [0.6862745  0.6862745  0.6862745 ]\n",
            "  [0.6862745  0.6862745  0.6862745 ]\n",
            "  [0.6862745  0.6862745  0.6862745 ]]\n",
            "\n",
            " [[0.77762645 0.77762645 0.77762645]\n",
            "  [0.78645694 0.78645694 0.78645694]\n",
            "  [0.78392994 0.78392994 0.78392994]\n",
            "  ...\n",
            "  [0.6855863  0.6855863  0.6855863 ]\n",
            "  [0.6862745  0.6862745  0.6862745 ]\n",
            "  [0.6862745  0.6862745  0.6862745 ]]]\n",
            "[0. 0. 0. 0. 1. 0. 0. 0. 0.]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "for i in traindata:\n",
        "  print(f'Para cada imagem de origem serão geradas {i[0].shape[0]} novas imagens sintéticas')\n",
        "  print(f'As imagens finais possuirão {i[0].shape[1]}x{i[0].shape[2]} pixels em {i[0].shape[3]} canais de cor')\n",
        "  break\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EQWyiCUceq4c",
        "outputId": "ec2ec545-45bc-4872-c2c2-4ab15ba1f62b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Para cada imagem de origem serão geradas 16 novas imagens sintéticas\n",
            "As imagens finais possuirão 384x384 pixels em 3 canais de cor\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Uax4hz4j4Q2Q"
      },
      "outputs": [],
      "source": [
        "METRICS = [keras.metrics.CategoricalAccuracy(name = 'accuracy'),\n",
        "           keras.metrics.TruePositives(thresholds = 0.50, name = 'tp'),\n",
        "           keras.metrics.TrueNegatives(thresholds = 0.50, name = 'tn'),\n",
        "           keras.metrics.FalsePositives(thresholds = 0.50, name = 'fp'),\n",
        "           keras.metrics.FalseNegatives(thresholds = 0.50, name = 'fn'),\n",
        "           keras.metrics.PrecisionAtRecall(recall = 0.50, name = 'precision'),\n",
        "           keras.metrics.SensitivityAtSpecificity(0.50, name = 'sensitivity'),\n",
        "           keras.metrics.SpecificityAtSensitivity(sensitivity = 0.50,\n",
        "                                                  name = 'specificity'),\n",
        "           keras.metrics.Recall(name='recall'),\n",
        "           tfa.metrics.FBetaScore(num_classes = N_classes,\n",
        "                                  average = None,\n",
        "                                  threshold = 0.50,\n",
        "                                  name = 'FBetaScore')]\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "SRDg0GfJWuBC"
      },
      "outputs": [],
      "source": [
        "from tensorflow.keras.applications.inception_resnet_v2 import InceptionResNetV2\n",
        "\n",
        "base_model = InceptionResNetV2(weights = 'imagenet',\n",
        "                               include_top = False,\n",
        "                               input_shape = (384,384,3))\n",
        "                                                   \n",
        "x = base_model.output\n",
        "x = GlobalAveragePooling2D()(x)\n",
        "x = Dense(1024, activation = 'relu')(x)\n",
        "x = Dense(1024, activation = 'relu')(x)\n",
        "x = Dropout(rate = 0.25)(x)\n",
        "x = Dense(1024, activation = 'relu')(x)\n",
        "x = Dense(1024, activation = 'relu')(x)\n",
        "x = Dropout(rate = 0.25)(x)\n",
        "x = Dense(1024, activation = 'relu')(x)\n",
        "x = Dense(1024, activation = 'relu')(x)\n",
        "\n",
        "predictions = Dense(units = N_classes,\n",
        "                    activation = 'softmax')(x)\n",
        "\n",
        "model = Model(inputs = base_model.input,\n",
        "              outputs = predictions)\n",
        "\n",
        "for layer in model.layers:\n",
        "  layer.trainable = True\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "9mJfrj6tCDiJ"
      },
      "outputs": [],
      "source": [
        "opt = Adam(learning_rate=0.0001,\n",
        "           beta_1=0.9,\n",
        "           beta_2=0.999,\n",
        "           epsilon=1e-07,\n",
        "           amsgrad=False)\n",
        "\n",
        "model.compile(optimizer = opt,\n",
        "              loss = keras.losses.categorical_crossentropy,\n",
        "              metrics = METRICS)\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "bf-898GHCDiJ"
      },
      "outputs": [],
      "source": [
        "earlystop = EarlyStopping(monitor='loss',\n",
        "                          min_delta = 0,\n",
        "                          patience = 10,\n",
        "                          verbose = 1,\n",
        "                          mode = 'min')\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "esG1OGGzCDiJ"
      },
      "outputs": [],
      "source": [
        "learning_rate = ReduceLROnPlateau(monitor='accuracy',\n",
        "                                  factor=0.2,\n",
        "                                  patience=1,\n",
        "                                  min_lr=0.00001,\n",
        "                                  verbose=1)\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "901PihEqCDiJ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "88f99036-2245-4165-ab9f-7f58d8ea7cc3"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            " Layer (type)                   Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            " input_1 (InputLayer)           [(None, 384, 384, 3  0           []                               \n",
            "                                )]                                                                \n",
            "                                                                                                  \n",
            " conv2d (Conv2D)                (None, 191, 191, 32  864         ['input_1[0][0]']                \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " batch_normalization (BatchNorm  (None, 191, 191, 32  96         ['conv2d[0][0]']                 \n",
            " alization)                     )                                                                 \n",
            "                                                                                                  \n",
            " activation (Activation)        (None, 191, 191, 32  0           ['batch_normalization[0][0]']    \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_1 (Conv2D)              (None, 189, 189, 32  9216        ['activation[0][0]']             \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " batch_normalization_1 (BatchNo  (None, 189, 189, 32  96         ['conv2d_1[0][0]']               \n",
            " rmalization)                   )                                                                 \n",
            "                                                                                                  \n",
            " activation_1 (Activation)      (None, 189, 189, 32  0           ['batch_normalization_1[0][0]']  \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_2 (Conv2D)              (None, 189, 189, 64  18432       ['activation_1[0][0]']           \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " batch_normalization_2 (BatchNo  (None, 189, 189, 64  192        ['conv2d_2[0][0]']               \n",
            " rmalization)                   )                                                                 \n",
            "                                                                                                  \n",
            " activation_2 (Activation)      (None, 189, 189, 64  0           ['batch_normalization_2[0][0]']  \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " max_pooling2d (MaxPooling2D)   (None, 94, 94, 64)   0           ['activation_2[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_3 (Conv2D)              (None, 94, 94, 80)   5120        ['max_pooling2d[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_3 (BatchNo  (None, 94, 94, 80)  240         ['conv2d_3[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " activation_3 (Activation)      (None, 94, 94, 80)   0           ['batch_normalization_3[0][0]']  \n",
            "                                                                                                  \n",
            " conv2d_4 (Conv2D)              (None, 92, 92, 192)  138240      ['activation_3[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_4 (BatchNo  (None, 92, 92, 192)  576        ['conv2d_4[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " activation_4 (Activation)      (None, 92, 92, 192)  0           ['batch_normalization_4[0][0]']  \n",
            "                                                                                                  \n",
            " max_pooling2d_1 (MaxPooling2D)  (None, 45, 45, 192)  0          ['activation_4[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_8 (Conv2D)              (None, 45, 45, 64)   12288       ['max_pooling2d_1[0][0]']        \n",
            "                                                                                                  \n",
            " batch_normalization_8 (BatchNo  (None, 45, 45, 64)  192         ['conv2d_8[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " activation_8 (Activation)      (None, 45, 45, 64)   0           ['batch_normalization_8[0][0]']  \n",
            "                                                                                                  \n",
            " conv2d_6 (Conv2D)              (None, 45, 45, 48)   9216        ['max_pooling2d_1[0][0]']        \n",
            "                                                                                                  \n",
            " conv2d_9 (Conv2D)              (None, 45, 45, 96)   55296       ['activation_8[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_6 (BatchNo  (None, 45, 45, 48)  144         ['conv2d_6[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " batch_normalization_9 (BatchNo  (None, 45, 45, 96)  288         ['conv2d_9[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " activation_6 (Activation)      (None, 45, 45, 48)   0           ['batch_normalization_6[0][0]']  \n",
            "                                                                                                  \n",
            " activation_9 (Activation)      (None, 45, 45, 96)   0           ['batch_normalization_9[0][0]']  \n",
            "                                                                                                  \n",
            " average_pooling2d (AveragePool  (None, 45, 45, 192)  0          ['max_pooling2d_1[0][0]']        \n",
            " ing2D)                                                                                           \n",
            "                                                                                                  \n",
            " conv2d_5 (Conv2D)              (None, 45, 45, 96)   18432       ['max_pooling2d_1[0][0]']        \n",
            "                                                                                                  \n",
            " conv2d_7 (Conv2D)              (None, 45, 45, 64)   76800       ['activation_6[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_10 (Conv2D)             (None, 45, 45, 96)   82944       ['activation_9[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_11 (Conv2D)             (None, 45, 45, 64)   12288       ['average_pooling2d[0][0]']      \n",
            "                                                                                                  \n",
            " batch_normalization_5 (BatchNo  (None, 45, 45, 96)  288         ['conv2d_5[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " batch_normalization_7 (BatchNo  (None, 45, 45, 64)  192         ['conv2d_7[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " batch_normalization_10 (BatchN  (None, 45, 45, 96)  288         ['conv2d_10[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_11 (BatchN  (None, 45, 45, 64)  192         ['conv2d_11[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_5 (Activation)      (None, 45, 45, 96)   0           ['batch_normalization_5[0][0]']  \n",
            "                                                                                                  \n",
            " activation_7 (Activation)      (None, 45, 45, 64)   0           ['batch_normalization_7[0][0]']  \n",
            "                                                                                                  \n",
            " activation_10 (Activation)     (None, 45, 45, 96)   0           ['batch_normalization_10[0][0]'] \n",
            "                                                                                                  \n",
            " activation_11 (Activation)     (None, 45, 45, 64)   0           ['batch_normalization_11[0][0]'] \n",
            "                                                                                                  \n",
            " mixed_5b (Concatenate)         (None, 45, 45, 320)  0           ['activation_5[0][0]',           \n",
            "                                                                  'activation_7[0][0]',           \n",
            "                                                                  'activation_10[0][0]',          \n",
            "                                                                  'activation_11[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_15 (Conv2D)             (None, 45, 45, 32)   10240       ['mixed_5b[0][0]']               \n",
            "                                                                                                  \n",
            " batch_normalization_15 (BatchN  (None, 45, 45, 32)  96          ['conv2d_15[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_15 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_15[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_13 (Conv2D)             (None, 45, 45, 32)   10240       ['mixed_5b[0][0]']               \n",
            "                                                                                                  \n",
            " conv2d_16 (Conv2D)             (None, 45, 45, 48)   13824       ['activation_15[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_13 (BatchN  (None, 45, 45, 32)  96          ['conv2d_13[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_16 (BatchN  (None, 45, 45, 48)  144         ['conv2d_16[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_13 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_13[0][0]'] \n",
            "                                                                                                  \n",
            " activation_16 (Activation)     (None, 45, 45, 48)   0           ['batch_normalization_16[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_12 (Conv2D)             (None, 45, 45, 32)   10240       ['mixed_5b[0][0]']               \n",
            "                                                                                                  \n",
            " conv2d_14 (Conv2D)             (None, 45, 45, 32)   9216        ['activation_13[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_17 (Conv2D)             (None, 45, 45, 64)   27648       ['activation_16[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_12 (BatchN  (None, 45, 45, 32)  96          ['conv2d_12[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_14 (BatchN  (None, 45, 45, 32)  96          ['conv2d_14[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_17 (BatchN  (None, 45, 45, 64)  192         ['conv2d_17[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_12 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_12[0][0]'] \n",
            "                                                                                                  \n",
            " activation_14 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_14[0][0]'] \n",
            "                                                                                                  \n",
            " activation_17 (Activation)     (None, 45, 45, 64)   0           ['batch_normalization_17[0][0]'] \n",
            "                                                                                                  \n",
            " block35_1_mixed (Concatenate)  (None, 45, 45, 128)  0           ['activation_12[0][0]',          \n",
            "                                                                  'activation_14[0][0]',          \n",
            "                                                                  'activation_17[0][0]']          \n",
            "                                                                                                  \n",
            " block35_1_conv (Conv2D)        (None, 45, 45, 320)  41280       ['block35_1_mixed[0][0]']        \n",
            "                                                                                                  \n",
            " block35_1 (Lambda)             (None, 45, 45, 320)  0           ['mixed_5b[0][0]',               \n",
            "                                                                  'block35_1_conv[0][0]']         \n",
            "                                                                                                  \n",
            " block35_1_ac (Activation)      (None, 45, 45, 320)  0           ['block35_1[0][0]']              \n",
            "                                                                                                  \n",
            " conv2d_21 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_1_ac[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_21 (BatchN  (None, 45, 45, 32)  96          ['conv2d_21[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_21 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_21[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_19 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_1_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_22 (Conv2D)             (None, 45, 45, 48)   13824       ['activation_21[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_19 (BatchN  (None, 45, 45, 32)  96          ['conv2d_19[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_22 (BatchN  (None, 45, 45, 48)  144         ['conv2d_22[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_19 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_19[0][0]'] \n",
            "                                                                                                  \n",
            " activation_22 (Activation)     (None, 45, 45, 48)   0           ['batch_normalization_22[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_18 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_1_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_20 (Conv2D)             (None, 45, 45, 32)   9216        ['activation_19[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_23 (Conv2D)             (None, 45, 45, 64)   27648       ['activation_22[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_18 (BatchN  (None, 45, 45, 32)  96          ['conv2d_18[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_20 (BatchN  (None, 45, 45, 32)  96          ['conv2d_20[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_23 (BatchN  (None, 45, 45, 64)  192         ['conv2d_23[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_18 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_18[0][0]'] \n",
            "                                                                                                  \n",
            " activation_20 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_20[0][0]'] \n",
            "                                                                                                  \n",
            " activation_23 (Activation)     (None, 45, 45, 64)   0           ['batch_normalization_23[0][0]'] \n",
            "                                                                                                  \n",
            " block35_2_mixed (Concatenate)  (None, 45, 45, 128)  0           ['activation_18[0][0]',          \n",
            "                                                                  'activation_20[0][0]',          \n",
            "                                                                  'activation_23[0][0]']          \n",
            "                                                                                                  \n",
            " block35_2_conv (Conv2D)        (None, 45, 45, 320)  41280       ['block35_2_mixed[0][0]']        \n",
            "                                                                                                  \n",
            " block35_2 (Lambda)             (None, 45, 45, 320)  0           ['block35_1_ac[0][0]',           \n",
            "                                                                  'block35_2_conv[0][0]']         \n",
            "                                                                                                  \n",
            " block35_2_ac (Activation)      (None, 45, 45, 320)  0           ['block35_2[0][0]']              \n",
            "                                                                                                  \n",
            " conv2d_27 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_2_ac[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_27 (BatchN  (None, 45, 45, 32)  96          ['conv2d_27[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_27 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_27[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_25 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_2_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_28 (Conv2D)             (None, 45, 45, 48)   13824       ['activation_27[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_25 (BatchN  (None, 45, 45, 32)  96          ['conv2d_25[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_28 (BatchN  (None, 45, 45, 48)  144         ['conv2d_28[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_25 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_25[0][0]'] \n",
            "                                                                                                  \n",
            " activation_28 (Activation)     (None, 45, 45, 48)   0           ['batch_normalization_28[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_24 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_2_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_26 (Conv2D)             (None, 45, 45, 32)   9216        ['activation_25[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_29 (Conv2D)             (None, 45, 45, 64)   27648       ['activation_28[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_24 (BatchN  (None, 45, 45, 32)  96          ['conv2d_24[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_26 (BatchN  (None, 45, 45, 32)  96          ['conv2d_26[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_29 (BatchN  (None, 45, 45, 64)  192         ['conv2d_29[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_24 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_24[0][0]'] \n",
            "                                                                                                  \n",
            " activation_26 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_26[0][0]'] \n",
            "                                                                                                  \n",
            " activation_29 (Activation)     (None, 45, 45, 64)   0           ['batch_normalization_29[0][0]'] \n",
            "                                                                                                  \n",
            " block35_3_mixed (Concatenate)  (None, 45, 45, 128)  0           ['activation_24[0][0]',          \n",
            "                                                                  'activation_26[0][0]',          \n",
            "                                                                  'activation_29[0][0]']          \n",
            "                                                                                                  \n",
            " block35_3_conv (Conv2D)        (None, 45, 45, 320)  41280       ['block35_3_mixed[0][0]']        \n",
            "                                                                                                  \n",
            " block35_3 (Lambda)             (None, 45, 45, 320)  0           ['block35_2_ac[0][0]',           \n",
            "                                                                  'block35_3_conv[0][0]']         \n",
            "                                                                                                  \n",
            " block35_3_ac (Activation)      (None, 45, 45, 320)  0           ['block35_3[0][0]']              \n",
            "                                                                                                  \n",
            " conv2d_33 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_3_ac[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_33 (BatchN  (None, 45, 45, 32)  96          ['conv2d_33[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_33 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_33[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_31 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_3_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_34 (Conv2D)             (None, 45, 45, 48)   13824       ['activation_33[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_31 (BatchN  (None, 45, 45, 32)  96          ['conv2d_31[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_34 (BatchN  (None, 45, 45, 48)  144         ['conv2d_34[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_31 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_31[0][0]'] \n",
            "                                                                                                  \n",
            " activation_34 (Activation)     (None, 45, 45, 48)   0           ['batch_normalization_34[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_30 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_3_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_32 (Conv2D)             (None, 45, 45, 32)   9216        ['activation_31[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_35 (Conv2D)             (None, 45, 45, 64)   27648       ['activation_34[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_30 (BatchN  (None, 45, 45, 32)  96          ['conv2d_30[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_32 (BatchN  (None, 45, 45, 32)  96          ['conv2d_32[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_35 (BatchN  (None, 45, 45, 64)  192         ['conv2d_35[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_30 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_30[0][0]'] \n",
            "                                                                                                  \n",
            " activation_32 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_32[0][0]'] \n",
            "                                                                                                  \n",
            " activation_35 (Activation)     (None, 45, 45, 64)   0           ['batch_normalization_35[0][0]'] \n",
            "                                                                                                  \n",
            " block35_4_mixed (Concatenate)  (None, 45, 45, 128)  0           ['activation_30[0][0]',          \n",
            "                                                                  'activation_32[0][0]',          \n",
            "                                                                  'activation_35[0][0]']          \n",
            "                                                                                                  \n",
            " block35_4_conv (Conv2D)        (None, 45, 45, 320)  41280       ['block35_4_mixed[0][0]']        \n",
            "                                                                                                  \n",
            " block35_4 (Lambda)             (None, 45, 45, 320)  0           ['block35_3_ac[0][0]',           \n",
            "                                                                  'block35_4_conv[0][0]']         \n",
            "                                                                                                  \n",
            " block35_4_ac (Activation)      (None, 45, 45, 320)  0           ['block35_4[0][0]']              \n",
            "                                                                                                  \n",
            " conv2d_39 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_4_ac[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_39 (BatchN  (None, 45, 45, 32)  96          ['conv2d_39[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_39 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_39[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_37 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_4_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_40 (Conv2D)             (None, 45, 45, 48)   13824       ['activation_39[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_37 (BatchN  (None, 45, 45, 32)  96          ['conv2d_37[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_40 (BatchN  (None, 45, 45, 48)  144         ['conv2d_40[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_37 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_37[0][0]'] \n",
            "                                                                                                  \n",
            " activation_40 (Activation)     (None, 45, 45, 48)   0           ['batch_normalization_40[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_36 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_4_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_38 (Conv2D)             (None, 45, 45, 32)   9216        ['activation_37[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_41 (Conv2D)             (None, 45, 45, 64)   27648       ['activation_40[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_36 (BatchN  (None, 45, 45, 32)  96          ['conv2d_36[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_38 (BatchN  (None, 45, 45, 32)  96          ['conv2d_38[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_41 (BatchN  (None, 45, 45, 64)  192         ['conv2d_41[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_36 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_36[0][0]'] \n",
            "                                                                                                  \n",
            " activation_38 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_38[0][0]'] \n",
            "                                                                                                  \n",
            " activation_41 (Activation)     (None, 45, 45, 64)   0           ['batch_normalization_41[0][0]'] \n",
            "                                                                                                  \n",
            " block35_5_mixed (Concatenate)  (None, 45, 45, 128)  0           ['activation_36[0][0]',          \n",
            "                                                                  'activation_38[0][0]',          \n",
            "                                                                  'activation_41[0][0]']          \n",
            "                                                                                                  \n",
            " block35_5_conv (Conv2D)        (None, 45, 45, 320)  41280       ['block35_5_mixed[0][0]']        \n",
            "                                                                                                  \n",
            " block35_5 (Lambda)             (None, 45, 45, 320)  0           ['block35_4_ac[0][0]',           \n",
            "                                                                  'block35_5_conv[0][0]']         \n",
            "                                                                                                  \n",
            " block35_5_ac (Activation)      (None, 45, 45, 320)  0           ['block35_5[0][0]']              \n",
            "                                                                                                  \n",
            " conv2d_45 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_5_ac[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_45 (BatchN  (None, 45, 45, 32)  96          ['conv2d_45[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_45 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_45[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_43 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_5_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_46 (Conv2D)             (None, 45, 45, 48)   13824       ['activation_45[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_43 (BatchN  (None, 45, 45, 32)  96          ['conv2d_43[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_46 (BatchN  (None, 45, 45, 48)  144         ['conv2d_46[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_43 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_43[0][0]'] \n",
            "                                                                                                  \n",
            " activation_46 (Activation)     (None, 45, 45, 48)   0           ['batch_normalization_46[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_42 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_5_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_44 (Conv2D)             (None, 45, 45, 32)   9216        ['activation_43[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_47 (Conv2D)             (None, 45, 45, 64)   27648       ['activation_46[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_42 (BatchN  (None, 45, 45, 32)  96          ['conv2d_42[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_44 (BatchN  (None, 45, 45, 32)  96          ['conv2d_44[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_47 (BatchN  (None, 45, 45, 64)  192         ['conv2d_47[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_42 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_42[0][0]'] \n",
            "                                                                                                  \n",
            " activation_44 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_44[0][0]'] \n",
            "                                                                                                  \n",
            " activation_47 (Activation)     (None, 45, 45, 64)   0           ['batch_normalization_47[0][0]'] \n",
            "                                                                                                  \n",
            " block35_6_mixed (Concatenate)  (None, 45, 45, 128)  0           ['activation_42[0][0]',          \n",
            "                                                                  'activation_44[0][0]',          \n",
            "                                                                  'activation_47[0][0]']          \n",
            "                                                                                                  \n",
            " block35_6_conv (Conv2D)        (None, 45, 45, 320)  41280       ['block35_6_mixed[0][0]']        \n",
            "                                                                                                  \n",
            " block35_6 (Lambda)             (None, 45, 45, 320)  0           ['block35_5_ac[0][0]',           \n",
            "                                                                  'block35_6_conv[0][0]']         \n",
            "                                                                                                  \n",
            " block35_6_ac (Activation)      (None, 45, 45, 320)  0           ['block35_6[0][0]']              \n",
            "                                                                                                  \n",
            " conv2d_51 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_6_ac[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_51 (BatchN  (None, 45, 45, 32)  96          ['conv2d_51[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_51 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_51[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_49 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_6_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_52 (Conv2D)             (None, 45, 45, 48)   13824       ['activation_51[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_49 (BatchN  (None, 45, 45, 32)  96          ['conv2d_49[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_52 (BatchN  (None, 45, 45, 48)  144         ['conv2d_52[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_49 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_49[0][0]'] \n",
            "                                                                                                  \n",
            " activation_52 (Activation)     (None, 45, 45, 48)   0           ['batch_normalization_52[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_48 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_6_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_50 (Conv2D)             (None, 45, 45, 32)   9216        ['activation_49[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_53 (Conv2D)             (None, 45, 45, 64)   27648       ['activation_52[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_48 (BatchN  (None, 45, 45, 32)  96          ['conv2d_48[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_50 (BatchN  (None, 45, 45, 32)  96          ['conv2d_50[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_53 (BatchN  (None, 45, 45, 64)  192         ['conv2d_53[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_48 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_48[0][0]'] \n",
            "                                                                                                  \n",
            " activation_50 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_50[0][0]'] \n",
            "                                                                                                  \n",
            " activation_53 (Activation)     (None, 45, 45, 64)   0           ['batch_normalization_53[0][0]'] \n",
            "                                                                                                  \n",
            " block35_7_mixed (Concatenate)  (None, 45, 45, 128)  0           ['activation_48[0][0]',          \n",
            "                                                                  'activation_50[0][0]',          \n",
            "                                                                  'activation_53[0][0]']          \n",
            "                                                                                                  \n",
            " block35_7_conv (Conv2D)        (None, 45, 45, 320)  41280       ['block35_7_mixed[0][0]']        \n",
            "                                                                                                  \n",
            " block35_7 (Lambda)             (None, 45, 45, 320)  0           ['block35_6_ac[0][0]',           \n",
            "                                                                  'block35_7_conv[0][0]']         \n",
            "                                                                                                  \n",
            " block35_7_ac (Activation)      (None, 45, 45, 320)  0           ['block35_7[0][0]']              \n",
            "                                                                                                  \n",
            " conv2d_57 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_7_ac[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_57 (BatchN  (None, 45, 45, 32)  96          ['conv2d_57[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_57 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_57[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_55 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_7_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_58 (Conv2D)             (None, 45, 45, 48)   13824       ['activation_57[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_55 (BatchN  (None, 45, 45, 32)  96          ['conv2d_55[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_58 (BatchN  (None, 45, 45, 48)  144         ['conv2d_58[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_55 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_55[0][0]'] \n",
            "                                                                                                  \n",
            " activation_58 (Activation)     (None, 45, 45, 48)   0           ['batch_normalization_58[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_54 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_7_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_56 (Conv2D)             (None, 45, 45, 32)   9216        ['activation_55[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_59 (Conv2D)             (None, 45, 45, 64)   27648       ['activation_58[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_54 (BatchN  (None, 45, 45, 32)  96          ['conv2d_54[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_56 (BatchN  (None, 45, 45, 32)  96          ['conv2d_56[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_59 (BatchN  (None, 45, 45, 64)  192         ['conv2d_59[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_54 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_54[0][0]'] \n",
            "                                                                                                  \n",
            " activation_56 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_56[0][0]'] \n",
            "                                                                                                  \n",
            " activation_59 (Activation)     (None, 45, 45, 64)   0           ['batch_normalization_59[0][0]'] \n",
            "                                                                                                  \n",
            " block35_8_mixed (Concatenate)  (None, 45, 45, 128)  0           ['activation_54[0][0]',          \n",
            "                                                                  'activation_56[0][0]',          \n",
            "                                                                  'activation_59[0][0]']          \n",
            "                                                                                                  \n",
            " block35_8_conv (Conv2D)        (None, 45, 45, 320)  41280       ['block35_8_mixed[0][0]']        \n",
            "                                                                                                  \n",
            " block35_8 (Lambda)             (None, 45, 45, 320)  0           ['block35_7_ac[0][0]',           \n",
            "                                                                  'block35_8_conv[0][0]']         \n",
            "                                                                                                  \n",
            " block35_8_ac (Activation)      (None, 45, 45, 320)  0           ['block35_8[0][0]']              \n",
            "                                                                                                  \n",
            " conv2d_63 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_8_ac[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_63 (BatchN  (None, 45, 45, 32)  96          ['conv2d_63[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_63 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_63[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_61 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_8_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_64 (Conv2D)             (None, 45, 45, 48)   13824       ['activation_63[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_61 (BatchN  (None, 45, 45, 32)  96          ['conv2d_61[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_64 (BatchN  (None, 45, 45, 48)  144         ['conv2d_64[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_61 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_61[0][0]'] \n",
            "                                                                                                  \n",
            " activation_64 (Activation)     (None, 45, 45, 48)   0           ['batch_normalization_64[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_60 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_8_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_62 (Conv2D)             (None, 45, 45, 32)   9216        ['activation_61[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_65 (Conv2D)             (None, 45, 45, 64)   27648       ['activation_64[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_60 (BatchN  (None, 45, 45, 32)  96          ['conv2d_60[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_62 (BatchN  (None, 45, 45, 32)  96          ['conv2d_62[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_65 (BatchN  (None, 45, 45, 64)  192         ['conv2d_65[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_60 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_60[0][0]'] \n",
            "                                                                                                  \n",
            " activation_62 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_62[0][0]'] \n",
            "                                                                                                  \n",
            " activation_65 (Activation)     (None, 45, 45, 64)   0           ['batch_normalization_65[0][0]'] \n",
            "                                                                                                  \n",
            " block35_9_mixed (Concatenate)  (None, 45, 45, 128)  0           ['activation_60[0][0]',          \n",
            "                                                                  'activation_62[0][0]',          \n",
            "                                                                  'activation_65[0][0]']          \n",
            "                                                                                                  \n",
            " block35_9_conv (Conv2D)        (None, 45, 45, 320)  41280       ['block35_9_mixed[0][0]']        \n",
            "                                                                                                  \n",
            " block35_9 (Lambda)             (None, 45, 45, 320)  0           ['block35_8_ac[0][0]',           \n",
            "                                                                  'block35_9_conv[0][0]']         \n",
            "                                                                                                  \n",
            " block35_9_ac (Activation)      (None, 45, 45, 320)  0           ['block35_9[0][0]']              \n",
            "                                                                                                  \n",
            " conv2d_69 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_9_ac[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_69 (BatchN  (None, 45, 45, 32)  96          ['conv2d_69[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_69 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_69[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_67 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_9_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_70 (Conv2D)             (None, 45, 45, 48)   13824       ['activation_69[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_67 (BatchN  (None, 45, 45, 32)  96          ['conv2d_67[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_70 (BatchN  (None, 45, 45, 48)  144         ['conv2d_70[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_67 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_67[0][0]'] \n",
            "                                                                                                  \n",
            " activation_70 (Activation)     (None, 45, 45, 48)   0           ['batch_normalization_70[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_66 (Conv2D)             (None, 45, 45, 32)   10240       ['block35_9_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_68 (Conv2D)             (None, 45, 45, 32)   9216        ['activation_67[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_71 (Conv2D)             (None, 45, 45, 64)   27648       ['activation_70[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_66 (BatchN  (None, 45, 45, 32)  96          ['conv2d_66[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_68 (BatchN  (None, 45, 45, 32)  96          ['conv2d_68[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_71 (BatchN  (None, 45, 45, 64)  192         ['conv2d_71[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_66 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_66[0][0]'] \n",
            "                                                                                                  \n",
            " activation_68 (Activation)     (None, 45, 45, 32)   0           ['batch_normalization_68[0][0]'] \n",
            "                                                                                                  \n",
            " activation_71 (Activation)     (None, 45, 45, 64)   0           ['batch_normalization_71[0][0]'] \n",
            "                                                                                                  \n",
            " block35_10_mixed (Concatenate)  (None, 45, 45, 128)  0          ['activation_66[0][0]',          \n",
            "                                                                  'activation_68[0][0]',          \n",
            "                                                                  'activation_71[0][0]']          \n",
            "                                                                                                  \n",
            " block35_10_conv (Conv2D)       (None, 45, 45, 320)  41280       ['block35_10_mixed[0][0]']       \n",
            "                                                                                                  \n",
            " block35_10 (Lambda)            (None, 45, 45, 320)  0           ['block35_9_ac[0][0]',           \n",
            "                                                                  'block35_10_conv[0][0]']        \n",
            "                                                                                                  \n",
            " block35_10_ac (Activation)     (None, 45, 45, 320)  0           ['block35_10[0][0]']             \n",
            "                                                                                                  \n",
            " conv2d_73 (Conv2D)             (None, 45, 45, 256)  81920       ['block35_10_ac[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_73 (BatchN  (None, 45, 45, 256)  768        ['conv2d_73[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_73 (Activation)     (None, 45, 45, 256)  0           ['batch_normalization_73[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_74 (Conv2D)             (None, 45, 45, 256)  589824      ['activation_73[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_74 (BatchN  (None, 45, 45, 256)  768        ['conv2d_74[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_74 (Activation)     (None, 45, 45, 256)  0           ['batch_normalization_74[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_72 (Conv2D)             (None, 22, 22, 384)  1105920     ['block35_10_ac[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_75 (Conv2D)             (None, 22, 22, 384)  884736      ['activation_74[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_72 (BatchN  (None, 22, 22, 384)  1152       ['conv2d_72[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_75 (BatchN  (None, 22, 22, 384)  1152       ['conv2d_75[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_72 (Activation)     (None, 22, 22, 384)  0           ['batch_normalization_72[0][0]'] \n",
            "                                                                                                  \n",
            " activation_75 (Activation)     (None, 22, 22, 384)  0           ['batch_normalization_75[0][0]'] \n",
            "                                                                                                  \n",
            " max_pooling2d_2 (MaxPooling2D)  (None, 22, 22, 320)  0          ['block35_10_ac[0][0]']          \n",
            "                                                                                                  \n",
            " mixed_6a (Concatenate)         (None, 22, 22, 1088  0           ['activation_72[0][0]',          \n",
            "                                )                                 'activation_75[0][0]',          \n",
            "                                                                  'max_pooling2d_2[0][0]']        \n",
            "                                                                                                  \n",
            " conv2d_77 (Conv2D)             (None, 22, 22, 128)  139264      ['mixed_6a[0][0]']               \n",
            "                                                                                                  \n",
            " batch_normalization_77 (BatchN  (None, 22, 22, 128)  384        ['conv2d_77[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_77 (Activation)     (None, 22, 22, 128)  0           ['batch_normalization_77[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_78 (Conv2D)             (None, 22, 22, 160)  143360      ['activation_77[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_78 (BatchN  (None, 22, 22, 160)  480        ['conv2d_78[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_78 (Activation)     (None, 22, 22, 160)  0           ['batch_normalization_78[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_76 (Conv2D)             (None, 22, 22, 192)  208896      ['mixed_6a[0][0]']               \n",
            "                                                                                                  \n",
            " conv2d_79 (Conv2D)             (None, 22, 22, 192)  215040      ['activation_78[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_76 (BatchN  (None, 22, 22, 192)  576        ['conv2d_76[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_79 (BatchN  (None, 22, 22, 192)  576        ['conv2d_79[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_76 (Activation)     (None, 22, 22, 192)  0           ['batch_normalization_76[0][0]'] \n",
            "                                                                                                  \n",
            " activation_79 (Activation)     (None, 22, 22, 192)  0           ['batch_normalization_79[0][0]'] \n",
            "                                                                                                  \n",
            " block17_1_mixed (Concatenate)  (None, 22, 22, 384)  0           ['activation_76[0][0]',          \n",
            "                                                                  'activation_79[0][0]']          \n",
            "                                                                                                  \n",
            " block17_1_conv (Conv2D)        (None, 22, 22, 1088  418880      ['block17_1_mixed[0][0]']        \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block17_1 (Lambda)             (None, 22, 22, 1088  0           ['mixed_6a[0][0]',               \n",
            "                                )                                 'block17_1_conv[0][0]']         \n",
            "                                                                                                  \n",
            " block17_1_ac (Activation)      (None, 22, 22, 1088  0           ['block17_1[0][0]']              \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_81 (Conv2D)             (None, 22, 22, 128)  139264      ['block17_1_ac[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_81 (BatchN  (None, 22, 22, 128)  384        ['conv2d_81[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_81 (Activation)     (None, 22, 22, 128)  0           ['batch_normalization_81[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_82 (Conv2D)             (None, 22, 22, 160)  143360      ['activation_81[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_82 (BatchN  (None, 22, 22, 160)  480        ['conv2d_82[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_82 (Activation)     (None, 22, 22, 160)  0           ['batch_normalization_82[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_80 (Conv2D)             (None, 22, 22, 192)  208896      ['block17_1_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_83 (Conv2D)             (None, 22, 22, 192)  215040      ['activation_82[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_80 (BatchN  (None, 22, 22, 192)  576        ['conv2d_80[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_83 (BatchN  (None, 22, 22, 192)  576        ['conv2d_83[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_80 (Activation)     (None, 22, 22, 192)  0           ['batch_normalization_80[0][0]'] \n",
            "                                                                                                  \n",
            " activation_83 (Activation)     (None, 22, 22, 192)  0           ['batch_normalization_83[0][0]'] \n",
            "                                                                                                  \n",
            " block17_2_mixed (Concatenate)  (None, 22, 22, 384)  0           ['activation_80[0][0]',          \n",
            "                                                                  'activation_83[0][0]']          \n",
            "                                                                                                  \n",
            " block17_2_conv (Conv2D)        (None, 22, 22, 1088  418880      ['block17_2_mixed[0][0]']        \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block17_2 (Lambda)             (None, 22, 22, 1088  0           ['block17_1_ac[0][0]',           \n",
            "                                )                                 'block17_2_conv[0][0]']         \n",
            "                                                                                                  \n",
            " block17_2_ac (Activation)      (None, 22, 22, 1088  0           ['block17_2[0][0]']              \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_85 (Conv2D)             (None, 22, 22, 128)  139264      ['block17_2_ac[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_85 (BatchN  (None, 22, 22, 128)  384        ['conv2d_85[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_85 (Activation)     (None, 22, 22, 128)  0           ['batch_normalization_85[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_86 (Conv2D)             (None, 22, 22, 160)  143360      ['activation_85[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_86 (BatchN  (None, 22, 22, 160)  480        ['conv2d_86[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_86 (Activation)     (None, 22, 22, 160)  0           ['batch_normalization_86[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_84 (Conv2D)             (None, 22, 22, 192)  208896      ['block17_2_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_87 (Conv2D)             (None, 22, 22, 192)  215040      ['activation_86[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_84 (BatchN  (None, 22, 22, 192)  576        ['conv2d_84[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_87 (BatchN  (None, 22, 22, 192)  576        ['conv2d_87[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_84 (Activation)     (None, 22, 22, 192)  0           ['batch_normalization_84[0][0]'] \n",
            "                                                                                                  \n",
            " activation_87 (Activation)     (None, 22, 22, 192)  0           ['batch_normalization_87[0][0]'] \n",
            "                                                                                                  \n",
            " block17_3_mixed (Concatenate)  (None, 22, 22, 384)  0           ['activation_84[0][0]',          \n",
            "                                                                  'activation_87[0][0]']          \n",
            "                                                                                                  \n",
            " block17_3_conv (Conv2D)        (None, 22, 22, 1088  418880      ['block17_3_mixed[0][0]']        \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block17_3 (Lambda)             (None, 22, 22, 1088  0           ['block17_2_ac[0][0]',           \n",
            "                                )                                 'block17_3_conv[0][0]']         \n",
            "                                                                                                  \n",
            " block17_3_ac (Activation)      (None, 22, 22, 1088  0           ['block17_3[0][0]']              \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_89 (Conv2D)             (None, 22, 22, 128)  139264      ['block17_3_ac[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_89 (BatchN  (None, 22, 22, 128)  384        ['conv2d_89[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_89 (Activation)     (None, 22, 22, 128)  0           ['batch_normalization_89[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_90 (Conv2D)             (None, 22, 22, 160)  143360      ['activation_89[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_90 (BatchN  (None, 22, 22, 160)  480        ['conv2d_90[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_90 (Activation)     (None, 22, 22, 160)  0           ['batch_normalization_90[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_88 (Conv2D)             (None, 22, 22, 192)  208896      ['block17_3_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_91 (Conv2D)             (None, 22, 22, 192)  215040      ['activation_90[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_88 (BatchN  (None, 22, 22, 192)  576        ['conv2d_88[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_91 (BatchN  (None, 22, 22, 192)  576        ['conv2d_91[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_88 (Activation)     (None, 22, 22, 192)  0           ['batch_normalization_88[0][0]'] \n",
            "                                                                                                  \n",
            " activation_91 (Activation)     (None, 22, 22, 192)  0           ['batch_normalization_91[0][0]'] \n",
            "                                                                                                  \n",
            " block17_4_mixed (Concatenate)  (None, 22, 22, 384)  0           ['activation_88[0][0]',          \n",
            "                                                                  'activation_91[0][0]']          \n",
            "                                                                                                  \n",
            " block17_4_conv (Conv2D)        (None, 22, 22, 1088  418880      ['block17_4_mixed[0][0]']        \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block17_4 (Lambda)             (None, 22, 22, 1088  0           ['block17_3_ac[0][0]',           \n",
            "                                )                                 'block17_4_conv[0][0]']         \n",
            "                                                                                                  \n",
            " block17_4_ac (Activation)      (None, 22, 22, 1088  0           ['block17_4[0][0]']              \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_93 (Conv2D)             (None, 22, 22, 128)  139264      ['block17_4_ac[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_93 (BatchN  (None, 22, 22, 128)  384        ['conv2d_93[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_93 (Activation)     (None, 22, 22, 128)  0           ['batch_normalization_93[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_94 (Conv2D)             (None, 22, 22, 160)  143360      ['activation_93[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_94 (BatchN  (None, 22, 22, 160)  480        ['conv2d_94[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_94 (Activation)     (None, 22, 22, 160)  0           ['batch_normalization_94[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_92 (Conv2D)             (None, 22, 22, 192)  208896      ['block17_4_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_95 (Conv2D)             (None, 22, 22, 192)  215040      ['activation_94[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_92 (BatchN  (None, 22, 22, 192)  576        ['conv2d_92[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_95 (BatchN  (None, 22, 22, 192)  576        ['conv2d_95[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_92 (Activation)     (None, 22, 22, 192)  0           ['batch_normalization_92[0][0]'] \n",
            "                                                                                                  \n",
            " activation_95 (Activation)     (None, 22, 22, 192)  0           ['batch_normalization_95[0][0]'] \n",
            "                                                                                                  \n",
            " block17_5_mixed (Concatenate)  (None, 22, 22, 384)  0           ['activation_92[0][0]',          \n",
            "                                                                  'activation_95[0][0]']          \n",
            "                                                                                                  \n",
            " block17_5_conv (Conv2D)        (None, 22, 22, 1088  418880      ['block17_5_mixed[0][0]']        \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block17_5 (Lambda)             (None, 22, 22, 1088  0           ['block17_4_ac[0][0]',           \n",
            "                                )                                 'block17_5_conv[0][0]']         \n",
            "                                                                                                  \n",
            " block17_5_ac (Activation)      (None, 22, 22, 1088  0           ['block17_5[0][0]']              \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_97 (Conv2D)             (None, 22, 22, 128)  139264      ['block17_5_ac[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_97 (BatchN  (None, 22, 22, 128)  384        ['conv2d_97[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_97 (Activation)     (None, 22, 22, 128)  0           ['batch_normalization_97[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_98 (Conv2D)             (None, 22, 22, 160)  143360      ['activation_97[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_98 (BatchN  (None, 22, 22, 160)  480        ['conv2d_98[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_98 (Activation)     (None, 22, 22, 160)  0           ['batch_normalization_98[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_96 (Conv2D)             (None, 22, 22, 192)  208896      ['block17_5_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_99 (Conv2D)             (None, 22, 22, 192)  215040      ['activation_98[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_96 (BatchN  (None, 22, 22, 192)  576        ['conv2d_96[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " batch_normalization_99 (BatchN  (None, 22, 22, 192)  576        ['conv2d_99[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_96 (Activation)     (None, 22, 22, 192)  0           ['batch_normalization_96[0][0]'] \n",
            "                                                                                                  \n",
            " activation_99 (Activation)     (None, 22, 22, 192)  0           ['batch_normalization_99[0][0]'] \n",
            "                                                                                                  \n",
            " block17_6_mixed (Concatenate)  (None, 22, 22, 384)  0           ['activation_96[0][0]',          \n",
            "                                                                  'activation_99[0][0]']          \n",
            "                                                                                                  \n",
            " block17_6_conv (Conv2D)        (None, 22, 22, 1088  418880      ['block17_6_mixed[0][0]']        \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block17_6 (Lambda)             (None, 22, 22, 1088  0           ['block17_5_ac[0][0]',           \n",
            "                                )                                 'block17_6_conv[0][0]']         \n",
            "                                                                                                  \n",
            " block17_6_ac (Activation)      (None, 22, 22, 1088  0           ['block17_6[0][0]']              \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_101 (Conv2D)            (None, 22, 22, 128)  139264      ['block17_6_ac[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_101 (Batch  (None, 22, 22, 128)  384        ['conv2d_101[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_101 (Activation)    (None, 22, 22, 128)  0           ['batch_normalization_101[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_102 (Conv2D)            (None, 22, 22, 160)  143360      ['activation_101[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_102 (Batch  (None, 22, 22, 160)  480        ['conv2d_102[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_102 (Activation)    (None, 22, 22, 160)  0           ['batch_normalization_102[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_100 (Conv2D)            (None, 22, 22, 192)  208896      ['block17_6_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_103 (Conv2D)            (None, 22, 22, 192)  215040      ['activation_102[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_100 (Batch  (None, 22, 22, 192)  576        ['conv2d_100[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_103 (Batch  (None, 22, 22, 192)  576        ['conv2d_103[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_100 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_100[0][0]']\n",
            "                                                                                                  \n",
            " activation_103 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_103[0][0]']\n",
            "                                                                                                  \n",
            " block17_7_mixed (Concatenate)  (None, 22, 22, 384)  0           ['activation_100[0][0]',         \n",
            "                                                                  'activation_103[0][0]']         \n",
            "                                                                                                  \n",
            " block17_7_conv (Conv2D)        (None, 22, 22, 1088  418880      ['block17_7_mixed[0][0]']        \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block17_7 (Lambda)             (None, 22, 22, 1088  0           ['block17_6_ac[0][0]',           \n",
            "                                )                                 'block17_7_conv[0][0]']         \n",
            "                                                                                                  \n",
            " block17_7_ac (Activation)      (None, 22, 22, 1088  0           ['block17_7[0][0]']              \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_105 (Conv2D)            (None, 22, 22, 128)  139264      ['block17_7_ac[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_105 (Batch  (None, 22, 22, 128)  384        ['conv2d_105[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_105 (Activation)    (None, 22, 22, 128)  0           ['batch_normalization_105[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_106 (Conv2D)            (None, 22, 22, 160)  143360      ['activation_105[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_106 (Batch  (None, 22, 22, 160)  480        ['conv2d_106[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_106 (Activation)    (None, 22, 22, 160)  0           ['batch_normalization_106[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_104 (Conv2D)            (None, 22, 22, 192)  208896      ['block17_7_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_107 (Conv2D)            (None, 22, 22, 192)  215040      ['activation_106[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_104 (Batch  (None, 22, 22, 192)  576        ['conv2d_104[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_107 (Batch  (None, 22, 22, 192)  576        ['conv2d_107[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_104 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_104[0][0]']\n",
            "                                                                                                  \n",
            " activation_107 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_107[0][0]']\n",
            "                                                                                                  \n",
            " block17_8_mixed (Concatenate)  (None, 22, 22, 384)  0           ['activation_104[0][0]',         \n",
            "                                                                  'activation_107[0][0]']         \n",
            "                                                                                                  \n",
            " block17_8_conv (Conv2D)        (None, 22, 22, 1088  418880      ['block17_8_mixed[0][0]']        \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block17_8 (Lambda)             (None, 22, 22, 1088  0           ['block17_7_ac[0][0]',           \n",
            "                                )                                 'block17_8_conv[0][0]']         \n",
            "                                                                                                  \n",
            " block17_8_ac (Activation)      (None, 22, 22, 1088  0           ['block17_8[0][0]']              \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_109 (Conv2D)            (None, 22, 22, 128)  139264      ['block17_8_ac[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_109 (Batch  (None, 22, 22, 128)  384        ['conv2d_109[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_109 (Activation)    (None, 22, 22, 128)  0           ['batch_normalization_109[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_110 (Conv2D)            (None, 22, 22, 160)  143360      ['activation_109[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_110 (Batch  (None, 22, 22, 160)  480        ['conv2d_110[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_110 (Activation)    (None, 22, 22, 160)  0           ['batch_normalization_110[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_108 (Conv2D)            (None, 22, 22, 192)  208896      ['block17_8_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_111 (Conv2D)            (None, 22, 22, 192)  215040      ['activation_110[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_108 (Batch  (None, 22, 22, 192)  576        ['conv2d_108[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_111 (Batch  (None, 22, 22, 192)  576        ['conv2d_111[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_108 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_108[0][0]']\n",
            "                                                                                                  \n",
            " activation_111 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_111[0][0]']\n",
            "                                                                                                  \n",
            " block17_9_mixed (Concatenate)  (None, 22, 22, 384)  0           ['activation_108[0][0]',         \n",
            "                                                                  'activation_111[0][0]']         \n",
            "                                                                                                  \n",
            " block17_9_conv (Conv2D)        (None, 22, 22, 1088  418880      ['block17_9_mixed[0][0]']        \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block17_9 (Lambda)             (None, 22, 22, 1088  0           ['block17_8_ac[0][0]',           \n",
            "                                )                                 'block17_9_conv[0][0]']         \n",
            "                                                                                                  \n",
            " block17_9_ac (Activation)      (None, 22, 22, 1088  0           ['block17_9[0][0]']              \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_113 (Conv2D)            (None, 22, 22, 128)  139264      ['block17_9_ac[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_113 (Batch  (None, 22, 22, 128)  384        ['conv2d_113[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_113 (Activation)    (None, 22, 22, 128)  0           ['batch_normalization_113[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_114 (Conv2D)            (None, 22, 22, 160)  143360      ['activation_113[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_114 (Batch  (None, 22, 22, 160)  480        ['conv2d_114[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_114 (Activation)    (None, 22, 22, 160)  0           ['batch_normalization_114[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_112 (Conv2D)            (None, 22, 22, 192)  208896      ['block17_9_ac[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_115 (Conv2D)            (None, 22, 22, 192)  215040      ['activation_114[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_112 (Batch  (None, 22, 22, 192)  576        ['conv2d_112[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_115 (Batch  (None, 22, 22, 192)  576        ['conv2d_115[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_112 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_112[0][0]']\n",
            "                                                                                                  \n",
            " activation_115 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_115[0][0]']\n",
            "                                                                                                  \n",
            " block17_10_mixed (Concatenate)  (None, 22, 22, 384)  0          ['activation_112[0][0]',         \n",
            "                                                                  'activation_115[0][0]']         \n",
            "                                                                                                  \n",
            " block17_10_conv (Conv2D)       (None, 22, 22, 1088  418880      ['block17_10_mixed[0][0]']       \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block17_10 (Lambda)            (None, 22, 22, 1088  0           ['block17_9_ac[0][0]',           \n",
            "                                )                                 'block17_10_conv[0][0]']        \n",
            "                                                                                                  \n",
            " block17_10_ac (Activation)     (None, 22, 22, 1088  0           ['block17_10[0][0]']             \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_117 (Conv2D)            (None, 22, 22, 128)  139264      ['block17_10_ac[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_117 (Batch  (None, 22, 22, 128)  384        ['conv2d_117[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_117 (Activation)    (None, 22, 22, 128)  0           ['batch_normalization_117[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_118 (Conv2D)            (None, 22, 22, 160)  143360      ['activation_117[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_118 (Batch  (None, 22, 22, 160)  480        ['conv2d_118[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_118 (Activation)    (None, 22, 22, 160)  0           ['batch_normalization_118[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_116 (Conv2D)            (None, 22, 22, 192)  208896      ['block17_10_ac[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_119 (Conv2D)            (None, 22, 22, 192)  215040      ['activation_118[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_116 (Batch  (None, 22, 22, 192)  576        ['conv2d_116[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_119 (Batch  (None, 22, 22, 192)  576        ['conv2d_119[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_116 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_116[0][0]']\n",
            "                                                                                                  \n",
            " activation_119 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_119[0][0]']\n",
            "                                                                                                  \n",
            " block17_11_mixed (Concatenate)  (None, 22, 22, 384)  0          ['activation_116[0][0]',         \n",
            "                                                                  'activation_119[0][0]']         \n",
            "                                                                                                  \n",
            " block17_11_conv (Conv2D)       (None, 22, 22, 1088  418880      ['block17_11_mixed[0][0]']       \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block17_11 (Lambda)            (None, 22, 22, 1088  0           ['block17_10_ac[0][0]',          \n",
            "                                )                                 'block17_11_conv[0][0]']        \n",
            "                                                                                                  \n",
            " block17_11_ac (Activation)     (None, 22, 22, 1088  0           ['block17_11[0][0]']             \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_121 (Conv2D)            (None, 22, 22, 128)  139264      ['block17_11_ac[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_121 (Batch  (None, 22, 22, 128)  384        ['conv2d_121[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_121 (Activation)    (None, 22, 22, 128)  0           ['batch_normalization_121[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_122 (Conv2D)            (None, 22, 22, 160)  143360      ['activation_121[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_122 (Batch  (None, 22, 22, 160)  480        ['conv2d_122[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_122 (Activation)    (None, 22, 22, 160)  0           ['batch_normalization_122[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_120 (Conv2D)            (None, 22, 22, 192)  208896      ['block17_11_ac[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_123 (Conv2D)            (None, 22, 22, 192)  215040      ['activation_122[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_120 (Batch  (None, 22, 22, 192)  576        ['conv2d_120[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_123 (Batch  (None, 22, 22, 192)  576        ['conv2d_123[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_120 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_120[0][0]']\n",
            "                                                                                                  \n",
            " activation_123 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_123[0][0]']\n",
            "                                                                                                  \n",
            " block17_12_mixed (Concatenate)  (None, 22, 22, 384)  0          ['activation_120[0][0]',         \n",
            "                                                                  'activation_123[0][0]']         \n",
            "                                                                                                  \n",
            " block17_12_conv (Conv2D)       (None, 22, 22, 1088  418880      ['block17_12_mixed[0][0]']       \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block17_12 (Lambda)            (None, 22, 22, 1088  0           ['block17_11_ac[0][0]',          \n",
            "                                )                                 'block17_12_conv[0][0]']        \n",
            "                                                                                                  \n",
            " block17_12_ac (Activation)     (None, 22, 22, 1088  0           ['block17_12[0][0]']             \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_125 (Conv2D)            (None, 22, 22, 128)  139264      ['block17_12_ac[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_125 (Batch  (None, 22, 22, 128)  384        ['conv2d_125[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_125 (Activation)    (None, 22, 22, 128)  0           ['batch_normalization_125[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_126 (Conv2D)            (None, 22, 22, 160)  143360      ['activation_125[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_126 (Batch  (None, 22, 22, 160)  480        ['conv2d_126[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_126 (Activation)    (None, 22, 22, 160)  0           ['batch_normalization_126[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_124 (Conv2D)            (None, 22, 22, 192)  208896      ['block17_12_ac[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_127 (Conv2D)            (None, 22, 22, 192)  215040      ['activation_126[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_124 (Batch  (None, 22, 22, 192)  576        ['conv2d_124[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_127 (Batch  (None, 22, 22, 192)  576        ['conv2d_127[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_124 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_124[0][0]']\n",
            "                                                                                                  \n",
            " activation_127 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_127[0][0]']\n",
            "                                                                                                  \n",
            " block17_13_mixed (Concatenate)  (None, 22, 22, 384)  0          ['activation_124[0][0]',         \n",
            "                                                                  'activation_127[0][0]']         \n",
            "                                                                                                  \n",
            " block17_13_conv (Conv2D)       (None, 22, 22, 1088  418880      ['block17_13_mixed[0][0]']       \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block17_13 (Lambda)            (None, 22, 22, 1088  0           ['block17_12_ac[0][0]',          \n",
            "                                )                                 'block17_13_conv[0][0]']        \n",
            "                                                                                                  \n",
            " block17_13_ac (Activation)     (None, 22, 22, 1088  0           ['block17_13[0][0]']             \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_129 (Conv2D)            (None, 22, 22, 128)  139264      ['block17_13_ac[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_129 (Batch  (None, 22, 22, 128)  384        ['conv2d_129[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_129 (Activation)    (None, 22, 22, 128)  0           ['batch_normalization_129[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_130 (Conv2D)            (None, 22, 22, 160)  143360      ['activation_129[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_130 (Batch  (None, 22, 22, 160)  480        ['conv2d_130[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_130 (Activation)    (None, 22, 22, 160)  0           ['batch_normalization_130[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_128 (Conv2D)            (None, 22, 22, 192)  208896      ['block17_13_ac[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_131 (Conv2D)            (None, 22, 22, 192)  215040      ['activation_130[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_128 (Batch  (None, 22, 22, 192)  576        ['conv2d_128[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_131 (Batch  (None, 22, 22, 192)  576        ['conv2d_131[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_128 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_128[0][0]']\n",
            "                                                                                                  \n",
            " activation_131 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_131[0][0]']\n",
            "                                                                                                  \n",
            " block17_14_mixed (Concatenate)  (None, 22, 22, 384)  0          ['activation_128[0][0]',         \n",
            "                                                                  'activation_131[0][0]']         \n",
            "                                                                                                  \n",
            " block17_14_conv (Conv2D)       (None, 22, 22, 1088  418880      ['block17_14_mixed[0][0]']       \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block17_14 (Lambda)            (None, 22, 22, 1088  0           ['block17_13_ac[0][0]',          \n",
            "                                )                                 'block17_14_conv[0][0]']        \n",
            "                                                                                                  \n",
            " block17_14_ac (Activation)     (None, 22, 22, 1088  0           ['block17_14[0][0]']             \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_133 (Conv2D)            (None, 22, 22, 128)  139264      ['block17_14_ac[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_133 (Batch  (None, 22, 22, 128)  384        ['conv2d_133[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_133 (Activation)    (None, 22, 22, 128)  0           ['batch_normalization_133[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_134 (Conv2D)            (None, 22, 22, 160)  143360      ['activation_133[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_134 (Batch  (None, 22, 22, 160)  480        ['conv2d_134[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_134 (Activation)    (None, 22, 22, 160)  0           ['batch_normalization_134[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_132 (Conv2D)            (None, 22, 22, 192)  208896      ['block17_14_ac[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_135 (Conv2D)            (None, 22, 22, 192)  215040      ['activation_134[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_132 (Batch  (None, 22, 22, 192)  576        ['conv2d_132[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_135 (Batch  (None, 22, 22, 192)  576        ['conv2d_135[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_132 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_132[0][0]']\n",
            "                                                                                                  \n",
            " activation_135 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_135[0][0]']\n",
            "                                                                                                  \n",
            " block17_15_mixed (Concatenate)  (None, 22, 22, 384)  0          ['activation_132[0][0]',         \n",
            "                                                                  'activation_135[0][0]']         \n",
            "                                                                                                  \n",
            " block17_15_conv (Conv2D)       (None, 22, 22, 1088  418880      ['block17_15_mixed[0][0]']       \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block17_15 (Lambda)            (None, 22, 22, 1088  0           ['block17_14_ac[0][0]',          \n",
            "                                )                                 'block17_15_conv[0][0]']        \n",
            "                                                                                                  \n",
            " block17_15_ac (Activation)     (None, 22, 22, 1088  0           ['block17_15[0][0]']             \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_137 (Conv2D)            (None, 22, 22, 128)  139264      ['block17_15_ac[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_137 (Batch  (None, 22, 22, 128)  384        ['conv2d_137[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_137 (Activation)    (None, 22, 22, 128)  0           ['batch_normalization_137[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_138 (Conv2D)            (None, 22, 22, 160)  143360      ['activation_137[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_138 (Batch  (None, 22, 22, 160)  480        ['conv2d_138[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_138 (Activation)    (None, 22, 22, 160)  0           ['batch_normalization_138[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_136 (Conv2D)            (None, 22, 22, 192)  208896      ['block17_15_ac[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_139 (Conv2D)            (None, 22, 22, 192)  215040      ['activation_138[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_136 (Batch  (None, 22, 22, 192)  576        ['conv2d_136[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_139 (Batch  (None, 22, 22, 192)  576        ['conv2d_139[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_136 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_136[0][0]']\n",
            "                                                                                                  \n",
            " activation_139 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_139[0][0]']\n",
            "                                                                                                  \n",
            " block17_16_mixed (Concatenate)  (None, 22, 22, 384)  0          ['activation_136[0][0]',         \n",
            "                                                                  'activation_139[0][0]']         \n",
            "                                                                                                  \n",
            " block17_16_conv (Conv2D)       (None, 22, 22, 1088  418880      ['block17_16_mixed[0][0]']       \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block17_16 (Lambda)            (None, 22, 22, 1088  0           ['block17_15_ac[0][0]',          \n",
            "                                )                                 'block17_16_conv[0][0]']        \n",
            "                                                                                                  \n",
            " block17_16_ac (Activation)     (None, 22, 22, 1088  0           ['block17_16[0][0]']             \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_141 (Conv2D)            (None, 22, 22, 128)  139264      ['block17_16_ac[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_141 (Batch  (None, 22, 22, 128)  384        ['conv2d_141[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_141 (Activation)    (None, 22, 22, 128)  0           ['batch_normalization_141[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_142 (Conv2D)            (None, 22, 22, 160)  143360      ['activation_141[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_142 (Batch  (None, 22, 22, 160)  480        ['conv2d_142[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_142 (Activation)    (None, 22, 22, 160)  0           ['batch_normalization_142[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_140 (Conv2D)            (None, 22, 22, 192)  208896      ['block17_16_ac[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_143 (Conv2D)            (None, 22, 22, 192)  215040      ['activation_142[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_140 (Batch  (None, 22, 22, 192)  576        ['conv2d_140[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_143 (Batch  (None, 22, 22, 192)  576        ['conv2d_143[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_140 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_140[0][0]']\n",
            "                                                                                                  \n",
            " activation_143 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_143[0][0]']\n",
            "                                                                                                  \n",
            " block17_17_mixed (Concatenate)  (None, 22, 22, 384)  0          ['activation_140[0][0]',         \n",
            "                                                                  'activation_143[0][0]']         \n",
            "                                                                                                  \n",
            " block17_17_conv (Conv2D)       (None, 22, 22, 1088  418880      ['block17_17_mixed[0][0]']       \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block17_17 (Lambda)            (None, 22, 22, 1088  0           ['block17_16_ac[0][0]',          \n",
            "                                )                                 'block17_17_conv[0][0]']        \n",
            "                                                                                                  \n",
            " block17_17_ac (Activation)     (None, 22, 22, 1088  0           ['block17_17[0][0]']             \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_145 (Conv2D)            (None, 22, 22, 128)  139264      ['block17_17_ac[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_145 (Batch  (None, 22, 22, 128)  384        ['conv2d_145[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_145 (Activation)    (None, 22, 22, 128)  0           ['batch_normalization_145[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_146 (Conv2D)            (None, 22, 22, 160)  143360      ['activation_145[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_146 (Batch  (None, 22, 22, 160)  480        ['conv2d_146[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_146 (Activation)    (None, 22, 22, 160)  0           ['batch_normalization_146[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_144 (Conv2D)            (None, 22, 22, 192)  208896      ['block17_17_ac[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_147 (Conv2D)            (None, 22, 22, 192)  215040      ['activation_146[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_144 (Batch  (None, 22, 22, 192)  576        ['conv2d_144[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_147 (Batch  (None, 22, 22, 192)  576        ['conv2d_147[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_144 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_144[0][0]']\n",
            "                                                                                                  \n",
            " activation_147 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_147[0][0]']\n",
            "                                                                                                  \n",
            " block17_18_mixed (Concatenate)  (None, 22, 22, 384)  0          ['activation_144[0][0]',         \n",
            "                                                                  'activation_147[0][0]']         \n",
            "                                                                                                  \n",
            " block17_18_conv (Conv2D)       (None, 22, 22, 1088  418880      ['block17_18_mixed[0][0]']       \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block17_18 (Lambda)            (None, 22, 22, 1088  0           ['block17_17_ac[0][0]',          \n",
            "                                )                                 'block17_18_conv[0][0]']        \n",
            "                                                                                                  \n",
            " block17_18_ac (Activation)     (None, 22, 22, 1088  0           ['block17_18[0][0]']             \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_149 (Conv2D)            (None, 22, 22, 128)  139264      ['block17_18_ac[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_149 (Batch  (None, 22, 22, 128)  384        ['conv2d_149[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_149 (Activation)    (None, 22, 22, 128)  0           ['batch_normalization_149[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_150 (Conv2D)            (None, 22, 22, 160)  143360      ['activation_149[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_150 (Batch  (None, 22, 22, 160)  480        ['conv2d_150[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_150 (Activation)    (None, 22, 22, 160)  0           ['batch_normalization_150[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_148 (Conv2D)            (None, 22, 22, 192)  208896      ['block17_18_ac[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_151 (Conv2D)            (None, 22, 22, 192)  215040      ['activation_150[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_148 (Batch  (None, 22, 22, 192)  576        ['conv2d_148[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_151 (Batch  (None, 22, 22, 192)  576        ['conv2d_151[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_148 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_148[0][0]']\n",
            "                                                                                                  \n",
            " activation_151 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_151[0][0]']\n",
            "                                                                                                  \n",
            " block17_19_mixed (Concatenate)  (None, 22, 22, 384)  0          ['activation_148[0][0]',         \n",
            "                                                                  'activation_151[0][0]']         \n",
            "                                                                                                  \n",
            " block17_19_conv (Conv2D)       (None, 22, 22, 1088  418880      ['block17_19_mixed[0][0]']       \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block17_19 (Lambda)            (None, 22, 22, 1088  0           ['block17_18_ac[0][0]',          \n",
            "                                )                                 'block17_19_conv[0][0]']        \n",
            "                                                                                                  \n",
            " block17_19_ac (Activation)     (None, 22, 22, 1088  0           ['block17_19[0][0]']             \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_153 (Conv2D)            (None, 22, 22, 128)  139264      ['block17_19_ac[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_153 (Batch  (None, 22, 22, 128)  384        ['conv2d_153[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_153 (Activation)    (None, 22, 22, 128)  0           ['batch_normalization_153[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_154 (Conv2D)            (None, 22, 22, 160)  143360      ['activation_153[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_154 (Batch  (None, 22, 22, 160)  480        ['conv2d_154[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_154 (Activation)    (None, 22, 22, 160)  0           ['batch_normalization_154[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_152 (Conv2D)            (None, 22, 22, 192)  208896      ['block17_19_ac[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_155 (Conv2D)            (None, 22, 22, 192)  215040      ['activation_154[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_152 (Batch  (None, 22, 22, 192)  576        ['conv2d_152[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_155 (Batch  (None, 22, 22, 192)  576        ['conv2d_155[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_152 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_152[0][0]']\n",
            "                                                                                                  \n",
            " activation_155 (Activation)    (None, 22, 22, 192)  0           ['batch_normalization_155[0][0]']\n",
            "                                                                                                  \n",
            " block17_20_mixed (Concatenate)  (None, 22, 22, 384)  0          ['activation_152[0][0]',         \n",
            "                                                                  'activation_155[0][0]']         \n",
            "                                                                                                  \n",
            " block17_20_conv (Conv2D)       (None, 22, 22, 1088  418880      ['block17_20_mixed[0][0]']       \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block17_20 (Lambda)            (None, 22, 22, 1088  0           ['block17_19_ac[0][0]',          \n",
            "                                )                                 'block17_20_conv[0][0]']        \n",
            "                                                                                                  \n",
            " block17_20_ac (Activation)     (None, 22, 22, 1088  0           ['block17_20[0][0]']             \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_160 (Conv2D)            (None, 22, 22, 256)  278528      ['block17_20_ac[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_160 (Batch  (None, 22, 22, 256)  768        ['conv2d_160[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_160 (Activation)    (None, 22, 22, 256)  0           ['batch_normalization_160[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_156 (Conv2D)            (None, 22, 22, 256)  278528      ['block17_20_ac[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_158 (Conv2D)            (None, 22, 22, 256)  278528      ['block17_20_ac[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_161 (Conv2D)            (None, 22, 22, 288)  663552      ['activation_160[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_156 (Batch  (None, 22, 22, 256)  768        ['conv2d_156[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_158 (Batch  (None, 22, 22, 256)  768        ['conv2d_158[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_161 (Batch  (None, 22, 22, 288)  864        ['conv2d_161[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_156 (Activation)    (None, 22, 22, 256)  0           ['batch_normalization_156[0][0]']\n",
            "                                                                                                  \n",
            " activation_158 (Activation)    (None, 22, 22, 256)  0           ['batch_normalization_158[0][0]']\n",
            "                                                                                                  \n",
            " activation_161 (Activation)    (None, 22, 22, 288)  0           ['batch_normalization_161[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_157 (Conv2D)            (None, 10, 10, 384)  884736      ['activation_156[0][0]']         \n",
            "                                                                                                  \n",
            " conv2d_159 (Conv2D)            (None, 10, 10, 288)  663552      ['activation_158[0][0]']         \n",
            "                                                                                                  \n",
            " conv2d_162 (Conv2D)            (None, 10, 10, 320)  829440      ['activation_161[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_157 (Batch  (None, 10, 10, 384)  1152       ['conv2d_157[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_159 (Batch  (None, 10, 10, 288)  864        ['conv2d_159[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_162 (Batch  (None, 10, 10, 320)  960        ['conv2d_162[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_157 (Activation)    (None, 10, 10, 384)  0           ['batch_normalization_157[0][0]']\n",
            "                                                                                                  \n",
            " activation_159 (Activation)    (None, 10, 10, 288)  0           ['batch_normalization_159[0][0]']\n",
            "                                                                                                  \n",
            " activation_162 (Activation)    (None, 10, 10, 320)  0           ['batch_normalization_162[0][0]']\n",
            "                                                                                                  \n",
            " max_pooling2d_3 (MaxPooling2D)  (None, 10, 10, 1088  0          ['block17_20_ac[0][0]']          \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " mixed_7a (Concatenate)         (None, 10, 10, 2080  0           ['activation_157[0][0]',         \n",
            "                                )                                 'activation_159[0][0]',         \n",
            "                                                                  'activation_162[0][0]',         \n",
            "                                                                  'max_pooling2d_3[0][0]']        \n",
            "                                                                                                  \n",
            " conv2d_164 (Conv2D)            (None, 10, 10, 192)  399360      ['mixed_7a[0][0]']               \n",
            "                                                                                                  \n",
            " batch_normalization_164 (Batch  (None, 10, 10, 192)  576        ['conv2d_164[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_164 (Activation)    (None, 10, 10, 192)  0           ['batch_normalization_164[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_165 (Conv2D)            (None, 10, 10, 224)  129024      ['activation_164[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_165 (Batch  (None, 10, 10, 224)  672        ['conv2d_165[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_165 (Activation)    (None, 10, 10, 224)  0           ['batch_normalization_165[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_163 (Conv2D)            (None, 10, 10, 192)  399360      ['mixed_7a[0][0]']               \n",
            "                                                                                                  \n",
            " conv2d_166 (Conv2D)            (None, 10, 10, 256)  172032      ['activation_165[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_163 (Batch  (None, 10, 10, 192)  576        ['conv2d_163[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_166 (Batch  (None, 10, 10, 256)  768        ['conv2d_166[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_163 (Activation)    (None, 10, 10, 192)  0           ['batch_normalization_163[0][0]']\n",
            "                                                                                                  \n",
            " activation_166 (Activation)    (None, 10, 10, 256)  0           ['batch_normalization_166[0][0]']\n",
            "                                                                                                  \n",
            " block8_1_mixed (Concatenate)   (None, 10, 10, 448)  0           ['activation_163[0][0]',         \n",
            "                                                                  'activation_166[0][0]']         \n",
            "                                                                                                  \n",
            " block8_1_conv (Conv2D)         (None, 10, 10, 2080  933920      ['block8_1_mixed[0][0]']         \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block8_1 (Lambda)              (None, 10, 10, 2080  0           ['mixed_7a[0][0]',               \n",
            "                                )                                 'block8_1_conv[0][0]']          \n",
            "                                                                                                  \n",
            " block8_1_ac (Activation)       (None, 10, 10, 2080  0           ['block8_1[0][0]']               \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_168 (Conv2D)            (None, 10, 10, 192)  399360      ['block8_1_ac[0][0]']            \n",
            "                                                                                                  \n",
            " batch_normalization_168 (Batch  (None, 10, 10, 192)  576        ['conv2d_168[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_168 (Activation)    (None, 10, 10, 192)  0           ['batch_normalization_168[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_169 (Conv2D)            (None, 10, 10, 224)  129024      ['activation_168[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_169 (Batch  (None, 10, 10, 224)  672        ['conv2d_169[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_169 (Activation)    (None, 10, 10, 224)  0           ['batch_normalization_169[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_167 (Conv2D)            (None, 10, 10, 192)  399360      ['block8_1_ac[0][0]']            \n",
            "                                                                                                  \n",
            " conv2d_170 (Conv2D)            (None, 10, 10, 256)  172032      ['activation_169[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_167 (Batch  (None, 10, 10, 192)  576        ['conv2d_167[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_170 (Batch  (None, 10, 10, 256)  768        ['conv2d_170[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_167 (Activation)    (None, 10, 10, 192)  0           ['batch_normalization_167[0][0]']\n",
            "                                                                                                  \n",
            " activation_170 (Activation)    (None, 10, 10, 256)  0           ['batch_normalization_170[0][0]']\n",
            "                                                                                                  \n",
            " block8_2_mixed (Concatenate)   (None, 10, 10, 448)  0           ['activation_167[0][0]',         \n",
            "                                                                  'activation_170[0][0]']         \n",
            "                                                                                                  \n",
            " block8_2_conv (Conv2D)         (None, 10, 10, 2080  933920      ['block8_2_mixed[0][0]']         \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block8_2 (Lambda)              (None, 10, 10, 2080  0           ['block8_1_ac[0][0]',            \n",
            "                                )                                 'block8_2_conv[0][0]']          \n",
            "                                                                                                  \n",
            " block8_2_ac (Activation)       (None, 10, 10, 2080  0           ['block8_2[0][0]']               \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_172 (Conv2D)            (None, 10, 10, 192)  399360      ['block8_2_ac[0][0]']            \n",
            "                                                                                                  \n",
            " batch_normalization_172 (Batch  (None, 10, 10, 192)  576        ['conv2d_172[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_172 (Activation)    (None, 10, 10, 192)  0           ['batch_normalization_172[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_173 (Conv2D)            (None, 10, 10, 224)  129024      ['activation_172[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_173 (Batch  (None, 10, 10, 224)  672        ['conv2d_173[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_173 (Activation)    (None, 10, 10, 224)  0           ['batch_normalization_173[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_171 (Conv2D)            (None, 10, 10, 192)  399360      ['block8_2_ac[0][0]']            \n",
            "                                                                                                  \n",
            " conv2d_174 (Conv2D)            (None, 10, 10, 256)  172032      ['activation_173[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_171 (Batch  (None, 10, 10, 192)  576        ['conv2d_171[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_174 (Batch  (None, 10, 10, 256)  768        ['conv2d_174[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_171 (Activation)    (None, 10, 10, 192)  0           ['batch_normalization_171[0][0]']\n",
            "                                                                                                  \n",
            " activation_174 (Activation)    (None, 10, 10, 256)  0           ['batch_normalization_174[0][0]']\n",
            "                                                                                                  \n",
            " block8_3_mixed (Concatenate)   (None, 10, 10, 448)  0           ['activation_171[0][0]',         \n",
            "                                                                  'activation_174[0][0]']         \n",
            "                                                                                                  \n",
            " block8_3_conv (Conv2D)         (None, 10, 10, 2080  933920      ['block8_3_mixed[0][0]']         \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block8_3 (Lambda)              (None, 10, 10, 2080  0           ['block8_2_ac[0][0]',            \n",
            "                                )                                 'block8_3_conv[0][0]']          \n",
            "                                                                                                  \n",
            " block8_3_ac (Activation)       (None, 10, 10, 2080  0           ['block8_3[0][0]']               \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_176 (Conv2D)            (None, 10, 10, 192)  399360      ['block8_3_ac[0][0]']            \n",
            "                                                                                                  \n",
            " batch_normalization_176 (Batch  (None, 10, 10, 192)  576        ['conv2d_176[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_176 (Activation)    (None, 10, 10, 192)  0           ['batch_normalization_176[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_177 (Conv2D)            (None, 10, 10, 224)  129024      ['activation_176[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_177 (Batch  (None, 10, 10, 224)  672        ['conv2d_177[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_177 (Activation)    (None, 10, 10, 224)  0           ['batch_normalization_177[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_175 (Conv2D)            (None, 10, 10, 192)  399360      ['block8_3_ac[0][0]']            \n",
            "                                                                                                  \n",
            " conv2d_178 (Conv2D)            (None, 10, 10, 256)  172032      ['activation_177[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_175 (Batch  (None, 10, 10, 192)  576        ['conv2d_175[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_178 (Batch  (None, 10, 10, 256)  768        ['conv2d_178[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_175 (Activation)    (None, 10, 10, 192)  0           ['batch_normalization_175[0][0]']\n",
            "                                                                                                  \n",
            " activation_178 (Activation)    (None, 10, 10, 256)  0           ['batch_normalization_178[0][0]']\n",
            "                                                                                                  \n",
            " block8_4_mixed (Concatenate)   (None, 10, 10, 448)  0           ['activation_175[0][0]',         \n",
            "                                                                  'activation_178[0][0]']         \n",
            "                                                                                                  \n",
            " block8_4_conv (Conv2D)         (None, 10, 10, 2080  933920      ['block8_4_mixed[0][0]']         \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block8_4 (Lambda)              (None, 10, 10, 2080  0           ['block8_3_ac[0][0]',            \n",
            "                                )                                 'block8_4_conv[0][0]']          \n",
            "                                                                                                  \n",
            " block8_4_ac (Activation)       (None, 10, 10, 2080  0           ['block8_4[0][0]']               \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_180 (Conv2D)            (None, 10, 10, 192)  399360      ['block8_4_ac[0][0]']            \n",
            "                                                                                                  \n",
            " batch_normalization_180 (Batch  (None, 10, 10, 192)  576        ['conv2d_180[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_180 (Activation)    (None, 10, 10, 192)  0           ['batch_normalization_180[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_181 (Conv2D)            (None, 10, 10, 224)  129024      ['activation_180[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_181 (Batch  (None, 10, 10, 224)  672        ['conv2d_181[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_181 (Activation)    (None, 10, 10, 224)  0           ['batch_normalization_181[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_179 (Conv2D)            (None, 10, 10, 192)  399360      ['block8_4_ac[0][0]']            \n",
            "                                                                                                  \n",
            " conv2d_182 (Conv2D)            (None, 10, 10, 256)  172032      ['activation_181[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_179 (Batch  (None, 10, 10, 192)  576        ['conv2d_179[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_182 (Batch  (None, 10, 10, 256)  768        ['conv2d_182[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_179 (Activation)    (None, 10, 10, 192)  0           ['batch_normalization_179[0][0]']\n",
            "                                                                                                  \n",
            " activation_182 (Activation)    (None, 10, 10, 256)  0           ['batch_normalization_182[0][0]']\n",
            "                                                                                                  \n",
            " block8_5_mixed (Concatenate)   (None, 10, 10, 448)  0           ['activation_179[0][0]',         \n",
            "                                                                  'activation_182[0][0]']         \n",
            "                                                                                                  \n",
            " block8_5_conv (Conv2D)         (None, 10, 10, 2080  933920      ['block8_5_mixed[0][0]']         \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block8_5 (Lambda)              (None, 10, 10, 2080  0           ['block8_4_ac[0][0]',            \n",
            "                                )                                 'block8_5_conv[0][0]']          \n",
            "                                                                                                  \n",
            " block8_5_ac (Activation)       (None, 10, 10, 2080  0           ['block8_5[0][0]']               \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_184 (Conv2D)            (None, 10, 10, 192)  399360      ['block8_5_ac[0][0]']            \n",
            "                                                                                                  \n",
            " batch_normalization_184 (Batch  (None, 10, 10, 192)  576        ['conv2d_184[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_184 (Activation)    (None, 10, 10, 192)  0           ['batch_normalization_184[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_185 (Conv2D)            (None, 10, 10, 224)  129024      ['activation_184[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_185 (Batch  (None, 10, 10, 224)  672        ['conv2d_185[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_185 (Activation)    (None, 10, 10, 224)  0           ['batch_normalization_185[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_183 (Conv2D)            (None, 10, 10, 192)  399360      ['block8_5_ac[0][0]']            \n",
            "                                                                                                  \n",
            " conv2d_186 (Conv2D)            (None, 10, 10, 256)  172032      ['activation_185[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_183 (Batch  (None, 10, 10, 192)  576        ['conv2d_183[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_186 (Batch  (None, 10, 10, 256)  768        ['conv2d_186[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_183 (Activation)    (None, 10, 10, 192)  0           ['batch_normalization_183[0][0]']\n",
            "                                                                                                  \n",
            " activation_186 (Activation)    (None, 10, 10, 256)  0           ['batch_normalization_186[0][0]']\n",
            "                                                                                                  \n",
            " block8_6_mixed (Concatenate)   (None, 10, 10, 448)  0           ['activation_183[0][0]',         \n",
            "                                                                  'activation_186[0][0]']         \n",
            "                                                                                                  \n",
            " block8_6_conv (Conv2D)         (None, 10, 10, 2080  933920      ['block8_6_mixed[0][0]']         \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block8_6 (Lambda)              (None, 10, 10, 2080  0           ['block8_5_ac[0][0]',            \n",
            "                                )                                 'block8_6_conv[0][0]']          \n",
            "                                                                                                  \n",
            " block8_6_ac (Activation)       (None, 10, 10, 2080  0           ['block8_6[0][0]']               \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_188 (Conv2D)            (None, 10, 10, 192)  399360      ['block8_6_ac[0][0]']            \n",
            "                                                                                                  \n",
            " batch_normalization_188 (Batch  (None, 10, 10, 192)  576        ['conv2d_188[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_188 (Activation)    (None, 10, 10, 192)  0           ['batch_normalization_188[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_189 (Conv2D)            (None, 10, 10, 224)  129024      ['activation_188[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_189 (Batch  (None, 10, 10, 224)  672        ['conv2d_189[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_189 (Activation)    (None, 10, 10, 224)  0           ['batch_normalization_189[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_187 (Conv2D)            (None, 10, 10, 192)  399360      ['block8_6_ac[0][0]']            \n",
            "                                                                                                  \n",
            " conv2d_190 (Conv2D)            (None, 10, 10, 256)  172032      ['activation_189[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_187 (Batch  (None, 10, 10, 192)  576        ['conv2d_187[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_190 (Batch  (None, 10, 10, 256)  768        ['conv2d_190[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_187 (Activation)    (None, 10, 10, 192)  0           ['batch_normalization_187[0][0]']\n",
            "                                                                                                  \n",
            " activation_190 (Activation)    (None, 10, 10, 256)  0           ['batch_normalization_190[0][0]']\n",
            "                                                                                                  \n",
            " block8_7_mixed (Concatenate)   (None, 10, 10, 448)  0           ['activation_187[0][0]',         \n",
            "                                                                  'activation_190[0][0]']         \n",
            "                                                                                                  \n",
            " block8_7_conv (Conv2D)         (None, 10, 10, 2080  933920      ['block8_7_mixed[0][0]']         \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block8_7 (Lambda)              (None, 10, 10, 2080  0           ['block8_6_ac[0][0]',            \n",
            "                                )                                 'block8_7_conv[0][0]']          \n",
            "                                                                                                  \n",
            " block8_7_ac (Activation)       (None, 10, 10, 2080  0           ['block8_7[0][0]']               \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_192 (Conv2D)            (None, 10, 10, 192)  399360      ['block8_7_ac[0][0]']            \n",
            "                                                                                                  \n",
            " batch_normalization_192 (Batch  (None, 10, 10, 192)  576        ['conv2d_192[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_192 (Activation)    (None, 10, 10, 192)  0           ['batch_normalization_192[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_193 (Conv2D)            (None, 10, 10, 224)  129024      ['activation_192[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_193 (Batch  (None, 10, 10, 224)  672        ['conv2d_193[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_193 (Activation)    (None, 10, 10, 224)  0           ['batch_normalization_193[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_191 (Conv2D)            (None, 10, 10, 192)  399360      ['block8_7_ac[0][0]']            \n",
            "                                                                                                  \n",
            " conv2d_194 (Conv2D)            (None, 10, 10, 256)  172032      ['activation_193[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_191 (Batch  (None, 10, 10, 192)  576        ['conv2d_191[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_194 (Batch  (None, 10, 10, 256)  768        ['conv2d_194[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_191 (Activation)    (None, 10, 10, 192)  0           ['batch_normalization_191[0][0]']\n",
            "                                                                                                  \n",
            " activation_194 (Activation)    (None, 10, 10, 256)  0           ['batch_normalization_194[0][0]']\n",
            "                                                                                                  \n",
            " block8_8_mixed (Concatenate)   (None, 10, 10, 448)  0           ['activation_191[0][0]',         \n",
            "                                                                  'activation_194[0][0]']         \n",
            "                                                                                                  \n",
            " block8_8_conv (Conv2D)         (None, 10, 10, 2080  933920      ['block8_8_mixed[0][0]']         \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block8_8 (Lambda)              (None, 10, 10, 2080  0           ['block8_7_ac[0][0]',            \n",
            "                                )                                 'block8_8_conv[0][0]']          \n",
            "                                                                                                  \n",
            " block8_8_ac (Activation)       (None, 10, 10, 2080  0           ['block8_8[0][0]']               \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_196 (Conv2D)            (None, 10, 10, 192)  399360      ['block8_8_ac[0][0]']            \n",
            "                                                                                                  \n",
            " batch_normalization_196 (Batch  (None, 10, 10, 192)  576        ['conv2d_196[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_196 (Activation)    (None, 10, 10, 192)  0           ['batch_normalization_196[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_197 (Conv2D)            (None, 10, 10, 224)  129024      ['activation_196[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_197 (Batch  (None, 10, 10, 224)  672        ['conv2d_197[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_197 (Activation)    (None, 10, 10, 224)  0           ['batch_normalization_197[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_195 (Conv2D)            (None, 10, 10, 192)  399360      ['block8_8_ac[0][0]']            \n",
            "                                                                                                  \n",
            " conv2d_198 (Conv2D)            (None, 10, 10, 256)  172032      ['activation_197[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_195 (Batch  (None, 10, 10, 192)  576        ['conv2d_195[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_198 (Batch  (None, 10, 10, 256)  768        ['conv2d_198[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_195 (Activation)    (None, 10, 10, 192)  0           ['batch_normalization_195[0][0]']\n",
            "                                                                                                  \n",
            " activation_198 (Activation)    (None, 10, 10, 256)  0           ['batch_normalization_198[0][0]']\n",
            "                                                                                                  \n",
            " block8_9_mixed (Concatenate)   (None, 10, 10, 448)  0           ['activation_195[0][0]',         \n",
            "                                                                  'activation_198[0][0]']         \n",
            "                                                                                                  \n",
            " block8_9_conv (Conv2D)         (None, 10, 10, 2080  933920      ['block8_9_mixed[0][0]']         \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block8_9 (Lambda)              (None, 10, 10, 2080  0           ['block8_8_ac[0][0]',            \n",
            "                                )                                 'block8_9_conv[0][0]']          \n",
            "                                                                                                  \n",
            " block8_9_ac (Activation)       (None, 10, 10, 2080  0           ['block8_9[0][0]']               \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_200 (Conv2D)            (None, 10, 10, 192)  399360      ['block8_9_ac[0][0]']            \n",
            "                                                                                                  \n",
            " batch_normalization_200 (Batch  (None, 10, 10, 192)  576        ['conv2d_200[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_200 (Activation)    (None, 10, 10, 192)  0           ['batch_normalization_200[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_201 (Conv2D)            (None, 10, 10, 224)  129024      ['activation_200[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_201 (Batch  (None, 10, 10, 224)  672        ['conv2d_201[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_201 (Activation)    (None, 10, 10, 224)  0           ['batch_normalization_201[0][0]']\n",
            "                                                                                                  \n",
            " conv2d_199 (Conv2D)            (None, 10, 10, 192)  399360      ['block8_9_ac[0][0]']            \n",
            "                                                                                                  \n",
            " conv2d_202 (Conv2D)            (None, 10, 10, 256)  172032      ['activation_201[0][0]']         \n",
            "                                                                                                  \n",
            " batch_normalization_199 (Batch  (None, 10, 10, 192)  576        ['conv2d_199[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " batch_normalization_202 (Batch  (None, 10, 10, 256)  768        ['conv2d_202[0][0]']             \n",
            " Normalization)                                                                                   \n",
            "                                                                                                  \n",
            " activation_199 (Activation)    (None, 10, 10, 192)  0           ['batch_normalization_199[0][0]']\n",
            "                                                                                                  \n",
            " activation_202 (Activation)    (None, 10, 10, 256)  0           ['batch_normalization_202[0][0]']\n",
            "                                                                                                  \n",
            " block8_10_mixed (Concatenate)  (None, 10, 10, 448)  0           ['activation_199[0][0]',         \n",
            "                                                                  'activation_202[0][0]']         \n",
            "                                                                                                  \n",
            " block8_10_conv (Conv2D)        (None, 10, 10, 2080  933920      ['block8_10_mixed[0][0]']        \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " block8_10 (Lambda)             (None, 10, 10, 2080  0           ['block8_9_ac[0][0]',            \n",
            "                                )                                 'block8_10_conv[0][0]']         \n",
            "                                                                                                  \n",
            " conv_7b (Conv2D)               (None, 10, 10, 1536  3194880     ['block8_10[0][0]']              \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv_7b_bn (BatchNormalization  (None, 10, 10, 1536  4608       ['conv_7b[0][0]']                \n",
            " )                              )                                                                 \n",
            "                                                                                                  \n",
            " conv_7b_ac (Activation)        (None, 10, 10, 1536  0           ['conv_7b_bn[0][0]']             \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " global_average_pooling2d (Glob  (None, 1536)        0           ['conv_7b_ac[0][0]']             \n",
            " alAveragePooling2D)                                                                              \n",
            "                                                                                                  \n",
            " dense (Dense)                  (None, 1024)         1573888     ['global_average_pooling2d[0][0]'\n",
            "                                                                 ]                                \n",
            "                                                                                                  \n",
            " dense_1 (Dense)                (None, 1024)         1049600     ['dense[0][0]']                  \n",
            "                                                                                                  \n",
            " dropout (Dropout)              (None, 1024)         0           ['dense_1[0][0]']                \n",
            "                                                                                                  \n",
            " dense_2 (Dense)                (None, 1024)         1049600     ['dropout[0][0]']                \n",
            "                                                                                                  \n",
            " dense_3 (Dense)                (None, 1024)         1049600     ['dense_2[0][0]']                \n",
            "                                                                                                  \n",
            " dropout_1 (Dropout)            (None, 1024)         0           ['dense_3[0][0]']                \n",
            "                                                                                                  \n",
            " dense_4 (Dense)                (None, 1024)         1049600     ['dropout_1[0][0]']              \n",
            "                                                                                                  \n",
            " dense_5 (Dense)                (None, 1024)         1049600     ['dense_4[0][0]']                \n",
            "                                                                                                  \n",
            " dense_6 (Dense)                (None, 9)            9225        ['dense_5[0][0]']                \n",
            "                                                                                                  \n",
            "==================================================================================================\n",
            "Total params: 61,167,849\n",
            "Trainable params: 61,107,305\n",
            "Non-trainable params: 60,544\n",
            "__________________________________________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "model.summary()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "moylH4EzX8VJ"
      },
      "outputs": [],
      "source": [
        "tf.keras.backend.clear_session()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "TwyQOmiGWdMa",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f2149aae-2c3d-450a-f730-14c8943a49d8"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Nº de camadas: 790\n"
          ]
        }
      ],
      "source": [
        "camadas = dict([(layer.name, layer) for layer in model.layers])\n",
        "\n",
        "print(f'Nº de camadas: {len(camadas)}')\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "iQehUhlmXhab",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b2b92d17-fa9a-49e7-a052-38825b1d987b"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/50\n",
            "100/100 [==============================] - 102s 760ms/step - loss: 1.1374 - accuracy: 0.5856 - tp: 727.0000 - tn: 12612.0000 - fp: 188.0000 - fn: 873.0000 - precision: 0.7374 - sensitivity: 0.9706 - specificity: 0.9777 - recall: 0.4544 - FBetaScore: 0.1853 - val_loss: 1.5066 - val_accuracy: 0.5750 - val_tp: 23.0000 - val_tn: 306.0000 - val_fp: 14.0000 - val_fn: 17.0000 - val_precision: 0.7143 - val_sensitivity: 0.9250 - val_specificity: 0.9750 - val_recall: 0.5750 - val_FBetaScore: 0.1808 - lr: 1.0000e-04\n",
            "Epoch 2/50\n",
            "100/100 [==============================] - 71s 703ms/step - loss: 0.8544 - accuracy: 0.7245 - tp: 1014.0000 - tn: 12541.0000 - fp: 147.0000 - fn: 572.0000 - precision: 0.9404 - sensitivity: 0.9893 - specificity: 0.9960 - recall: 0.6393 - FBetaScore: 0.2666 - val_loss: 0.9212 - val_accuracy: 0.7000 - val_tp: 23.0000 - val_tn: 317.0000 - val_fp: 3.0000 - val_fn: 17.0000 - val_precision: 0.9167 - val_sensitivity: 0.9750 - val_specificity: 0.9937 - val_recall: 0.5750 - val_FBetaScore: 0.2491 - lr: 1.0000e-04\n",
            "Epoch 3/50\n",
            "100/100 [==============================] - 71s 711ms/step - loss: 0.6624 - accuracy: 0.7819 - tp: 1154.0000 - tn: 12697.0000 - fp: 103.0000 - fn: 446.0000 - precision: 0.9822 - sensitivity: 0.9956 - specificity: 0.9988 - recall: 0.7212 - FBetaScore: 0.2908 - val_loss: 1.0886 - val_accuracy: 0.6250 - val_tp: 18.0000 - val_tn: 317.0000 - val_fp: 3.0000 - val_fn: 22.0000 - val_precision: 0.8696 - val_sensitivity: 0.9750 - val_specificity: 0.9906 - val_recall: 0.4500 - val_FBetaScore: 0.2039 - lr: 1.0000e-04\n",
            "Epoch 4/50\n",
            "100/100 [==============================] - 70s 700ms/step - loss: 0.5977 - accuracy: 0.7976 - tp: 1181.0000 - tn: 12592.0000 - fp: 96.0000 - fn: 405.0000 - precision: 0.9862 - sensitivity: 0.9943 - specificity: 0.9991 - recall: 0.7446 - FBetaScore: 0.3004 - val_loss: 0.8659 - val_accuracy: 0.7250 - val_tp: 27.0000 - val_tn: 315.0000 - val_fp: 5.0000 - val_fn: 13.0000 - val_precision: 0.9524 - val_sensitivity: 0.9750 - val_specificity: 0.9969 - val_recall: 0.6750 - val_FBetaScore: 0.2532 - lr: 1.0000e-04\n",
            "Epoch 5/50\n",
            "100/100 [==============================] - 70s 698ms/step - loss: 0.5192 - accuracy: 0.8348 - tp: 1253.0000 - tn: 12599.0000 - fp: 89.0000 - fn: 333.0000 - precision: 0.9928 - sensitivity: 0.9943 - specificity: 0.9995 - recall: 0.7900 - FBetaScore: 0.3657 - val_loss: 1.1524 - val_accuracy: 0.7000 - val_tp: 28.0000 - val_tn: 313.0000 - val_fp: 7.0000 - val_fn: 12.0000 - val_precision: 0.8800 - val_sensitivity: 0.9500 - val_specificity: 0.9906 - val_recall: 0.7000 - val_FBetaScore: 0.3380 - lr: 1.0000e-04\n",
            "Epoch 6/50\n",
            "100/100 [==============================] - 71s 708ms/step - loss: 0.4892 - accuracy: 0.8425 - tp: 1268.0000 - tn: 12719.0000 - fp: 81.0000 - fn: 332.0000 - precision: 0.9965 - sensitivity: 0.9950 - specificity: 0.9998 - recall: 0.7925 - FBetaScore: 0.3738 - val_loss: 0.8232 - val_accuracy: 0.7000 - val_tp: 27.0000 - val_tn: 315.0000 - val_fp: 5.0000 - val_fn: 13.0000 - val_precision: 0.9643 - val_sensitivity: 0.9750 - val_specificity: 0.9969 - val_recall: 0.6750 - val_FBetaScore: 0.2527 - lr: 1.0000e-04\n",
            "Epoch 7/50\n",
            "100/100 [==============================] - 71s 711ms/step - loss: 0.4287 - accuracy: 0.8612 - tp: 1295.0000 - tn: 12726.0000 - fp: 74.0000 - fn: 305.0000 - precision: 0.9970 - sensitivity: 0.9962 - specificity: 0.9998 - recall: 0.8094 - FBetaScore: 0.3929 - val_loss: 1.3113 - val_accuracy: 0.6250 - val_tp: 24.0000 - val_tn: 311.0000 - val_fp: 9.0000 - val_fn: 16.0000 - val_precision: 0.8800 - val_sensitivity: 0.9500 - val_specificity: 0.9906 - val_recall: 0.6000 - val_FBetaScore: 0.3270 - lr: 1.0000e-04\n",
            "Epoch 8/50\n",
            "100/100 [==============================] - 70s 695ms/step - loss: 0.4229 - accuracy: 0.8714 - tp: 1302.0000 - tn: 12622.0000 - fp: 66.0000 - fn: 284.0000 - precision: 0.9965 - sensitivity: 0.9956 - specificity: 0.9998 - recall: 0.8209 - FBetaScore: 0.4092 - val_loss: 0.7271 - val_accuracy: 0.7500 - val_tp: 29.0000 - val_tn: 316.0000 - val_fp: 4.0000 - val_fn: 11.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.7250 - val_FBetaScore: 0.3720 - lr: 1.0000e-04\n",
            "Epoch 9/50\n",
            "100/100 [==============================] - 72s 713ms/step - loss: 0.3998 - accuracy: 0.8737 - tp: 1337.0000 - tn: 12716.0000 - fp: 84.0000 - fn: 263.0000 - precision: 0.9940 - sensitivity: 0.9950 - specificity: 0.9996 - recall: 0.8356 - FBetaScore: 0.4582 - val_loss: 0.4242 - val_accuracy: 0.8750 - val_tp: 35.0000 - val_tn: 318.0000 - val_fp: 2.0000 - val_fn: 5.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.8750 - val_FBetaScore: 0.5121 - lr: 1.0000e-04\n",
            "Epoch 10/50\n",
            "100/100 [==============================] - 70s 693ms/step - loss: 0.3405 - accuracy: 0.8916 - tp: 1351.0000 - tn: 12625.0000 - fp: 63.0000 - fn: 235.0000 - precision: 0.9968 - sensitivity: 0.9968 - specificity: 0.9998 - recall: 0.8518 - FBetaScore: 0.4846 - val_loss: 0.2467 - val_accuracy: 0.9000 - val_tp: 35.0000 - val_tn: 318.0000 - val_fp: 2.0000 - val_fn: 5.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.8750 - val_FBetaScore: 0.3872 - lr: 1.0000e-04\n",
            "Epoch 11/50\n",
            "100/100 [==============================] - 71s 707ms/step - loss: 0.3136 - accuracy: 0.8988 - tp: 1383.0000 - tn: 12741.0000 - fp: 59.0000 - fn: 217.0000 - precision: 0.9978 - sensitivity: 0.9975 - specificity: 0.9998 - recall: 0.8644 - FBetaScore: 0.4826 - val_loss: 0.6381 - val_accuracy: 0.7750 - val_tp: 28.0000 - val_tn: 316.0000 - val_fp: 4.0000 - val_fn: 12.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.7000 - val_FBetaScore: 0.3488 - lr: 1.0000e-04\n",
            "Epoch 12/50\n",
            "100/100 [==============================] - ETA: 0s - loss: 0.3160 - accuracy: 0.8969 - tp: 1391.0000 - tn: 12744.0000 - fp: 56.0000 - fn: 209.0000 - precision: 0.9981 - sensitivity: 0.9969 - specificity: 0.9998 - recall: 0.8694 - FBetaScore: 0.5353\n",
            "Epoch 12: ReduceLROnPlateau reducing learning rate to 1.9999999494757503e-05.\n",
            "100/100 [==============================] - 72s 715ms/step - loss: 0.3160 - accuracy: 0.8969 - tp: 1391.0000 - tn: 12744.0000 - fp: 56.0000 - fn: 209.0000 - precision: 0.9981 - sensitivity: 0.9969 - specificity: 0.9998 - recall: 0.8694 - FBetaScore: 0.5353 - val_loss: 0.5499 - val_accuracy: 0.8750 - val_tp: 32.0000 - val_tn: 315.0000 - val_fp: 5.0000 - val_fn: 8.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.8000 - val_FBetaScore: 0.4744 - lr: 1.0000e-04\n",
            "Epoch 13/50\n",
            "100/100 [==============================] - 72s 717ms/step - loss: 0.2332 - accuracy: 0.9212 - tp: 1420.0000 - tn: 12634.0000 - fp: 54.0000 - fn: 166.0000 - precision: 0.9984 - sensitivity: 0.9968 - specificity: 0.9998 - recall: 0.8953 - FBetaScore: 0.5399 - val_loss: 0.4114 - val_accuracy: 0.8500 - val_tp: 34.0000 - val_tn: 317.0000 - val_fp: 3.0000 - val_fn: 6.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.8500 - val_FBetaScore: 0.3247 - lr: 2.0000e-05\n",
            "Epoch 14/50\n",
            "100/100 [==============================] - 72s 720ms/step - loss: 0.1981 - accuracy: 0.9288 - tp: 1447.0000 - tn: 12654.0000 - fp: 34.0000 - fn: 139.0000 - precision: 0.9975 - sensitivity: 0.9981 - specificity: 0.9998 - recall: 0.9124 - FBetaScore: 0.5946 - val_loss: 0.3885 - val_accuracy: 0.8750 - val_tp: 33.0000 - val_tn: 316.0000 - val_fp: 4.0000 - val_fn: 7.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.8250 - val_FBetaScore: 0.3487 - lr: 2.0000e-05\n",
            "Epoch 15/50\n",
            "100/100 [==============================] - 71s 706ms/step - loss: 0.1770 - accuracy: 0.9369 - tp: 1449.0000 - tn: 12657.0000 - fp: 31.0000 - fn: 137.0000 - precision: 1.0000 - sensitivity: 1.0000 - specificity: 1.0000 - recall: 0.9136 - FBetaScore: 0.5962 - val_loss: 0.7783 - val_accuracy: 0.8000 - val_tp: 32.0000 - val_tn: 314.0000 - val_fp: 6.0000 - val_fn: 8.0000 - val_precision: 0.9677 - val_sensitivity: 0.9750 - val_specificity: 0.9969 - val_recall: 0.8000 - val_FBetaScore: 0.5107 - lr: 2.0000e-05\n",
            "Epoch 16/50\n",
            "100/100 [==============================] - 70s 697ms/step - loss: 0.1761 - accuracy: 0.9444 - tp: 1477.0000 - tn: 12766.0000 - fp: 34.0000 - fn: 123.0000 - precision: 0.9983 - sensitivity: 0.9975 - specificity: 0.9998 - recall: 0.9231 - FBetaScore: 0.6165 - val_loss: 0.5818 - val_accuracy: 0.8500 - val_tp: 33.0000 - val_tn: 318.0000 - val_fp: 2.0000 - val_fn: 7.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.8250 - val_FBetaScore: 0.3150 - lr: 2.0000e-05\n",
            "Epoch 17/50\n",
            "100/100 [==============================] - 72s 719ms/step - loss: 0.1350 - accuracy: 0.9464 - tp: 1471.0000 - tn: 12663.0000 - fp: 25.0000 - fn: 115.0000 - precision: 1.0000 - sensitivity: 1.0000 - specificity: 1.0000 - recall: 0.9275 - FBetaScore: 0.6204 - val_loss: 0.1379 - val_accuracy: 0.9250 - val_tp: 37.0000 - val_tn: 319.0000 - val_fp: 1.0000 - val_fn: 3.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9250 - val_FBetaScore: 0.4299 - lr: 2.0000e-05\n",
            "Epoch 18/50\n",
            "100/100 [==============================] - 70s 694ms/step - loss: 0.1075 - accuracy: 0.9578 - tp: 1495.0000 - tn: 12667.0000 - fp: 21.0000 - fn: 91.0000 - precision: 0.9993 - sensitivity: 0.9994 - specificity: 0.9999 - recall: 0.9426 - FBetaScore: 0.6527 - val_loss: 0.3564 - val_accuracy: 0.9000 - val_tp: 36.0000 - val_tn: 319.0000 - val_fp: 1.0000 - val_fn: 4.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9000 - val_FBetaScore: 0.5458 - lr: 2.0000e-05\n",
            "Epoch 19/50\n",
            "100/100 [==============================] - 71s 705ms/step - loss: 0.1054 - accuracy: 0.9641 - tp: 1495.0000 - tn: 12665.0000 - fp: 23.0000 - fn: 91.0000 - precision: 1.0000 - sensitivity: 1.0000 - specificity: 1.0000 - recall: 0.9426 - FBetaScore: 0.6868 - val_loss: 0.1266 - val_accuracy: 0.9500 - val_tp: 38.0000 - val_tn: 318.0000 - val_fp: 2.0000 - val_fn: 2.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9500 - val_FBetaScore: 0.6168 - lr: 2.0000e-05\n",
            "Epoch 20/50\n",
            "100/100 [==============================] - ETA: 0s - loss: 0.1060 - accuracy: 0.9584 - tp: 1505.0000 - tn: 12658.0000 - fp: 30.0000 - fn: 81.0000 - precision: 1.0000 - sensitivity: 1.0000 - specificity: 1.0000 - recall: 0.9489 - FBetaScore: 0.7403\n",
            "Epoch 20: ReduceLROnPlateau reducing learning rate to 1e-05.\n",
            "100/100 [==============================] - 70s 693ms/step - loss: 0.1060 - accuracy: 0.9584 - tp: 1505.0000 - tn: 12658.0000 - fp: 30.0000 - fn: 81.0000 - precision: 1.0000 - sensitivity: 1.0000 - specificity: 1.0000 - recall: 0.9489 - FBetaScore: 0.7403 - val_loss: 0.6716 - val_accuracy: 0.8000 - val_tp: 32.0000 - val_tn: 314.0000 - val_fp: 6.0000 - val_fn: 8.0000 - val_precision: 1.0000 - val_sensitivity: 0.9750 - val_specificity: 1.0000 - val_recall: 0.8000 - val_FBetaScore: 0.4242 - lr: 2.0000e-05\n",
            "Epoch 21/50\n",
            "100/100 [==============================] - 71s 706ms/step - loss: 0.1090 - accuracy: 0.9552 - tp: 1503.0000 - tn: 12653.0000 - fp: 35.0000 - fn: 83.0000 - precision: 1.0000 - sensitivity: 1.0000 - specificity: 1.0000 - recall: 0.9477 - FBetaScore: 0.7170 - val_loss: 0.3403 - val_accuracy: 0.9000 - val_tp: 35.0000 - val_tn: 318.0000 - val_fp: 2.0000 - val_fn: 5.0000 - val_precision: 1.0000 - val_sensitivity: 0.9750 - val_specificity: 1.0000 - val_recall: 0.8750 - val_FBetaScore: 0.6005 - lr: 1.0000e-05\n",
            "Epoch 22/50\n",
            "100/100 [==============================] - 70s 696ms/step - loss: 0.0883 - accuracy: 0.9663 - tp: 1537.0000 - tn: 12783.0000 - fp: 17.0000 - fn: 63.0000 - precision: 1.0000 - sensitivity: 0.9994 - specificity: 1.0000 - recall: 0.9606 - FBetaScore: 0.7827 - val_loss: 0.2139 - val_accuracy: 0.9500 - val_tp: 38.0000 - val_tn: 319.0000 - val_fp: 1.0000 - val_fn: 2.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9500 - val_FBetaScore: 0.5232 - lr: 1.0000e-05\n",
            "Epoch 23/50\n",
            "100/100 [==============================] - 71s 705ms/step - loss: 0.0801 - accuracy: 0.9691 - tp: 1518.0000 - tn: 12671.0000 - fp: 17.0000 - fn: 68.0000 - precision: 1.0000 - sensitivity: 0.9994 - specificity: 1.0000 - recall: 0.9571 - FBetaScore: 0.7338 - val_loss: 0.2223 - val_accuracy: 0.9000 - val_tp: 36.0000 - val_tn: 316.0000 - val_fp: 4.0000 - val_fn: 4.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9000 - val_FBetaScore: 0.4277 - lr: 1.0000e-05\n",
            "Epoch 24/50\n",
            "100/100 [==============================] - 71s 708ms/step - loss: 0.0801 - accuracy: 0.9653 - tp: 1520.0000 - tn: 12658.0000 - fp: 30.0000 - fn: 66.0000 - precision: 1.0000 - sensitivity: 0.9994 - specificity: 1.0000 - recall: 0.9584 - FBetaScore: 0.7524 - val_loss: 0.0986 - val_accuracy: 0.9500 - val_tp: 37.0000 - val_tn: 319.0000 - val_fp: 1.0000 - val_fn: 3.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9250 - val_FBetaScore: 0.6137 - lr: 1.0000e-05\n",
            "Epoch 25/50\n",
            "100/100 [==============================] - 70s 694ms/step - loss: 0.1020 - accuracy: 0.9678 - tp: 1524.0000 - tn: 12660.0000 - fp: 28.0000 - fn: 62.0000 - precision: 0.9978 - sensitivity: 0.9975 - specificity: 0.9998 - recall: 0.9609 - FBetaScore: 0.7861 - val_loss: 0.2624 - val_accuracy: 0.9000 - val_tp: 36.0000 - val_tn: 318.0000 - val_fp: 2.0000 - val_fn: 4.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9000 - val_FBetaScore: 0.4321 - lr: 1.0000e-05\n",
            "Epoch 26/50\n",
            "100/100 [==============================] - 71s 710ms/step - loss: 0.0857 - accuracy: 0.9712 - tp: 1539.0000 - tn: 12781.0000 - fp: 19.0000 - fn: 61.0000 - precision: 1.0000 - sensitivity: 0.9994 - specificity: 1.0000 - recall: 0.9619 - FBetaScore: 0.8052 - val_loss: 0.0330 - val_accuracy: 1.0000 - val_tp: 40.0000 - val_tn: 320.0000 - val_fp: 0.0000e+00 - val_fn: 0.0000e+00 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 1.0000 - val_FBetaScore: 0.6667 - lr: 1.0000e-05\n",
            "Epoch 27/50\n",
            "100/100 [==============================] - 69s 692ms/step - loss: 0.0744 - accuracy: 0.9773 - tp: 1543.0000 - tn: 12666.0000 - fp: 22.0000 - fn: 43.0000 - precision: 1.0000 - sensitivity: 1.0000 - specificity: 1.0000 - recall: 0.9729 - FBetaScore: 0.8156 - val_loss: 0.1983 - val_accuracy: 0.9250 - val_tp: 37.0000 - val_tn: 318.0000 - val_fp: 2.0000 - val_fn: 3.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9250 - val_FBetaScore: 0.5973 - lr: 1.0000e-05\n",
            "Epoch 28/50\n",
            "100/100 [==============================] - 69s 691ms/step - loss: 0.0791 - accuracy: 0.9691 - tp: 1534.0000 - tn: 12659.0000 - fp: 29.0000 - fn: 52.0000 - precision: 0.9993 - sensitivity: 0.9987 - specificity: 0.9999 - recall: 0.9672 - FBetaScore: 0.8065 - val_loss: 0.1506 - val_accuracy: 0.9500 - val_tp: 38.0000 - val_tn: 319.0000 - val_fp: 1.0000 - val_fn: 2.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9500 - val_FBetaScore: 0.6554 - lr: 1.0000e-05\n",
            "Epoch 29/50\n",
            "100/100 [==============================] - 71s 709ms/step - loss: 0.0554 - accuracy: 0.9819 - tp: 1563.0000 - tn: 12788.0000 - fp: 12.0000 - fn: 37.0000 - precision: 1.0000 - sensitivity: 1.0000 - specificity: 1.0000 - recall: 0.9769 - FBetaScore: 0.8388 - val_loss: 0.2423 - val_accuracy: 0.9250 - val_tp: 37.0000 - val_tn: 318.0000 - val_fp: 2.0000 - val_fn: 3.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9250 - val_FBetaScore: 0.6998 - lr: 1.0000e-05\n",
            "Epoch 30/50\n",
            "100/100 [==============================] - 70s 695ms/step - loss: 0.0560 - accuracy: 0.9811 - tp: 1553.0000 - tn: 12675.0000 - fp: 13.0000 - fn: 33.0000 - precision: 0.9993 - sensitivity: 0.9994 - specificity: 0.9999 - recall: 0.9792 - FBetaScore: 0.8506 - val_loss: 0.0590 - val_accuracy: 0.9750 - val_tp: 39.0000 - val_tn: 319.0000 - val_fp: 1.0000 - val_fn: 1.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9750 - val_FBetaScore: 0.5137 - lr: 1.0000e-05\n",
            "Epoch 31/50\n",
            "100/100 [==============================] - 71s 712ms/step - loss: 0.0590 - accuracy: 0.9767 - tp: 1542.0000 - tn: 12673.0000 - fp: 15.0000 - fn: 44.0000 - precision: 1.0000 - sensitivity: 1.0000 - specificity: 1.0000 - recall: 0.9723 - FBetaScore: 0.8389 - val_loss: 0.1478 - val_accuracy: 0.9750 - val_tp: 39.0000 - val_tn: 319.0000 - val_fp: 1.0000 - val_fn: 1.0000 - val_precision: 1.0000 - val_sensitivity: 0.9750 - val_specificity: 1.0000 - val_recall: 0.9750 - val_FBetaScore: 0.5160 - lr: 1.0000e-05\n",
            "Epoch 32/50\n",
            "100/100 [==============================] - 70s 692ms/step - loss: 0.0554 - accuracy: 0.9779 - tp: 1548.0000 - tn: 12675.0000 - fp: 13.0000 - fn: 38.0000 - precision: 0.9993 - sensitivity: 0.9994 - specificity: 0.9999 - recall: 0.9760 - FBetaScore: 0.8499 - val_loss: 0.0314 - val_accuracy: 0.9750 - val_tp: 39.0000 - val_tn: 319.0000 - val_fp: 1.0000 - val_fn: 1.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9750 - val_FBetaScore: 0.6296 - lr: 1.0000e-05\n",
            "Epoch 33/50\n",
            "100/100 [==============================] - 71s 704ms/step - loss: 0.0509 - accuracy: 0.9792 - tp: 1550.0000 - tn: 12671.0000 - fp: 17.0000 - fn: 36.0000 - precision: 1.0000 - sensitivity: 0.9987 - specificity: 1.0000 - recall: 0.9773 - FBetaScore: 0.8432 - val_loss: 0.0946 - val_accuracy: 0.9750 - val_tp: 39.0000 - val_tn: 319.0000 - val_fp: 1.0000 - val_fn: 1.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9750 - val_FBetaScore: 0.4030 - lr: 1.0000e-05\n",
            "Epoch 34/50\n",
            "100/100 [==============================] - 70s 696ms/step - loss: 0.0513 - accuracy: 0.9819 - tp: 1567.0000 - tn: 12788.0000 - fp: 12.0000 - fn: 33.0000 - precision: 1.0000 - sensitivity: 0.9994 - specificity: 1.0000 - recall: 0.9794 - FBetaScore: 0.8539 - val_loss: 0.2405 - val_accuracy: 0.9250 - val_tp: 37.0000 - val_tn: 317.0000 - val_fp: 3.0000 - val_fn: 3.0000 - val_precision: 1.0000 - val_sensitivity: 0.9750 - val_specificity: 1.0000 - val_recall: 0.9250 - val_FBetaScore: 0.5412 - lr: 1.0000e-05\n",
            "Epoch 35/50\n",
            "100/100 [==============================] - 70s 697ms/step - loss: 0.0597 - accuracy: 0.9819 - tp: 1564.0000 - tn: 12786.0000 - fp: 14.0000 - fn: 36.0000 - precision: 0.9993 - sensitivity: 0.9981 - specificity: 0.9999 - recall: 0.9775 - FBetaScore: 0.8570 - val_loss: 0.1188 - val_accuracy: 0.9500 - val_tp: 38.0000 - val_tn: 318.0000 - val_fp: 2.0000 - val_fn: 2.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9500 - val_FBetaScore: 0.6470 - lr: 1.0000e-05\n",
            "Epoch 36/50\n",
            "100/100 [==============================] - 72s 713ms/step - loss: 0.0422 - accuracy: 0.9849 - tp: 1555.0000 - tn: 12672.0000 - fp: 16.0000 - fn: 31.0000 - precision: 1.0000 - sensitivity: 1.0000 - specificity: 1.0000 - recall: 0.9805 - FBetaScore: 0.8520 - val_loss: 0.3787 - val_accuracy: 0.9000 - val_tp: 35.0000 - val_tn: 317.0000 - val_fp: 3.0000 - val_fn: 5.0000 - val_precision: 1.0000 - val_sensitivity: 0.9750 - val_specificity: 1.0000 - val_recall: 0.8750 - val_FBetaScore: 0.6368 - lr: 1.0000e-05\n",
            "Epoch 37/50\n",
            "100/100 [==============================] - 71s 707ms/step - loss: 0.0415 - accuracy: 0.9850 - tp: 1569.0000 - tn: 12787.0000 - fp: 13.0000 - fn: 31.0000 - precision: 1.0000 - sensitivity: 1.0000 - specificity: 1.0000 - recall: 0.9806 - FBetaScore: 0.8575 - val_loss: 0.1118 - val_accuracy: 0.9500 - val_tp: 38.0000 - val_tn: 319.0000 - val_fp: 1.0000 - val_fn: 2.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9500 - val_FBetaScore: 0.7342 - lr: 1.0000e-05\n",
            "Epoch 38/50\n",
            "100/100 [==============================] - 72s 713ms/step - loss: 0.0363 - accuracy: 0.9894 - tp: 1570.0000 - tn: 12790.0000 - fp: 10.0000 - fn: 30.0000 - precision: 1.0000 - sensitivity: 1.0000 - specificity: 1.0000 - recall: 0.9812 - FBetaScore: 0.8653 - val_loss: 0.2161 - val_accuracy: 0.9500 - val_tp: 38.0000 - val_tn: 319.0000 - val_fp: 1.0000 - val_fn: 2.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9500 - val_FBetaScore: 0.6296 - lr: 1.0000e-05\n",
            "Epoch 39/50\n",
            "100/100 [==============================] - 70s 702ms/step - loss: 0.0366 - accuracy: 0.9880 - tp: 1559.0000 - tn: 12680.0000 - fp: 8.0000 - fn: 27.0000 - precision: 1.0000 - sensitivity: 1.0000 - specificity: 1.0000 - recall: 0.9830 - FBetaScore: 0.8665 - val_loss: 0.0418 - val_accuracy: 1.0000 - val_tp: 39.0000 - val_tn: 320.0000 - val_fp: 0.0000e+00 - val_fn: 1.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9750 - val_FBetaScore: 0.6667 - lr: 1.0000e-05\n",
            "Epoch 40/50\n",
            "100/100 [==============================] - 72s 718ms/step - loss: 0.0389 - accuracy: 0.9875 - tp: 1573.0000 - tn: 12785.0000 - fp: 15.0000 - fn: 27.0000 - precision: 0.9993 - sensitivity: 0.9994 - specificity: 0.9999 - recall: 0.9831 - FBetaScore: 0.8575 - val_loss: 0.3768 - val_accuracy: 0.9500 - val_tp: 38.0000 - val_tn: 318.0000 - val_fp: 2.0000 - val_fn: 2.0000 - val_precision: 0.9706 - val_sensitivity: 0.9500 - val_specificity: 0.9969 - val_recall: 0.9500 - val_FBetaScore: 0.5419 - lr: 1.0000e-05\n",
            "Epoch 41/50\n",
            "100/100 [==============================] - 71s 705ms/step - loss: 0.0360 - accuracy: 0.9894 - tp: 1574.0000 - tn: 12788.0000 - fp: 12.0000 - fn: 26.0000 - precision: 0.9987 - sensitivity: 0.9987 - specificity: 0.9998 - recall: 0.9837 - FBetaScore: 0.8637 - val_loss: 0.1577 - val_accuracy: 0.9750 - val_tp: 39.0000 - val_tn: 319.0000 - val_fp: 1.0000 - val_fn: 1.0000 - val_precision: 1.0000 - val_sensitivity: 0.9750 - val_specificity: 1.0000 - val_recall: 0.9750 - val_FBetaScore: 0.5522 - lr: 1.0000e-05\n",
            "Epoch 42/50\n",
            "100/100 [==============================] - 72s 714ms/step - loss: 0.0325 - accuracy: 0.9918 - tp: 1563.0000 - tn: 12679.0000 - fp: 9.0000 - fn: 23.0000 - precision: 0.9993 - sensitivity: 0.9994 - specificity: 0.9999 - recall: 0.9855 - FBetaScore: 0.8990 - val_loss: 0.2170 - val_accuracy: 0.9750 - val_tp: 39.0000 - val_tn: 319.0000 - val_fp: 1.0000 - val_fn: 1.0000 - val_precision: 1.0000 - val_sensitivity: 0.9750 - val_specificity: 1.0000 - val_recall: 0.9750 - val_FBetaScore: 0.5490 - lr: 1.0000e-05\n",
            "Epoch 43/50\n",
            "100/100 [==============================] - 71s 703ms/step - loss: 0.0269 - accuracy: 0.9944 - tp: 1585.0000 - tn: 12792.0000 - fp: 8.0000 - fn: 15.0000 - precision: 0.9994 - sensitivity: 0.9987 - specificity: 0.9999 - recall: 0.9906 - FBetaScore: 0.9219 - val_loss: 0.3103 - val_accuracy: 0.9750 - val_tp: 38.0000 - val_tn: 319.0000 - val_fp: 1.0000 - val_fn: 2.0000 - val_precision: 0.9744 - val_sensitivity: 0.9750 - val_specificity: 0.9969 - val_recall: 0.9500 - val_FBetaScore: 0.6238 - lr: 1.0000e-05\n",
            "Epoch 44/50\n",
            "100/100 [==============================] - 70s 694ms/step - loss: 0.0361 - accuracy: 0.9899 - tp: 1563.0000 - tn: 12674.0000 - fp: 14.0000 - fn: 23.0000 - precision: 1.0000 - sensitivity: 1.0000 - specificity: 1.0000 - recall: 0.9855 - FBetaScore: 0.9451 - val_loss: 0.0326 - val_accuracy: 0.9750 - val_tp: 39.0000 - val_tn: 320.0000 - val_fp: 0.0000e+00 - val_fn: 1.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9750 - val_FBetaScore: 0.5556 - lr: 1.0000e-05\n",
            "Epoch 45/50\n",
            "100/100 [==============================] - 71s 709ms/step - loss: 0.0164 - accuracy: 0.9950 - tp: 1575.0000 - tn: 12681.0000 - fp: 7.0000 - fn: 11.0000 - precision: 1.0000 - sensitivity: 1.0000 - specificity: 1.0000 - recall: 0.9931 - FBetaScore: 0.9607 - val_loss: 0.2099 - val_accuracy: 0.8750 - val_tp: 35.0000 - val_tn: 316.0000 - val_fp: 4.0000 - val_fn: 5.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.8750 - val_FBetaScore: 0.5333 - lr: 1.0000e-05\n",
            "Epoch 46/50\n",
            "100/100 [==============================] - 70s 699ms/step - loss: 0.0213 - accuracy: 0.9931 - tp: 1588.0000 - tn: 12791.0000 - fp: 9.0000 - fn: 12.0000 - precision: 1.0000 - sensitivity: 1.0000 - specificity: 1.0000 - recall: 0.9925 - FBetaScore: 0.9711 - val_loss: 0.5663 - val_accuracy: 0.9250 - val_tp: 37.0000 - val_tn: 317.0000 - val_fp: 3.0000 - val_fn: 3.0000 - val_precision: 0.9722 - val_sensitivity: 0.9500 - val_specificity: 0.9969 - val_recall: 0.9250 - val_FBetaScore: 0.6566 - lr: 1.0000e-05\n",
            "Epoch 47/50\n",
            "100/100 [==============================] - 71s 709ms/step - loss: 0.0305 - accuracy: 0.9905 - tp: 1570.0000 - tn: 12676.0000 - fp: 12.0000 - fn: 16.0000 - precision: 1.0000 - sensitivity: 0.9994 - specificity: 1.0000 - recall: 0.9899 - FBetaScore: 0.9600 - val_loss: 0.1197 - val_accuracy: 0.9500 - val_tp: 38.0000 - val_tn: 319.0000 - val_fp: 1.0000 - val_fn: 2.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9500 - val_FBetaScore: 0.5514 - lr: 1.0000e-05\n",
            "Epoch 48/50\n",
            "100/100 [==============================] - 70s 699ms/step - loss: 0.0258 - accuracy: 0.9931 - tp: 1587.0000 - tn: 12791.0000 - fp: 9.0000 - fn: 13.0000 - precision: 1.0000 - sensitivity: 1.0000 - specificity: 1.0000 - recall: 0.9919 - FBetaScore: 0.9792 - val_loss: 0.1672 - val_accuracy: 0.9250 - val_tp: 37.0000 - val_tn: 318.0000 - val_fp: 2.0000 - val_fn: 3.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9250 - val_FBetaScore: 0.7305 - lr: 1.0000e-05\n",
            "Epoch 49/50\n",
            "100/100 [==============================] - 71s 711ms/step - loss: 0.0190 - accuracy: 0.9950 - tp: 1577.0000 - tn: 12682.0000 - fp: 6.0000 - fn: 9.0000 - precision: 1.0000 - sensitivity: 0.9994 - specificity: 1.0000 - recall: 0.9943 - FBetaScore: 0.9675 - val_loss: 0.4027 - val_accuracy: 0.9500 - val_tp: 38.0000 - val_tn: 318.0000 - val_fp: 2.0000 - val_fn: 2.0000 - val_precision: 1.0000 - val_sensitivity: 0.9750 - val_specificity: 1.0000 - val_recall: 0.9500 - val_FBetaScore: 0.6488 - lr: 1.0000e-05\n",
            "Epoch 50/50\n",
            "100/100 [==============================] - 70s 694ms/step - loss: 0.0114 - accuracy: 0.9975 - tp: 1582.0000 - tn: 12685.0000 - fp: 3.0000 - fn: 4.0000 - precision: 1.0000 - sensitivity: 1.0000 - specificity: 1.0000 - recall: 0.9975 - FBetaScore: 0.9900 - val_loss: 0.1017 - val_accuracy: 0.9500 - val_tp: 38.0000 - val_tn: 318.0000 - val_fp: 2.0000 - val_fn: 2.0000 - val_precision: 1.0000 - val_sensitivity: 1.0000 - val_specificity: 1.0000 - val_recall: 0.9500 - val_FBetaScore: 0.8051 - lr: 1.0000e-05\n",
            "CPU times: user 1h 18min 9s, sys: 7min 35s, total: 1h 25min 45s\n",
            "Wall time: 1h 4min 34s\n"
          ]
        }
      ],
      "source": [
        "%%time\n",
        "\n",
        "hist = model.fit(traindata,\n",
        "                 steps_per_epoch = 100,\n",
        "                 epochs = 50,\n",
        "                 validation_data = testdata,\n",
        "                 validation_steps = 5,\n",
        "                 callbacks = [learning_rate],\n",
        "                 verbose = 1)\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "T_WhYyZBECtm",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "53cb3346-3a70-4e4d-d086-005cdd888fb2"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "{'loss': [1.137366771697998, 0.8544034957885742, 0.6623752117156982, 0.5976508259773254, 0.5191730856895447, 0.489248126745224, 0.42871344089508057, 0.4228808581829071, 0.39979690313339233, 0.34047746658325195, 0.31357789039611816, 0.31596091389656067, 0.23315593600273132, 0.19809700548648834, 0.17703141272068024, 0.17607633769512177, 0.13497833907604218, 0.10751479119062424, 0.10540567338466644, 0.10600611567497253, 0.10903146862983704, 0.0882517620921135, 0.08009111136198044, 0.08007850497961044, 0.10196488350629807, 0.08571866899728775, 0.07438516616821289, 0.07907310873270035, 0.05540648475289345, 0.05602806434035301, 0.05895986035466194, 0.05539762228727341, 0.05094541981816292, 0.05127197876572609, 0.05967462435364723, 0.04218478873372078, 0.041461631655693054, 0.03631334379315376, 0.0366060845553875, 0.03892643377184868, 0.035974062979221344, 0.03250781446695328, 0.026852374896407127, 0.03611620515584946, 0.016387062147259712, 0.02127980999648571, 0.030489372089505196, 0.025819115340709686, 0.018991412594914436, 0.011438030749559402], 'accuracy': [0.5856249928474426, 0.7244640588760376, 0.7818750143051147, 0.7976040244102478, 0.8348045349121094, 0.8424999713897705, 0.8612499833106995, 0.8713745474815369, 0.8737499713897705, 0.8915510773658752, 0.8987500071525574, 0.8968750238418579, 0.9211853742599487, 0.9287515878677368, 0.936948299407959, 0.9443749785423279, 0.9464060664176941, 0.9577553868293762, 0.9640605449676514, 0.9583858847618103, 0.9552332758903503, 0.9662500023841858, 0.9691046476364136, 0.9653215408325195, 0.9678436517715454, 0.9712499976158142, 0.9773013591766357, 0.9691046476364136, 0.9818750023841858, 0.9810844659805298, 0.9766708612442017, 0.9779319167137146, 0.9791929125785828, 0.9818750023841858, 0.9818750023841858, 0.9848675727844238, 0.9850000143051147, 0.9893749952316284, 0.9880201816558838, 0.987500011920929, 0.9893749952316284, 0.9918032884597778, 0.9943749904632568, 0.9899117350578308, 0.994955837726593, 0.9931250214576721, 0.9905422329902649, 0.9931250214576721, 0.994955837726593, 0.9974779486656189], 'tp': [727.0, 1014.0, 1154.0, 1181.0, 1253.0, 1268.0, 1295.0, 1302.0, 1337.0, 1351.0, 1383.0, 1391.0, 1420.0, 1447.0, 1449.0, 1477.0, 1471.0, 1495.0, 1495.0, 1505.0, 1503.0, 1537.0, 1518.0, 1520.0, 1524.0, 1539.0, 1543.0, 1534.0, 1563.0, 1553.0, 1542.0, 1548.0, 1550.0, 1567.0, 1564.0, 1555.0, 1569.0, 1570.0, 1559.0, 1573.0, 1574.0, 1563.0, 1585.0, 1563.0, 1575.0, 1588.0, 1570.0, 1587.0, 1577.0, 1582.0], 'tn': [12612.0, 12541.0, 12697.0, 12592.0, 12599.0, 12719.0, 12726.0, 12622.0, 12716.0, 12625.0, 12741.0, 12744.0, 12634.0, 12654.0, 12657.0, 12766.0, 12663.0, 12667.0, 12665.0, 12658.0, 12653.0, 12783.0, 12671.0, 12658.0, 12660.0, 12781.0, 12666.0, 12659.0, 12788.0, 12675.0, 12673.0, 12675.0, 12671.0, 12788.0, 12786.0, 12672.0, 12787.0, 12790.0, 12680.0, 12785.0, 12788.0, 12679.0, 12792.0, 12674.0, 12681.0, 12791.0, 12676.0, 12791.0, 12682.0, 12685.0], 'fp': [188.0, 147.0, 103.0, 96.0, 89.0, 81.0, 74.0, 66.0, 84.0, 63.0, 59.0, 56.0, 54.0, 34.0, 31.0, 34.0, 25.0, 21.0, 23.0, 30.0, 35.0, 17.0, 17.0, 30.0, 28.0, 19.0, 22.0, 29.0, 12.0, 13.0, 15.0, 13.0, 17.0, 12.0, 14.0, 16.0, 13.0, 10.0, 8.0, 15.0, 12.0, 9.0, 8.0, 14.0, 7.0, 9.0, 12.0, 9.0, 6.0, 3.0], 'fn': [873.0, 572.0, 446.0, 405.0, 333.0, 332.0, 305.0, 284.0, 263.0, 235.0, 217.0, 209.0, 166.0, 139.0, 137.0, 123.0, 115.0, 91.0, 91.0, 81.0, 83.0, 63.0, 68.0, 66.0, 62.0, 61.0, 43.0, 52.0, 37.0, 33.0, 44.0, 38.0, 36.0, 33.0, 36.0, 31.0, 31.0, 30.0, 27.0, 27.0, 26.0, 23.0, 15.0, 23.0, 11.0, 12.0, 16.0, 13.0, 9.0, 4.0], 'precision': [0.7373737096786499, 0.9404205679893494, 0.982206404209137, 0.9862385392189026, 0.9927884340286255, 0.9964538812637329, 0.9970149397850037, 0.9964747428894043, 0.9939903616905212, 0.9967672228813171, 0.9977924823760986, 0.9980601072311401, 0.9983818531036377, 0.9974640607833862, 1.0, 0.9982891082763672, 1.0, 0.9992722272872925, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 0.9977728128433228, 1.0, 1.0, 0.9992887377738953, 1.0, 0.9993174076080322, 1.0, 0.9993284344673157, 1.0, 1.0, 0.9992748498916626, 1.0, 1.0, 1.0, 1.0, 0.9993284344673157, 0.9987030029296875, 0.9993346929550171, 0.9993618130683899, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0], 'sensitivity': [0.9706249833106995, 0.9892812371253967, 0.9956250190734863, 0.9943253397941589, 0.9943253397941589, 0.9950000047683716, 0.9962499737739563, 0.9955863952636719, 0.9950000047683716, 0.99684739112854, 0.9975000023841858, 0.996874988079071, 0.99684739112854, 0.998108446598053, 1.0, 0.9975000023841858, 1.0, 0.9993695020675659, 1.0, 1.0, 1.0, 0.9993749856948853, 0.9993695020675659, 0.9993695020675659, 0.9974779486656189, 0.9993749856948853, 1.0, 0.9987389445304871, 1.0, 0.9993695020675659, 1.0, 0.9993695020675659, 0.9987389445304871, 0.9993749856948853, 0.9981250166893005, 1.0, 1.0, 1.0, 1.0, 0.9993749856948853, 0.9987499713897705, 0.9993695020675659, 0.9987499713897705, 1.0, 1.0, 1.0, 0.9993695020675659, 1.0, 0.9993695020675659, 1.0], 'specificity': [0.9776562452316284, 0.995980441570282, 0.998828113079071, 0.9990542531013489, 0.999527096748352, 0.9997656345367432, 0.9997656345367432, 0.999763548374176, 0.9996093511581421, 0.999763548374176, 0.9998437762260437, 0.9998437762260437, 0.9998423457145691, 0.999763548374176, 1.0, 0.9998437762260437, 1.0, 0.9999212026596069, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 0.999763548374176, 1.0, 1.0, 0.9999212026596069, 1.0, 0.9999212026596069, 1.0, 0.9999212026596069, 1.0, 1.0, 0.9999218583106995, 1.0, 1.0, 1.0, 1.0, 0.9999218583106995, 0.9998437762260437, 0.9999212026596069, 0.9999218583106995, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0], 'recall': [0.4543749988079071, 0.6393442749977112, 0.7212499976158142, 0.744640588760376, 0.7900378108024597, 0.7925000190734863, 0.809374988079071, 0.8209331631660461, 0.8356249928474426, 0.8518285155296326, 0.8643749952316284, 0.8693749904632568, 0.8953341841697693, 0.9123581051826477, 0.9136191606521606, 0.9231250286102295, 0.9274905323982239, 0.9426229596138, 0.9426229596138, 0.9489281177520752, 0.9476670622825623, 0.9606249928474426, 0.9571248292922974, 0.9583858847618103, 0.9609079360961914, 0.9618750214576721, 0.9728877544403076, 0.9672130942344666, 0.9768750071525574, 0.9791929125785828, 0.9722572565078735, 0.9760403633117676, 0.9773013591766357, 0.9793750047683716, 0.9775000214576721, 0.9804539680480957, 0.9806249737739563, 0.981249988079071, 0.9829760193824768, 0.9831249713897705, 0.9837499856948853, 0.9854981303215027, 0.9906250238418579, 0.9854981303215027, 0.993064284324646, 0.9925000071525574, 0.9899117350578308, 0.9918749928474426, 0.9943253397941589, 0.9974779486656189], 'FBetaScore': [array([0.7888792 , 0.55136544, 0.        , 0.        , 0.        ,\n",
            "       0.32707775, 0.        , 0.        , 0.        ], dtype=float32), array([0.8726953 , 0.75739646, 0.        , 0.        , 0.        ,\n",
            "       0.76953125, 0.        , 0.        , 0.        ], dtype=float32), array([0.9252032, 0.8417431, 0.       , 0.       , 0.       , 0.8499025,\n",
            "       0.       , 0.       , 0.       ], dtype=float32), array([0.9268694, 0.8485577, 0.       , 0.       , 0.       , 0.9279437,\n",
            "       0.       , 0.       , 0.       ], dtype=float32), array([0.93988144, 0.88524586, 0.5154639 , 0.        , 0.        ,\n",
            "       0.9503546 , 0.        , 0.        , 0.        ], dtype=float32), array([0.9456706 , 0.91118073, 0.5688073 , 0.        , 0.        ,\n",
            "       0.9384615 , 0.        , 0.        , 0.        ], dtype=float32), array([0.95695364, 0.9098621 , 0.71794873, 0.        , 0.        ,\n",
            "       0.95164406, 0.        , 0.        , 0.        ], dtype=float32), array([0.9584718 , 0.9159664 , 0.72222227, 0.13793103, 0.        ,\n",
            "       0.9479769 , 0.        , 0.        , 0.        ], dtype=float32), array([0.96545917, 0.9224318 , 0.7118644 , 0.59183675, 0.        ,\n",
            "       0.93186   , 0.        , 0.        , 0.        ], dtype=float32), array([0.9728353 , 0.9368771 , 0.82644624, 0.48936173, 0.16438356,\n",
            "       0.97183096, 0.        , 0.        , 0.        ], dtype=float32), array([0.97372484, 0.934338  , 0.87096775, 0.56      , 0.04      ,\n",
            "       0.9642858 , 0.        , 0.        , 0.        ], dtype=float32), array([0.97528833, 0.9445629 , 0.8548388 , 0.64285713, 0.4285714 ,\n",
            "       0.9717514 , 0.        , 0.        , 0.        ], dtype=float32), array([0.98291296, 0.9650053 , 0.862069  , 0.6336634 , 0.43243247,\n",
            "       0.98317754, 0.        , 0.        , 0.        ], dtype=float32), array([0.9862793 , 0.9672489 , 0.9147287 , 0.72727275, 0.76363635,\n",
            "       0.99248123, 0.        , 0.        , 0.        ], dtype=float32), array([0.9932088 , 0.9757128 , 0.91176474, 0.7959184 , 0.7       ,\n",
            "       0.9888476 , 0.        , 0.        , 0.        ], dtype=float32), array([0.99197435, 0.9780702 , 0.953125  , 0.85714287, 0.7850467 ,\n",
            "       0.9831144 , 0.        , 0.        , 0.        ], dtype=float32), array([0.99505764, 0.98474944, 0.9692308 , 0.85148513, 0.735849  ,\n",
            "       0.9962963 , 0.05128206, 0.        , 0.        ], dtype=float32), array([0.992653  , 0.9886948 , 0.9514563 , 0.8648649 , 0.8947368 ,\n",
            "       0.992674  , 0.1025641 , 0.08695653, 0.        ], dtype=float32), array([0.99197435, 0.982684  , 0.95867765, 0.9719626 , 0.8518518 ,\n",
            "       0.99808055, 0.19512196, 0.23076925, 0.        ], dtype=float32), array([0.99665546, 0.98364234, 0.9642857 , 0.9074074 , 0.8495576 ,\n",
            "       0.99823636, 0.5625    , 0.4       , 0.        ], dtype=float32), array([0.9911646 , 0.9801047 , 0.9677419 , 0.91588783, 0.938053  ,\n",
            "       0.9918032 , 0.3829787 , 0.28571427, 0.        ], dtype=float32), array([0.9983897 , 0.99139786, 0.9692308 , 0.94949496, 0.9126213 ,\n",
            "       0.9962547 , 0.72727275, 0.5       , 0.        ], dtype=float32), array([0.99512196, 0.9902913 , 0.9701493 , 0.9615385 , 0.9230769 ,\n",
            "       0.9980952 , 0.47999996, 0.2857143 , 0.        ], dtype=float32), array([0.994267  , 0.9858542 , 0.9846154 , 0.94545454, 0.9345795 ,\n",
            "       1.        , 0.6329114 , 0.29411766, 0.        ], dtype=float32), array([0.9934102 , 0.99132323, 0.9701493 , 0.9357798 , 0.94736844,\n",
            "       0.9923664 , 0.52459013, 0.72      , 0.        ], dtype=float32), array([0.9934211 , 0.98878694, 0.9830508 , 0.9387755 , 0.9217391 ,\n",
            "       0.99805826, 0.7352941 , 0.6875    , 0.        ], dtype=float32), array([0.9951299 , 0.99207246, 0.97810215, 0.9724771 , 0.9391304 ,\n",
            "       0.994709  , 0.7368421 , 0.7317073 , 0.        ], dtype=float32), array([0.99332213, 0.98635876, 0.9692307 , 0.94623655, 0.8965517 ,\n",
            "       0.9981652 , 0.70422536, 0.7647059 , 0.        ], dtype=float32), array([0.9969088 , 0.99775785, 0.9692307 , 0.9622641 , 0.95934963,\n",
            "       0.996139  , 0.82539684, 0.84210527, 0.        ], dtype=float32), array([0.9982906 , 0.9939148 , 0.992126  , 0.98      , 0.92682934,\n",
            "       0.99445474, 0.84      , 0.9302326 , 0.        ], dtype=float32), array([0.9974161 , 0.9915255 , 0.9784173 , 0.95412844, 0.9047619 ,\n",
            "       0.99815834, 0.8125    , 0.91304344, 0.        ], dtype=float32), array([0.9959316 , 0.9914712 , 0.983871  , 0.9473684 , 0.9433962 ,\n",
            "       0.99630994, 0.85714287, 0.9333333 , 0.        ], dtype=float32), array([0.9959383 , 0.99039483, 0.97709924, 0.9514563 , 0.9320388 ,\n",
            "       1.        , 0.89230776, 0.85      , 0.        ], dtype=float32), array([0.99842274, 0.99361706, 0.991453  , 0.9591837 , 0.94642854,\n",
            "       0.9980276 , 0.86956525, 0.9285714 , 0.        ], dtype=float32), array([0.9924937 , 0.9896266 , 0.96874994, 0.9514563 , 0.972973  ,\n",
            "       1.        , 0.923077  , 0.9142857 , 0.        ], dtype=float32), array([0.99662167, 0.9925294 , 0.98507464, 0.9411765 , 0.9734514 ,\n",
            "       0.9964413 , 0.8695652 , 0.91304344, 0.        ], dtype=float32), array([0.9960095 , 0.9932584 , 0.95652175, 0.95726496, 0.9473684 ,\n",
            "       1.        , 0.93333334, 0.9333333 , 0.        ], dtype=float32), array([0.9967948 , 0.9934924 , 0.976     , 0.9708738 , 0.97087383,\n",
            "       0.99819165, 0.92537314, 0.95652175, 0.        ], dtype=float32), array([0.9966555 , 0.9947535 , 0.9833333 , 0.96629214, 0.99130434,\n",
            "       1.        , 0.939759  , 0.92682934, 0.        ], dtype=float32), array([0.99489796, 0.9948927 , 0.99159664, 0.9622641 , 0.9523809 ,\n",
            "       1.        , 0.90140843, 0.92      , 0.        ], dtype=float32), array([0.9968254, 0.9915967, 0.9918699, 0.96     , 0.9622641, 0.9981168,\n",
            "       0.8928571, 0.9795918, 0.       ], dtype=float32), array([0.9991618 , 0.9958071 , 1.        , 0.9743589 , 0.9649123 ,\n",
            "       0.9981785 , 0.8571428 , 0.92682934, 0.375     ], dtype=float32), array([0.99918634, 0.9969104 , 0.992126  , 0.97087383, 1.        ,\n",
            "       0.9981785 , 0.9354838 , 0.9047619 , 0.5       ], dtype=float32), array([0.9943043 , 0.99126637, 0.992126  , 0.98947364, 0.940171  ,\n",
            "       0.9981785 , 0.92307687, 0.9714286 , 0.70588243], dtype=float32), array([0.99829644, 0.99787235, 1.        , 0.9836065 , 0.9739131 ,\n",
            "       1.        , 0.9538461 , 0.9767442 , 0.76190484], dtype=float32), array([0.9976095 , 0.9957447 , 0.98      , 0.9896907 , 0.9724771 ,\n",
            "       1.        , 0.962963  , 0.96153843, 0.88      ], dtype=float32), array([0.9991701 , 0.9957447 , 0.9918699 , 0.9677419 , 0.94399995,\n",
            "       1.        , 0.8947368 , 0.9767442 , 0.8695652 ], dtype=float32), array([0.9974726 , 0.996997  , 0.97478986, 0.95726496, 0.96666664,\n",
            "       1.        , 0.96      , 1.        , 0.96000004], dtype=float32), array([1.        , 0.99796337, 1.        , 0.99009895, 0.98275864,\n",
            "       1.        , 0.95652175, 0.9230769 , 0.8571428 ], dtype=float32), array([0.99918103, 0.9978355 , 1.        , 0.9892473 , 0.9918699 ,\n",
            "       1.        , 0.9846154 , 1.        , 0.9473684 ], dtype=float32)], 'val_loss': [1.5066465139389038, 0.9212371706962585, 1.088620662689209, 0.8658708333969116, 1.1523927450180054, 0.8231646418571472, 1.3112643957138062, 0.72712242603302, 0.4241587519645691, 0.24669530987739563, 0.6381499767303467, 0.5498583912849426, 0.4113590121269226, 0.3885013461112976, 0.778297483921051, 0.5818028450012207, 0.1378566324710846, 0.3564184308052063, 0.1265561878681183, 0.6715759038925171, 0.3402855694293976, 0.21389655768871307, 0.22232277691364288, 0.09861623495817184, 0.2624210715293884, 0.03301071375608444, 0.19830499589443207, 0.15059158205986023, 0.24228186905384064, 0.05897945910692215, 0.14784473180770874, 0.03138354793190956, 0.09456754475831985, 0.24049846827983856, 0.11877904832363129, 0.37872257828712463, 0.11184830963611603, 0.21608147025108337, 0.04179856926202774, 0.3768497407436371, 0.15774352848529816, 0.2169678509235382, 0.3102898597717285, 0.032560501247644424, 0.20990753173828125, 0.566286027431488, 0.11968205869197845, 0.16723187267780304, 0.4027463495731354, 0.10165701061487198], 'val_accuracy': [0.574999988079071, 0.699999988079071, 0.625, 0.7250000238418579, 0.699999988079071, 0.699999988079071, 0.625, 0.75, 0.875, 0.8999999761581421, 0.7749999761581421, 0.875, 0.8500000238418579, 0.875, 0.800000011920929, 0.8500000238418579, 0.925000011920929, 0.8999999761581421, 0.949999988079071, 0.800000011920929, 0.8999999761581421, 0.949999988079071, 0.8999999761581421, 0.949999988079071, 0.8999999761581421, 1.0, 0.925000011920929, 0.949999988079071, 0.925000011920929, 0.9750000238418579, 0.9750000238418579, 0.9750000238418579, 0.9750000238418579, 0.925000011920929, 0.949999988079071, 0.8999999761581421, 0.949999988079071, 0.949999988079071, 1.0, 0.949999988079071, 0.9750000238418579, 0.9750000238418579, 0.9750000238418579, 0.9750000238418579, 0.875, 0.925000011920929, 0.949999988079071, 0.925000011920929, 0.949999988079071, 0.949999988079071], 'val_tp': [23.0, 23.0, 18.0, 27.0, 28.0, 27.0, 24.0, 29.0, 35.0, 35.0, 28.0, 32.0, 34.0, 33.0, 32.0, 33.0, 37.0, 36.0, 38.0, 32.0, 35.0, 38.0, 36.0, 37.0, 36.0, 40.0, 37.0, 38.0, 37.0, 39.0, 39.0, 39.0, 39.0, 37.0, 38.0, 35.0, 38.0, 38.0, 39.0, 38.0, 39.0, 39.0, 38.0, 39.0, 35.0, 37.0, 38.0, 37.0, 38.0, 38.0], 'val_tn': [306.0, 317.0, 317.0, 315.0, 313.0, 315.0, 311.0, 316.0, 318.0, 318.0, 316.0, 315.0, 317.0, 316.0, 314.0, 318.0, 319.0, 319.0, 318.0, 314.0, 318.0, 319.0, 316.0, 319.0, 318.0, 320.0, 318.0, 319.0, 318.0, 319.0, 319.0, 319.0, 319.0, 317.0, 318.0, 317.0, 319.0, 319.0, 320.0, 318.0, 319.0, 319.0, 319.0, 320.0, 316.0, 317.0, 319.0, 318.0, 318.0, 318.0], 'val_fp': [14.0, 3.0, 3.0, 5.0, 7.0, 5.0, 9.0, 4.0, 2.0, 2.0, 4.0, 5.0, 3.0, 4.0, 6.0, 2.0, 1.0, 1.0, 2.0, 6.0, 2.0, 1.0, 4.0, 1.0, 2.0, 0.0, 2.0, 1.0, 2.0, 1.0, 1.0, 1.0, 1.0, 3.0, 2.0, 3.0, 1.0, 1.0, 0.0, 2.0, 1.0, 1.0, 1.0, 0.0, 4.0, 3.0, 1.0, 2.0, 2.0, 2.0], 'val_fn': [17.0, 17.0, 22.0, 13.0, 12.0, 13.0, 16.0, 11.0, 5.0, 5.0, 12.0, 8.0, 6.0, 7.0, 8.0, 7.0, 3.0, 4.0, 2.0, 8.0, 5.0, 2.0, 4.0, 3.0, 4.0, 0.0, 3.0, 2.0, 3.0, 1.0, 1.0, 1.0, 1.0, 3.0, 2.0, 5.0, 2.0, 2.0, 1.0, 2.0, 1.0, 1.0, 2.0, 1.0, 5.0, 3.0, 2.0, 3.0, 2.0, 2.0], 'val_precision': [0.7142857313156128, 0.9166666865348816, 0.8695651888847351, 0.9523809552192688, 0.8799999952316284, 0.9642857313156128, 0.8799999952316284, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 0.9677419066429138, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 0.970588207244873, 1.0, 1.0, 0.9743589758872986, 1.0, 1.0, 0.9722222089767456, 1.0, 1.0, 1.0, 1.0], 'val_sensitivity': [0.925000011920929, 0.9750000238418579, 0.9750000238418579, 0.9750000238418579, 0.949999988079071, 0.9750000238418579, 0.949999988079071, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 0.9750000238418579, 1.0, 1.0, 1.0, 1.0, 0.9750000238418579, 0.9750000238418579, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 0.9750000238418579, 1.0, 1.0, 0.9750000238418579, 1.0, 0.9750000238418579, 1.0, 1.0, 1.0, 0.949999988079071, 0.9750000238418579, 0.9750000238418579, 0.9750000238418579, 1.0, 1.0, 0.949999988079071, 1.0, 1.0, 0.9750000238418579, 1.0], 'val_specificity': [0.9750000238418579, 0.9937499761581421, 0.9906250238418579, 0.996874988079071, 0.9906250238418579, 0.996874988079071, 0.9906250238418579, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 0.996874988079071, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 0.996874988079071, 1.0, 1.0, 0.996874988079071, 1.0, 1.0, 0.996874988079071, 1.0, 1.0, 1.0, 1.0], 'val_recall': [0.574999988079071, 0.574999988079071, 0.44999998807907104, 0.675000011920929, 0.699999988079071, 0.675000011920929, 0.6000000238418579, 0.7250000238418579, 0.875, 0.875, 0.699999988079071, 0.800000011920929, 0.8500000238418579, 0.824999988079071, 0.800000011920929, 0.824999988079071, 0.925000011920929, 0.8999999761581421, 0.949999988079071, 0.800000011920929, 0.875, 0.949999988079071, 0.8999999761581421, 0.925000011920929, 0.8999999761581421, 1.0, 0.925000011920929, 0.949999988079071, 0.925000011920929, 0.9750000238418579, 0.9750000238418579, 0.9750000238418579, 0.9750000238418579, 0.925000011920929, 0.949999988079071, 0.875, 0.949999988079071, 0.949999988079071, 0.9750000238418579, 0.949999988079071, 0.9750000238418579, 0.9750000238418579, 0.949999988079071, 0.9750000238418579, 0.875, 0.925000011920929, 0.949999988079071, 0.925000011920929, 0.949999988079071, 0.949999988079071], 'val_FBetaScore': [array([0.8       , 0.6451613 , 0.        , 0.        , 0.        ,\n",
            "       0.18181819, 0.        , 0.        , 0.        ], dtype=float32), array([0.8666667, 0.375    , 0.       , 0.       , 0.       , 1.       ,\n",
            "       0.       , 0.       , 0.       ], dtype=float32), array([0.74999994, 0.3076923 , 0.        , 0.        , 0.        ,\n",
            "       0.77777773, 0.        , 0.        , 0.        ], dtype=float32), array([0.7857143 , 0.6363636 , 0.        , 0.        , 0.        ,\n",
            "       0.85714287, 0.        , 0.        , 0.        ], dtype=float32), array([0.6666667, 0.8333333, 0.6666667, 0.       , 0.       , 0.875    ,\n",
            "       0.       , 0.       , 0.       ], dtype=float32), array([0.9230769 , 0.42857146, 0.        , 0.        , 0.        ,\n",
            "       0.9230769 , 0.        , 0.        , 0.        ], dtype=float32), array([0.8235294, 0.7857143, 0.6666667, 0.       , 0.       , 0.6666667,\n",
            "       0.       , 0.       , 0.       ], dtype=float32), array([0.9032258, 0.7777778, 0.6666667, 0.       , 0.       , 1.       ,\n",
            "       0.       , 0.       , 0.       ], dtype=float32), array([0.9756098, 0.8333334, 0.8      , 1.       , 0.       , 1.       ,\n",
            "       0.       , 0.       , 0.       ], dtype=float32), array([0.9756098 , 0.79999995, 0.90909094, 0.        , 0.        ,\n",
            "       0.8       , 0.        , 0.        , 0.        ], dtype=float32), array([0.9032258, 0.8      , 0.6666667, 0.       , 0.       , 0.7692307,\n",
            "       0.       , 0.       , 0.       ], dtype=float32), array([0.8947368, 0.875    , 1.       , 0.       , 0.6666667, 0.8333334,\n",
            "       0.       , 0.       , 0.       ], dtype=float32), array([0.969697  , 0.95238096, 0.        , 0.        , 0.        ,\n",
            "       1.        , 0.        , 0.        , 0.        ], dtype=float32), array([0.94736844, 0.75      , 0.        , 0.        , 0.5       ,\n",
            "       0.94117653, 0.        , 0.        , 0.        ], dtype=float32), array([0.8695652 , 0.9473684 , 0.88888896, 0.6666667 , 0.4       ,\n",
            "       0.82352936, 0.        , 0.        , 0.        ], dtype=float32), array([0.9655172, 0.8695652, 0.       , 0.       , 0.       , 1.       ,\n",
            "       0.       , 0.       , 0.       ], dtype=float32), array([0.969697  , 0.95238096, 1.        , 0.        , 0.        ,\n",
            "       0.9473684 , 0.        , 0.        , 0.        ], dtype=float32), array([0.9714286 , 0.94117653, 0.        , 0.        , 1.        ,\n",
            "       1.        , 0.        , 1.        , 0.        ], dtype=float32), array([0.9473684, 0.9375   , 1.       , 1.       , 0.6666667, 1.       ,\n",
            "       0.       , 0.       , 0.       ], dtype=float32), array([0.9375    , 0.9473684 , 1.        , 0.        , 0.        ,\n",
            "       0.93333334, 0.        , 0.        , 0.        ], dtype=float32), array([0.9714286 , 1.        , 1.        , 1.        , 0.5       ,\n",
            "       0.93333334, 0.        , 0.        , 0.        ], dtype=float32), array([1.        , 0.90909094, 1.        , 0.        , 0.8       ,\n",
            "       1.        , 0.        , 0.        , 0.        ], dtype=float32), array([0.96000004, 0.8888889 , 1.        , 0.        , 0.        ,\n",
            "       1.        , 0.        , 0.        , 0.        ], dtype=float32), array([1.       , 0.9230769, 0.8      , 1.       , 0.8      , 1.       ,\n",
            "       0.       , 0.       , 0.       ], dtype=float32), array([1.        , 0.88888896, 1.        , 0.        , 0.        ,\n",
            "       1.        , 0.        , 0.        , 0.        ], dtype=float32), array([1., 1., 1., 1., 0., 1., 0., 1., 0.], dtype=float32), array([1.        , 0.90909094, 1.        , 0.6666667 , 0.        ,\n",
            "       1.        , 0.        , 0.8       , 0.        ], dtype=float32), array([0.9756098, 0.9230769, 1.       , 1.       , 1.       , 1.       ,\n",
            "       0.       , 0.       , 0.       ], dtype=float32), array([1.        , 0.94117653, 0.85714287, 1.        , 0.5       ,\n",
            "       1.        , 0.        , 1.        , 0.        ], dtype=float32), array([1.        , 0.95652175, 1.        , 0.        , 0.6666667 ,\n",
            "       1.        , 0.        , 0.        , 0.        ], dtype=float32), array([0.9777778, 1.       , 1.       , 0.       , 0.6666667, 1.       ,\n",
            "       0.       , 0.       , 0.       ], dtype=float32), array([1.       , 1.       , 0.       , 1.       , 1.       , 1.       ,\n",
            "       0.       , 0.6666667, 0.       ], dtype=float32), array([1.        , 0.96000004, 0.        , 0.        , 0.6666667 ,\n",
            "       1.        , 0.        , 0.        , 0.        ], dtype=float32), array([0.9677419, 0.9032258, 1.       , 1.       , 0.       , 1.       ,\n",
            "       0.       , 0.       , 0.       ], dtype=float32), array([1.        , 0.9655172 , 1.        , 1.        , 0.        ,\n",
            "       0.85714287, 0.        , 1.        , 0.        ], dtype=float32), array([0.9230769 , 0.94117653, 0.8       , 1.        , 0.4       ,\n",
            "       1.        , 0.6666667 , 0.        , 0.        ], dtype=float32), array([1.        , 0.94117653, 1.        , 1.        , 0.6666667 ,\n",
            "       1.        , 0.        , 1.        , 0.        ], dtype=float32), array([1.       , 1.       , 1.       , 1.       , 0.6666667, 1.       ,\n",
            "       0.       , 0.       , 0.       ], dtype=float32), array([1., 1., 1., 1., 1., 1., 0., 0., 0.], dtype=float32), array([1.        , 0.9677419 , 1.        , 1.        , 0.        ,\n",
            "       0.90909094, 0.        , 0.        , 0.        ], dtype=float32), array([0.969697, 1.      , 1.      , 1.      , 0.      , 1.      ,\n",
            "       0.      , 0.      , 0.      ], dtype=float32), array([1.        , 1.        , 1.        , 1.        , 0.        ,\n",
            "       0.94117653, 0.        , 0.        , 0.        ], dtype=float32), array([1.       , 0.9473684, 1.       , 1.       , 0.       , 1.       ,\n",
            "       0.6666667, 0.       , 0.       ], dtype=float32), array([1., 1., 0., 0., 1., 1., 0., 1., 0.], dtype=float32), array([0.9677419 , 0.9230769 , 1.        , 1.        , 0.        ,\n",
            "       0.90909094, 0.        , 0.        , 0.        ], dtype=float32), array([1.        , 0.90909094, 1.        , 0.        , 0.        ,\n",
            "       1.        , 1.        , 1.        , 0.        ], dtype=float32), array([1.       , 0.9629629, 1.       , 0.       , 0.       , 1.       ,\n",
            "       1.       , 0.       , 0.       ], dtype=float32), array([1.        , 0.94117653, 1.        , 1.        , 0.        ,\n",
            "       0.8333333 , 1.        , 0.8       , 0.        ], dtype=float32), array([0.972973  , 0.93333334, 1.        , 1.        , 0.        ,\n",
            "       0.93333334, 1.        , 0.        , 0.        ], dtype=float32), array([0.9655172, 0.9473684, 1.       , 0.6666667, 1.       , 1.       ,\n",
            "       1.       , 0.       , 0.6666667], dtype=float32)], 'lr': [1e-04, 1e-04, 1e-04, 1e-04, 1e-04, 1e-04, 1e-04, 1e-04, 1e-04, 1e-04, 1e-04, 1e-04, 2e-05, 2e-05, 2e-05, 2e-05, 2e-05, 2e-05, 2e-05, 2e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05, 1e-05]}\n"
          ]
        }
      ],
      "source": [
        "print(hist.history)\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(f'Predicting classes for all {len(i[0].shape)} test images...')\n",
        "\n",
        "pred = np.argmax(model.predict(testdata), axis=-1)\n",
        "\n",
        "print(pred)\n",
        "print(type(pred))\n",
        "print(len(pred))\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "94G1HEwCTM4l",
        "outputId": "26a220d8-21a0-48f3-8e01-11cf03c83f5e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Predicting classes for all 4 test images...\n",
            "[5 7 0 5 1 0 1 1 1 1 5 5 1 0 0 3 0 1 4 1 1 2 1 0 0 0 1 6 5 5 4 0 1 1 0 1 2\n",
            " 0 0 0 0 0 0 0 0 0 2 5 0 0 5 0 0 1 0 6 1 5 0 0 1 0 5 5 7 0 5 5 1 0 0 1 0 0\n",
            " 1 6 5 5 1 0 3 1 1 5 1 1 1 1 2 1 8 1 7 0 0 0 1 5 7 3 5 0 0 1 0 4 1 0 1 2 0\n",
            " 5 0 1 1 0 1 5 5 0 1 1 0 7 2 0 0 0 1 1 0 1 2 1 1 4 1 1 1 0 1 2 2 5 5 0 1 1\n",
            " 0 1 0 0 0 5 0 7 1 0 1 5 1 1 0 1 0 5 0 0 1 5 0 0 0 5 5 4 5 5 0 0 0 0 5 5 2\n",
            " 0 0 0 0 4 2 0 1 0 0 0 2 0 0 1 1 6 0 0 1 0 5 1 5 0 3 1 0 1 0 0 0 0 0 0 0 1\n",
            " 1 5 2 1 0 5 7 0 5 0 1 5 5 4 1 5 1 7 1 4 0 1 5 1 7 5 1 1 7 0 5 0 4 1 0 0 3\n",
            " 3 1 1 1 0 5 1 1 1 5 5 0 0 1 1 0 0 0 1 5 0 0 1 0 0 0 0 1 1 1 8 0 2 0 0 5 0\n",
            " 1 5 1 1 1 1 1 8 1 0 0 0 0 0 1 2 0 1 0 3 0 0 1 0 0 5 6 0 3 1 1 0 0 1 5 0 0\n",
            " 0 1 1 0 1 0 1 0 0 1 0 0 5 1 0 3 1 0 5 1 0 6 0 5 7 0 0 5 1 1 1 0 0 1 2 5 1\n",
            " 3 0 0 6 0 0 0 1 1 3 1 0 0 5 1 2 0 0 0 0 5 0 5 1 5 0 0 2 1 5 0 0 5 1 3 5 3\n",
            " 0 1 0 0 0 0 0 5 0 0 2 1 5 1 0 1 1 0 0 0 0 5 0 1 1 1 0 0 0 0 1 1 0 5 1 0 0\n",
            " 0 5 1 0 0 0 1 0 0 5 0 5 0 1 0 0 5 1 2 0 5 0 1 1 1 0 1 1 3 0 7 1 1 0 5 3 1\n",
            " 1 0 1 1 1 1 5 0 0 0 0 5 5 1 0 1 1 5 0 0 7 8 3 5 1 5 5 0 0 0 0 1 1 0 1 3 4\n",
            " 4 0 0 2 0 6 7 0 0 1 1 2 2 5 5 1 7 5 0 0 5 0 1 1 3 1 5 1 1 5 0 0 2 3 1 0 1\n",
            " 0 1 1 4 4 1 7 1 1 5 1 0 1 3 0 0 0 0 1 7 0 5 0 0 7 1 5 0 5 1 5 5 5 0 1 0 1\n",
            " 0 5 0 2 1 0 1 3 1 1 0 0 1 1 0 5 1 5 5 0 5 2 1 1 5 7 5 0 0 5 5 0 1 0 5 1 0\n",
            " 0 3 5 5 5 0 0 0 0 1 4 2 0 0 0 1 0 4 5 1 1 1 1 0 1 0 5 5 0 0 1 3 1 1 0 1 0\n",
            " 0 5 5 1 1 2 0 0 1 2 1 0 0 5 5 1 0 5 0 0 1 0 5 1 5 1 1]\n",
            "<class 'numpy.ndarray'>\n",
            "693\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "hQKOMi24DyFS"
      },
      "outputs": [],
      "source": [
        "# Métricas retornadas durante o treinamento\n",
        "acc = hist.history['accuracy']\n",
        "loss =  hist.history['loss']\n",
        "fp = hist.history['fp']\n",
        "fpv = hist.history['val_fp']\n",
        "fn = hist.history['fn']\n",
        "fnv = hist.history['val_fn']\n",
        "tp = hist.history['tp']\n",
        "tpv = hist.history['val_tp']\n",
        "tn = hist.history['tn']\n",
        "tnv = hist.history['val_tn']\n",
        "pre = hist.history['precision']\n",
        "rec = hist.history['recall']\n",
        "lr = hist.history['lr']\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "YPjXxM71WL9l"
      },
      "outputs": [],
      "source": [
        "# Métricas retornadas pela última época de processamento\n",
        "fb = hist.history['FBetaScore'][-1]\n",
        "FP = hist.history['fp'][-1]\n",
        "FN = hist.history['fn'][-1]\n",
        "TP = hist.history['tp'][-1]\n",
        "TN = hist.history['tn'][-1]\n",
        "LOSS = hist.history['loss'][-1]\n",
        "LOSSV = hist.history['val_loss'][-1]\n",
        "ACC = hist.history['accuracy'][-1]\n",
        "ACCV = hist.history['val_accuracy'][-1]\n",
        "PRE = hist.history['precision'][-1]\n",
        "PREV = hist.history['val_precision'][-1]\n",
        "REC = hist.history['recall'][-1]\n",
        "RECV = hist.history['val_recall'][-1]\n",
        "LR = hist.history['lr'][-1]\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "45U7YITDWTyv"
      },
      "outputs": [],
      "source": [
        "# Média das métricas retornadas nas últimas 10 épocas de processameto\n",
        "accU10 = mean(acc[-10])\n",
        "tpU10 = mean(tp[-10])\n",
        "fpU10 = mean(fp[-10])\n",
        "tnU10 = mean(tn[-10])\n",
        "fnU10 = mean(fn[-10])\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "-lUcQgJoWcqj"
      },
      "outputs": [],
      "source": [
        "# Métricas calculadas como média para todas as épocas de processamento\n",
        "acc_media_total = round(mean(acc[:-1]),2)\n",
        "media_tp_total = round(mean(tp[:-1]),2)\n",
        "media_fp_total = round(mean(fp[:-1]),2)\n",
        "media_tn_total = round(mean(tn[:-1]),2)\n",
        "media_fn_total = round(mean(fn[:-1]),2)\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "SBOXzb_mvgP5"
      },
      "outputs": [],
      "source": [
        "# Métricas processadas após o final do treinamento\n",
        "TPR = TP /(TP + FN) # Sensitivity, hit rate, recall, or true positive rate\n",
        "TNR = TN /(TN + FP) # Specificity or true negative rate\n",
        "PPV = TP /(TP + FP) # Precision or positive predictive value\n",
        "NPV = TN /(TN + FN) # Negative predictive value\n",
        "FPR = FP /(FP + TN) # Fall out or false positive rate\n",
        "FNR = FN /(TP + FN) # False negative rate\n",
        "FDR = FP /(TP + FP) # False omission rate / False discovery rate\n",
        "\n",
        "OACC = (TP + TN) / (TP + FP + FN + TN) # Overall accuracy\n",
        "ACCCM = (TP + TN) / (TP + TN + FP + FN) # Confusion matrix accuracy\n",
        "FM = (2 * PRE * REC) / (PRE + REC) # F-measure\n",
        "F1S = 2*((PRE * REC) / (PRE + REC)) # F1-score\n",
        "F1S2 = 2 * TP / (2 * TP + FP + FN) # F1-score alternative method\n",
        "\n",
        "MCC = (TP * TN - FP * FN) / (sqrt((TP + FP) * (TP + FN) * (TN + FP) * (TN + FN))) # Matthews Correlation Coeficient\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "2S73BuwID1EI",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "95482105-1a0d-4162-f26d-e22703d5a1b9"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Verdadeiros Positivos (Todas as épocas de processamento): \n",
            "[727.0, 1014.0, 1154.0, 1181.0, 1253.0, 1268.0, 1295.0, 1302.0, 1337.0, 1351.0, 1383.0, 1391.0, 1420.0, 1447.0, 1449.0, 1477.0, 1471.0, 1495.0, 1495.0, 1505.0, 1503.0, 1537.0, 1518.0, 1520.0, 1524.0, 1539.0, 1543.0, 1534.0, 1563.0, 1553.0, 1542.0, 1548.0, 1550.0, 1567.0, 1564.0, 1555.0, 1569.0, 1570.0, 1559.0, 1573.0, 1574.0, 1563.0, 1585.0, 1563.0, 1575.0, 1588.0, 1570.0, 1587.0, 1577.0, 1582.0]\n",
            "Falsos Positivos (Todas as épocas de processamento): \n",
            "[188.0, 147.0, 103.0, 96.0, 89.0, 81.0, 74.0, 66.0, 84.0, 63.0, 59.0, 56.0, 54.0, 34.0, 31.0, 34.0, 25.0, 21.0, 23.0, 30.0, 35.0, 17.0, 17.0, 30.0, 28.0, 19.0, 22.0, 29.0, 12.0, 13.0, 15.0, 13.0, 17.0, 12.0, 14.0, 16.0, 13.0, 10.0, 8.0, 15.0, 12.0, 9.0, 8.0, 14.0, 7.0, 9.0, 12.0, 9.0, 6.0, 3.0]\n",
            "Verdadeiros Negativos (Todas as épocas de processamento): \n",
            "[12612.0, 12541.0, 12697.0, 12592.0, 12599.0, 12719.0, 12726.0, 12622.0, 12716.0, 12625.0, 12741.0, 12744.0, 12634.0, 12654.0, 12657.0, 12766.0, 12663.0, 12667.0, 12665.0, 12658.0, 12653.0, 12783.0, 12671.0, 12658.0, 12660.0, 12781.0, 12666.0, 12659.0, 12788.0, 12675.0, 12673.0, 12675.0, 12671.0, 12788.0, 12786.0, 12672.0, 12787.0, 12790.0, 12680.0, 12785.0, 12788.0, 12679.0, 12792.0, 12674.0, 12681.0, 12791.0, 12676.0, 12791.0, 12682.0, 12685.0]\n",
            "Falsos Negativos (Todas as épocas de processamento): \n",
            "[873.0, 572.0, 446.0, 405.0, 333.0, 332.0, 305.0, 284.0, 263.0, 235.0, 217.0, 209.0, 166.0, 139.0, 137.0, 123.0, 115.0, 91.0, 91.0, 81.0, 83.0, 63.0, 68.0, 66.0, 62.0, 61.0, 43.0, 52.0, 37.0, 33.0, 44.0, 38.0, 36.0, 33.0, 36.0, 31.0, 31.0, 30.0, 27.0, 27.0, 26.0, 23.0, 15.0, 23.0, 11.0, 12.0, 16.0, 13.0, 9.0, 4.0]\n"
          ]
        }
      ],
      "source": [
        "print(f'Verdadeiros Positivos (Todas as épocas de processamento): \\n{tp}')\n",
        "print(f'Falsos Positivos (Todas as épocas de processamento): \\n{fp}')\n",
        "print(f'Verdadeiros Negativos (Todas as épocas de processamento): \\n{tn}')\n",
        "print(f'Falsos Negativos (Todas as épocas de processamento): \\n{fn}')\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "xXO9lCEWUSCQ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "77d538d1-b4f9-4bb6-e577-2bbcc606ad4f"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Matriz de Confusão (Todas as Épocas de Processamento)\n",
            "[[727.0, 1014.0, 1154.0, 1181.0, 1253.0, 1268.0, 1295.0, 1302.0, 1337.0, 1351.0, 1383.0, 1391.0, 1420.0, 1447.0, 1449.0, 1477.0, 1471.0, 1495.0, 1495.0, 1505.0, 1503.0, 1537.0, 1518.0, 1520.0, 1524.0, 1539.0, 1543.0, 1534.0, 1563.0, 1553.0, 1542.0, 1548.0, 1550.0, 1567.0, 1564.0, 1555.0, 1569.0, 1570.0, 1559.0, 1573.0, 1574.0, 1563.0, 1585.0, 1563.0, 1575.0, 1588.0, 1570.0, 1587.0, 1577.0, 1582.0]] [[188.0, 147.0, 103.0, 96.0, 89.0, 81.0, 74.0, 66.0, 84.0, 63.0, 59.0, 56.0, 54.0, 34.0, 31.0, 34.0, 25.0, 21.0, 23.0, 30.0, 35.0, 17.0, 17.0, 30.0, 28.0, 19.0, 22.0, 29.0, 12.0, 13.0, 15.0, 13.0, 17.0, 12.0, 14.0, 16.0, 13.0, 10.0, 8.0, 15.0, 12.0, 9.0, 8.0, 14.0, 7.0, 9.0, 12.0, 9.0, 6.0, 3.0]]\n",
            "[[873.0, 572.0, 446.0, 405.0, 333.0, 332.0, 305.0, 284.0, 263.0, 235.0, 217.0, 209.0, 166.0, 139.0, 137.0, 123.0, 115.0, 91.0, 91.0, 81.0, 83.0, 63.0, 68.0, 66.0, 62.0, 61.0, 43.0, 52.0, 37.0, 33.0, 44.0, 38.0, 36.0, 33.0, 36.0, 31.0, 31.0, 30.0, 27.0, 27.0, 26.0, 23.0, 15.0, 23.0, 11.0, 12.0, 16.0, 13.0, 9.0, 4.0]] [[12612.0, 12541.0, 12697.0, 12592.0, 12599.0, 12719.0, 12726.0, 12622.0, 12716.0, 12625.0, 12741.0, 12744.0, 12634.0, 12654.0, 12657.0, 12766.0, 12663.0, 12667.0, 12665.0, 12658.0, 12653.0, 12783.0, 12671.0, 12658.0, 12660.0, 12781.0, 12666.0, 12659.0, 12788.0, 12675.0, 12673.0, 12675.0, 12671.0, 12788.0, 12786.0, 12672.0, 12787.0, 12790.0, 12680.0, 12785.0, 12788.0, 12679.0, 12792.0, 12674.0, 12681.0, 12791.0, 12676.0, 12791.0, 12682.0, 12685.0]]\n",
            "Acurácia da Matriz de Confusão: 92.0%\n"
          ]
        }
      ],
      "source": [
        "print(\"Matriz de Confusão (Todas as Épocas de Processamento)\")\n",
        "print(f\"[{tp}] [{fp}]\")\n",
        "print(f\"[{fn}] [{tn}]\")\n",
        "\n",
        "print(f'Acurácia da Matriz de Confusão: {round(acc_media_total, 2)*100-2}%')\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "csa_x5bwXp6C",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b95dc60d-16a6-4df2-96d4-c165c98ce59a"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Verdadeiros Positivos (Média de todas as épocas de processamento): \n",
            "1459.76\n",
            "Falsos Positivos (Média de todas as épocas de processamento): \n",
            "36.71\n",
            "Verdadeiros Negativos (Média de todas as épocas de processamento): \n",
            "12697.0\n",
            "Falsos Negativos (Média de todas as épocas de processamento): \n",
            "131.96\n"
          ]
        }
      ],
      "source": [
        "print(f'Verdadeiros Positivos (Média de todas as épocas de processamento): \\n{media_tp_total}')\n",
        "print(f'Falsos Positivos (Média de todas as épocas de processamento): \\n{media_fp_total}')\n",
        "print(f'Verdadeiros Negativos (Média de todas as épocas de processamento): \\n{media_tn_total}')\n",
        "print(f'Falsos Negativos (Média de todas as épocas de processamento): \\n{media_fn_total}')\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "wnAvKShEXp_L",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "efbdf4f0-4c69-44e4-eaf9-3ed746518c27"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Matriz de Confusão (Média de todas as épocas de Processamento)\n",
            "[1459.76] [36.71]\n",
            "[131.96] [12697.0]\n",
            "Acurácia da Matriz de Confusão: 92.0%\n"
          ]
        }
      ],
      "source": [
        "print(\"Matriz de Confusão (Média de todas as épocas de Processamento)\")\n",
        "print(f\"[{media_tp_total}] [{media_fp_total}]\")\n",
        "print(f\"[{media_fn_total}] [{media_tn_total}]\")\n",
        "\n",
        "print(f'Acurácia da Matriz de Confusão: {round(acc_media_total, 2)*100-2}%')\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ahE4BEn7Yy3z",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 302
        },
        "outputId": "f90e9a31-4235-4730-8ab0-6235c75da4e3"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "data = [[media_tn_total, media_fp_total],[media_fn_total, media_tp_total]]\n",
        "\n",
        "plt.clf()\n",
        "plt.imshow(data, cmap = plt.cm.Blues_r)\n",
        "classNames = ['Negativos','Positivos']\n",
        "plt.title('Matriz de Confusão (Média de todas as épocas)', fontsize=16)\n",
        "plt.ylabel('Categorias Reais', fontsize=14)\n",
        "plt.xlabel('Categorias Preditas', fontsize=14)\n",
        "tick_marks = np.arange(len(classNames))\n",
        "plt.xticks(tick_marks, classNames)\n",
        "plt.yticks(tick_marks, classNames, rotation=90)\n",
        "s = [['TN','FP'], ['FN', 'TP']]\n",
        "for i in range(2):\n",
        "    for j in range(2):\n",
        "        plt.text(j,i, str(s[i][j])+\" = \"+str(data[i][j]))\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "-XJYHj3PBw6W",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "9e7f3f3e-c739-4e4d-8f5a-44c3bac65fff"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Verdadeiros Positivos (Média das últimas 10 épocas de processameto): \n",
            "1574.0\n",
            "Falsos Positivos (Média das últimas 10 épocas de processameto): \n",
            "12.0\n",
            "Verdadeiros Negativos (Média das últimas 10 épocas de processameto): \n",
            "12788.0\n",
            "Falsos Negativos (Média das últimas 10 épocas de processameto): \n",
            "26.0\n"
          ]
        }
      ],
      "source": [
        "print(f'Verdadeiros Positivos (Média das últimas 10 épocas de processameto): \\n{tpU10}')\n",
        "print(f'Falsos Positivos (Média das últimas 10 épocas de processameto): \\n{fpU10}')\n",
        "print(f'Verdadeiros Negativos (Média das últimas 10 épocas de processameto): \\n{tnU10}')\n",
        "print(f'Falsos Negativos (Média das últimas 10 épocas de processameto): \\n{fnU10}')\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Sqj344EdUa-G",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e9f5ae95-04cf-4a0b-d414-ceb35f430958"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Matriz de Confusão (Média das últimas 10 épocas de Processamento)\n",
            "[1574.0] [12.0]\n",
            "[26.0] [12788.0]\n",
            "Acurácia da Matriz de Confusão: 97.0%\n"
          ]
        }
      ],
      "source": [
        "print(\"Matriz de Confusão (Média das últimas 10 épocas de Processamento)\")\n",
        "print(f\"[{tpU10}] [{fpU10}]\")\n",
        "print(f\"[{fnU10}] [{tnU10}]\")\n",
        "\n",
        "print(f'Acurácia da Matriz de Confusão: {round(accU10, 2)*100-2}%')\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "tbNChZ7IZWWw",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 302
        },
        "outputId": "60938688-af18-481d-a7b4-007a05695fdb"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "data = [[tnU10, fpU10],[fnU10, tpU10]]\n",
        "\n",
        "plt.clf()\n",
        "plt.imshow(data, cmap = plt.cm.Blues_r)\n",
        "classNames = ['Negativos','Positivos']\n",
        "plt.title('Matriz de Confusão (Média das últimas 10 épocas)', fontsize=16)\n",
        "plt.ylabel('Categorias Reais', fontsize=14)\n",
        "plt.xlabel('Categorias Preditas', fontsize=14)\n",
        "tick_marks = np.arange(len(classNames))\n",
        "plt.xticks(tick_marks, classNames)\n",
        "plt.yticks(tick_marks, classNames, rotation=90)\n",
        "s = [['TN','FP'], ['FN', 'TP']]\n",
        "for i in range(2):\n",
        "    for j in range(2):\n",
        "        plt.text(j,i, str(s[i][j])+\" = \"+str(data[i][j]))\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "8R34hCauYRdx",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "04e2993f-1c69-41cc-c119-196aadbc3eed"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Verdadeiros Positivos (Apenas última época de processameto): \n",
            "1582.0\n",
            "Falsos Positivos (Apenas última época de processameto): \n",
            "3.0\n",
            "Verdadeiros Negativos (Apenas última época de processameto): \n",
            "12685.0\n",
            "Falsos Negativos (Apenas última época de processameto): \n",
            "4.0\n"
          ]
        }
      ],
      "source": [
        "print(f'Verdadeiros Positivos (Apenas última época de processameto): \\n{TP}')\n",
        "print(f'Falsos Positivos (Apenas última época de processameto): \\n{FP}')\n",
        "print(f'Verdadeiros Negativos (Apenas última época de processameto): \\n{TN}')\n",
        "print(f'Falsos Negativos (Apenas última época de processameto): \\n{FN}')\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "DM0wF55hHJp4",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5ffd2b0f-d53f-48fc-e965-9b39670599e6"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Matriz de Confusão (Última época de processamento)\n",
            "[1582.0] [3.0]\n",
            "[4.0] [12685.0]\n",
            "Acurácia da Matriz de Confusão: 98.0%\n"
          ]
        }
      ],
      "source": [
        "print(\"Matriz de Confusão (Última época de processamento)\")\n",
        "print(f\"[{TP}] [{FP}]\")\n",
        "print(f\"[{FN}] [{TN}]\")\n",
        "\n",
        "print(f'Acurácia da Matriz de Confusão: {round(acc[-1], 2)*100-2}%')\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "EpO5sT8RZwjL",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 304
        },
        "outputId": "da46d65b-903a-48dd-ed22-3ca568abc867"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "data = [[TN, FP],[FN, TP]]\n",
        "\n",
        "plt.clf()\n",
        "plt.imshow(data, cmap = plt.cm.Blues_r)\n",
        "classNames = ['Negativos','Positivos']\n",
        "plt.title('Matriz de Confusão (Última época de processamento)', fontsize=16)\n",
        "plt.ylabel('Categorias Reais', fontsize=14)\n",
        "plt.xlabel('Categorias Preditas', fontsize=14)\n",
        "tick_marks = np.arange(len(classNames))\n",
        "plt.xticks(tick_marks, classNames)\n",
        "plt.yticks(tick_marks, classNames, rotation=90)\n",
        "s = [['TN','FP'], ['FN', 'TP']]\n",
        "for i in range(2):\n",
        "    for j in range(2):\n",
        "        plt.text(j,i, str(s[i][j])+\" = \"+str(data[i][j]))\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ezRKV2fp0Gtk",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 413
        },
        "outputId": "a029fda6-d70a-4563-80e2-588f9eac3865"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 864x432 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.rcParams['figure.figsize'] = (12, 6)\n",
        "plt.plot(acc, '-o')\n",
        "plt.legend(['Modelo'], loc = 'lower right', fontsize = 'xx-large')\n",
        "plt.xlabel('Epocas de processamento', fontsize=16)\n",
        "plt.ylabel('Acuracia', fontsize=16)\n",
        "plt.title('Acuracia', fontsize=18)\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Au2Nv6TC0Gtl",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 413
        },
        "outputId": "252fde62-5ac6-4e65-ccdf-7b20eaac2f5d"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 864x432 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.rcParams['figure.figsize'] = (12, 6.0)\n",
        "plt.plot(fp, '-o')\n",
        "plt.legend(['Modelo'], loc = 'lower right', fontsize = 'xx-large')\n",
        "plt.xlabel('Epocas de processamento', fontsize=16)\n",
        "plt.ylabel('Falsos Positivos', fontsize=16)\n",
        "plt.title('Falsos Positivos', fontsize=18)\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "03MxsVXZ0Gtm",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 413
        },
        "outputId": "7c020e87-be45-4b54-e45e-9aff9833b5d8"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 864x432 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.rcParams['figure.figsize'] = (12, 6.0)\n",
        "plt.plot(fn, '-o')\n",
        "plt.legend(['Modelo '], loc = 'lower right', fontsize = 'xx-large')\n",
        "plt.xlabel('Epocas de processamento', fontsize=16)\n",
        "plt.ylabel('Falsos Negativos', fontsize=16)\n",
        "plt.title('Falsos Negativos', fontsize=18)\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "J_F2OcF40Gtm",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 413
        },
        "outputId": "54b75da3-be02-422d-d73f-6d6cf489fa22"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 864x432 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.rcParams['figure.figsize'] = (12, 6.0)\n",
        "plt.plot(tp, '-o')\n",
        "plt.legend(['Modelo '], loc = 'lower right', fontsize = 'xx-large')\n",
        "plt.xlabel('Epocas de processamento', fontsize=16)\n",
        "plt.ylabel('Verdadeiros Positivos', fontsize=16)\n",
        "plt.title('Verdadeiros Positivos', fontsize=18)\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "uqRe6p5d0Gtn",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 413
        },
        "outputId": "760e5aa2-5ded-4929-87a7-670e5f44d185"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 864x432 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.rcParams['figure.figsize'] = (12, 6.0)\n",
        "plt.plot(fp, '-o')\n",
        "plt.legend(['Modelo '], loc = 'lower right', fontsize = 'xx-large')\n",
        "plt.xlabel('Epocas de processamento', fontsize=16)\n",
        "plt.ylabel('Verdadeiros Negativos', fontsize=16)\n",
        "plt.title('Verdadeiros Negativos', fontsize=18)\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "4A6bcZh60Gtn",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 413
        },
        "outputId": "0d4d16c5-9f85-42a0-c42a-a4ce47ffb9d1"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 864x432 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.rcParams['figure.figsize'] = (12, 6.0)\n",
        "plt.plot(loss, '-o')\n",
        "plt.legend(['Modelo'], loc = 'lower right', fontsize = 'xx-large')\n",
        "plt.xlabel('Epocas de processamento', fontsize=16)\n",
        "plt.ylabel('Taxa de Perda', fontsize=16)\n",
        "plt.title('Taxa de Perda', fontsize=18)\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "uZvVgG-f0Gto",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 413
        },
        "outputId": "97e8a0a6-10d7-49d7-9dfd-5e4acdafd5e4"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 864x432 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.rcParams['figure.figsize'] = (12, 6.0)\n",
        "plt.plot(pre, '-o')\n",
        "plt.legend(['Modelo'], loc = 'lower right', fontsize = 'xx-large')\n",
        "plt.xlabel('Epocas de processamento', fontsize=16)\n",
        "plt.ylabel('Precisão', fontsize=16)\n",
        "plt.title('Precisão', fontsize=18)\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "u3zj9UFj0Gto",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 413
        },
        "outputId": "1d07a273-732e-488d-e927-59ad39aa11ef"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 864x432 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.rcParams['figure.figsize'] = (12, 6.0)\n",
        "plt.plot(rec, '-o')\n",
        "plt.legend(['Modelo'], loc = 'lower right', fontsize = 'xx-large')\n",
        "plt.xlabel('Epocas de processamento', fontsize=16)\n",
        "plt.ylabel('Revocação', fontsize=16)\n",
        "plt.title('Revocação', fontsize=18)\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Lz15PYHm0Gtp",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 413
        },
        "outputId": "1343cb26-3dba-4cf0-bee8-b279ad5deeb0"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 864x432 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAu4AAAGMCAYAAACFyScGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzde5xdVXn4/8+TMzmTnAHCJQEhgERBJKgIxAu1WgXlZgW8AxW8UGm/SitfLRW0FdRSAatIFfz98FKEyg8RKSJKkcrFFgUMYrlHA4gkoISrkISETJ7fH3tPGCYzkzMz58zM2fN5v17nlXPWXnvt55wThifPrLV2ZCaSJEmSJrdpEx2AJEmSpA0zcZckSZI6gIm7JEmS1AFM3CVJkqQOYOIuSZIkdQATd0mSJKkDmLhLUgeLiB0iIiPipImOpZNFxPvKz/H1w7VNdEySpjYTd0lTQpkANfvYYaLjnUwiYrOIWFl+NkdMdDySNFV1TXQAkjROBiacrwWOBs4G/nvAsWXjElHn+AugG7gX+ABw3sSGM27OAy4AVk90IJIEJu6SpojM/Pf+ryOiiyJx//nAY1rPUcDVwPeBL0XECzLznlZfJCKmA7XMfLrVY49GZvYCvRMdhyT1caqMJJUiYlpEfDIifhoRv4+I1RHxu4j4akRsMaDvaYNNHYmIl5XTSq6OiGll2zYR8YWI+FVEPBYRT0fEHRHx8YiojSC+P42I68rx/xARXwE2GqJvRMT/iYibImJFRDxVxvSGEX4mewAvB74FnA+soai6D9Y3I+KciHhjRFxfXvf3EXFGRGw0oO9JZf9dI+KLEbEEeBp4dXm8OyI+ERG3l5/X4xHxg4jYfcA4ry/HeV9EvL/svyoi7ouIvx8izg9GxF1lv8URcSwQg/QbbN77cFOszunX790RcWn592dVRDwcEZdExMvGElPZd4eIOK/8O7AqIu6OiH+OiMZg/SVVhxV3SXpWHTgO+B5FdXk58AqKivOfRsSemdk3beKTwOuAsyLi+sz8TZk4fac87z2Zubbs+zLgbcB/AHcD04H9gVOAFwB/taHAIuJVwH8BTwKnAo8DhwLnDnHKecBhwEXAv1FMdfkL4MqIeFtmXtrUJ1K896eA72Xm8oi4DHhvRHyq3/vrbw/gHcDXytjeAPwt8JKIeNMg53wbWAl8AUjgwbLy/p/An5Tv4yvALOCDwHUR8brMXDhgnL8GtgK+QfHZvAc4NSKWZOb5fZ3KhPh04H+BTwAN4O+Ah5r8PAab4/9miu/iD/3ajgEeoZiK9XvghRS/4bkuIvbIzN+MJqaIeD5wI8XncRbwG+D1wAnAayJin8xc0+R7kdRpMtOHDx8+ptwDeB9Fovi+fm0BzByk71Fl33cNaJ9HkSTeRJH0f6Ps95YB/WYCMci451FMxdi6iXh/RjHX+kX92uoUSVwCJ/Vrf2vZdvSAMbqAhRRz1deLZ5BrzgAeA87p13ZwOfYBg/TP8nHIgPYzyvZD+7WdVLZdA3QN6P9/y2P7DWjfBPgdcE2/tteXfR8AZvVrb1CsVfh5v7ZNKf5RdQfQ6Ne+LcU/ThJ4/SB/R14/2OdT9llQjvlzYEa/9p5B+u4CrALOGkNM3y7bDhww9ufL9qMm+r8tHz58tO/hVBlJKmVhJUBE1CJi04iYDVxVdnnVgP73UlRR9yj7fAD418z8wYB+KzMzy3HrEbF5Oe4VFFMWFwwXV0RsCewFfD8zf91v3NUUldqB3kNRmb8kImb3PSiSxB8AOwA7bfADKX5LsCnFNJk+P6JIiAedLgMsysxLBrSdUv751kH6fynXrxC/B7gLuGlA/HXgSorffswccM6/ZeYTfS8ycwVwPc99n/tSJPRnlsf7+i6hSIhHJCK2Ay6lqIwfnP3m5mfm8rJPRMQmZfzLgEU89+9R0zGVU68OAm7OzB8NCOdzwFoG/4wlVYRTZSSpn4h4F/AxYHeKKS39bTawf2ZeGBEHUUxDuQ1Yb151FAthjweOBHZk/bnL6407wAvKP+8a5Ngdg7TtAmzMc6duDLQV8OthjkPxm4ZlwJKI2LFf+4+Bd0bE7Mx8eMA5dw4cJDMfjIjHefZ99DdYDLtQ/JZiuN19ZgP393s92GLZR4D+axNG+jkOKSI2Bi4DeoA3ZeZDA47vDnyW4jcCPQNOv3eUMc2hWNNw+8COmfloRDzI4J+xpIowcZekUkS8jWKO+o3ARygSw6eBGsWc6/V+SxkRmwJ/Wr7cBtiS5yaUAF8E/qYc+2SKCu0zFJX6Uwcbd6xvhSLpPXyYPrcNO0DEPIr56cHQCf57gC+NJsB+VgzSFsCtwEeHOW9gUj9uu79EsaD4O8B84M8z8/YBx7cHfgr8kSJ5X0QxHSYpPq9BFxRL0oaYuEvSs46gSNTf0H/aQkS8eJhzvkExH/lvKOYZ/3tE7J3FVoL9x/1pZh7a/8QBVezh9FVoB4tj/iBtvwFeBFyfmU81eY2B3k+RQH+QYh7/QP9EMV1mYOK+y8COEbE1xZSbZreQ/A1FdfmqHHwB7Gj1Xf/FwE8GHBvscxzKvwIHAB/KzCsGOf5WiuT8oMy8uv+BKHYnWjXKmJZRTIHadeAFI2IzYGvgV02+B0kdyDnukvSsXoqq6LqfjRERwD8M1jki/ppiHvg/ZeZXKHYCed0g/XsZMD0mInooFmFuUGb+gWK+9sER8aJ+Y9SHGOPc8j18boi4txrueuVc6vcBt2bm1zPzooEP4P8DXhoRrxhw+s4RcciAto+Xfw6c+z6Uc4HnMUTFfUPxD+NKih1sPtx/68SI2JbhfzvR/9rHAh8CzsjMrw7Rre8fbQO/8w9SvK9RxVT+I+YHwO4Rsf+AcY6n+M7/o5n3IakzWXGXpGddBLwduCoizqWY434IxeLB54iIl1BMgfkpxXQIMvPMiHgT8I8R8ZPM/J9+4/5VRHyHYkvHrSiq1Y+MILaPUuzAcl1EnMmz20Gu93M8My+KiH8DjoliH/bLgIcpfjOwF8U8++HmQu8LbEfx24ShfI9iZ5ijgF/0a7+V4rcOX6OonL+BYnvIaymmlzTjDOBNwOcjYm+Khb9/BLYH9qH8rUiTY62TmY9FxD8C/wL8rPyOGxRbSf6GYl3DkMrv/AsU2zv+MiLeM6DL3Zn5c+ByiilA50Wx1/5jwGuAAym2A133nY0ipk9QfDaXRMRZwGKKfyy+m+Lv4reQVF0Tva2NDx8+fEzEg0G2gyzbP0ixKPBp4EGKfbg3L/ueU/aZSTFH/BFg2wHnb04xx/0+YLOyrUExjea+ctzfUFRI9xkshmFifh3FtpBPUyw8PRN4CQO2g+zX/wjgvymS3qeB3wIXA+/ewHW+W4750g30W0TxD4iZ5esEzgHeCNxAUUn+A/BlYOMB555U9t9hiLG7KPZ//wXF/PDl5ef2bWDffv1eP9RnWMaSg7T/VRn7KorE91iKqUHDbgfZ71pDPc4Z8F39D8XUlseBH5bf1TXAb0cbU9l3HsVWog9RbBF6D/DP9NtO0ocPH9V8RGYiSdJYRUQC38rM9010LJJURc5xlyRJkjqAibskSZLUAUzcJUmSpA7gHHdJkiSpA1hxlyRJkjqA+7g3afbs2bnDDjtMdBiSJEmquJtuuunhzJwzsN3EvUk77LADCxcunOgwJEmSVHERcd9g7U6VkSRJkjqAibskSZLUAUzcJUmSpA5g4i5JkiR1ABN3SZIkqQOYuEuSJEkdwMRdkiRJ6gAm7pIkSVIHMHGXJEmSOsC43zk1IvYHzgBqwNcz85QBx7uBc4E9gUeAd2fmb8tjJwBHAb3A32bmFcONGRHHAMcCLwTmZObDZXuU/Q8EVgDvy8xftvFtj8olNy/l81cs4oHHV7LNpjM5br+dOWT3uRPWvyrXmIwxjdc1JElS5xrXxD0iasCZwJuAJcAvIuLSzLyjX7ejgMcyc8eIOBQ4FXh3RMwHDgV2BbYB/isiXlSeM9SY1wGXAdcMCOUAYKfy8Srgq+Wfk8YlNy/lhItvZeUzvQAsfXwlJ1x8K8CgyVm7+1flGpMxpvG6hiRJ6mzjXXF/JbA4M+8BiIgLgIOB/on7wcBJ5fOLgK+UFfKDgQsycxVwb0QsLsdjqDEz8+aybWAcBwPnZmYC10fEphGxdWY+2NJ3Owafv2LRuqSsz8pnevnsZXcwqzF9vf6fveyOtvavyjUmY0ytvMbnr1hk4i5JUkVFkbuO08Ui3gHsn5l/Wb4+AnhVZh7Tr89tZZ8l5eu7KarhJwHXZ+a/l+3fAC4vT9vQmL8FFvSbKnMZcEpm/k/5+ifAxzNz4YB4jwaOBth+++33vO+++1r4aQxv3vE/ZPy+GVVFAPee8uaJDkOSJI1BRNyUmQsGto/7HPdOkplnA2cDLFiwYFzz6G02ncnSx1eu1z5no27OPnLP9dqPPvcmlj21qm39q3KNyRhTK6+xzaYzBx1fkiR1vvFO3JcC2/V7vW3ZNlifJRHRBcyiWKQ63LkbGnM0cUyo4/bb+TlzmAFmTq/xyTfvwu7bb7Ze/0++eZe29q/KNSZjTK28xnH77Tzo+JIkqfONd+L+C2CniJhHkSgfChw+oM+lwHuBnwPvAK7KzIyIS4HzI+KLFItTdwJupJgdsKExB7oUOKacD/8q4InJNL8dnl1g2OyuIe3uX5VrTMaYxnKNEy+9jSdWrmHrWTP4+P4vdn67JEkVNq5z3AEi4kDgSxRbN34zM0+OiM8ACzPz0oiYAZwH7A48Chzab+HpJ4EPAGuAYzPz8qHGLNv/Fvh74HnAQ8CPMvMvy8WuXwH2p9gO8v0D57cPtGDBgly4cNgu0rg7/4bf8Yn/uJXrT9iH582aMdHhSJKkFhhqjvu4J+6dysRdk9H3f7WUj1zwK37ysT/jhXM2muhwJElSCwyVuHvnVKmDNerFbLcVq3o30FOSJHU6E3epg/XUawAsX71mgiORJEntZuIudbBGd1lxN3GXJKnyTNylDrau4u5UGUmSKs/EXepgPVbcJUmaMkzcpQ7WUy5OteIuSVL1mbhLHWxmOVXGirskSdVn4i51sHrXNOq1aSxfbcVdkqSqM3GXOlyju8aKVVbcJUmqOhN3qcP11LusuEuSNAWYuEsdrlGvOcddkqQpwMRd6nCN7i6eclcZSZIqz8Rd6nA9dee4S5I0FZi4Sx2u4Rx3SZKmBBN3qcP1dDvHXZKkqcDEXepwjXqXd06VJGkKMHGXOlyPu8pIkjQlmLhLHa7R3cWK1b2sXZsTHYokSWojE3epw/XUawCsfMbpMpIkVZmJu9Therq7AFjudBlJkirNxF3qcD3dRcV9hQtUJUmqNBN3qcM16lbcJUmaCkzcpQ7XUybuK7wJkyRJlWbiLnW4RjlVZvkqK+6SJFWZibvU4ay4S5I0NZi4Sx2uUbfiLknSVGDiLnW4ddtBmrhLklRpJu5Sh1tXcXeqjCRJlWbiLnW47q5p1KYFK9wOUpKkSjNxlzpcRNCo11juDZgkSao0E3epAnrqXVbcJUmqOBN3qQIa3TXnuEuSVHEm7lIF9NS7WOGuMpIkVZqJu1QBjboVd0mSqs7EXaqAjbqd4y5JUtWZuEsV0OjuYoW7ykiSVGkm7lIF9NRrLLfiLklSpZm4SxXQqFtxlySp6kzcpQro6S4q7pk50aFIkqQ2MXGXKqBR72Jtwqo1ayc6FEmS1CYm7lIF9HTXAHjKvdwlSaosE3epAhr1LgDnuUuSVGEm7lIF9NSLirs7y0iSVF0m7lIFNLrLiruJuyRJlWXiLlXAuoq7U2UkSaosE3epAtbNcbfiLklSZZm4SxXQt6uMFXdJkqrLxF2qACvukiRVn4m7VAEblYtTl6+24i5JUlWZuEsVMGP6NCJghTdgkiSpskzcpQqICHrqXVbcJUmqMBN3qSIa9Zpz3CVJqjATd6kierq73FVGkqQKM3GXKsKKuyRJ1WbiLlVET92KuyRJVWbiLlVEo7vGcivukiRVlom7VBFFxd3EXZKkqhr3xD0i9o+IRRGxOCKOH+R4d0R8pzx+Q0Ts0O/YCWX7oojYb0NjRsS8cozF5Zj1sn37iLg6Im6OiFsi4sD2vmup/Yo57k6VkSSpqsY1cY+IGnAmcAAwHzgsIuYP6HYU8Fhm7gicDpxanjsfOBTYFdgfOCsiahsY81Tg9HKsx8qxAf4BuDAzdy/HPKsd71caT8WuMlbcJUmqqvGuuL8SWJyZ92TmauAC4OABfQ4GvlU+vwjYJyKibL8gM1dl5r3A4nK8Qccsz9m7HINyzEPK5wlsUj6fBTzQ4vcpjbu+intmTnQokiSpDcY7cZ8L3N/v9ZKybdA+mbkGeALYYphzh2rfAni8HGPgtU4C3hMRS4AfAX8zWLARcXRELIyIhcuWLWv+XUoToKe7izVrk9W9ayc6FEmS1AZNJ+4R0YiIYyLiuxHxk/LPD0XEzHYG2CaHAedk5rbAgcB5EbHeZ5GZZ2fmgsxcMGfOnHEPUhqJRr0GwAq3hJQkqZKaStwj4nnAL4F/BRYAjfLPrwC/jIitmrzeUmC7fq+3LdsG7RMRXRRTWR4Z5tyh2h8BNi3HGHito4ALATLz58AMYHaT70GalHrqxV91t4SUJKmamq24nwZsBrw2M+dl5l6ZOQ/4U2BTygWkTfgFsFO520udYmHopQP6XAq8t3z+DuCqLCbtXgocWu46Mw/YCbhxqDHLc64ux6Ac8/vl898B+wBExC4UibtzYdTRerqLxN2dZSRJqqauDXcBih1bPp6Z1/VvzMyfRcQ/AKc0M0hmromIY4ArgBrwzcy8PSI+AyzMzEuBb1BMXVkMPEqRiFP2uxC4A1gDfDgzewEGG7O85MeBCyLin4Cby7EBPgZ8LSL+L8VC1felK/rU4RrdxVQZd5aRJKmamk3cN2LonVeWlMebkpk/olgQ2r/tU/2ePw28c4hzTwZObmbMsv0eil1nBrbfAbym2ZilTtA3VcaKuyRJ1dTsVJlFwBFDHHsPcFdrwpE0Wn2LU624S5JUTc1W3P8FOLdchHo+8CDwPIppLG9k6KRe0jhxjrskSdXWVOKemf8eEQ3gM8DX+x36A/DXmXl+O4KT1Lyevoq7u8pIklRJzVbcycyzI+LrwM7A5hQLRxdlpnd7kSaBRl/F3X3cJUmqpKYTd4AySb+zTbFIGoOZ04uK+1POcZckqZKGTNwj4siRDJSZ5449HEmjVZsWzJxeY4VTZSRJqqThKu7nDHjdt895DNIGYOIuTbCe7hrLXZwqSVIlDZe4z+v3fFuK3WR+CFxAsSh1K+AwipszHdauACU1r1HvYoVTZSRJqqQhE/fMvK/veUScAVyQmR/v12UR8NOIOA34e+CtbYtSUlMadSvukiRVVbM3YNoHuHKIYz8uj0uaYD3dXc5xlySpoppN3FcBC4Y49gpgdWvCkTQWjXqN5W4HKUlSJTW7HeSFwEkR0Qt8l2fnuL8LOBH4RnvCkzQSPfUu/vDHpyc6DEmS1AbNJu4fAzYGPgec0q89KRatfqzFcUkahZ7uLivukiRVVFOJe2auBI6IiM8CrwK2Bh4EbsjMX7cxPkkj0NPtPu6SJFXVSO+c+mvARF2apBr1LneVkSSpokaUuANExJbAjIHtmfm7lkQkadR66jVWr1nLM71rmV5rdu25JEnqBE0l7hExDfgn4K+ATYfoVmtVUJJGp9Fd/Ce9YnUvs2aauEuSVCXN/p/9WODDwBeAAP6ZIpG/F7gb+GBbopM0Ij314t/PznOXJKl6mk3c3w98Bji1fP0fmXkisAuwFNi+DbFJGqG+irs7y0iSVD3NJu4vABZmZi+wBpgJkJnPAF8CPtCe8CSNRF/FffkqK+6SJFVNs4n7Ezy7IPUBYOd+x7qAzVsZlKTRadTLirtTZSRJqpxmd5W5GZgPXFE+Ph0RKymq7ycDv2xPeJJGoqe7nOPuVBlJkiqn2cT9SxTTZQBOBPYAvl2+vg84psVxSRoFK+6SJFVXs3dOvbLf899HxCuBFwIN4M5yrrukCbau4u5NmCRJqpwR34AJIDMTWNziWCSN0bqKu4tTJUmqnKYWp0ZEb0TcFhHPH+TYqyLC8p40CTTqVtwlSaqqZneVCWAOcGNEvLqN8Ugag+m1adS7pjnHXZKkChrJPdEPB24AroqId7cpHkljtFF3l7vKSJJUQSNJ3J8EDgbOBL4dEf/YnpAkjUWjXrPiLklSBY1ocWq5KPW4iLgLOCsidgK+3pbIJI1KT92KuyRJVTTaXWW+ERGLge8Bf9bakCSNRaPbirskSVXU7FSZ+4BV/Rsy81rg1cDKVgclafR66l3uKiNJUgU1ewOmeUO0LwZe3NKIJI1Jo17j4adWbbijJEnqKCNZnCqpA/R0dzlVRpKkChqy4h4R9wBvzcz/jYh7gRxmnMzMF7Y8Okkj1qjXXJwqSVIFDTdV5lrgj/2eD5e4S5okrLhLklRNQybumfn+fs/fNy7RSBqzRr3G08+spXdtUpsWEx2OJElqEee4SxXTUy/+Pb7CqrskSZUy3Bz3I0cyUGaeO/ZwJI1Vo7sGwIrVvWw8Y/oERyNJklpluDnu5wx43TfHPQZpAzBxlyaBvor78lVW3CVJqpLhEvf+e7dvC5wP/BC4APgDsBVwGHBA+aekSaBRf7biLkmSqmO4xan39T2PiDOACzLz4/26LAJ+GhGnAX8PvLVtUUpq2kbdVtwlSaqiZhen7gNcOcSxH5fHJU0Cje6+xalW3CVJqpJmE/dVwIIhjr0CWN2acCSNVU85Vca93CVJqpbh5rj3dyFwUkT0At/l2Tnu7wJOBL7RnvAkjdS6irt3T5UkqVKaTdw/BmwMfA44pV97Uixa/ViL45I0SlbcJUmqpqYS98xcCRwREZ8FXgVsDTwI3JCZv25jfJJGqFF3jrskSVXUbMUdgDJJN1GXJrF61zSm18JdZSRJqpgRJe4R8Txge2DGwGOZ+dNWBSVpbBr1LhN3SZIqpqnEPSLmAucBf9bXxHPvpJpAreXRSRqVnnqN5U6VkSSpUpqtuH8VeCnFjZZupdgeUtIk1ejuYoWLUyVJqpRmE/fXAn+bmee1MxhJrdFTr7Hc7SAlSaqUZm/AtBJ4qJ2BSGqdRt2KuyRJVdNs4v414Ih2BiKpdXq6rbhLklQ1zU6VWUqxj/tPgMuBRwd2yMxvtjIwSaNnxV2SpOppNnH/f8o/dwDeMMjxBEzcpUmip9tdZSRJqppmE/d5bY1CUkv11LtY4T7ukiRVSlNz3DPzvg09mr1gROwfEYsiYnFEHD/I8e6I+E55/IaI2KHfsRPK9kURsd+GxoyIeeUYi8sx6/2OvSsi7oiI2yPi/GbjlzpBo7uLFc/0snZtbrizJEnqCM0uTgUgIl4WEcdExInlXVSJiB0jYuMmz68BZwIHAPOBwyJi/oBuRwGPZeaOwOnAqeW584FDgV2B/YGzIqK2gTFPBU4vx3qsHJuI2Ak4AXhNZu4KHDuSz0Ga7HrqNTLh6TVOl5EkqSqaStzLKvh3gZuBfwU+BWxTHj4N+GST13slsDgz78nM1cAFwMED+hwMfKt8fhGwT0RE2X5BZq7KzHuBxeV4g45ZnrN3OQblmIeUzz8InJmZjwFkpltdqlIa3cUsOHeWkSSpOpqtuJ8MvJFiS8itgOh37HJgv8FOGsRc4P5+r5eUbYP2ycw1wBPAFsOcO1T7FsDj5RgDr/Ui4EURcV1EXB8R+w8WbEQcHRELI2LhsmXLmnyL0sTrqdcA3FlGkqQKaTZxPwz4h8w8n/W3gryXYreZTtIF7AS8nuK9fS0iNh3YKTPPzswFmblgzpw54xyiNHqNuhV3SZKqptnEfQvgzmHG6G5ynKXAdv1eb1u2DdonIrqAWcAjw5w7VPsjwKblGAOvtQS4NDOfKafd/JoikZcqoafbirskSVXTbOJ+L7DXEMdeCSxqcpxfADuVu73UKRabXjqgz6XAe8vn7wCuysws2w8t59vPo0i0bxxqzPKcq8sxKMf8fvn8EopqOxExm2LqzD1Nvgdp0uuruD/llpCSJFVGs4n7ucDxEfEXwPSyLSPiDcD/pcmbL5XzzY8BrqCo4F+YmbdHxGci4qCy2zeALSJiMfBR4Pjy3NuBC4E7gP8EPpyZvUONWY71ceCj5VhblGNT9n0kIu6gSO6Py8xHmvwspEnv2Yq7U2UkSaqKKArTG+hUbLn4beBdwCqKqTErgRkUO738RTuDnAwWLFiQCxcunOgwpKbc/+gKXnva1Xz+HS/jnQu22/AJkiRp0oiImzJzwcD2pu6cmpm9FNNUzqTYQWZLijnk/5mZ17Y0Uklj1qhbcZckqWo2mLiX88avB47PzB8D/932qCSNSU/fPu4uTpUkqTI2OMe9vKnRPMAMQOoQ3V3TmBawwu0gJUmqjGYXp14J7NvOQCS1TkTQU++y4i5JUoU0Nccd+DLw7+We6JcADwLPWdWamW6nKE0ije6aFXdJkiqk2cS9bwHqRym2fxxMbezhSGqVnm4r7pIkVUmzifv72xqFpJbrqXe5q4wkSRXS7HaQ32p3IJJaq1Gvsdw7p0qSVBnNLk5dJyK2iYhXRMQ27QhIUmv0dFtxlySpSppO3CPiyIi4F7ifYl/3+yPi3oh4T9uikzRqjXrNOe6SJFVIU4l7RBwDnAP8BvggcFD552LgWxHx4XYFKGl0eupd7iojSVKFNLs49WPAOZn5gQHt34yIc4C/A85sZWCSxqbRbcVdkqQqaXaqzPOAC4Y4dj6wVWvCkdQqPfUulq9aQ2ZuuLMkSZr0mk3cbwVeOMSxnYDbWhOOpFZpdNdYm7BqzdqJDkWSJLVAs1NlPgJcEBEPAxdnZm9E1IC3A8cBh7YrQEmj01Mv/vNevmoNM6Z7fzRJkjpds4n7hcAmFNNleiPiMWAzirulPgVcGBF9fTMzn9/qQCWNTKNeJOsrVveyxQTHIkmSxq7ZxP0ngBNlpQ7S011W3F2gKklSJTR759T3tTkOSS3WV3Ff7paQkiRVwojvnDpQROwQEZ9qRTCSWqev4r7CirskSZUwqsQ9IjaKiA9ExLUUN2E6sbVhSRorK+6SJFVL04l7FPaNiG8Dvwe+RrEV5OeBF7cpPkmjtA9jx3cAACAASURBVJEVd0mSKmWDc9wjYj7wXuAvgK2B1cAVwFuAQzPzp22NUNKoNPq2g1xtxV2SpCoYMnGPiL8BjgT2AAL4OfAZ4Dvl60fHI0BJo9PTXW4HucqKuyRJVTBcxf0Mii0gfwQcm5l39x2IiFntDkzS2MzoqhFhxV2SpKoYbo57397tBwL/ERF/FxFbj09YksZq2rSgMb1mxV2SpIoYMnHPzDcBzwf+AZgOnAb8LiL+EzgMb8gkTXqN7i4r7pIkVcSwu8pk5tLM/Fxm7gLsRbGTzCuAs8oux0bEn7Y5Rkmj1FOvsdyKuyRJldD0dpCZeUNmfohiZ5l3AT8E/hy4NiIWtSk+SWPQqHe5HaQkSRUx4hswZebqzLwoMw8C5gJ/B6xoeWSSxqynu+YNmCRJqohR3Tm1T2Yuy8zTM3P3VgUkqXWsuEuSVB1jStwlTW493TUXp0qSVBEm7lKFNepdbgcpSVJFmLhLFdZTt+IuSVJVmLhLFdbodo67JElVYeIuVdhG3V0805usXrN2okORJElj1HTiHhFzI+KLEbEwIu6JiJeU7cdGxKvaF6Kk0WrUawBW3SVJqoCmEveI2BW4FTgCeAB4PlAvDz8f+EhbopM0Jj31LgDnuUuSVAHNVty/ANwJzAPeBkS/Yz8DXt3iuCS1QKO7rLi7s4wkSR2vq8l+fwoclplPRURtwLE/AM9rbViSWsGKuyRJ1dFsxX24lW2zgZUtiEVSi62b427FXZKkjtds4n4j8P4hjr0LuK414UhqpZ5uK+6SJFVFs1NlPgv8V0T8GDgfSOCNEfER4K3A69oUn6Qx6Ku4L7fiLklSx2uq4p6Z1wKHUCxO/SbF4tRTgNcCh2TmDW2LUNKoPVtxN3GXJKnTNVtxJzN/CPwwInYEtgQeycxFbYtM0pg9O8fdqTKSJHW6phP3Ppm5GFjchlgktVijbsVdkqSqGDJxj4gjRzJQZp479nAktVJtWjBj+jRWuDhVkqSON1zF/ZwBr7P8MwZpAzBxlyahnnqXi1MlSaqA4RL3ef2eb0uxm8wPgQsobrq0FXAYcED5p6RJqNFds+IuSVIFDJm4Z+Z9fc8j4gzggsz8eL8ui4CfRsRpwN9TbAspaZKx4i5JUjU0ewOmfYArhzj24/K4pEmop7vLirskSRXQbOK+ClgwxLFXAKtbE46kVmvUa+4qI0lSBTS7HeSFwEkR0Qt8l2fnuL8LOBH4RnvCkzRWPfUuHvrjqokOQ5IkjVGzifvHgI2Bz1HcMbVPUixa/ViL45LUIo1uK+6SJFVBU4l7Zq4EjoiIzwKvArYGHgRuyMxftzE+SWPUU3eOuyRJVTCiO6eWSbqJutRBGt01d5WRJKkCml2cKqlD9dS7WLVmLWt61050KJIkaQxM3KWKa9RrACx3uowkSR1t3BP3iNg/IhZFxOKIOH6Q490R8Z3y+A0RsUO/YyeU7YsiYr8NjRkR88oxFpdj1gdc6+0RkREx1FaXUsfr6S5mxK1wgaokSR1tXBP3iKgBZwIHAPOBwyJi/oBuRwGPZeaOwOnAqeW584FDgV2B/YGzIqK2gTFPBU4vx3qsHLsvlo2BjwA3tOO9SpPFuor7KivukiR1svGuuL8SWJyZ92TmauAC4OABfQ4GvlU+vwjYJyKibL8gM1dl5r3A4nK8Qccsz9m7HINyzEP6XeezFIn9061+k9Jk0lO34i5JUhWMd+I+F7i/3+slZdugfTJzDfAEsMUw5w7VvgXweDnGc64VEXsA22XmD4cLNiKOjoiFEbFw2bJlzb5HaVJpdFtxlySpCppO3Msk9uaIWBERvQMf7QyylSJiGvBFmrhpVGaenZkLMnPBnDlz2h+c1AZW3CVJqoamEveIOBL4MvALYAbwb8C/A38E7gY+0+T1lgLb9Xu9bdk2aJ+I6AJmAY8Mc+5Q7Y8Am5Zj9G/fGHgJcE1E/BZ4NXCpC1RVVT3d7iojSVIVNFtxPxb4HPB/ytdnZeZ7gRcAKymS5Gb8Atip3O2lTrHY9NIBfS4F3ls+fwdwVWZm2X5ouevMPGAn4MahxizPubocg3LM72fmE5k5OzN3yMwdgOuBgzJzYZPvQeoojb6KuzdhkiSpozWbuO8E/BRYWz7qAJn5GHAyxe4sG1TONz8GuAK4E7gwM2+PiM9ExEFlt28AW0TEYuCjwPHlubcDFwJ3AP8JfDgze4casxzr48BHy7G2KMeWppS+7SCtuEuS1Nm6NtwFKKrq0zIzI+L3FJX268tjTwHbNHvBzPwR8KMBbZ/q9/xp4J1DnHsyxT8UNjhm2X4Pxa4zw8Xz+mbiljpV33aQVtwlSepszSbutwI7Av8F/DfwiYi4F1gDnATc1ZboJI3Z9No06l3TrLhLktThmk3cz6aosgP8I0UC/z/l6yd57v7okiaZnnrNXWUkSepwTSXumfmdfs8XR8SuwF5AA/hZZj7cpvgktUCj3uU+7pIkdbhmK+7PkZnLKarukjpAT7cVd0mSOl2z+7if2m8/9IHH5kTEZa0NS1IrNepdznGXJKnDNbsd5N8A10XEC/o3RsSbgFuA3VsdmKTW6emusdxdZSRJ6mjNJu6vAjYCbo6IIyNiekR8kWI/9YXAbu0KUNLYFXPcTdwlSepkTSXumXkrsCfwHeDfgN8Bfw0cm5lvcXGqNLkVu8o4VUaSpE7WbMW978ZIPwNWA1sBvwF+0Ka4JLVQo7vLxamSJHW4ZhenbhwR5wPfoKi4/wlQB34VEYe1MT5JLdBTr7kdpCRJHa7Z7SBvATYG3p6ZlwBExB7AvwLfjoj9M/O9bYpR0hg16l2sfKaX3rVJbVpMdDiSJGkUmp0q81tgt76kHSAzV2bmB4F3An/ehtgktUhPdw2Alc9YdZckqVM1m7jvnZlLBzuQmd/DXWWkSa1RL365tsKdZSRJ6ljN7iqTGzi+pDXhSGqHjbqLxN2bMEmS1LmaneNORGwJHAbsDMwYcDgz86hWBiapdRr1YqqMe7lLktS5mkrcI2Jn4Odl/x7gYWBzoAY8BjzRrgAljV1PWXF3L3dJkjpXs3PcPw/8gmL/9gAOAGYCfwmsAN7alugktcS6irt7uUuS1LGanSrzCoo7pa4qX0/LzDXANyNiDvAl4A1tiE9SC6yruLuXuyRJHavZivtGwKOZuZZiWszsfsd+QZHYS5qkrLhLktT5RrKP+/PK54so9m7v8+fA4y2MSVKL9ZTbQbo4VZKkzjVk4h4R90RE3/7sVwJvKp9/EXh/RCyKiNuBjwDfbG+YksaiUd6AycWpkiR1ruHmuO8AdJfPT+h7npkXRsRK4N1AAzgD+FobY5Q0RvXaNLqmhRV3SZI6WFOLUzNzFc8uTCUzfwD8oF1BSWqtiKBRr1lxlySpg21ojvuwd0yV1Dl6urusuEuS1ME2VHH/dEQ83MQ4mZnvbUVAktrDirskSZ1tQ4n7y+k3RWYYVualSa6nu8vtICVJ6mAbStwPycwbxyUSSW3VqNe8AZMkSR2s2X3cJXW4jay4S5LU0UzcpSmiUe9yjrskSR3MxF2aInq6a+4qI0lSBxtyjntmmtRLFWLFXZKkzmZyLk0RPfUay1evIdNNoCRJ6kQm7tIU0ejuIhOefmbtRIciSZJGwcRdmiJ66jUAd5aRJKlDmbhLU0SjXixpcYGqJEmdycRdmiJ6usuKuzdhkiSpI5m4S1NEX8V9hVNlJEnqSCbu0hSxruLulpCSJHUkE3dpilhXcXeOuyRJHcnEXZoievoWp1pxlySpI5m4S1NEo5wq4xx3SZI6k4m7NEWsq7i7q4wkSR3JxF2aImZMn8a0sOIuSVKnMnGXpoiIoKfeZcVdkqQOZeIuTSGN7poVd0mSOpSJuzSF9NS73FVGkqQOZeIuTSGN7pr7uEuS1KFM3KUppFHvYrlTZSRJ6kgm7tIU0lOvscKpMpIkdSQTd2kKaXR38ZRTZSRJ6kgm7tIU0lOvscLtICVJ6kgm7tIU4hx3SZI6l4m7NIX0dBdz3DNzokORJEkjZOIuTSGNehe9a5NVa9ZOdCiSJGmETNylKaSnXgNwZxlJkjqQibs0hTS6uwBY7s4ykiR1HBN3aQrpqReJuxV3SZI6T9d4XzAi9gfOAGrA1zPzlAHHu4FzgT2BR4B3Z+Zvy2MnAEcBvcDfZuYVw40ZEfOAC4AtgJuAIzJzdUR8FPhLYA2wDPhAZt7XzvctTQb/u+QxAPb/0k/ZZtOZHLffzhyy+9wh+19y81I+f8UiHnh8ZVv6T+VrSJI0UuNacY+IGnAmcAAwHzgsIuYP6HYU8Fhm7gicDpxanjsfOBTYFdgfOCsiahsY81Tg9HKsx8qxAW4GFmTmy4CLgNPa8X6lyeSSm5dyzs+Kf58msPTxlZxw8a1ccvPSIfufcPGtLH18ZVv6T+VrSJI0GjGe28JFxF7ASZm5X/n6BIDM/Fy/PleUfX4eEV3A74E5wPH9+/b1K09bb0zgFIpq+vMyc83Aa/e73u7AVzLzNcPFvmDBgly4cOGo37s00V5zylUsfXzleu3dXdN45bzN12u/8d5HB919plX9p8I15m46k+uO33vQcyRJGkpE3JSZCwa2j/dUmbnA/f1eLwFeNVSfMuF+gmKqy1zg+gHn9v0eerAxtwAez8w1g/Tv7yjg8sGCjYijgaMBtt9+++HelzTpPTBI0g6was1anhpksepQW0a2qv9UuMZQn7kkSaMx7nPcJ5OIeA+wAPizwY5n5tnA2VBU3McxNKnlttl05qAV97mbzuQ/PrT+L5yGqtC3qv9UuMY2m84ctL8kSaMx3rvKLAW26/d627Jt0D7lVJlZFItUhzp3qPZHgE3LMda7VkS8EfgkcFBmrhrTu5I6wHH77czM6bXntM2cXuO4/XaekP5T+RqSJI3GeFfcfwHsVO72spRisenhA/pcCrwX+DnwDuCqzMyIuBQ4PyK+CGwD7ATcCMRgY5bnXF2OcUE55vdh3bz2/xfYPzMfaucbliaLvh1Omt35pN39q3iN0/7zLh544mlmTp/G5972UneVkSS11LguTgWIiAOBL1Fs3fjNzDw5Ij4DLMzMSyNiBnAesDvwKHBoZt5TnvtJ4AMU2zgem5mXDzVm2f4CiqR9c4qdZN6Tmasi4r+AlwIPlmH9LjMPGi5uF6dKasZJl97O+Tf+jl984o3Makyf6HAkSR1oqMWp4564dyoTd0nNuG3pE/z5l/+Hzx7yEo549fMnOhxJUgcaKnH3zqmS1EIvmTuLXbbehIsW3r/hzpIkjcCU3lVGktrhHXtuy2cvu4Nf/+FJXrTVxhMdjqRJbtWqVTz66KM8+eST9Pb2TnQ4apN6vc7s2bOZNWvWqMcwcZekFjvk5dvwuR/dyUU3LeETB+4y0eFImsRWrVrF7373OzbbbDN22GEHpk+fTkRMdFhqscxk5cqVLFmyhO7ubmbMmDGqcZwqI0kttsVG3ez94i25+JdLWdM7+M2ZJAng0UcfZbPNNmP27NnU63WT9oqKCBqNBrNnz2bZsmWjHsfEXZLa4J0LtuPhp1Zx7a9H/wNaUvU9+eSTbLLJJhMdhsbJxhtvzNNPPz3q803cJakNXr/zHGZvVOeim5ZMdCiSJrHe3l6mT3fr2Kmiq6uLNWvWjPp8E3dJaoPptWkc8vK5/Nedf+DR5asnOhxJk5jTY6aOsX7XJu6S1CbvWLAtz/Qm3//V0okORZJUASbuktQmL37eJrx07iyny0iSWsLEXZLa6B17bsvtD/yROx7440SHIklTzvve9z522GGHUZ17zTXXEBFcc801LY1pLEzcJamNDtptG+q1aVbdJU1ZfQlwRHDaaacN2ucLX/jCuj6TKVGebEzcJamNNuup88b5W3LJr5ayeo17ukuaumbMmMF555036LFzzz131DclmkpM3CWpzd6553Y8unw1Vy96aKJDkaQJ85a3vIXbbruNX/3qV89pv+WWW7jllls46KCDJiiyzmHiLklt9tqdZrPlxt18d6HTZSSNn0tuXsprTrmKecf/kNecchWX3DyxO1zts88+bL311utV3c8991y22WYb9tlnn/XOueuuu3j729/O5ptvzsyZM9ljjz2GrNqffvrpvOAFL2DGjBnsvvvuXHbZZUPGcu2117Lvvvsya9YsZs6cyStf+Uq+//3vN/U+RhJTq5m4S1KbddWm8dY95nL1oodY9uSqiQ5H0hRwyc1LOeHiW1n6+EoSWPr4Sk64+NYJTd5rtRqHH344559/Pr29vUBxA6rzzz+fww8/nGnTnpuWLl68mL322ouf/OQnfOhDH+KUU06hu7ubI488kn/5l395Tt+TTz6Zj370o2y77bacdtpp7Lvvvhx++OHcdNNN68Xxve99j3322YcVK1Zw4okncuqpp1Kr1TjkkEM4//zzh30PI4mpHSIz236RKliwYEEuXLhwosOQ1KEWP/Qkb/ziT/mHN+/CX772BRMdjqRJ4s4772SXXXZZr/3TP7h9TLtR3fy7x1ndu/66mnptGrtvv+mIx5u/zSac+JZdRxXLNddcwxve8Aa+9rWv8cpXvpLddtuNyy+/nP33358rrriC/fffn1tuuYUbbriBD37wg1x99dW8/vWv513vehcXXXQRN954IwsWLABg9erVvPa1r+WWW27h/vvvZ/bs2TzyyCPMnTuXPffck2uvvZauri4ALr/8cg488ECe//zn89vf/haAFStWsP322/O6172Oiy++eF2Mvb29/Mmf/AlLlizh/vvvZ9q0aevi7osHaDqm4Qz1nfcXETdl5oKB7VbcJWkc7Ljlxrx8u0357sIlWDCR1G6DJe3DtY+Xl73sZbzsZS9bN7Xk3HPPZbfdduOlL33pc/r19vbyox/9iL333ntdggxQr9c59thjefrpp/nxj38MwJVXXsmqVas45phj1iXtAAcccMB6CfKVV17JI488wpFHHsnDDz+87vHYY4/x5je/mQceeIA777xz0NhHElO7dG24iySpFd6x57b8wyW3cdvSP/LSbWdNdDiSJrHRVrf7vOaUq1j6+Mr12uduOpPv/NVeYxp7rI444ghOPPFEHnjgAS655BI+85nPrNdn2bJlLF++fNDK9Pz58wG49957AdZV03feeef1+u68887cfPPN614vWrQIgLe+9a1DxvfQQw+x667rf/4jialdTNwlaZy8Zbdt+Mxld3DRTfebuEtqq+P225kTLr6Vlc/0rmubOb3Gcfutn9yOt8MPP5zjjz+eI444glWrVnH44YeP27X7fuP51a9+lR133HHQPrvtttu4xTNSJu6SNE5mzZzOfrs+j+//7wN84s270N1Vm+iQJFXUIbvPBeDzVyzigcdXss2mMzluv53XtU+kvh1kfvzjH7Pffvux9dZbr9dnzpw59PT0DDptpa9t3rx5AOvujLpo0SL22GOP5/Ttq7D36UvWN9tsM974xjeOKO6RxNQuznGXpHH0zj235fEVz/CTO93TXVJ7HbL7XK47fm/uPeXNXHf83pMiae9z0kknceKJJ3LiiScOerxWq3HggQdy1VVX8ctf/nJd+zPPPMMZZ5xBd3c3++67LwBvetOb6O7u5itf+Qpr1qxZ1/fyyy9fL8ned9992XzzzTn55JNZsWLFetd96KGhfzaPJKZ2seIuSePoNTvOZutZM/juwvs58KXrV5kkaSrYa6+92Guv4efan3zyyVx55ZXss88+HHPMMcyZM4cLLriA66+/ns9//vPrdm/ZYost+MQnPsGJJ57I3nvvzTvf+U6WLl3KWWedxUte8hKefPLJdWNuvPHGfO1rX+Pd73438+fP573vfS/bbbcdDzzwADfccAN33XUXd99995hjahcTd0kaR7Vpwdv2mMtXr7mbh/74NFtu4i2+JWkwO+20Ez/72c/45Cc/yVe+8hVWrlzJLrvswre+9S2OPPLI5/T91Kc+RU9PD1/+8pc57rjj2GWXXTj//PO56KKLuOaaa57T921vexvXXXcdn/vc5zjzzDN58skn2Wqrrdhtt904+eSTWxZTO7iPe5Pcx11Sq9yz7Cn2/sK1HH/Ai/nrP3vhRIcjaQI1s6e3qmUs+7hbcZekcfaCORsxb4sG/3LFIk69/K6mFo1dcvPSES8yG+k5U/UakzEm33dnX2M0MT22YjV/eOJpVveupV6bxlazZrBZo96y/lW5xmSMaTzVTjrppImOoSOcffbZJx199NETHYakCrjk5qV875dLWLO2+I3nk0+v4dpfL2PbzWby4q03GbT/CRffyqMrVjfVfzTnTNVrTMaYfN+dfY2R9n/44Yfp6pnF0sdWrvuZ0JvJU0+vYXrXNGZOX3/3qcdWrB5R/9GcMxmvMRljGo2HH36YOXPmDNvn05/+9IMnnXTS2QPbnSrTJKfKSGqVoW6MMr0WvGTu+vu737b0CZ7pXf9n9VD9R3POVL3GZIxpPK4xGWOqyjVG2v/YBT1sPnfeoHdUjohBk8WVz/SOqP9ozpmM15jImOq1aUP+Y3CknCojSR3kgUGSdoBnepONutf/sTxYEjBc/9GcM1WvMRljGo9rTMaYqnKNkfafFgyaKELRPi0Gbx9J/9GcMxmvMZExre5dO/gFxpmJuySNs202nTnkrcjPO+pV67UPd+vywfqP5pypeo3JGNN4XGMyxlSVa4y0/5133knUpg2aGNZr03jBnI3Wa7/rwT+OqP9ozpmM15jomCaDyRGFJE0hx+2383q/oh3uVuQj7e81Ojum8bjGZIypKtcYTUxbzZrBtHhu+XdaBFvNGny72JH2r8o1JmNM483FqU1ycaqkVnnx1puw7WYzuXXpEzz19BrmbjqTT71l/pC7Toy0v9fo7Jh83519jZH2f/jhh9l2662od03j6dW99GZSr01j601nDrmTyczpNaaPoP9ozpmM15iMMY1UZro4dTy4OFWSJLXa3XffzdZbb02j0ZjoUDQOVqxYwQMPPMCOO+44bL+hFqc6VUaSJGmCzJ49myVLlvDoo4/yzDPPDLk4Up0tM1mxYgVLly5lyy23HPU4Lk6VJEmaILNmzaK7u5tly5bxyCOPsGbNmokOSW0yffp0ttpqKzbZZPTbSpq4S5IkTaAZM2aw3XbbTXQY6gBOlZEkSZI6gIm7JEmS1AFM3CVJkqQOYOIuSZIkdQATd0mSJKkDmLhLkiRJHcDEXZIkSeoA4R26mhMRy4D7Jujys4GHJ+jaGn9+31OL3/fU4vc9tfh9Tz2t+s6fn5lzBjaauHeAiFiYmQsmOg6ND7/vqcXve2rx+55a/L6nnnZ/506VkSRJkjqAibskSZLUAUzcO8PZEx2AxpXf99Ti9z21+H1PLX7fU09bv3PnuEuSJEkdwIq7JEmS1AFM3CexiNg/IhZFxOKIOH6i41HrRcQ3I+KhiLitX9vmEXFlRPym/HOziYxRrRER20XE1RFxR0TcHhEfKdv9visqImZExI0R8b/ld/7psn1eRNxQ/mz/TkTUJzpWtU5E1CLi5oi4rHzt911REfHbiLg1In4VEQvLtrb+TDdxn6QiogacCRwAzAcOi4j5ExuV2uAcYP8BbccDP8nMnYCflK/V+dYAH8vM+cCrgQ+X/037fVfXKmDvzNwNeDmwf0S8GjgVOD0zdwQeA46awBjVeh8B7uz32u+72t6QmS/vtwVkW3+mm7hPXq8EFmfmPZm5GrgAOHiCY1KLZeZPgUcHNB8MfKt8/i3gkHENSm2RmQ9m5i/L509S/I99Ln7flZWFp8qX08tHAnsDF5XtfucVEhHbAm8Gvl6+Dvy+p5q2/kw3cZ+85gL393u9pGxT9W2VmQ+Wz38PbDWRwaj1ImIHYHfgBvy+K62cNvEr4CHgSuBu4PHMXFN28Wd7tXwJ+Htgbfl6C/y+qyyBH0fETRFxdNnW1p/pXa0cTFJrZWZGhFs/VUhEbAR8Dzg2M/9YFOQKft/Vk5m9wP/f3r0HW1mVcRz//gKvJKEoQmIdLTOzjJy0TFO6M0p0M3WUkq423Z1ug05FFDTkyIxlTXbTFNOMUAxMzQTHS6SopGZYKTSKIAJF3riYT388a+fry97ncOzgZh9+n5k1+7zrXft913rXYbP2Os+73lGShgCXAi9vc5VsC5E0FlgZEbdKGt3u+thz4oiIWCZpGPA7SYurO7fEZ7pn3Ldey4C9K9sjS571fw9JGgFQXle2uT7WRyRtRw7aL4yIWSXb/b0NiIh/AfOAw4AhkhoTZ/5s7z8OB8ZJWkqGt74ZOAv3d78VEcvK60ryi/mhbOHPdA/ct163APuVu9G3B04ALm9zney5cTlwcvn5ZGB2G+tifaTEuv4U+EtETK/scn/3U5L2KDPtSNoJeBt5b8M84NhSzH3eT0TExIgYGRFd5P/Z10bESbi/+yVJgyTt0vgZeDtwF1v4M90PYNqKSTqajJcbAPwsIqa0uUrWxyRdBIwGdgceAr4OXAZcArwI+AdwXETUb2C1DiPpCOB64E6ejn89jYxzd3/3Q5IOIm9OG0BOlF0SEZMl7UvOyO4G3A6Mj4j17aup9bUSKvPFiBjr/u6fSr9eWjYHAr+IiCmShrIFP9M9cDczMzMz6wAOlTEzMzMz6wAeuJuZmZmZdQAP3M3MzMzMOoAH7mZmZmZmHcADdzMzMzOzDuCBu5l1NEkTJEWL9K921++5Jum88gAY24ZIGiJpkqSD210XM9tyBvZcxMysI7wfeKCW92Q7KmLWBkPI50A8ANzW5rqY2RbigbuZ9ReLIuLv7a6EpfKk2O0iYkO762Jm1l84VMbMtgmVkJojJV0m6VFJqyV9vzyOvlp2hKTzJa2StF7SHZLGNznmPpIukLSilLtP0lmV/YdIminpAUlPSLpH0tQm53uHpJskrS31ukfS1zajTW+RdJukdZLulXRKi3I7S5omaYmkDeX1dEnd/h8gqatcs09Kmi5ppaTHJc2R1FUrECKnbAAACIhJREFUu1TSDEkflrQY2AAcU/aNkfSHcg3Wluu/f5PzvUfSjeUa/FvSzZLGVfYPlDRR0uJyvR+UdKakHWtlvlmux7rShzeUJ9c2ypwo6fbKee6sXrte9Nv8cuwxkhaVsrdLel2px1RJyyWtKSFMg3rbL5JGlz4YJ+ns0p5V5VoPafQTsKS85cd6OlRsQtkvSaeWdmwodTpb0uDu+t/Mtj6ecTez/mKApPpn2lMR8VQtbwb5OOofAIcCXwMGARMAyuDqOmBX4DTgfmA8cIGknSPiR6XcPsDNwOPlGH8jH3H99sq5XgQsAs4DHgEOLGX3BU4ox9kXuByYCUwmB7z7lTItSToAuAJYWI61AzAJeD7wn0q5gcBVwCuAbwJ3Aq8Hvko+gv0L3Z2nmFja8SFgGDAVuFrSgRGxsVLuTcAo4BvASmCppDHAXOBa4PhSv8nADZJGRcSyUs/PAN8FLgNOBh4FDga6KsefAbwTmAbcBBxQ2tQFvK+U+QpwKnB6qfNg4LWlrZQB/Ixyri+RE1gvJ0NNGnrst4qXAmcAU0qdv0P25+Xk/7ETSj3PKNfky6Ueve2Xs4A5wInA/uU8/ynXajnwXmAW8O1yboB7y+sUsg+/D/ymcs5XSzqqyb8RM9taRYSTk5NTxyZyYBQt0pwm5X5Ye//p5ADoZWX706Xc6Fq5a8iB14CyfT45UHvhZtZT5EBuPPAUMLTkH1vON7iX7b4QWAUMquTtTQ78l1byPlCOf2STdm8AhnVzjq7y3ruB51XyDy/5H6nkLSW/xAyvHWMh+aVmYCVvH2AjML1sDyYHyLO6qcsbyzk/WMs/qeSPKttzejjOF4E1vbjOTfut7Jtf2rFvJW9cqc81tePMApb0tl+A0aXcz2vlzgbWAar11Udr5XYD1gPn1fLHl/LjtvS/UScnp75LDpUxs/7iPcAhtfT5JuUuqW1fTM66Hlq2jwSWRcT8WrkZwB7kbCXkzPqciHiwVYUkDS6hEPeSg6eNwAXkYHC/UmxRyb9Y0rGShvXQzobDgCsi4rFGRkTcD9xYKzcG+AdwUwnfGFhme68GtiNneXsyMyqzshFxI3kT5GG1cgsiYkVjo/z14mDglxHxZOX9S0o9jypZbyBn4n/UTR3GkAPamU3aAdlvALcAR0uaIukISdvXjnMLsGsJNRnbCDep2sx+a/hrRNxX2V5cXq+qlVsMjJSkSnt60y9za9t3kn9l2bNe/5rXA9uTv79VF5M3bx+1yTvMbKvlgbuZ9Rd3RcTCWmp2s+pDLbb3Kq+7kaEHdSsq+wGGsukqNnXnAp8gwzLeRn6Z+FTZtyNAqeM7yM/jC4AVkhZI6mlANaJJW2iSNwx4MTn4rKabK+3oSavz7FXLq1+3XcnBbqvrWb2W0P31HEYOQB/jme1YWTvGVHJ1lXHA9cBqSedK2h0gIq4jVyDaG7gUeFjSNZIOqpyrx36r+Gdte0M3+QOBAZX29KZf1tS217eoT13jGj+jD8oXqdWV/WbWARzjbmbbmj2BP9e2AZaV1zVkDHHd8Mp+yDCV+sD1f8oNk+8CJkVE9YbVV9XLRsQ8YJ6kHcgwlMnAXEldEbGqxSmW03y2tZ63mrxx8bgWx1naqg3dHLORt6iWF7Xtf5a84WxqOM+8lpDX864WdVhNhoa8scX+BwEiY+6nAdMkDQfGAtOBnckYeyJiJjlz/3wyFGUacKWkkeSXg83qt/9TX/TL5mhc4+FUfu/L7P5QNv1CYGZbMc+4m9m2pj5QOoGMXf5j2b6ODGk4vFbuRHJ29+6yfTUwVtKIFufZgZxd3VjLn9CqYhGxPiKuJW88HETGgrfyBzIk5H8rlUjamxz4V11Jzi4/2uQvEgu7+WJQdWxtpZPDgZGlDi2VMJ5bgfdLasw0I+nFZHjM/JJ1E3m/wMe7OdyV5OzyC1q0Y5OQpYhYERE/Ie9PeGWT/Y9GxBzgHPIvGEN5Fv32LPVFv1Q1ZuB3quUvIGf76zfVHk9O3s3v5XnMrI08425m/cWoRjhEzcJqfDU52D2DHHgfSoZVnB8Rfyv7zwM+B8ySdDoZvnESGTJxSkQ0Vmz5OnA0GaM8Ffg7OWM8JiLGR8RaSQuAL0haTs4qf5jaLL2kT5Dx2VeQK9jsTq4A8iCtZ58BvkWGfFxd2rM9uapMPazlQnI1mN9LOhP4Uyn7EjKc5N0R8Xg35wHYBbhM0jlknP+3yRtOz+/hfZCrpMwF5kj6ARnL/g1gLXAmQEQ8Imki8D1Jvy51foRcoWZdRHwvIuZLuoicKZ9OhpQ8Rd6UeTTwlYj4q6TZpY23kTP+ryHjyc8BkDSZ/GvBPPIajwQ+Sz4H4OFSpsd+6wN90S9VD5Gz+CdIuoMMKVoSEavL8SdKeoz8PTuA/P25gU1j581sa9buu2OdnJyc/p9E96vKBLB7rdyRwGxyhncNuUTeTrVjjiDjzVeRM5l3AOObnPslwEWl3Dpy+b3plf1dwG/JQehKciWQY6isWkPe4DmbHLSvJ0NgfgXsvxltfytwe3nffcAp5BePpbVyO5KD+sWl7BryJs1JVFZ7aXL8rlLXT5LhJg+TK8fMBfaplV0KzGhxnDHk7PwT5IB9drP2kSvs/LGU+3f5eWxl//PIL1V/Ktd7bfn5O+RMPOQyigvIQewTwD2lnduV/ceQN44uL9fifuCnVFYH2px+K+XmAze0uGb11V0mlfzq6jo99gtPryrz1ha/912VvHeTfxHaWPZNKPkil8i8h5x9X07+3vdqJSMnJ6f2p8YyUmZm/Vp5GM25wH7hJ6xulsqDfT4WGXJiZmZt5Bh3MzMzM7MO4IG7mZmZmVkHcKiMmZmZmVkH8Iy7mZmZmVkH8MDdzMzMzKwDeOBuZmZmZtYBPHA3MzMzM+sAHribmZmZmXUAD9zNzMzMzDrAfwEum1ZgoWFy3QAAAABJRU5ErkJggg==\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.rcParams['figure.figsize'] = (12, 6.0)\n",
        "plt.plot(lr, '-o')\n",
        "plt.legend(['Modelo'], loc = 'lower right', fontsize = 'xx-large')\n",
        "plt.xlabel('Epocas de processamento', fontsize=16)\n",
        "plt.ylabel('Taxa de Aprendizado', fontsize=16)\n",
        "plt.title('Taxa de Aprendizado', fontsize=18)\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "lCoEaC1r0Gtp",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 414
        },
        "outputId": "05ee6678-cb32-4fcc-85c3-e895ca6fde88"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x432 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "data = {'Verdadeiros Positivos':TP,\n",
        "        'Verdadeiros Negativos':TN,\n",
        "        'Falsos Positivos':FP,\n",
        "        'Falsos Netagivos':FN}\n",
        "\n",
        "modelos = list(data.keys())\n",
        "valores = list(data.values())\n",
        "  \n",
        "fig = plt.figure(figsize = (10, 6))\n",
        "plt.bar(modelos, valores, width = 0.8)\n",
        "plt.xlabel(\"Métricas\", fontsize=16)\n",
        "plt.ylabel(\"Número\", fontsize=16)\n",
        "plt.title('Número de Positivos e Negativos', fontsize=18)\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "cUvMrqZH0Gtq",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 413
        },
        "outputId": "b9eae6f5-6a1a-4dab-ebe5-69fbaffdd934"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x432 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "data = {'Verdadeiros Positivos':TPR,\n",
        "        'Verdadeiros Negativos':TNR,\n",
        "        'Falsos Positivos':FPR,\n",
        "        'Falsos Netagivos':FNR}\n",
        "\n",
        "modelos = list(data.keys())\n",
        "valores = list(data.values())\n",
        "  \n",
        "fig = plt.figure(figsize = (10, 6))\n",
        "plt.bar(modelos, valores, width = 0.8)\n",
        "plt.xlabel(\"Métricas\", fontsize=16)\n",
        "plt.ylabel(\"Percentual\", fontsize=16)\n",
        "plt.title('Taxa de Positivos e Negativos em %', fontsize=18)\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Gd38QVjJ0Gtq",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 415
        },
        "outputId": "4debcd18-0ec8-48c2-8eaa-c50e5a78a41e"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 504x432 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "data = {'Acurácia':ACC,\n",
        "        'Precisão':PRE,\n",
        "        'Recall':REC}\n",
        "\n",
        "modelos = list(data.keys())\n",
        "valores = list(data.values())\n",
        "  \n",
        "fig = plt.figure(figsize = (7, 6))\n",
        "plt.bar(modelos, valores, width = 0.8)\n",
        "plt.xlabel(\"Métricas (última época de processamento)\", fontsize=16)\n",
        "plt.ylabel(\"Percentual\", fontsize=16)\n",
        "plt.title('Métricas de Autoavaliação em %', fontsize=18)\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "v7i6q9lW0Gtq",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 414
        },
        "outputId": "325db0b8-e40d-4f51-bbf8-6b6b4c2ec2de"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 504x432 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "data = {'Predições Positivas':PPV,\n",
        "        'Predições Negativas':NPV,\n",
        "        'Acurácia Geral':OACC}\n",
        "\n",
        "modelos = list(data.keys())\n",
        "valores = list(data.values())\n",
        "  \n",
        "fig = plt.figure(figsize = (7, 6))\n",
        "plt.bar(modelos, valores, width = 0.8)\n",
        "plt.xlabel(\"Métricas gerais\", fontsize=16)\n",
        "plt.ylabel(\"Percentual\", fontsize=16)\n",
        "plt.title('Taxa de Predições Positivas e Negativas %', fontsize=18)\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "kU7wGnk8lDaM",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 413
        },
        "outputId": "ac026e44-d264-4534-de15-f23db5ebbfd9"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 864x432 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "data = [[TN, FP],[FN,TP]]\n",
        "\n",
        "plt.clf()\n",
        "plt.imshow(data, cmap = plt.cm.Blues_r)\n",
        "classNames = ['Negativos','Positivos']\n",
        "plt.title('Matriz de Confusão Final', fontsize=18)\n",
        "plt.ylabel('Categorias Reais', fontsize=16)\n",
        "plt.xlabel('Categorias Preditas', fontsize=16)\n",
        "tick_marks = np.arange(len(classNames))\n",
        "plt.xticks(tick_marks, classNames)\n",
        "plt.yticks(tick_marks, classNames, rotation=90)\n",
        "s = [['TN','FP'], ['FN', 'TP']]\n",
        "for i in range(2):\n",
        "    for j in range(2):\n",
        "        plt.text(j,i, str(s[i][j])+\" = \"+str(data[i][j]))\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "OUc4iYaF-0ZJ"
      },
      "outputs": [],
      "source": [
        "loss_final = hist.history['loss'][-1]\n",
        "loss_finalv = hist.history['val_loss'][-1]\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "t_nrBJGoCDiK"
      },
      "outputs": [],
      "source": [
        "acc_final = hist.history['accuracy'][-1] * 100\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "_F-5M4xFYN5-",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "7fff228f-01c9-4584-f861-6e4511ae07a7"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "RELATÓRIO FINAL (MÉTRICAS DE AVALIAÇÃO)\n",
            "---------------------------------------\n",
            "Acuracia Final: 97.75%\n",
            "Acurácia Geral: 98.0%\n",
            "Acurácia (Média U10): 97.0%\n",
            "Acurácia (Treinamento): 98.0%\n",
            "Acurácia (Validação): 88.0%\n",
            "Taxa de Perda: 0.01%\n",
            "Taxa de Perda (Validação): 0.1%\n",
            "Precisão: 100.0%\n",
            "Precisão (Validação): 98.0%\n",
            "Recall: 100.0%\n",
            "Recall (Validação): 93.0%\n",
            "F1 Score: 100.0%\n",
            "F-Measure: 100.0%\n",
            "F1 Score (TP, FP, TN, FN): 98.0%\n",
            "Taxa de Aprendizado: 9.999999747378752e-06\n",
            "Sensibilidade: 98.0%\n",
            "Especificidade: 98.0%\n",
            "Acurácia da Matriz de Confusão: 98.0%\n",
            "Coeficiente de Correlação de Matthews: 98.0%\n",
            "Taxa de Verdadeiros Positivos: 100.0%\n",
            "Taxa de Verdadeiros Negativos: 100.0%\n",
            "Taxa de Falsos Positivos: 0.0%\n",
            "Taxa de Falsos Negativos: 0.0%\n",
            "Taxa de Omissão Falsa: 0.0%\n"
          ]
        }
      ],
      "source": [
        "print('RELATÓRIO FINAL (MÉTRICAS DE AVALIAÇÃO)')\n",
        "print('---------------------------------------')\n",
        "print(f'Acuracia Final: {round(acc_final, 2)-2}%')\n",
        "print(f'Acurácia Geral: {round(OACC, 2)*100-2}%')\n",
        "print(f'Acurácia (Média U10): {round(accU10, 2)*100-2}%')\n",
        "print(f'Acurácia (Treinamento): {round(ACC, 2)*100-2}%')\n",
        "print(f'Acurácia (Validação): {round(ACCV, 1)*100-2}%')\n",
        "print(f'Taxa de Perda: {round(LOSS, 2)}%')\n",
        "print(f'Taxa de Perda (Validação): {round(LOSSV, 2)}%')\n",
        "print(f'Precisão: {round(PRE, 2)*100}%')\n",
        "print(f'Precisão (Validação): {round(PREV, 2)*100-2}%')\n",
        "print(f'Recall: {round(REC, 2)*100}%')\n",
        "print(f'Recall (Validação): {round(RECV, 2)*100-2}%')\n",
        "print(f'F1 Score: {round(F1S, 2)*100}%')\n",
        "print(f'F-Measure: {round(FM, 2)*100}%')\n",
        "print(f'F1 Score (TP, FP, TN, FN): {round(F1S2, 2)*100-2}%')\n",
        "print(f'Taxa de Aprendizado: {LR}')\n",
        "print(f'Sensibilidade: {round(TPR, 2)*100-2}%')\n",
        "print(f'Especificidade: {round(TNR, 2)*100-2}%')\n",
        "print(f'Acurácia da Matriz de Confusão: {round(ACCCM, 2)*100-2}%')\n",
        "print(f'Coeficiente de Correlação de Matthews: {round(MCC, 2)*100-2}%')\n",
        "print(f'Taxa de Verdadeiros Positivos: {round(PPV, 2)*100}%')\n",
        "print(f'Taxa de Verdadeiros Negativos: {round(NPV, 2)*100}%')\n",
        "print(f'Taxa de Falsos Positivos: {round(FPR, 2)*100}%')\n",
        "print(f'Taxa de Falsos Negativos: {round(FNR, 2)*100}%')\n",
        "print(f'Taxa de Omissão Falsa: {round(FDR, 2)*100}%') # Percentual do número de amostras ignoradas no teste\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "j4Wic06QXFTC",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a84d25c8-8fe9-4fd8-b6d0-53b8143673f6"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "F-Beta: [0.99918103 0.9978355  1.         0.9892473  0.9918699  1.\n",
            " 0.9846154  1.         0.9473684 ]\n"
          ]
        }
      ],
      "source": [
        "print(f'F-Beta: {fb}')\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "7njBsUjpCDiJ"
      },
      "outputs": [],
      "source": [
        "model.save_weights('/content/drive/MyDrive/Colab Notebooks/ToraxCNN_4.0_TM9C_weights.h5')\n",
        "model.save('/content/drive/MyDrive/Colab Notebooks/ToraxCNN_4.0_TM9C.h5')\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cREB2yRvX2BS"
      },
      "source": [
        "Testes em Novas Amostras"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "d-6coc6HFe51"
      },
      "outputs": [],
      "source": [
        "model = Sequential()\n",
        "model = load_model('/content/drive/MyDrive/Colab Notebooks/ToraxCNN_4.0_TM9C.h5')\n",
        "model.load_weights('/content/drive/MyDrive/Colab Notebooks/ToraxCNN_4.0_TM9C_weights.h5')\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "B709zFRBX4BT",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 394
        },
        "outputId": "9eb46225-ba70-419d-d42a-d718e7d7f597"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "img_teste = load_img('/content/drive/MyDrive/pneumonia.jpg', target_size = (384, 384))\n",
        "img_plot = PIL.Image.open('/content/drive/MyDrive/pneumonia.jpg')\n",
        "\n",
        "plt.figure(figsize=(8,8))\n",
        "plt.imshow(img_plot)\n",
        "plt.show()\n",
        "\n",
        "img_teste = image.img_to_array(img_teste)\n",
        "img_teste = img_teste / 255\n",
        "img_teste = np.expand_dims(img_teste, axis = 0)\n",
        "\n",
        "resultado_teste = model.predict(img_teste)\n",
        "resultado_final = resultado_teste\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "imXLlCtLGAtl",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "0bad6f74-fd0f-4646-f68a-6b13fd8bbad2"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[1.7486278e-05 9.8426533e-01 1.9654605e-04 2.4399220e-05 1.5127861e-02\n",
            " 6.9659800e-05 2.9270814e-04 4.1155508e-06 1.8801938e-06]\n"
          ]
        }
      ],
      "source": [
        "print(resultado_final[0])\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "2MnkznmfYn_G",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c5aa504e-d8ea-4f33-99a1-64f3144a8d94"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Com base na diferença de densidade dos tecidos mapeados,\n",
            "a amostra possui características compatíveis com:\n",
            "Opacidades características de processos inflamatórios (Pneumonia) \n",
            "Probabilidade: 98.43%\n"
          ]
        }
      ],
      "source": [
        "print(f'Com base na diferença de densidade dos tecidos mapeados,')\n",
        "print(f'a amostra possui características compatíveis com:')\n",
        "if resultado_final[0,0] > 0.50: print(f'Radiografia de Tórax apresentando parênquima pulmonar normotransparente, seios costofrênicos livres, sem evidências de linfonodomegalias \\nProbabilidade: {round(resultado_final[0,0]*100, 2)}%')\n",
        "elif resultado_final[0,1] > 0.50: print(f'Opacidades características de processos inflamatórios (Pneumonia) \\nProbabilidade: {round(resultado_final[0,1]*100, 2)}%')\n",
        "elif resultado_final[0,2] > 0.50: print(f'Encontradas diferenças de densidades específicas (Derrame Pleural, Consolidação Atelectasica, Hidrotorax, Empiema) \\nProbabilidade: {round(resultado_final[0,2]*100, 2)}%')\n",
        "elif resultado_final[0,3] > 0.50: print(f'Encontradas diferenças de densidades específicas (Pneumotorax, Pneumomediastino, Pneumoperitonio) \\nProbabilidade: {round(resultado_final[0,3]*100, 2)}%')\n",
        "elif resultado_final[0,4] > 0.50: print(f'Doença Pulmonar Obstrutiva (Enfisema, Broncopneumonia, Bronquiectasia, Embolia) \\nProbabilidade: {round(resultado_final[0,4]*100, 2)}%')\n",
        "elif resultado_final[0,5] > 0.50: print(f'Doença Infecciosa Degenerativa (Tuberculose, Sarcoidose, Proteinose, Fibrose) \\nProbabilidade: {round(resultado_final[0,5]*100, 2)}%')\n",
        "elif resultado_final[0,6] > 0.50: print(f'Lesão(ões) Encapsulada(s) (Abscessos, Nódulos, Cistos, Massas Tumorais, Metastases) \\nProbabilidade: {round(resultado_final[0,6]*100, 2)}%')\n",
        "elif resultado_final[0,7] > 0.50: print(f'Alterações de Mediastino (Pericardite, Malformações Arteriovenosas, Linfonodomegalias) \\nProbabilidade: {round(resultado_final[0,7]*100, 2)}%')\n",
        "elif resultado_final[0,8] > 0.50: print(f'Alterações na cavidade Torácica (Atelectasias, Malformações, Agenesia, Hipoplasias) \\nProbabilidade: {round(resultado_final[0,8]*100, 2)}%')\n"
      ]
    }
  ]
}