3.Image recognition_Other split.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Model_construction_for_image_pattern_recognition_17Nov2021+15Dec_2nd different_split_of_data.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "0pIA_GPjAZqN",
        "sTQAgmxDAuhZ",
        "F1EmdPK7A5Tj",
        "edLwUcxm0Bad",
        "yTIU2YsUChDj",
        "b-9AUv_sCzHK",
        "MBk9xGPbC3Bn"
      ],
      "toc_visible": true
    },
    "kernelspec": {
      "display_name": "Python 3 (ipykernel)",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.8.11"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/gbouzioto/chemical_polution/blob/main/notebooks/model_exploration.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0pIA_GPjAZqN"
      },
      "source": [
        "## Import Statements"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_g8VpZLbPRAZ",
        "outputId": "4d06e905-447e-4f1f-b3a8-976beb553de4"
      },
      "source": [
        "pip install tensorflow_addons"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting tensorflow_addons\n",
            "  Downloading tensorflow_addons-0.15.0-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.whl (1.1 MB)\n",
            "\u001b[?25l\r\u001b[K     |▎                               | 10 kB 22.0 MB/s eta 0:00:01\r\u001b[K     |▋                               | 20 kB 25.6 MB/s eta 0:00:01\r\u001b[K     |▉                               | 30 kB 29.8 MB/s eta 0:00:01\r\u001b[K     |█▏                              | 40 kB 31.5 MB/s eta 0:00:01\r\u001b[K     |█▌                              | 51 kB 33.0 MB/s eta 0:00:01\r\u001b[K     |█▊                              | 61 kB 34.6 MB/s eta 0:00:01\r\u001b[K     |██                              | 71 kB 29.0 MB/s eta 0:00:01\r\u001b[K     |██▍                             | 81 kB 30.8 MB/s eta 0:00:01\r\u001b[K     |██▋                             | 92 kB 31.7 MB/s eta 0:00:01\r\u001b[K     |███                             | 102 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███▎                            | 112 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███▌                            | 122 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███▉                            | 133 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████▏                           | 143 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████▍                           | 153 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████▊                           | 163 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████                           | 174 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████▎                          | 184 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████▋                          | 194 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████▉                          | 204 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████▏                         | 215 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████▌                         | 225 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████▊                         | 235 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████                         | 245 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████▍                        | 256 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████▋                        | 266 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████                        | 276 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████▎                       | 286 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████▌                       | 296 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████▉                       | 307 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████████                       | 317 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████████▍                      | 327 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████████▊                      | 337 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████                      | 348 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████▎                     | 358 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████▋                     | 368 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████▉                     | 378 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████████▏                    | 389 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████████▌                    | 399 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████████▊                    | 409 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████████                    | 419 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████████▍                   | 430 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████████▋                   | 440 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████████████                   | 450 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████████████▎                  | 460 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████████████▌                  | 471 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████████████▉                  | 481 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████████                  | 491 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████████▍                 | 501 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████████▊                 | 512 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████████████                 | 522 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████████████▎                | 532 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████████████▋                | 542 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████████████▉                | 552 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████████████▏               | 563 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████████████▌               | 573 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████████████▊               | 583 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████████████████               | 593 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████████████████▍              | 604 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████████████████▋              | 614 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████████████              | 624 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████████████▏             | 634 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████████████▌             | 645 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████████████▉             | 655 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████████████████             | 665 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████████████████▍            | 675 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████████████████▊            | 686 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████████████████            | 696 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████████████████▎           | 706 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████████████████▋           | 716 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████████████████▉           | 727 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████████████████████▏          | 737 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████████████████████▌          | 747 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████████████████████▊          | 757 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████████████████          | 768 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████████████████▎         | 778 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████████████████▋         | 788 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████████████████████         | 798 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████████████████████▏        | 808 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████████████████████▌        | 819 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████████████████████▉        | 829 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████████████████████        | 839 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████████████████████▍       | 849 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████████████████████▊       | 860 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████████████████████████       | 870 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████████████████████████▎      | 880 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████████████████████████▋      | 890 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████████████████████████▉      | 901 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████▏     | 911 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████▌     | 921 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████▊     | 931 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████████████████████████     | 942 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████████████████████████▎    | 952 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████████████████████████▋    | 962 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████████████████████████    | 972 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████████████████████████▏   | 983 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████████████████████████▌   | 993 kB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████████████████████████▉   | 1.0 MB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████████████████████████████   | 1.0 MB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████████████████████████████▍  | 1.0 MB 29.9 MB/s eta 0:00:01\r\u001b[K     |█████████████████████████████▊  | 1.0 MB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████████  | 1.0 MB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████████▎ | 1.1 MB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████████▋ | 1.1 MB 29.9 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████████▉ | 1.1 MB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████████████████████████████▏| 1.1 MB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████████████████████████████▍| 1.1 MB 29.9 MB/s eta 0:00:01\r\u001b[K     |███████████████████████████████▊| 1.1 MB 29.9 MB/s eta 0:00:01\r\u001b[K     |████████████████████████████████| 1.1 MB 29.9 MB/s \n",
            "\u001b[?25hRequirement already satisfied: typeguard>=2.7 in /usr/local/lib/python3.7/dist-packages (from tensorflow_addons) (2.7.1)\n",
            "Installing collected packages: tensorflow-addons\n",
            "Successfully installed tensorflow-addons-0.15.0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XbyZT9iGxPdK"
      },
      "source": [
        "from keras import backend as K\n",
        "def recall_m(y_true, y_pred):\n",
        "    true_positives = K.sum(K.round(K.clip(y_true * y_pred, 0, 1)))\n",
        "    possible_positives = K.sum(K.round(K.clip(y_true, 0, 1)))\n",
        "    recall = true_positives / (possible_positives + K.epsilon())\n",
        "    return recall\n",
        "\n",
        "def precision_m(y_true, y_pred):\n",
        "    true_positives = K.sum(K.round(K.clip(y_true * y_pred, 0, 1)))\n",
        "    predicted_positives = K.sum(K.round(K.clip(y_pred, 0, 1)))\n",
        "    precision = true_positives / (predicted_positives + K.epsilon())\n",
        "    return precision\n",
        "\n",
        "def f1_m(y_true, y_pred):\n",
        "    precision = precision_m(y_true, y_pred)\n",
        "    recall = recall_m(y_true, y_pred)\n",
        "    return 2*((precision*recall)/(precision+recall+K.epsilon()))"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z9qaEwiW5_KT"
      },
      "source": [
        "import os\n",
        "import numpy\n",
        "numpy.random.seed(1337)\n",
        "import urllib\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "import zipfile\n",
        "import tensorflow as tf\n",
        "import tensorflow_addons as tfa\n",
        "import os\n",
        "import PIL\n",
        "\n",
        "from tensorflow import keras\n",
        "from tensorflow.keras.applications import InceptionV3, MobileNetV2, ResNet50\n",
        "from tensorflow.keras.preprocessing import image\n",
        "from tensorflow.keras.models import Model\n",
        "from tensorflow.keras.layers import Dense, GlobalAveragePooling2D, Dropout\n",
        "from tensorflow.keras.utils import get_file\n",
        "from tensorflow import expand_dims\n",
        "from tensorflow.nn import softmax\n",
        "from tensorflow.keras.models import Sequential, save_model, load_model\n",
        "from sklearn.metrics import classification_report, confusion_matrix\n",
        "\n",
        "from tensorflow import keras\n",
        "from tensorflow.keras import layers\n",
        "from tensorflow.keras.models import Sequential\n",
        "#Tutorial https://www.tensorflow.org/tutorials/images/classification"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sTQAgmxDAuhZ"
      },
      "source": [
        "## Load Dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lOCgLHS3AuB7"
      },
      "source": [
        "project_dir = os.path.join('/', 'content')\n",
        "dataset_file = os.path.join(project_dir, 'BlackSea.zip')\n",
        "dataset_file_path = os.path.join(project_dir, 'BlackSea')\n",
        "\n",
        "if not os.path.isdir(dataset_file_path):\n",
        "    # download dataset if it does not exist\n",
        "    url = 'http://dsfp.chem.uoa.gr/BlackSea/Predicted%20spatial%20distribution%20dataset_different%20split%202.zip'\n",
        "    urllib.request.urlretrieve(url, dataset_file)\n",
        "    # extract images\n",
        "    with zipfile.ZipFile(dataset_file, 'r') as zip_ref:\n",
        "        zip_ref.extractall(dataset_file_path)\n",
        "\n",
        "dataset_train = os.path.join(dataset_file_path, 'train')\n",
        "dataset_evaluation = os.path.join(dataset_file_path, 'evaluation')"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "k3fKEBZS0BaQ",
        "outputId": "f2929160-f455-4fda-fcd9-593586cb94c9"
      },
      "source": [
        "dataset_train"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'/content/BlackSea/train'"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "c7bUOEw30BaR",
        "outputId": "e541a7b2-9cec-40a9-d3d3-48a305ffe9af"
      },
      "source": [
        "dataset_evaluation"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'/content/BlackSea/evaluation'"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "F1EmdPK7A5Tj"
      },
      "source": [
        "## Train, Validation and Test Generators"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GOzbLuAW2AG2"
      },
      "source": [
        "batch_size = 32\n",
        "img_height = 150\n",
        "img_width = 150"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lzdCiOx4B9vh"
      },
      "source": [
        "datagen = image.ImageDataGenerator(\n",
        "    rescale=1./255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    validation_split=0.2)"
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Yf1__QmcA4qa",
        "outputId": "36d20423-e076-456e-d5b4-e9654f83f67f"
      },
      "source": [
        "train_generator = datagen.flow_from_directory(\n",
        "    dataset_train, \n",
        "    target_size=(img_height, img_width), \n",
        "    batch_size=batch_size, \n",
        "    class_mode='categorical', \n",
        "    subset='training')"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 1126 images belonging to 3 classes.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "48IOOsS84H0R",
        "outputId": "d315634a-a4f6-4949-c438-a587341ab192"
      },
      "source": [
        "train_ds = tf.keras.preprocessing.image_dataset_from_directory(\n",
        "  'BlackSea/train',\n",
        "  validation_split=0.2,\n",
        "  subset=\"training\",\n",
        "  seed=123,\n",
        "  image_size=(img_height, img_width),\n",
        "  batch_size=batch_size)"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 1406 files belonging to 3 classes.\n",
            "Using 1125 files for training.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "cBmbVCBi7VxS",
        "outputId": "bf973d8d-c834-4713-ba9f-415ec39bfb79"
      },
      "source": [
        "class_names = train_ds.class_names\n",
        "print(class_names)\n",
        "\n",
        "plt.figure(figsize=(18, 18))\n",
        "for images, labels in train_ds.take(1):\n",
        "  for i in range(3):\n",
        "    ax = plt.subplot(3, 3, i + 1)\n",
        "    plt.imshow(images[i].numpy().astype(\"uint8\"))\n",
        "    plt.title(class_names[labels[i]])\n",
        "    plt.axis(\"off\")"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "['Danube', 'Dnieper', 'UnknownOrigin']\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1296x1296 with 3 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "L-i9sojx0BaU",
        "outputId": "36a30d25-45c1-429b-9e2f-e14cdfcb3359"
      },
      "source": [
        "validation_generator = datagen.flow_from_directory(\n",
        "    dataset_train, \n",
        "    target_size=(img_height, img_width), \n",
        "    batch_size=batch_size, \n",
        "    class_mode='categorical', \n",
        "    subset='validation')"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 280 images belonging to 3 classes.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WWhKPEQH4MyV",
        "outputId": "6603abc9-aedd-485b-be1e-7663519a8751"
      },
      "source": [
        "val_ds = tf.keras.preprocessing.image_dataset_from_directory(\n",
        "  'BlackSea/train',\n",
        "  validation_split=0.2,\n",
        "  subset=\"validation\",\n",
        "  seed=123,\n",
        "  image_size=(img_height, img_width),\n",
        "  batch_size=batch_size)"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 1406 files belonging to 3 classes.\n",
            "Using 281 files for validation.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PGS9G1yY7lTF"
      },
      "source": [
        "#AUTOTUNE = tf.data.AUTOTUNE\n",
        "#\n",
        "#train_ds = train_ds.cache().shuffle(1000).prefetch(buffer_size=AUTOTUNE)\n",
        "#val_ds = val_ds.cache().prefetch(buffer_size=AUTOTUNE)"
      ],
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EFlYlQW-0BaU",
        "outputId": "b0092c3d-7dac-4e2a-a2d7-a25e48641fe9"
      },
      "source": [
        "test_generator = datagen.flow_from_directory(\n",
        "    dataset_evaluation, \n",
        "    target_size=(img_height, img_height), \n",
        "    batch_size=batch_size, \n",
        "    shuffle = False,\n",
        "    class_mode='categorical')"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 224 images belonging to 3 classes.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "U1R-mmwD4ahk",
        "outputId": "042e25ca-a684-4454-de1d-6bb79a53dd3d"
      },
      "source": [
        "test_ds = tf.keras.preprocessing.image_dataset_from_directory(\n",
        "  'BlackSea/evaluation',\n",
        "  seed=123,\n",
        "  image_size=(img_height, img_width),\n",
        "  shuffle = False,\n",
        "  batch_size=batch_size)"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 224 files belonging to 3 classes.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TJSJFYZb0BaV"
      },
      "source": [
        "The number of images are correct. We have in total 1406 which were split in training set (1126 images) and validation set (280 images). Finally there are 224 images as the test set."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KCOzteRGBUAd",
        "outputId": "7fd3910a-43ef-447d-f2cd-87e2a8eedb89"
      },
      "source": [
        "test_steps_per_epoch = numpy.math.ceil(test_generator.samples / test_generator.batch_size)\n",
        "print(test_steps_per_epoch)"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "7\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AN3P4Glv7-14"
      },
      "source": [
        "num_classes = 3"
      ],
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gaNIZX4b0BaW"
      },
      "source": [
        "# 0.Baseline model implementation"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EWB7tgFm0BaW"
      },
      "source": [
        "Set the architecture of the neural network model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4L-P1FVy0BaW"
      },
      "source": [
        "model = tf.keras.models.Sequential([\n",
        "  tf.keras.layers.Flatten(),\n",
        "  tf.keras.layers.Dense(1500, activation='relu'),\n",
        "  tf.keras.layers.Dense(3, activation='softmax')\n",
        "])"
      ],
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tvckU7V90BaW"
      },
      "source": [
        "Set the loss and the optimizer function"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9tVN-cST0BaW"
      },
      "source": [
        "model.compile(loss=\"categorical_crossentropy\", optimizer=\"adam\", metrics=f1_m) #[tfa.metrics.F1Score(num_classes=3)])"
      ],
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "b0owkNts3OEr"
      },
      "source": [
        "callback = tf.keras.callbacks.EarlyStopping(monitor='f1_m', patience=10)"
      ],
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lx-k3KEE0BaX",
        "outputId": "9682285b-60bf-467f-af63-05e83f4c4a94"
      },
      "source": [
        "epochs=100\n",
        "history = model.fit(\n",
        "    train_generator,\n",
        "    validation_data=validation_generator,\n",
        "    epochs=epochs,\n",
        "    callbacks=[callback]\n",
        ")"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "36/36 [==============================] - 48s 1s/step - loss: 59.2379 - f1_m: 0.6803 - val_loss: 1.1708 - val_f1_m: 0.9306\n",
            "Epoch 2/100\n",
            "36/36 [==============================] - 31s 860ms/step - loss: 1.4324 - f1_m: 0.9444 - val_loss: 0.2418 - val_f1_m: 0.9826\n",
            "Epoch 3/100\n",
            "36/36 [==============================] - 32s 882ms/step - loss: 1.3017 - f1_m: 0.9346 - val_loss: 0.3212 - val_f1_m: 0.9792\n",
            "Epoch 4/100\n",
            "36/36 [==============================] - 31s 851ms/step - loss: 0.7339 - f1_m: 0.9575 - val_loss: 0.0713 - val_f1_m: 0.9931\n",
            "Epoch 5/100\n",
            "36/36 [==============================] - 31s 850ms/step - loss: 0.9390 - f1_m: 0.9514 - val_loss: 0.2559 - val_f1_m: 0.9896\n",
            "Epoch 6/100\n",
            "36/36 [==============================] - 32s 881ms/step - loss: 0.5350 - f1_m: 0.9606 - val_loss: 0.3439 - val_f1_m: 0.9711\n",
            "Epoch 7/100\n",
            "36/36 [==============================] - 31s 849ms/step - loss: 0.4964 - f1_m: 0.9731 - val_loss: 0.1802 - val_f1_m: 0.9931\n",
            "Epoch 8/100\n",
            "36/36 [==============================] - 31s 851ms/step - loss: 0.3460 - f1_m: 0.9792 - val_loss: 0.1483 - val_f1_m: 0.9861\n",
            "Epoch 9/100\n",
            "36/36 [==============================] - 32s 883ms/step - loss: 0.4122 - f1_m: 0.9757 - val_loss: 0.0092 - val_f1_m: 0.9907\n",
            "Epoch 10/100\n",
            "36/36 [==============================] - 31s 848ms/step - loss: 0.5122 - f1_m: 0.9609 - val_loss: 0.1176 - val_f1_m: 0.9931\n",
            "Epoch 11/100\n",
            "36/36 [==============================] - 31s 848ms/step - loss: 0.5902 - f1_m: 0.9653 - val_loss: 0.9238 - val_f1_m: 0.8993\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "26QTRdxm0BaX",
        "outputId": "80132aff-c580-4abc-d02d-cf709574ea4e"
      },
      "source": [
        "score = model.evaluate(test_generator, verbose=0)\n",
        "score"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.9017603993415833, 0.9062499403953552]"
            ]
          },
          "metadata": {},
          "execution_count": 23
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BRSE0C9-0BaX",
        "outputId": "93a17ee9-bb59-4da9-a1f9-0d627f64089e"
      },
      "source": [
        "print(\"Test loss:\", score[0])\n",
        "print(\"Test accuracy:\", score[1])"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Test loss: 0.9017603993415833\n",
            "Test accuracy: 0.9062499403953552\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dkGmNUnc6HAS",
        "outputId": "9fbe65a5-76ba-4f10-9e59-73d08f9f638a"
      },
      "source": [
        "# Save the model\n",
        "filepath = './saved_model_simple'\n",
        "save_model(model, filepath)"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: ./saved_model_simple/assets\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KxYYYFpG6Jx3",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "cfbbb14b-0600-40ce-a7e1-aa60db2ba908"
      },
      "source": [
        "# Load the model\n",
        "model = load_model(filepath, compile = False)"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "WARNING:tensorflow:Unable to restore custom metric. Please ensure that the layer implements `get_config` and `from_config` when saving. In addition, please use the `custom_objects` arg when calling `load_model()`.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 499
        },
        "id": "jr0tTEr-6N2k",
        "outputId": "02127e90-72e0-44c7-a7f5-7a91308d4038"
      },
      "source": [
        "acc = history.history['val_f1_m']\n",
        "val_acc = history.history['val_f1_m']\n",
        "\n",
        "loss = history.history['loss']\n",
        "val_loss = history.history['val_loss']\n",
        "\n",
        "epochs_range = range(len(acc))\n",
        "epochs_range\n",
        "\n",
        "plt.figure(figsize=(8, 8))\n",
        "plt.subplot(1, 2, 1)\n",
        "plt.plot(epochs_range, acc, label='Training Accuracy')\n",
        "plt.plot(epochs_range, val_acc, label='Validation Accuracy')\n",
        "plt.legend(loc='lower right')\n",
        "plt.title('Training and Validation Accuracy')\n",
        "\n",
        "\n",
        "plt.subplot(1, 2, 2)\n",
        "plt.plot(epochs_range, loss, label='Training Loss')\n",
        "plt.plot(epochs_range, val_loss, label='Validation Loss')\n",
        "plt.legend(loc='upper right')\n",
        "plt.title('Training and Validation Loss')\n",
        "plt.show()"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x576 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4af5sCR46Rgc",
        "outputId": "58306cc8-265d-4e6e-a88c-a7f757ecf44d"
      },
      "source": [
        "test_generator"
      ],
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.preprocessing.image.DirectoryIterator at 0x7fcbcd419a50>"
            ]
          },
          "metadata": {},
          "execution_count": 28
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "scrolled": true,
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Rb6reyWL0BaY",
        "outputId": "d481a636-2d72-489d-dd19-6610bd593ec4"
      },
      "source": [
        "predictions = model.predict(test_generator)\n",
        "predictions[range(5)]"
      ],
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[1.0000000e+00, 2.5107149e-13, 5.5869887e-14],\n",
              "       [9.9963140e-01, 3.6825531e-04, 3.7259065e-07],\n",
              "       [9.1976833e-01, 3.2986546e-04, 7.9901889e-02],\n",
              "       [1.0000000e+00, 6.3528862e-19, 2.8529748e-16],\n",
              "       [9.9997711e-01, 6.8303316e-06, 1.6088699e-05]], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 29
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zCoS1Wfh0BaY",
        "outputId": "264473d5-a030-4a31-a7eb-83f368f0f7e1"
      },
      "source": [
        "predicted_classes= numpy.argmax(predictions, axis=1)\n",
        "predicted_classes"
      ],
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,\n",
              "       0, 0, 2, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 1, 1, 1, 1, 2, 1,\n",
              "       2, 1, 1, 1, 2, 2, 2, 1, 1, 2, 2, 1, 1, 2, 1, 1, 1, 2, 1, 2, 1, 2,\n",
              "       1, 2, 1, 1])"
            ]
          },
          "metadata": {},
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "085EHm0V0BaY",
        "outputId": "7807866c-4a90-4c41-a98a-b6e92e1bd654"
      },
      "source": [
        "true_classes = test_generator.classes\n",
        "true_classes"
      ],
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2], dtype=int32)"
            ]
          },
          "metadata": {},
          "execution_count": 31
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pBX2s8WS0BaY",
        "outputId": "0c088944-8cf5-4205-e858-b3c8d79729a1"
      },
      "source": [
        "class_labels = list(test_generator.class_indices.keys())\n",
        "class_labels"
      ],
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['Danube', 'Dnieper', 'UnknownOrigin']"
            ]
          },
          "metadata": {},
          "execution_count": 32
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dCeN1xmx0BaZ",
        "outputId": "76f395c0-aa2c-4192-ebf8-401e2e2fe69f"
      },
      "source": [
        "report = classification_report(true_classes, predicted_classes, target_names=class_labels, zero_division=0)\n",
        "print(report)"
      ],
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "               precision    recall  f1-score   support\n",
            "\n",
            "       Danube       1.00      0.93      0.96       109\n",
            "      Dnieper       0.72      1.00      0.84        78\n",
            "UnknownOrigin       0.93      0.38      0.54        37\n",
            "\n",
            "     accuracy                           0.86       224\n",
            "    macro avg       0.89      0.77      0.78       224\n",
            " weighted avg       0.89      0.86      0.85       224\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 458
        },
        "id": "-rvHbWGU0BaZ",
        "outputId": "215948eb-21b3-4be2-c5aa-81e3a1da14e1"
      },
      "source": [
        "cmrforest = confusion_matrix(true_classes, predicted_classes)\n",
        "plt.figure(figsize=(10,7))\n",
        "sns.heatmap(cmrforest ,annot=True , fmt = 'd')\n",
        "plt.xlabel('Predicted')\n",
        "plt.ylabel('Truth')\n",
        "plt.title('Confusion Report')\n",
        "plt.show()"
      ],
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x504 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0egjiA5w0BaZ"
      },
      "source": [
        "# 1.Basic tensorflow CNN"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WEyY0lrz0Bab"
      },
      "source": [
        "model_CNN1 = Sequential([\n",
        "  layers.experimental.preprocessing.Rescaling(1./255, input_shape=(img_height, img_width, 3)),\n",
        "  layers.Conv2D(16, 3, padding='same', activation='relu'),\n",
        "  layers.MaxPooling2D(),\n",
        "  layers.Conv2D(32, 3, padding='same', activation='relu'),\n",
        "  layers.MaxPooling2D(),\n",
        "  layers.Dropout(0.2),\n",
        "  layers.Flatten(),\n",
        "  layers.Dense(128, activation='relu'),\n",
        "  layers.Dense(num_classes)\n",
        "])"
      ],
      "execution_count": 35,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-iRqtbGO0Bab"
      },
      "source": [
        "model_CNN1.compile(optimizer='adam',\n",
        "              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
        "               metrics=['acc', f1_m, precision_m, recall_m])"
      ],
      "execution_count": 36,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "emVI8BQU0Bab",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8b348f28-afb1-4ad4-9f03-5a534a660cc2"
      },
      "source": [
        "model_CNN1.summary()"
      ],
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " rescaling (Rescaling)       (None, 150, 150, 3)       0         \n",
            "                                                                 \n",
            " conv2d (Conv2D)             (None, 150, 150, 16)      448       \n",
            "                                                                 \n",
            " max_pooling2d (MaxPooling2D  (None, 75, 75, 16)       0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " conv2d_1 (Conv2D)           (None, 75, 75, 32)        4640      \n",
            "                                                                 \n",
            " max_pooling2d_1 (MaxPooling  (None, 37, 37, 32)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 37, 37, 32)        0         \n",
            "                                                                 \n",
            " flatten_1 (Flatten)         (None, 43808)             0         \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 128)               5607552   \n",
            "                                                                 \n",
            " dense_3 (Dense)             (None, 3)                 387       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 5,613,027\n",
            "Trainable params: 5,613,027\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2Iyobyk_0Bac",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "da347541-c6a4-4e33-d0dd-76d8dace8d05"
      },
      "source": [
        "history_CNN1 = model_CNN1.fit(train_ds,\n",
        "                    validation_data=val_ds,\n",
        "                    callbacks=[callback],\n",
        "                    epochs=epochs)"
      ],
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "36/36 [==============================] - 24s 633ms/step - loss: 0.6758 - acc: 0.8462 - f1_m: 0.7455 - precision_m: 0.5666 - recall_m: 1.1285 - val_loss: 0.0259 - val_acc: 0.9929 - val_f1_m: 0.6879 - val_precision_m: 0.5156 - val_recall_m: 1.0533\n",
            "Epoch 2/100\n",
            "36/36 [==============================] - 23s 624ms/step - loss: 0.0500 - acc: 0.9884 - f1_m: 0.7071 - precision_m: 0.5476 - recall_m: 1.0127 - val_loss: 0.0281 - val_acc: 0.9893 - val_f1_m: 0.6686 - val_precision_m: 0.5061 - val_recall_m: 1.0000\n",
            "Epoch 3/100\n",
            "36/36 [==============================] - 25s 666ms/step - loss: 0.0461 - acc: 0.9867 - f1_m: 0.7115 - precision_m: 0.5544 - recall_m: 1.0058 - val_loss: 0.0148 - val_acc: 0.9964 - val_f1_m: 0.6639 - val_precision_m: 0.5067 - val_recall_m: 0.9926\n",
            "Epoch 4/100\n",
            "36/36 [==============================] - 23s 627ms/step - loss: 0.0456 - acc: 0.9858 - f1_m: 0.7012 - precision_m: 0.5435 - recall_m: 1.0090 - val_loss: 0.0147 - val_acc: 0.9929 - val_f1_m: 0.6663 - val_precision_m: 0.5045 - val_recall_m: 1.0134\n",
            "Epoch 5/100\n",
            "36/36 [==============================] - 23s 626ms/step - loss: 0.0401 - acc: 0.9884 - f1_m: 0.7069 - precision_m: 0.5536 - recall_m: 1.0015 - val_loss: 0.0148 - val_acc: 0.9929 - val_f1_m: 0.6684 - val_precision_m: 0.5024 - val_recall_m: 1.0060\n",
            "Epoch 6/100\n",
            "36/36 [==============================] - 23s 624ms/step - loss: 0.0537 - acc: 0.9822 - f1_m: 0.7112 - precision_m: 0.5508 - recall_m: 1.0108 - val_loss: 0.0167 - val_acc: 1.0000 - val_f1_m: 0.6655 - val_precision_m: 0.5054 - val_recall_m: 0.9926\n",
            "Epoch 7/100\n",
            "36/36 [==============================] - 23s 624ms/step - loss: 0.0446 - acc: 0.9893 - f1_m: 0.6994 - precision_m: 0.5430 - recall_m: 0.9972 - val_loss: 0.0267 - val_acc: 0.9964 - val_f1_m: 0.6739 - val_precision_m: 0.5152 - val_recall_m: 0.9842\n",
            "Epoch 8/100\n",
            "36/36 [==============================] - 23s 626ms/step - loss: 0.0522 - acc: 0.9858 - f1_m: 0.7111 - precision_m: 0.5507 - recall_m: 1.0148 - val_loss: 0.0393 - val_acc: 0.9893 - val_f1_m: 0.6696 - val_precision_m: 0.5078 - val_recall_m: 1.0284\n",
            "Epoch 9/100\n",
            "36/36 [==============================] - 23s 625ms/step - loss: 0.0539 - acc: 0.9867 - f1_m: 0.6967 - precision_m: 0.5405 - recall_m: 0.9974 - val_loss: 0.0134 - val_acc: 0.9964 - val_f1_m: 0.6632 - val_precision_m: 0.5039 - val_recall_m: 0.9907\n",
            "Epoch 10/100\n",
            "36/36 [==============================] - 23s 624ms/step - loss: 0.0392 - acc: 0.9876 - f1_m: 0.6918 - precision_m: 0.5364 - recall_m: 0.9970 - val_loss: 0.0092 - val_acc: 0.9964 - val_f1_m: 0.6622 - val_precision_m: 0.5003 - val_recall_m: 1.0003\n",
            "Epoch 11/100\n",
            "36/36 [==============================] - 23s 628ms/step - loss: 0.0390 - acc: 0.9858 - f1_m: 0.6982 - precision_m: 0.5436 - recall_m: 0.9985 - val_loss: 0.0119 - val_acc: 0.9964 - val_f1_m: 0.6591 - val_precision_m: 0.4976 - val_recall_m: 1.0014\n",
            "Epoch 12/100\n",
            "36/36 [==============================] - 23s 626ms/step - loss: 0.0413 - acc: 0.9867 - f1_m: 0.7078 - precision_m: 0.5531 - recall_m: 1.0035 - val_loss: 0.0140 - val_acc: 0.9964 - val_f1_m: 0.6579 - val_precision_m: 0.4983 - val_recall_m: 1.0108\n",
            "Epoch 13/100\n",
            "36/36 [==============================] - 23s 627ms/step - loss: 0.0400 - acc: 0.9858 - f1_m: 0.7102 - precision_m: 0.5510 - recall_m: 1.0074 - val_loss: 0.0150 - val_acc: 0.9964 - val_f1_m: 0.6644 - val_precision_m: 0.5048 - val_recall_m: 1.0033\n",
            "Epoch 14/100\n",
            "36/36 [==============================] - 23s 626ms/step - loss: 0.0362 - acc: 0.9884 - f1_m: 0.6995 - precision_m: 0.5460 - recall_m: 0.9993 - val_loss: 0.0111 - val_acc: 0.9964 - val_f1_m: 0.6653 - val_precision_m: 0.5012 - val_recall_m: 1.0000\n",
            "Epoch 15/100\n",
            "36/36 [==============================] - 23s 626ms/step - loss: 0.0380 - acc: 0.9893 - f1_m: 0.6957 - precision_m: 0.5403 - recall_m: 0.9991 - val_loss: 0.0137 - val_acc: 0.9964 - val_f1_m: 0.6619 - val_precision_m: 0.5033 - val_recall_m: 1.0065\n",
            "Epoch 16/100\n",
            "36/36 [==============================] - 23s 626ms/step - loss: 0.0402 - acc: 0.9893 - f1_m: 0.7060 - precision_m: 0.5496 - recall_m: 1.0049 - val_loss: 0.0183 - val_acc: 0.9964 - val_f1_m: 0.6605 - val_precision_m: 0.4992 - val_recall_m: 1.0000\n",
            "Epoch 17/100\n",
            "36/36 [==============================] - 23s 622ms/step - loss: 0.0355 - acc: 0.9876 - f1_m: 0.7009 - precision_m: 0.5440 - recall_m: 0.9991 - val_loss: 0.0088 - val_acc: 0.9964 - val_f1_m: 0.6591 - val_precision_m: 0.4965 - val_recall_m: 0.9931\n",
            "Epoch 18/100\n",
            "36/36 [==============================] - 23s 625ms/step - loss: 0.0307 - acc: 0.9893 - f1_m: 0.6988 - precision_m: 0.5443 - recall_m: 0.9963 - val_loss: 0.0129 - val_acc: 0.9964 - val_f1_m: 0.6657 - val_precision_m: 0.5055 - val_recall_m: 1.0062\n",
            "Epoch 19/100\n",
            "36/36 [==============================] - 23s 630ms/step - loss: 0.0362 - acc: 0.9876 - f1_m: 0.7015 - precision_m: 0.5465 - recall_m: 0.9959 - val_loss: 0.0104 - val_acc: 0.9964 - val_f1_m: 0.6601 - val_precision_m: 0.4978 - val_recall_m: 0.9988\n",
            "Epoch 20/100\n",
            "36/36 [==============================] - 23s 625ms/step - loss: 0.0341 - acc: 0.9884 - f1_m: 0.6988 - precision_m: 0.5428 - recall_m: 0.9943 - val_loss: 0.0093 - val_acc: 0.9964 - val_f1_m: 0.6563 - val_precision_m: 0.4977 - val_recall_m: 0.9899\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vr39Ffrf0goB",
        "outputId": "9b008a93-e392-4fbc-ad4c-e83093055658"
      },
      "source": [
        "score = model_CNN1.evaluate(test_ds, verbose=0)\n",
        "print(\"Test loss:\", score[0])\n",
        "print(\"Test accuracy:\", score[1])\n",
        "\n",
        "\n",
        "# Save the model\n",
        "filepath = './model_CNN1'\n",
        "save_model(model, filepath)\n",
        "\n",
        "# Load the model\n",
        "model = load_model(filepath, compile = False)"
      ],
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Test loss: 0.025816572830080986\n",
            "Test accuracy: 0.9955357313156128\n",
            "WARNING:tensorflow:Compiled the loaded model, but the compiled metrics have yet to be built. `model.compile_metrics` will be empty until you train or evaluate the model.\n",
            "INFO:tensorflow:Assets written to: ./model_CNN1/assets\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sOt2_T1F-vAJ",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 499
        },
        "outputId": "5b2cb629-fa2d-4539-d9de-5989459357ee"
      },
      "source": [
        "acc = history_CNN1.history['f1_m']\n",
        "val_acc = history_CNN1.history['val_f1_m']\n",
        "\n",
        "loss = history_CNN1.history['loss']\n",
        "val_loss = history_CNN1.history['val_loss']\n",
        "\n",
        "epochs_range = range(len(acc))\n",
        "\n",
        "plt.figure(figsize=(8, 8))\n",
        "plt.subplot(1, 2, 1)\n",
        "plt.plot(epochs_range, acc, label='Training Accuracy')\n",
        "plt.plot(epochs_range, val_acc, label='Validation Accuracy')\n",
        "plt.legend(loc='lower right')\n",
        "plt.title('Training and Validation Accuracy')\n",
        "\n",
        "plt.subplot(1, 2, 2)\n",
        "plt.plot(epochs_range, loss, label='Training Loss')\n",
        "plt.plot(epochs_range, val_loss, label='Validation Loss')\n",
        "plt.legend(loc='upper right')\n",
        "plt.title('Training and Validation Loss')\n",
        "plt.show()"
      ],
      "execution_count": 40,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x576 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9QnrNrEr0Bac",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "409cccae-046a-42a3-f107-7493647cd24b"
      },
      "source": [
        "predictions = model_CNN1.predict(test_ds, steps=test_steps_per_epoch)\n",
        "# Get most likely class\n",
        "predicted_classes = numpy.argmax(predictions, axis=1)\n",
        "predicted_classes"
      ],
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2])"
            ]
          },
          "metadata": {},
          "execution_count": 41
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JzK37VbIE2bW",
        "outputId": "52af2be9-0a42-4520-e0a3-8ca71a4cb3c3"
      },
      "source": [
        "true_classes = test_generator.classes #true_classes = test_ds.class_names\n",
        "true_classes"
      ],
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2], dtype=int32)"
            ]
          },
          "metadata": {},
          "execution_count": 42
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vM5U0Mc90Bac",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c268f89e-0b06-4496-f4ff-b7da89bc6913"
      },
      "source": [
        "class_labels = list(test_generator.class_indices.keys())\n",
        "class_labels"
      ],
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['Danube', 'Dnieper', 'UnknownOrigin']"
            ]
          },
          "metadata": {},
          "execution_count": 43
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ItJdCNnF0Bac",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 631
        },
        "outputId": "0e87bcc8-b399-43e9-fe97-76ec88773206"
      },
      "source": [
        "report = classification_report(true_classes, predicted_classes, target_names=class_labels, zero_division=0)\n",
        "print(report)\n",
        "cmrforest = confusion_matrix(true_classes, predicted_classes)\n",
        "plt.figure(figsize=(10,7))\n",
        "sns.heatmap(cmrforest ,annot=True , fmt = 'd')\n",
        "plt.xlabel('Predicted')\n",
        "plt.ylabel('Truth')\n",
        "plt.title('Confusion Report')\n",
        "plt.show()"
      ],
      "execution_count": 44,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "               precision    recall  f1-score   support\n",
            "\n",
            "       Danube       1.00      1.00      1.00       109\n",
            "      Dnieper       1.00      0.99      0.99        78\n",
            "UnknownOrigin       0.97      1.00      0.99        37\n",
            "\n",
            "     accuracy                           1.00       224\n",
            "    macro avg       0.99      1.00      0.99       224\n",
            " weighted avg       1.00      1.00      1.00       224\n",
            "\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x504 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sv8vSqWOFlba"
      },
      "source": [
        "# 2.VGG Net"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "c3FmYSoeFbYS",
        "outputId": "1ac23312-3e8f-4813-862d-0b74858c61b0"
      },
      "source": [
        "from keras.models import Sequential\n",
        "from keras.layers import Dense, Conv2D, MaxPooling2D\n",
        "from keras.layers import Flatten, Activation\n",
        "\n",
        "custom_vgg = Sequential([\n",
        "layers.Conv2D(32, (3, 3), strides = 1, padding = \"same\", activation = \"relu\", input_shape = (img_height, img_width, 3)),\n",
        "layers.Dropout(0.4),\n",
        "layers.Conv2D(32, (3, 3), strides = 1, padding = \"same\", activation = \"relu\"),\n",
        "layers.Dropout(0.4),\n",
        "layers.MaxPooling2D((2, 2)),\n",
        "\n",
        "layers.Conv2D(64, (3, 3), strides = 1, padding = \"same\", activation = \"relu\"),\n",
        "layers.Dropout(0.4),\n",
        "layers.Conv2D(64, (3, 3), strides = 1, padding = \"same\", activation = \"relu\"),\n",
        "layers.Dropout(0.4),\n",
        "layers.MaxPooling2D((2, 2)),\n",
        "\n",
        "layers.Conv2D(128, (3, 3), strides = 1, padding = \"same\", activation = \"relu\"),\n",
        "layers.Dropout(0.4),\n",
        "layers.Conv2D(128, (3, 3), strides = 1, padding = \"same\", activation = \"relu\"),\n",
        "layers.Dropout(0.4),\n",
        "layers.MaxPooling2D((2, 2)),\n",
        "\n",
        "layers.Flatten(),\n",
        "layers.Dense(num_classes, activation = \"relu\")\n",
        "])\n",
        "\n",
        "custom_vgg.compile(optimizer = \"adam\",\n",
        "                   loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
        "                   metrics=['acc', f1_m, precision_m, recall_m])\n",
        "\n",
        "custom_vgg.summary()"
      ],
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_2\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " conv2d_2 (Conv2D)           (None, 150, 150, 32)      896       \n",
            "                                                                 \n",
            " dropout_1 (Dropout)         (None, 150, 150, 32)      0         \n",
            "                                                                 \n",
            " conv2d_3 (Conv2D)           (None, 150, 150, 32)      9248      \n",
            "                                                                 \n",
            " dropout_2 (Dropout)         (None, 150, 150, 32)      0         \n",
            "                                                                 \n",
            " max_pooling2d_2 (MaxPooling  (None, 75, 75, 32)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_4 (Conv2D)           (None, 75, 75, 64)        18496     \n",
            "                                                                 \n",
            " dropout_3 (Dropout)         (None, 75, 75, 64)        0         \n",
            "                                                                 \n",
            " conv2d_5 (Conv2D)           (None, 75, 75, 64)        36928     \n",
            "                                                                 \n",
            " dropout_4 (Dropout)         (None, 75, 75, 64)        0         \n",
            "                                                                 \n",
            " max_pooling2d_3 (MaxPooling  (None, 37, 37, 64)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_6 (Conv2D)           (None, 37, 37, 128)       73856     \n",
            "                                                                 \n",
            " dropout_5 (Dropout)         (None, 37, 37, 128)       0         \n",
            "                                                                 \n",
            " conv2d_7 (Conv2D)           (None, 37, 37, 128)       147584    \n",
            "                                                                 \n",
            " dropout_6 (Dropout)         (None, 37, 37, 128)       0         \n",
            "                                                                 \n",
            " max_pooling2d_4 (MaxPooling  (None, 18, 18, 128)      0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " flatten_2 (Flatten)         (None, 41472)             0         \n",
            "                                                                 \n",
            " dense_4 (Dense)             (None, 3)                 124419    \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 411,427\n",
            "Trainable params: 411,427\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "L54jI3bxGaFy",
        "outputId": "fa70e580-b354-4f2b-f97c-e55d01f60125"
      },
      "source": [
        "history_vgg = custom_vgg.fit(train_ds,\n",
        "                    validation_data=val_ds,\n",
        "                    callbacks=[callback],\n",
        "                    epochs=epochs)"
      ],
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "36/36 [==============================] - 193s 5s/step - loss: 97.2803 - acc: 0.4356 - f1_m: 0.0569 - precision_m: 0.0414 - recall_m: 0.0920 - val_loss: 1.0986 - val_acc: 0.4982 - val_f1_m: 0.0000e+00 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
            "Epoch 2/100\n",
            "36/36 [==============================] - 191s 5s/step - loss: 1.0986 - acc: 0.4551 - f1_m: 0.0000e+00 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.0986 - val_acc: 0.4982 - val_f1_m: 0.0000e+00 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
            "Epoch 3/100\n",
            "36/36 [==============================] - 194s 5s/step - loss: 1.0986 - acc: 0.4551 - f1_m: 0.0000e+00 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.0986 - val_acc: 0.4982 - val_f1_m: 0.0000e+00 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
            "Epoch 4/100\n",
            "36/36 [==============================] - 193s 5s/step - loss: 1.0986 - acc: 0.4551 - f1_m: 0.0000e+00 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.0986 - val_acc: 0.4982 - val_f1_m: 0.0000e+00 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
            "Epoch 5/100\n",
            "36/36 [==============================] - 193s 5s/step - loss: 1.0986 - acc: 0.4551 - f1_m: 0.0000e+00 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.0986 - val_acc: 0.4982 - val_f1_m: 0.0000e+00 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
            "Epoch 6/100\n",
            "36/36 [==============================] - 195s 5s/step - loss: 1.0986 - acc: 0.4551 - f1_m: 0.0000e+00 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.0986 - val_acc: 0.4982 - val_f1_m: 0.0000e+00 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
            "Epoch 7/100\n",
            "36/36 [==============================] - 193s 5s/step - loss: 1.0986 - acc: 0.4551 - f1_m: 0.0000e+00 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.0986 - val_acc: 0.4982 - val_f1_m: 0.0000e+00 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
            "Epoch 8/100\n",
            "36/36 [==============================] - 193s 5s/step - loss: 1.0986 - acc: 0.4551 - f1_m: 0.0000e+00 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.0986 - val_acc: 0.4982 - val_f1_m: 0.0000e+00 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
            "Epoch 9/100\n",
            "36/36 [==============================] - 193s 5s/step - loss: 1.0986 - acc: 0.4551 - f1_m: 0.0000e+00 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.0986 - val_acc: 0.4982 - val_f1_m: 0.0000e+00 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
            "Epoch 10/100\n",
            "36/36 [==============================] - 194s 5s/step - loss: 1.0986 - acc: 0.4551 - f1_m: 0.0000e+00 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.0986 - val_acc: 0.4982 - val_f1_m: 0.0000e+00 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
            "Epoch 11/100\n",
            "36/36 [==============================] - 194s 5s/step - loss: 1.0986 - acc: 0.4551 - f1_m: 0.0000e+00 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.0986 - val_acc: 0.4982 - val_f1_m: 0.0000e+00 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n",
            "Epoch 12/100\n",
            "36/36 [==============================] - 194s 5s/step - loss: 1.0986 - acc: 0.4551 - f1_m: 0.0000e+00 - precision_m: 0.0000e+00 - recall_m: 0.0000e+00 - val_loss: 1.0986 - val_acc: 0.4982 - val_f1_m: 0.0000e+00 - val_precision_m: 0.0000e+00 - val_recall_m: 0.0000e+00\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Od15Nopy46pK",
        "outputId": "daa8b148-b271-41f0-ab22-b06c8ae9cc83"
      },
      "source": [
        "score = custom_vgg.evaluate(test_ds, verbose=0)\n",
        "print(\"Test loss:\", score[0])\n",
        "print(\"Test accuracy:\", score[1])\n",
        "\n",
        "\n",
        "# Save the model\n",
        "filepath = './model_vgg'\n",
        "save_model(model, filepath)\n",
        "\n",
        "# Load the model\n",
        "model = load_model(filepath, compile = False)"
      ],
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Test loss: 1.0986123085021973\n",
            "Test accuracy: 0.4866071343421936\n",
            "WARNING:tensorflow:Compiled the loaded model, but the compiled metrics have yet to be built. `model.compile_metrics` will be empty until you train or evaluate the model.\n",
            "INFO:tensorflow:Assets written to: ./model_vgg/assets\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YEWLTcwWMLuj"
      },
      "source": [
        "acc = history_vgg.history['f1_m']\n",
        "val_acc = history_vgg.history['val_f1_m']\n",
        "\n",
        "loss = history_vgg.history['loss']\n",
        "val_loss = history_vgg.history['val_loss']\n",
        "\n",
        "epochs_range = range(len(acc))"
      ],
      "execution_count": 48,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 499
        },
        "id": "VqB0QBPWML6l",
        "outputId": "e47d6d31-0f83-4cd7-9297-ea449d4355a2"
      },
      "source": [
        "plt.figure(figsize=(8, 8))\n",
        "plt.subplot(1, 2, 1)\n",
        "plt.plot(epochs_range, acc, label='Training Accuracy')\n",
        "plt.plot(epochs_range, val_acc, label='Validation Accuracy')\n",
        "plt.legend(loc='lower right')\n",
        "plt.title('Training and Validation Accuracy')\n",
        "\n",
        "plt.subplot(1, 2, 2)\n",
        "plt.plot(epochs_range, loss, label='Training Loss')\n",
        "plt.plot(epochs_range, val_loss, label='Validation Loss')\n",
        "plt.legend(loc='upper right')\n",
        "plt.title('Training and Validation Loss')\n",
        "plt.show()"
      ],
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x576 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8BanTOWLML-T",
        "outputId": "ccf0d509-ee88-4322-cd9a-64d1bcaec48c"
      },
      "source": [
        "predictions = custom_vgg.predict(test_ds, steps=test_steps_per_epoch)\n",
        "# Get most likely class\n",
        "predicted_classes = numpy.argmax(predictions, axis=1)\n",
        "predicted_classes"
      ],
      "execution_count": 50,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0])"
            ]
          },
          "metadata": {},
          "execution_count": 50
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vrh0A_29MMBR",
        "outputId": "11b1a908-f0fa-4328-a653-30ac615e01bc"
      },
      "source": [
        "true_classes = test_generator.classes #true_classes = test_ds.class_names\n",
        "true_classes"
      ],
      "execution_count": 51,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2], dtype=int32)"
            ]
          },
          "metadata": {},
          "execution_count": 51
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sZGWJorbMMEM",
        "outputId": "20b7f9a7-4b4a-4a26-f600-d4a8acd723f7"
      },
      "source": [
        "class_labels = list(test_generator.class_indices.keys())\n",
        "class_labels"
      ],
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['Danube', 'Dnieper', 'UnknownOrigin']"
            ]
          },
          "metadata": {},
          "execution_count": 52
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 631
        },
        "id": "hrbbFfhlMa8O",
        "outputId": "c6e43482-8ca3-4948-c9f5-f116f2de7760"
      },
      "source": [
        "report = classification_report(true_classes, predicted_classes, target_names=class_labels, zero_division=0)\n",
        "print(report)\n",
        "cmrforest = confusion_matrix(true_classes, predicted_classes)\n",
        "plt.figure(figsize=(10,7))\n",
        "sns.heatmap(cmrforest ,annot=True , fmt = 'd')\n",
        "plt.xlabel('Predicted')\n",
        "plt.ylabel('Truth')\n",
        "plt.title('Confusion Report')\n",
        "plt.show()"
      ],
      "execution_count": 53,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "               precision    recall  f1-score   support\n",
            "\n",
            "       Danube       0.49      1.00      0.65       109\n",
            "      Dnieper       0.00      0.00      0.00        78\n",
            "UnknownOrigin       0.00      0.00      0.00        37\n",
            "\n",
            "     accuracy                           0.49       224\n",
            "    macro avg       0.16      0.33      0.22       224\n",
            " weighted avg       0.24      0.49      0.32       224\n",
            "\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x504 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VLUElIA4NMj7"
      },
      "source": [
        "# 3.ResNet"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bTgp61f3NRff"
      },
      "source": [
        "stride = 1\n",
        "CHANNEL_AXIS = 3\n",
        "\n",
        "def res_layer(x, filters, pooling = False, dropout = 0.0):\n",
        "    temp = x\n",
        "    temp = tf.keras.layers.Conv2D(filters,(3,3),strides = stride,padding = \"same\")(temp)\n",
        "    temp = tf.keras.layers.BatchNormalization(axis = CHANNEL_AXIS)(temp)\n",
        "    temp = tf.keras.layers.Activation(\"relu\")(temp)\n",
        "    temp = tf.keras.layers.Conv2D(filters,(3,3),strides = stride,padding = \"same\")(temp)\n",
        "\n",
        "    x = tf.keras.layers.add([temp,tf.keras.layers.Conv2D(filters,(3,3),strides = stride,padding = \"same\")(x)])\n",
        "    if pooling:\n",
        "        x = tf.keras.layers.MaxPooling2D((2,2))(x)\n",
        "    if dropout != 0.0:\n",
        "        x = Dropout(dropout)(x)\n",
        "    x = tf.keras.layers.BatchNormalization(axis = CHANNEL_AXIS)(x)\n",
        "    x = tf.keras.layers.Activation(\"relu\")(x)\n",
        "    return x"
      ],
      "execution_count": 54,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "I_PLanL9NRoj",
        "outputId": "2a0e3471-cca4-4502-9149-b8f6c92449f1"
      },
      "source": [
        "inp = tf.keras.Input(shape = (img_height, img_width,3))\n",
        "x = inp\n",
        "x = tf.keras.layers.Conv2D(16,(3,3),strides = stride,padding = \"same\")(x)\n",
        "x = tf.keras.layers.BatchNormalization(axis = CHANNEL_AXIS)(x)\n",
        "x = tf.keras.layers.Activation(\"relu\")(x)\n",
        "\n",
        "x = res_layer(x,8,dropout = 0.2)\n",
        "x = res_layer(x,8,dropout = 0.3)\n",
        "x = res_layer(x,8,dropout = 0.4,pooling = True)\n",
        "x = res_layer(x,16,dropout = 0.2)\n",
        "x = res_layer(x,16,dropout = 0.2,pooling = True)\n",
        "x = res_layer(x,64,dropout = 0.4)\n",
        "x = tf.keras.layers.Flatten()(x)\n",
        "x = tf.keras.layers.Dropout(0.4)(x)\n",
        "x = tf.keras.layers.Dense(256,activation = \"relu\")(x)\n",
        "x = tf.keras.layers.Dropout(0.23)(x)\n",
        "x = tf.keras.layers.Dense(num_classes)(x)\n",
        "\n",
        "resnet_model = Model(inp,x,name = \"Resnet\")\n",
        "resnet_model.summary()"
      ],
      "execution_count": 55,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"Resnet\"\n",
            "__________________________________________________________________________________________________\n",
            " Layer (type)                   Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            " input_1 (InputLayer)           [(None, 150, 150, 3  0           []                               \n",
            "                                )]                                                                \n",
            "                                                                                                  \n",
            " conv2d_8 (Conv2D)              (None, 150, 150, 16  448         ['input_1[0][0]']                \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " batch_normalization (BatchNorm  (None, 150, 150, 16  64         ['conv2d_8[0][0]']               \n",
            " alization)                     )                                                                 \n",
            "                                                                                                  \n",
            " activation (Activation)        (None, 150, 150, 16  0           ['batch_normalization[0][0]']    \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_9 (Conv2D)              (None, 150, 150, 8)  1160        ['activation[0][0]']             \n",
            "                                                                                                  \n",
            " batch_normalization_1 (BatchNo  (None, 150, 150, 8)  32         ['conv2d_9[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " activation_1 (Activation)      (None, 150, 150, 8)  0           ['batch_normalization_1[0][0]']  \n",
            "                                                                                                  \n",
            " conv2d_10 (Conv2D)             (None, 150, 150, 8)  584         ['activation_1[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_11 (Conv2D)             (None, 150, 150, 8)  1160        ['activation[0][0]']             \n",
            "                                                                                                  \n",
            " add (Add)                      (None, 150, 150, 8)  0           ['conv2d_10[0][0]',              \n",
            "                                                                  'conv2d_11[0][0]']              \n",
            "                                                                                                  \n",
            " dropout_7 (Dropout)            (None, 150, 150, 8)  0           ['add[0][0]']                    \n",
            "                                                                                                  \n",
            " batch_normalization_2 (BatchNo  (None, 150, 150, 8)  32         ['dropout_7[0][0]']              \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " activation_2 (Activation)      (None, 150, 150, 8)  0           ['batch_normalization_2[0][0]']  \n",
            "                                                                                                  \n",
            " conv2d_12 (Conv2D)             (None, 150, 150, 8)  584         ['activation_2[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_3 (BatchNo  (None, 150, 150, 8)  32         ['conv2d_12[0][0]']              \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " activation_3 (Activation)      (None, 150, 150, 8)  0           ['batch_normalization_3[0][0]']  \n",
            "                                                                                                  \n",
            " conv2d_13 (Conv2D)             (None, 150, 150, 8)  584         ['activation_3[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_14 (Conv2D)             (None, 150, 150, 8)  584         ['activation_2[0][0]']           \n",
            "                                                                                                  \n",
            " add_1 (Add)                    (None, 150, 150, 8)  0           ['conv2d_13[0][0]',              \n",
            "                                                                  'conv2d_14[0][0]']              \n",
            "                                                                                                  \n",
            " dropout_8 (Dropout)            (None, 150, 150, 8)  0           ['add_1[0][0]']                  \n",
            "                                                                                                  \n",
            " batch_normalization_4 (BatchNo  (None, 150, 150, 8)  32         ['dropout_8[0][0]']              \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " activation_4 (Activation)      (None, 150, 150, 8)  0           ['batch_normalization_4[0][0]']  \n",
            "                                                                                                  \n",
            " conv2d_15 (Conv2D)             (None, 150, 150, 8)  584         ['activation_4[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_5 (BatchNo  (None, 150, 150, 8)  32         ['conv2d_15[0][0]']              \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " activation_5 (Activation)      (None, 150, 150, 8)  0           ['batch_normalization_5[0][0]']  \n",
            "                                                                                                  \n",
            " conv2d_16 (Conv2D)             (None, 150, 150, 8)  584         ['activation_5[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_17 (Conv2D)             (None, 150, 150, 8)  584         ['activation_4[0][0]']           \n",
            "                                                                                                  \n",
            " add_2 (Add)                    (None, 150, 150, 8)  0           ['conv2d_16[0][0]',              \n",
            "                                                                  'conv2d_17[0][0]']              \n",
            "                                                                                                  \n",
            " max_pooling2d_5 (MaxPooling2D)  (None, 75, 75, 8)   0           ['add_2[0][0]']                  \n",
            "                                                                                                  \n",
            " dropout_9 (Dropout)            (None, 75, 75, 8)    0           ['max_pooling2d_5[0][0]']        \n",
            "                                                                                                  \n",
            " batch_normalization_6 (BatchNo  (None, 75, 75, 8)   32          ['dropout_9[0][0]']              \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " activation_6 (Activation)      (None, 75, 75, 8)    0           ['batch_normalization_6[0][0]']  \n",
            "                                                                                                  \n",
            " conv2d_18 (Conv2D)             (None, 75, 75, 16)   1168        ['activation_6[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_7 (BatchNo  (None, 75, 75, 16)  64          ['conv2d_18[0][0]']              \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " activation_7 (Activation)      (None, 75, 75, 16)   0           ['batch_normalization_7[0][0]']  \n",
            "                                                                                                  \n",
            " conv2d_19 (Conv2D)             (None, 75, 75, 16)   2320        ['activation_7[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_20 (Conv2D)             (None, 75, 75, 16)   1168        ['activation_6[0][0]']           \n",
            "                                                                                                  \n",
            " add_3 (Add)                    (None, 75, 75, 16)   0           ['conv2d_19[0][0]',              \n",
            "                                                                  'conv2d_20[0][0]']              \n",
            "                                                                                                  \n",
            " dropout_10 (Dropout)           (None, 75, 75, 16)   0           ['add_3[0][0]']                  \n",
            "                                                                                                  \n",
            " batch_normalization_8 (BatchNo  (None, 75, 75, 16)  64          ['dropout_10[0][0]']             \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " activation_8 (Activation)      (None, 75, 75, 16)   0           ['batch_normalization_8[0][0]']  \n",
            "                                                                                                  \n",
            " conv2d_21 (Conv2D)             (None, 75, 75, 16)   2320        ['activation_8[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_9 (BatchNo  (None, 75, 75, 16)  64          ['conv2d_21[0][0]']              \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " activation_9 (Activation)      (None, 75, 75, 16)   0           ['batch_normalization_9[0][0]']  \n",
            "                                                                                                  \n",
            " conv2d_22 (Conv2D)             (None, 75, 75, 16)   2320        ['activation_9[0][0]']           \n",
            "                                                                                                  \n",
            " conv2d_23 (Conv2D)             (None, 75, 75, 16)   2320        ['activation_8[0][0]']           \n",
            "                                                                                                  \n",
            " add_4 (Add)                    (None, 75, 75, 16)   0           ['conv2d_22[0][0]',              \n",
            "                                                                  'conv2d_23[0][0]']              \n",
            "                                                                                                  \n",
            " max_pooling2d_6 (MaxPooling2D)  (None, 37, 37, 16)  0           ['add_4[0][0]']                  \n",
            "                                                                                                  \n",
            " dropout_11 (Dropout)           (None, 37, 37, 16)   0           ['max_pooling2d_6[0][0]']        \n",
            "                                                                                                  \n",
            " batch_normalization_10 (BatchN  (None, 37, 37, 16)  64          ['dropout_11[0][0]']             \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_10 (Activation)     (None, 37, 37, 16)   0           ['batch_normalization_10[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_24 (Conv2D)             (None, 37, 37, 64)   9280        ['activation_10[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_11 (BatchN  (None, 37, 37, 64)  256         ['conv2d_24[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_11 (Activation)     (None, 37, 37, 64)   0           ['batch_normalization_11[0][0]'] \n",
            "                                                                                                  \n",
            " conv2d_25 (Conv2D)             (None, 37, 37, 64)   36928       ['activation_11[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_26 (Conv2D)             (None, 37, 37, 64)   9280        ['activation_10[0][0]']          \n",
            "                                                                                                  \n",
            " add_5 (Add)                    (None, 37, 37, 64)   0           ['conv2d_25[0][0]',              \n",
            "                                                                  'conv2d_26[0][0]']              \n",
            "                                                                                                  \n",
            " dropout_12 (Dropout)           (None, 37, 37, 64)   0           ['add_5[0][0]']                  \n",
            "                                                                                                  \n",
            " batch_normalization_12 (BatchN  (None, 37, 37, 64)  256         ['dropout_12[0][0]']             \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_12 (Activation)     (None, 37, 37, 64)   0           ['batch_normalization_12[0][0]'] \n",
            "                                                                                                  \n",
            " flatten_3 (Flatten)            (None, 87616)        0           ['activation_12[0][0]']          \n",
            "                                                                                                  \n",
            " dropout_13 (Dropout)           (None, 87616)        0           ['flatten_3[0][0]']              \n",
            "                                                                                                  \n",
            " dense_5 (Dense)                (None, 256)          22429952    ['dropout_13[0][0]']             \n",
            "                                                                                                  \n",
            " dropout_14 (Dropout)           (None, 256)          0           ['dense_5[0][0]']                \n",
            "                                                                                                  \n",
            " dense_6 (Dense)                (None, 3)            771         ['dropout_14[0][0]']             \n",
            "                                                                                                  \n",
            "==================================================================================================\n",
            "Total params: 22,505,707\n",
            "Trainable params: 22,505,195\n",
            "Non-trainable params: 512\n",
            "__________________________________________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ph0kldXONRrF"
      },
      "source": [
        "resnet_model.compile(optimizer='adam',\n",
        "              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
        "               metrics=['acc', f1_m, precision_m, recall_m])"
      ],
      "execution_count": 56,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MsIHOUJANRtp",
        "outputId": "7f502c64-06f0-405f-9397-a7fd30982b49"
      },
      "source": [
        "history_resnet = resnet_model.fit(train_ds,\n",
        "                    validation_data=val_ds,\n",
        "                    callbacks=[callback],\n",
        "                    epochs=epochs)"
      ],
      "execution_count": 57,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "36/36 [==============================] - 212s 6s/step - loss: 4.1343 - acc: 0.9004 - f1_m: 0.7284 - precision_m: 0.5184 - recall_m: 1.2681 - val_loss: 0.8336 - val_acc: 0.5125 - val_f1_m: 0.5437 - val_precision_m: 0.3663 - val_recall_m: 1.1024\n",
            "Epoch 2/100\n",
            "36/36 [==============================] - 209s 6s/step - loss: 0.3246 - acc: 0.9787 - f1_m: 0.6859 - precision_m: 0.4795 - recall_m: 1.2391 - val_loss: 2.3185 - val_acc: 0.1388 - val_f1_m: 0.7946 - val_precision_m: 0.5013 - val_recall_m: 2.0000\n",
            "Epoch 3/100\n",
            "36/36 [==============================] - 212s 6s/step - loss: 0.5161 - acc: 0.9733 - f1_m: 0.7210 - precision_m: 0.5130 - recall_m: 1.2434 - val_loss: 2.3506 - val_acc: 0.1388 - val_f1_m: 0.6195 - val_precision_m: 0.4114 - val_recall_m: 1.2783\n",
            "Epoch 4/100\n",
            "36/36 [==============================] - 209s 6s/step - loss: 0.2414 - acc: 0.9867 - f1_m: 0.7164 - precision_m: 0.5205 - recall_m: 1.1836 - val_loss: 2.4094 - val_acc: 0.1388 - val_f1_m: 0.6672 - val_precision_m: 0.5032 - val_recall_m: 1.0000\n",
            "Epoch 5/100\n",
            "36/36 [==============================] - 227s 6s/step - loss: 0.1949 - acc: 0.9849 - f1_m: 0.7062 - precision_m: 0.5327 - recall_m: 1.0789 - val_loss: 2.0674 - val_acc: 0.1388 - val_f1_m: 0.8491 - val_precision_m: 0.6386 - val_recall_m: 1.2803\n",
            "Epoch 6/100\n",
            "36/36 [==============================] - 210s 6s/step - loss: 0.3025 - acc: 0.9689 - f1_m: 0.6602 - precision_m: 0.5195 - recall_m: 0.9216 - val_loss: 2.2694 - val_acc: 0.1388 - val_f1_m: 0.6613 - val_precision_m: 0.4952 - val_recall_m: 1.0000\n",
            "Epoch 7/100\n",
            "36/36 [==============================] - 208s 6s/step - loss: 0.2855 - acc: 0.9804 - f1_m: 0.7291 - precision_m: 0.5573 - recall_m: 1.0745 - val_loss: 1.8324 - val_acc: 0.3986 - val_f1_m: 0.6524 - val_precision_m: 0.4349 - val_recall_m: 1.3424\n",
            "Epoch 8/100\n",
            "36/36 [==============================] - 208s 6s/step - loss: 0.3060 - acc: 0.9804 - f1_m: 0.7134 - precision_m: 0.5327 - recall_m: 1.1237 - val_loss: 2.6375 - val_acc: 0.2989 - val_f1_m: 0.5822 - val_precision_m: 0.4143 - val_recall_m: 1.0000\n",
            "Epoch 9/100\n",
            "36/36 [==============================] - 208s 6s/step - loss: 0.0941 - acc: 0.9867 - f1_m: 0.6494 - precision_m: 0.5017 - recall_m: 0.9589 - val_loss: 3.3159 - val_acc: 0.2491 - val_f1_m: 0.5870 - val_precision_m: 0.4288 - val_recall_m: 1.0000\n",
            "Epoch 10/100\n",
            "36/36 [==============================] - 208s 6s/step - loss: 0.1525 - acc: 0.9867 - f1_m: 0.5897 - precision_m: 0.4675 - recall_m: 0.8265 - val_loss: 3.5212 - val_acc: 0.1815 - val_f1_m: 0.6003 - val_precision_m: 0.4330 - val_recall_m: 1.0000\n",
            "Epoch 11/100\n",
            "36/36 [==============================] - 208s 6s/step - loss: 0.0724 - acc: 0.9876 - f1_m: 0.6332 - precision_m: 0.5028 - recall_m: 0.8884 - val_loss: 3.3276 - val_acc: 0.2349 - val_f1_m: 0.5952 - val_precision_m: 0.4248 - val_recall_m: 1.0000\n",
            "Epoch 12/100\n",
            "36/36 [==============================] - 211s 6s/step - loss: 0.0936 - acc: 0.9867 - f1_m: 0.6173 - precision_m: 0.4873 - recall_m: 0.8655 - val_loss: 3.3859 - val_acc: 0.2420 - val_f1_m: 0.5872 - val_precision_m: 0.4247 - val_recall_m: 1.0000\n",
            "Epoch 13/100\n",
            "36/36 [==============================] - 210s 6s/step - loss: 0.1244 - acc: 0.9840 - f1_m: 0.5717 - precision_m: 0.4535 - recall_m: 0.8031 - val_loss: 3.4139 - val_acc: 0.2349 - val_f1_m: 0.5834 - val_precision_m: 0.4171 - val_recall_m: 1.0000\n",
            "Epoch 14/100\n",
            "36/36 [==============================] - 209s 6s/step - loss: 0.0992 - acc: 0.9867 - f1_m: 0.3207 - precision_m: 0.2925 - recall_m: 0.3813 - val_loss: 4.1575 - val_acc: 0.2491 - val_f1_m: 0.5857 - val_precision_m: 0.4191 - val_recall_m: 1.0000\n",
            "Epoch 15/100\n",
            "36/36 [==============================] - 218s 6s/step - loss: 0.0968 - acc: 0.9858 - f1_m: 0.3162 - precision_m: 0.2754 - recall_m: 0.3916 - val_loss: 3.7175 - val_acc: 0.2598 - val_f1_m: 0.5675 - val_precision_m: 0.4032 - val_recall_m: 1.0000\n",
            "Epoch 16/100\n",
            "36/36 [==============================] - 220s 6s/step - loss: 0.0572 - acc: 0.9858 - f1_m: 0.4591 - precision_m: 0.3758 - recall_m: 0.6176 - val_loss: 4.2240 - val_acc: 0.2242 - val_f1_m: 0.5690 - val_precision_m: 0.4005 - val_recall_m: 1.0000\n",
            "Epoch 17/100\n",
            "36/36 [==============================] - 210s 6s/step - loss: 0.0656 - acc: 0.9849 - f1_m: 0.5292 - precision_m: 0.4268 - recall_m: 0.7272 - val_loss: 4.0128 - val_acc: 0.2420 - val_f1_m: 0.6547 - val_precision_m: 0.4461 - val_recall_m: 1.2704\n",
            "Epoch 18/100\n",
            "36/36 [==============================] - 211s 6s/step - loss: 0.0411 - acc: 0.9858 - f1_m: 0.4170 - precision_m: 0.3498 - recall_m: 0.5461 - val_loss: 4.2038 - val_acc: 0.2313 - val_f1_m: 0.6555 - val_precision_m: 0.4515 - val_recall_m: 1.2600\n",
            "Epoch 19/100\n",
            "36/36 [==============================] - 223s 6s/step - loss: 0.0487 - acc: 0.9840 - f1_m: 0.4361 - precision_m: 0.3594 - recall_m: 0.5804 - val_loss: 4.3016 - val_acc: 0.2313 - val_f1_m: 0.7242 - val_precision_m: 0.4624 - val_recall_m: 1.6926\n",
            "Epoch 20/100\n",
            "36/36 [==============================] - 217s 6s/step - loss: 0.0404 - acc: 0.9902 - f1_m: 0.3914 - precision_m: 0.3291 - recall_m: 0.5076 - val_loss: 4.1918 - val_acc: 0.2633 - val_f1_m: 0.7436 - val_precision_m: 0.4715 - val_recall_m: 1.7764\n",
            "Epoch 21/100\n",
            "36/36 [==============================] - 211s 6s/step - loss: 0.0436 - acc: 0.9858 - f1_m: 0.4197 - precision_m: 0.3449 - recall_m: 0.5562 - val_loss: 4.2430 - val_acc: 0.2562 - val_f1_m: 0.6731 - val_precision_m: 0.4367 - val_recall_m: 1.5369\n",
            "Epoch 22/100\n",
            "36/36 [==============================] - 211s 6s/step - loss: 0.0525 - acc: 0.9840 - f1_m: 0.4313 - precision_m: 0.3681 - recall_m: 0.5365 - val_loss: 3.3812 - val_acc: 0.3238 - val_f1_m: 0.6851 - val_precision_m: 0.4448 - val_recall_m: 1.5786\n",
            "Epoch 23/100\n",
            "36/36 [==============================] - 211s 6s/step - loss: 0.0586 - acc: 0.9858 - f1_m: 0.6264 - precision_m: 0.4822 - recall_m: 0.9216 - val_loss: 4.2929 - val_acc: 0.2491 - val_f1_m: 0.7464 - val_precision_m: 0.4745 - val_recall_m: 1.7855\n",
            "Epoch 24/100\n",
            "36/36 [==============================] - 212s 6s/step - loss: 0.0517 - acc: 0.9796 - f1_m: 0.6910 - precision_m: 0.5219 - recall_m: 1.0477 - val_loss: 3.7035 - val_acc: 0.2705 - val_f1_m: 0.6821 - val_precision_m: 0.4398 - val_recall_m: 1.5478\n",
            "Epoch 25/100\n",
            "36/36 [==============================] - 211s 6s/step - loss: 0.0617 - acc: 0.9867 - f1_m: 0.7320 - precision_m: 0.5344 - recall_m: 1.1858 - val_loss: 4.4565 - val_acc: 0.2171 - val_f1_m: 0.6853 - val_precision_m: 0.4418 - val_recall_m: 1.5476\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qimzZUTAGk-f",
        "outputId": "0ba310fe-f953-43fd-a227-6ad7e78f24a5"
      },
      "source": [
        "score = resnet_model.evaluate(test_ds, verbose=0)\n",
        "print(\"Test loss:\", score[0])\n",
        "print(\"Test accuracy:\", score[1])\n",
        "\n",
        "\n",
        "# Save the model\n",
        "filepath = './model_resnet'\n",
        "save_model(model, filepath)\n",
        "\n",
        "# Load the model\n",
        "model = load_model(filepath, compile = False)"
      ],
      "execution_count": 58,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Test loss: 4.572501182556152\n",
            "Test accuracy: 0.1964285671710968\n",
            "WARNING:tensorflow:Compiled the loaded model, but the compiled metrics have yet to be built. `model.compile_metrics` will be empty until you train or evaluate the model.\n",
            "INFO:tensorflow:Assets written to: ./model_resnet/assets\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "I2MIrRIFsu7F"
      },
      "source": [
        "acc = history_resnet.history['f1_m']\n",
        "val_acc = history_resnet.history['val_f1_m']\n",
        "\n",
        "loss = history_resnet.history['loss']\n",
        "val_loss = history_resnet.history['val_loss']\n",
        "\n",
        "epochs_range = range(len(acc))"
      ],
      "execution_count": 59,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 499
        },
        "id": "juVkGZgBsvCJ",
        "outputId": "1480940d-7da0-45ab-90d1-160065888c0e"
      },
      "source": [
        "plt.figure(figsize=(8, 8))\n",
        "plt.subplot(1, 2, 1)\n",
        "plt.plot(epochs_range, acc, label='Training Accuracy')\n",
        "plt.plot(epochs_range, val_acc, label='Validation Accuracy')\n",
        "plt.legend(loc='lower right')\n",
        "plt.title('Training and Validation Accuracy')\n",
        "\n",
        "plt.subplot(1, 2, 2)\n",
        "plt.plot(epochs_range, loss, label='Training Loss')\n",
        "plt.plot(epochs_range, val_loss, label='Validation Loss')\n",
        "plt.legend(loc='upper right')\n",
        "plt.title('Training and Validation Loss')\n",
        "plt.show()"
      ],
      "execution_count": 60,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x576 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5Ew5kdV9svIL",
        "outputId": "d963ab9b-256c-4389-d985-59cce3ee1bed"
      },
      "source": [
        "predictions = resnet_model.predict(test_ds, steps=test_steps_per_epoch)\n",
        "# Get most likely class\n",
        "predicted_classes = numpy.argmax(predictions, axis=1)\n",
        "predicted_classes"
      ],
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 2, 2, 2, 2, 2, 0, 2, 2, 0, 2,\n",
              "       2, 2, 2, 2, 2, 2, 2, 2, 0, 2, 2, 2, 2, 2, 2, 2, 0, 2, 2, 2, 0, 2,\n",
              "       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2])"
            ]
          },
          "metadata": {},
          "execution_count": 61
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9sFQXxxRs9OX",
        "outputId": "1d4d1ad9-943c-4208-8b77-85bb51ca9dbb"
      },
      "source": [
        "true_classes = test_generator.classes #true_classes = test_ds.class_names\n",
        "true_classes"
      ],
      "execution_count": 62,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2], dtype=int32)"
            ]
          },
          "metadata": {},
          "execution_count": 62
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kAwDc0Jas9V6",
        "outputId": "a6c7105c-740c-4aaa-e0af-cfb19907ce7e"
      },
      "source": [
        "class_labels = list(test_generator.class_indices.keys())\n",
        "class_labels"
      ],
      "execution_count": 63,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['Danube', 'Dnieper', 'UnknownOrigin']"
            ]
          },
          "metadata": {},
          "execution_count": 63
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 631
        },
        "id": "abJ41UE-s9Zb",
        "outputId": "5d312cb3-fc1f-49b6-e24b-761024869546"
      },
      "source": [
        "report = classification_report(true_classes, predicted_classes, target_names=class_labels, zero_division=0)\n",
        "print(report)\n",
        "cmrforest = confusion_matrix(true_classes, predicted_classes)\n",
        "plt.figure(figsize=(10,7))\n",
        "sns.heatmap(cmrforest ,annot=True , fmt = 'd')\n",
        "plt.xlabel('Predicted')\n",
        "plt.ylabel('Truth')\n",
        "plt.title('Confusion Report')\n",
        "plt.show()"
      ],
      "execution_count": 64,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "               precision    recall  f1-score   support\n",
            "\n",
            "       Danube       1.00      0.06      0.12       109\n",
            "      Dnieper       0.00      0.00      0.00        78\n",
            "UnknownOrigin       0.17      1.00      0.29        37\n",
            "\n",
            "     accuracy                           0.20       224\n",
            "    macro avg       0.39      0.35      0.14       224\n",
            " weighted avg       0.51      0.20      0.11       224\n",
            "\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x504 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "edLwUcxm0Bad"
      },
      "source": [
        "# Transfer learning"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8sSs69tQAfXn"
      },
      "source": [
        "## ResNet50"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wb006oPsAjqi"
      },
      "source": [
        "### Initialization and Summary"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VmCa5UV4tquS",
        "outputId": "faf8e9f3-4d07-429b-eb4a-7c920f7b0043"
      },
      "source": [
        "# create the base pre-trained model\n",
        "base_model = ResNet50(weights='imagenet', include_top=False)\n",
        "base_model.trainable = False\n",
        "\n",
        "# add a global spatial average pooling layer\n",
        "x = base_model.output\n",
        "x = GlobalAveragePooling2D()(x)\n",
        "# add a fully-connected layer\n",
        "x = Dense(1024, activation='relu')(x)\n",
        "# add a dropout layer\n",
        "x = Dropout(0.2)(x)\n",
        "# and a logistic layer -- we have 3 classes\n",
        "predictions = Dense(3, activation='softmax')(x)\n",
        "\n",
        "# this is the model we will train\n",
        "model = Model(inputs=base_model.input, outputs=predictions)\n",
        "\n",
        "# compile the model (should be done *after* setting layers to non-trainable)\n",
        "model.compile(loss=\"categorical_crossentropy\", optimizer=\"adam\", metrics=[\"accuracy\"])\n",
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Model: \"model_3\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_4 (InputLayer)            [(None, None, None,  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "conv1_pad (ZeroPadding2D)       (None, None, None, 3 0           input_4[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "conv1_conv (Conv2D)             (None, None, None, 6 9472        conv1_pad[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "conv1_bn (BatchNormalization)   (None, None, None, 6 256         conv1_conv[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1_relu (Activation)         (None, None, None, 6 0           conv1_bn[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "pool1_pad (ZeroPadding2D)       (None, None, None, 6 0           conv1_relu[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "pool1_pool (MaxPooling2D)       (None, None, None, 6 0           pool1_pad[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_1_conv (Conv2D)    (None, None, None, 6 4160        pool1_pool[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_1_bn (BatchNormali (None, None, None, 6 256         conv2_block1_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_1_relu (Activation (None, None, None, 6 0           conv2_block1_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_2_conv (Conv2D)    (None, None, None, 6 36928       conv2_block1_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_2_bn (BatchNormali (None, None, None, 6 256         conv2_block1_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_2_relu (Activation (None, None, None, 6 0           conv2_block1_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_0_conv (Conv2D)    (None, None, None, 2 16640       pool1_pool[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_3_conv (Conv2D)    (None, None, None, 2 16640       conv2_block1_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_0_bn (BatchNormali (None, None, None, 2 1024        conv2_block1_0_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_3_bn (BatchNormali (None, None, None, 2 1024        conv2_block1_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_add (Add)          (None, None, None, 2 0           conv2_block1_0_bn[0][0]          \n",
            "                                                                 conv2_block1_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_out (Activation)   (None, None, None, 2 0           conv2_block1_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block2_1_conv (Conv2D)    (None, None, None, 6 16448       conv2_block1_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block2_1_bn (BatchNormali (None, None, None, 6 256         conv2_block2_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block2_1_relu (Activation (None, None, None, 6 0           conv2_block2_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block2_2_conv (Conv2D)    (None, None, None, 6 36928       conv2_block2_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block2_2_bn (BatchNormali (None, None, None, 6 256         conv2_block2_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block2_2_relu (Activation (None, None, None, 6 0           conv2_block2_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block2_3_conv (Conv2D)    (None, None, None, 2 16640       conv2_block2_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block2_3_bn (BatchNormali (None, None, None, 2 1024        conv2_block2_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block2_add (Add)          (None, None, None, 2 0           conv2_block1_out[0][0]           \n",
            "                                                                 conv2_block2_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block2_out (Activation)   (None, None, None, 2 0           conv2_block2_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block3_1_conv (Conv2D)    (None, None, None, 6 16448       conv2_block2_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block3_1_bn (BatchNormali (None, None, None, 6 256         conv2_block3_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block3_1_relu (Activation (None, None, None, 6 0           conv2_block3_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block3_2_conv (Conv2D)    (None, None, None, 6 36928       conv2_block3_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block3_2_bn (BatchNormali (None, None, None, 6 256         conv2_block3_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block3_2_relu (Activation (None, None, None, 6 0           conv2_block3_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block3_3_conv (Conv2D)    (None, None, None, 2 16640       conv2_block3_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block3_3_bn (BatchNormali (None, None, None, 2 1024        conv2_block3_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block3_add (Add)          (None, None, None, 2 0           conv2_block2_out[0][0]           \n",
            "                                                                 conv2_block3_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block3_out (Activation)   (None, None, None, 2 0           conv2_block3_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_1_conv (Conv2D)    (None, None, None, 1 32896       conv2_block3_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_1_bn (BatchNormali (None, None, None, 1 512         conv3_block1_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_1_relu (Activation (None, None, None, 1 0           conv3_block1_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_2_conv (Conv2D)    (None, None, None, 1 147584      conv3_block1_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_2_bn (BatchNormali (None, None, None, 1 512         conv3_block1_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_2_relu (Activation (None, None, None, 1 0           conv3_block1_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_0_conv (Conv2D)    (None, None, None, 5 131584      conv2_block3_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_3_conv (Conv2D)    (None, None, None, 5 66048       conv3_block1_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_0_bn (BatchNormali (None, None, None, 5 2048        conv3_block1_0_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_3_bn (BatchNormali (None, None, None, 5 2048        conv3_block1_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_add (Add)          (None, None, None, 5 0           conv3_block1_0_bn[0][0]          \n",
            "                                                                 conv3_block1_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_out (Activation)   (None, None, None, 5 0           conv3_block1_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block2_1_conv (Conv2D)    (None, None, None, 1 65664       conv3_block1_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block2_1_bn (BatchNormali (None, None, None, 1 512         conv3_block2_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block2_1_relu (Activation (None, None, None, 1 0           conv3_block2_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block2_2_conv (Conv2D)    (None, None, None, 1 147584      conv3_block2_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block2_2_bn (BatchNormali (None, None, None, 1 512         conv3_block2_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block2_2_relu (Activation (None, None, None, 1 0           conv3_block2_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block2_3_conv (Conv2D)    (None, None, None, 5 66048       conv3_block2_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block2_3_bn (BatchNormali (None, None, None, 5 2048        conv3_block2_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block2_add (Add)          (None, None, None, 5 0           conv3_block1_out[0][0]           \n",
            "                                                                 conv3_block2_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block2_out (Activation)   (None, None, None, 5 0           conv3_block2_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block3_1_conv (Conv2D)    (None, None, None, 1 65664       conv3_block2_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block3_1_bn (BatchNormali (None, None, None, 1 512         conv3_block3_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block3_1_relu (Activation (None, None, None, 1 0           conv3_block3_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block3_2_conv (Conv2D)    (None, None, None, 1 147584      conv3_block3_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block3_2_bn (BatchNormali (None, None, None, 1 512         conv3_block3_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block3_2_relu (Activation (None, None, None, 1 0           conv3_block3_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block3_3_conv (Conv2D)    (None, None, None, 5 66048       conv3_block3_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block3_3_bn (BatchNormali (None, None, None, 5 2048        conv3_block3_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block3_add (Add)          (None, None, None, 5 0           conv3_block2_out[0][0]           \n",
            "                                                                 conv3_block3_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block3_out (Activation)   (None, None, None, 5 0           conv3_block3_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block4_1_conv (Conv2D)    (None, None, None, 1 65664       conv3_block3_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block4_1_bn (BatchNormali (None, None, None, 1 512         conv3_block4_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block4_1_relu (Activation (None, None, None, 1 0           conv3_block4_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block4_2_conv (Conv2D)    (None, None, None, 1 147584      conv3_block4_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block4_2_bn (BatchNormali (None, None, None, 1 512         conv3_block4_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block4_2_relu (Activation (None, None, None, 1 0           conv3_block4_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block4_3_conv (Conv2D)    (None, None, None, 5 66048       conv3_block4_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block4_3_bn (BatchNormali (None, None, None, 5 2048        conv3_block4_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block4_add (Add)          (None, None, None, 5 0           conv3_block3_out[0][0]           \n",
            "                                                                 conv3_block4_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block4_out (Activation)   (None, None, None, 5 0           conv3_block4_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_1_conv (Conv2D)    (None, None, None, 2 131328      conv3_block4_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_1_bn (BatchNormali (None, None, None, 2 1024        conv4_block1_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_1_relu (Activation (None, None, None, 2 0           conv4_block1_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_2_conv (Conv2D)    (None, None, None, 2 590080      conv4_block1_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_2_bn (BatchNormali (None, None, None, 2 1024        conv4_block1_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_2_relu (Activation (None, None, None, 2 0           conv4_block1_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_0_conv (Conv2D)    (None, None, None, 1 525312      conv3_block4_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_3_conv (Conv2D)    (None, None, None, 1 263168      conv4_block1_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_0_bn (BatchNormali (None, None, None, 1 4096        conv4_block1_0_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_3_bn (BatchNormali (None, None, None, 1 4096        conv4_block1_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_add (Add)          (None, None, None, 1 0           conv4_block1_0_bn[0][0]          \n",
            "                                                                 conv4_block1_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_out (Activation)   (None, None, None, 1 0           conv4_block1_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block2_1_conv (Conv2D)    (None, None, None, 2 262400      conv4_block1_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block2_1_bn (BatchNormali (None, None, None, 2 1024        conv4_block2_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block2_1_relu (Activation (None, None, None, 2 0           conv4_block2_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block2_2_conv (Conv2D)    (None, None, None, 2 590080      conv4_block2_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block2_2_bn (BatchNormali (None, None, None, 2 1024        conv4_block2_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block2_2_relu (Activation (None, None, None, 2 0           conv4_block2_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block2_3_conv (Conv2D)    (None, None, None, 1 263168      conv4_block2_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block2_3_bn (BatchNormali (None, None, None, 1 4096        conv4_block2_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block2_add (Add)          (None, None, None, 1 0           conv4_block1_out[0][0]           \n",
            "                                                                 conv4_block2_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block2_out (Activation)   (None, None, None, 1 0           conv4_block2_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block3_1_conv (Conv2D)    (None, None, None, 2 262400      conv4_block2_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block3_1_bn (BatchNormali (None, None, None, 2 1024        conv4_block3_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block3_1_relu (Activation (None, None, None, 2 0           conv4_block3_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block3_2_conv (Conv2D)    (None, None, None, 2 590080      conv4_block3_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block3_2_bn (BatchNormali (None, None, None, 2 1024        conv4_block3_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block3_2_relu (Activation (None, None, None, 2 0           conv4_block3_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block3_3_conv (Conv2D)    (None, None, None, 1 263168      conv4_block3_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block3_3_bn (BatchNormali (None, None, None, 1 4096        conv4_block3_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block3_add (Add)          (None, None, None, 1 0           conv4_block2_out[0][0]           \n",
            "                                                                 conv4_block3_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block3_out (Activation)   (None, None, None, 1 0           conv4_block3_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block4_1_conv (Conv2D)    (None, None, None, 2 262400      conv4_block3_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block4_1_bn (BatchNormali (None, None, None, 2 1024        conv4_block4_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block4_1_relu (Activation (None, None, None, 2 0           conv4_block4_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block4_2_conv (Conv2D)    (None, None, None, 2 590080      conv4_block4_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block4_2_bn (BatchNormali (None, None, None, 2 1024        conv4_block4_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block4_2_relu (Activation (None, None, None, 2 0           conv4_block4_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block4_3_conv (Conv2D)    (None, None, None, 1 263168      conv4_block4_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block4_3_bn (BatchNormali (None, None, None, 1 4096        conv4_block4_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block4_add (Add)          (None, None, None, 1 0           conv4_block3_out[0][0]           \n",
            "                                                                 conv4_block4_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block4_out (Activation)   (None, None, None, 1 0           conv4_block4_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block5_1_conv (Conv2D)    (None, None, None, 2 262400      conv4_block4_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block5_1_bn (BatchNormali (None, None, None, 2 1024        conv4_block5_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block5_1_relu (Activation (None, None, None, 2 0           conv4_block5_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block5_2_conv (Conv2D)    (None, None, None, 2 590080      conv4_block5_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block5_2_bn (BatchNormali (None, None, None, 2 1024        conv4_block5_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block5_2_relu (Activation (None, None, None, 2 0           conv4_block5_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block5_3_conv (Conv2D)    (None, None, None, 1 263168      conv4_block5_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block5_3_bn (BatchNormali (None, None, None, 1 4096        conv4_block5_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block5_add (Add)          (None, None, None, 1 0           conv4_block4_out[0][0]           \n",
            "                                                                 conv4_block5_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block5_out (Activation)   (None, None, None, 1 0           conv4_block5_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block6_1_conv (Conv2D)    (None, None, None, 2 262400      conv4_block5_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block6_1_bn (BatchNormali (None, None, None, 2 1024        conv4_block6_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block6_1_relu (Activation (None, None, None, 2 0           conv4_block6_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block6_2_conv (Conv2D)    (None, None, None, 2 590080      conv4_block6_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block6_2_bn (BatchNormali (None, None, None, 2 1024        conv4_block6_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block6_2_relu (Activation (None, None, None, 2 0           conv4_block6_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block6_3_conv (Conv2D)    (None, None, None, 1 263168      conv4_block6_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block6_3_bn (BatchNormali (None, None, None, 1 4096        conv4_block6_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block6_add (Add)          (None, None, None, 1 0           conv4_block5_out[0][0]           \n",
            "                                                                 conv4_block6_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block6_out (Activation)   (None, None, None, 1 0           conv4_block6_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_1_conv (Conv2D)    (None, None, None, 5 524800      conv4_block6_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_1_bn (BatchNormali (None, None, None, 5 2048        conv5_block1_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_1_relu (Activation (None, None, None, 5 0           conv5_block1_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_2_conv (Conv2D)    (None, None, None, 5 2359808     conv5_block1_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_2_bn (BatchNormali (None, None, None, 5 2048        conv5_block1_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_2_relu (Activation (None, None, None, 5 0           conv5_block1_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_0_conv (Conv2D)    (None, None, None, 2 2099200     conv4_block6_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_3_conv (Conv2D)    (None, None, None, 2 1050624     conv5_block1_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_0_bn (BatchNormali (None, None, None, 2 8192        conv5_block1_0_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_3_bn (BatchNormali (None, None, None, 2 8192        conv5_block1_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_add (Add)          (None, None, None, 2 0           conv5_block1_0_bn[0][0]          \n",
            "                                                                 conv5_block1_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_out (Activation)   (None, None, None, 2 0           conv5_block1_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block2_1_conv (Conv2D)    (None, None, None, 5 1049088     conv5_block1_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block2_1_bn (BatchNormali (None, None, None, 5 2048        conv5_block2_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block2_1_relu (Activation (None, None, None, 5 0           conv5_block2_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block2_2_conv (Conv2D)    (None, None, None, 5 2359808     conv5_block2_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block2_2_bn (BatchNormali (None, None, None, 5 2048        conv5_block2_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block2_2_relu (Activation (None, None, None, 5 0           conv5_block2_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block2_3_conv (Conv2D)    (None, None, None, 2 1050624     conv5_block2_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block2_3_bn (BatchNormali (None, None, None, 2 8192        conv5_block2_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block2_add (Add)          (None, None, None, 2 0           conv5_block1_out[0][0]           \n",
            "                                                                 conv5_block2_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block2_out (Activation)   (None, None, None, 2 0           conv5_block2_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block3_1_conv (Conv2D)    (None, None, None, 5 1049088     conv5_block2_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block3_1_bn (BatchNormali (None, None, None, 5 2048        conv5_block3_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block3_1_relu (Activation (None, None, None, 5 0           conv5_block3_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block3_2_conv (Conv2D)    (None, None, None, 5 2359808     conv5_block3_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block3_2_bn (BatchNormali (None, None, None, 5 2048        conv5_block3_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block3_2_relu (Activation (None, None, None, 5 0           conv5_block3_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block3_3_conv (Conv2D)    (None, None, None, 2 1050624     conv5_block3_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block3_3_bn (BatchNormali (None, None, None, 2 8192        conv5_block3_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block3_add (Add)          (None, None, None, 2 0           conv5_block2_out[0][0]           \n",
            "                                                                 conv5_block3_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block3_out (Activation)   (None, None, None, 2 0           conv5_block3_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "global_average_pooling2d_3 (Glo (None, 2048)         0           conv5_block3_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "dense_6 (Dense)                 (None, 1024)         2098176     global_average_pooling2d_3[0][0] \n",
            "__________________________________________________________________________________________________\n",
            "dropout_3 (Dropout)             (None, 1024)         0           dense_6[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "dense_7 (Dense)                 (None, 3)            3075        dropout_3[0][0]                  \n",
            "==================================================================================================\n",
            "Total params: 25,688,963\n",
            "Trainable params: 2,101,251\n",
            "Non-trainable params: 23,587,712\n",
            "__________________________________________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UckqmxTMCa6n"
      },
      "source": [
        "### Model Training"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "07OlYETI9bnw",
        "outputId": "c3b843b9-3f40-4348-9242-4976e1644b20"
      },
      "source": [
        "model.fit(\n",
        "    train_generator,\n",
        "    validation_data=validation_generator,\n",
        "    steps_per_epoch=1126//32,\n",
        "    epochs=14,\n",
        "    validation_steps=280//32)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/14\n",
            "35/35 [==============================] - 30s 790ms/step - loss: 1.0122 - accuracy: 0.5338 - val_loss: 0.8317 - val_accuracy: 0.6367\n",
            "Epoch 2/14\n",
            "35/35 [==============================] - 26s 748ms/step - loss: 0.7628 - accuracy: 0.7349 - val_loss: 0.7381 - val_accuracy: 0.6094\n",
            "Epoch 3/14\n",
            "35/35 [==============================] - 26s 744ms/step - loss: 0.6284 - accuracy: 0.8035 - val_loss: 0.5965 - val_accuracy: 0.9102\n",
            "Epoch 4/14\n",
            "35/35 [==============================] - 26s 740ms/step - loss: 0.4988 - accuracy: 0.8565 - val_loss: 0.4244 - val_accuracy: 0.8672\n",
            "Epoch 5/14\n",
            "35/35 [==============================] - 26s 744ms/step - loss: 0.4386 - accuracy: 0.8565 - val_loss: 0.3724 - val_accuracy: 0.9258\n",
            "Epoch 6/14\n",
            "35/35 [==============================] - 26s 742ms/step - loss: 0.3496 - accuracy: 0.8967 - val_loss: 0.3009 - val_accuracy: 0.8906\n",
            "Epoch 7/14\n",
            "35/35 [==============================] - 26s 749ms/step - loss: 0.3159 - accuracy: 0.9113 - val_loss: 0.2576 - val_accuracy: 0.9258\n",
            "Epoch 8/14\n",
            "35/35 [==============================] - 26s 743ms/step - loss: 0.2759 - accuracy: 0.9168 - val_loss: 0.3165 - val_accuracy: 0.9141\n",
            "Epoch 9/14\n",
            "35/35 [==============================] - 26s 743ms/step - loss: 0.2585 - accuracy: 0.9232 - val_loss: 0.1777 - val_accuracy: 0.9688\n",
            "Epoch 10/14\n",
            "35/35 [==============================] - 26s 740ms/step - loss: 0.2352 - accuracy: 0.9324 - val_loss: 0.1647 - val_accuracy: 0.9648\n",
            "Epoch 11/14\n",
            "35/35 [==============================] - 26s 756ms/step - loss: 0.2363 - accuracy: 0.9351 - val_loss: 0.1913 - val_accuracy: 0.9727\n",
            "Epoch 12/14\n",
            "35/35 [==============================] - 26s 747ms/step - loss: 0.2170 - accuracy: 0.9360 - val_loss: 0.3334 - val_accuracy: 0.8984\n",
            "Epoch 13/14\n",
            "35/35 [==============================] - 26s 753ms/step - loss: 0.2338 - accuracy: 0.9214 - val_loss: 0.1834 - val_accuracy: 0.9766\n",
            "Epoch 14/14\n",
            "35/35 [==============================] - 26s 746ms/step - loss: 0.2275 - accuracy: 0.9250 - val_loss: 0.1252 - val_accuracy: 0.9688\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "<tensorflow.python.keras.callbacks.History at 0x7f61842de310>"
            ]
          },
          "execution_count": 21,
          "metadata": {
            "tags": []
          },
          "output_type": "execute_result"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iLO2rZn9ItGp"
      },
      "source": [
        "### Model Accuracy"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ocCyHDzJIrfy",
        "outputId": "42bd7974-c983-4685-a700-1cf21c612fc5"
      },
      "source": [
        "score = model.evaluate(test_generator, verbose=0)\n",
        "print(\"Test loss:\", score[0])\n",
        "print(\"Test accuracy:\", score[1])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Test loss: 0.2832943797111511\n",
            "Test accuracy: 0.8616071343421936\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "b2O4bhCPCcMp"
      },
      "source": [
        "### Classification Report and Confusion Matrix Results"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 628
        },
        "id": "vn6eUMPa_Yit",
        "outputId": "a031ce62-0152-439b-bc0a-dc59f6f2994d"
      },
      "source": [
        "predictions = model.predict(test_generator, steps=test_steps_per_epoch)\n",
        "# Get most likely class\n",
        "predicted_classes = numpy.argmax(predictions, axis=1)\n",
        "\n",
        "true_classes = test_generator.classes\n",
        "class_labels = list(test_generator.class_indices.keys())   \n",
        "\n",
        "report = classification_report(true_classes, predicted_classes, target_names=class_labels, zero_division=0)\n",
        "print(report)\n",
        "cmrforest = confusion_matrix(true_classes, predicted_classes)\n",
        "plt.figure(figsize=(10,7))\n",
        "sns.heatmap(cmrforest ,annot=True , fmt = 'd')\n",
        "plt.xlabel('Predicted')\n",
        "plt.ylabel('Truth')\n",
        "plt.title('Confusion Report')\n",
        "plt.show()\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "               precision    recall  f1-score   support\n",
            "\n",
            "       Danube       0.89      1.00      0.94       109\n",
            "      Dnieper       0.82      1.00      0.90        78\n",
            "UnknownOrigin       1.00      0.16      0.28        37\n",
            "\n",
            "     accuracy                           0.86       224\n",
            "    macro avg       0.90      0.72      0.71       224\n",
            " weighted avg       0.88      0.86      0.82       224\n",
            "\n"
          ]
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x504 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light",
            "tags": []
          },
          "output_type": "display_data"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yTIU2YsUChDj"
      },
      "source": [
        "## InceptionV3"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QuTwwwIPCksS"
      },
      "source": [
        "### Initialization and Summary"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "khlH3_RMzbfK",
        "outputId": "f8b9d2bf-be39-40a2-9b8c-1597c74f2c8f"
      },
      "source": [
        "base_model = InceptionV3(weights='imagenet', include_top=False)\n",
        "base_model.trainable = False\n",
        "\n",
        "# add a global spatial average pooling layer\n",
        "x = base_model.output\n",
        "x = GlobalAveragePooling2D()(x)\n",
        "# add a fully-connected layer\n",
        "x = Dense(1024, activation='relu')(x)\n",
        "# add a dropout layer\n",
        "x = Dropout(0.2)(x)\n",
        "# and a logistic layer -- we have 3 classes\n",
        "predictions = Dense(3, activation='softmax')(x)\n",
        "\n",
        "# this is the model we will train\n",
        "model = Model(inputs=base_model.input, outputs=predictions)\n",
        "\n",
        "# compile the model (should be done *after* setting layers to non-trainable)\n",
        "model.compile(loss=\"categorical_crossentropy\", optimizer=\"adam\", metrics=[\"accuracy\"])\n",
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/inception_v3/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "87916544/87910968 [==============================] - 0s 0us/step\n",
            "Model: \"model_1\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_2 (InputLayer)            [(None, None, None,  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "conv2d (Conv2D)                 (None, None, None, 3 864         input_2[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization (BatchNorma (None, None, None, 3 96          conv2d[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "activation (Activation)         (None, None, None, 3 0           batch_normalization[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_1 (Conv2D)               (None, None, None, 3 9216        activation[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_1 (BatchNor (None, None, None, 3 96          conv2d_1[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "activation_1 (Activation)       (None, None, None, 3 0           batch_normalization_1[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_2 (Conv2D)               (None, None, None, 6 18432       activation_1[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_2 (BatchNor (None, None, None, 6 192         conv2d_2[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "activation_2 (Activation)       (None, None, None, 6 0           batch_normalization_2[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling2d (MaxPooling2D)    (None, None, None, 6 0           activation_2[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_3 (Conv2D)               (None, None, None, 8 5120        max_pooling2d[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_3 (BatchNor (None, None, None, 8 240         conv2d_3[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "activation_3 (Activation)       (None, None, None, 8 0           batch_normalization_3[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_4 (Conv2D)               (None, None, None, 1 138240      activation_3[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_4 (BatchNor (None, None, None, 1 576         conv2d_4[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "activation_4 (Activation)       (None, None, None, 1 0           batch_normalization_4[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling2d_1 (MaxPooling2D)  (None, None, None, 1 0           activation_4[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_8 (Conv2D)               (None, None, None, 6 12288       max_pooling2d_1[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_8 (BatchNor (None, None, None, 6 192         conv2d_8[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "activation_8 (Activation)       (None, None, None, 6 0           batch_normalization_8[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_6 (Conv2D)               (None, None, None, 4 9216        max_pooling2d_1[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_9 (Conv2D)               (None, None, None, 9 55296       activation_8[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_6 (BatchNor (None, None, None, 4 144         conv2d_6[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_9 (BatchNor (None, None, None, 9 288         conv2d_9[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "activation_6 (Activation)       (None, None, None, 4 0           batch_normalization_6[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "activation_9 (Activation)       (None, None, None, 9 0           batch_normalization_9[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "average_pooling2d (AveragePooli (None, None, None, 1 0           max_pooling2d_1[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_5 (Conv2D)               (None, None, None, 6 12288       max_pooling2d_1[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_7 (Conv2D)               (None, None, None, 6 76800       activation_6[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_10 (Conv2D)              (None, None, None, 9 82944       activation_9[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_11 (Conv2D)              (None, None, None, 3 6144        average_pooling2d[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_5 (BatchNor (None, None, None, 6 192         conv2d_5[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_7 (BatchNor (None, None, None, 6 192         conv2d_7[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_10 (BatchNo (None, None, None, 9 288         conv2d_10[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_11 (BatchNo (None, None, None, 3 96          conv2d_11[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_5 (Activation)       (None, None, None, 6 0           batch_normalization_5[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "activation_7 (Activation)       (None, None, None, 6 0           batch_normalization_7[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "activation_10 (Activation)      (None, None, None, 9 0           batch_normalization_10[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_11 (Activation)      (None, None, None, 3 0           batch_normalization_11[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "mixed0 (Concatenate)            (None, None, None, 2 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, None, None, 6 16384       mixed0[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_15 (BatchNo (None, None, None, 6 192         conv2d_15[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_15 (Activation)      (None, None, None, 6 0           batch_normalization_15[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_13 (Conv2D)              (None, None, None, 4 12288       mixed0[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_16 (Conv2D)              (None, None, None, 9 55296       activation_15[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_13 (BatchNo (None, None, None, 4 144         conv2d_13[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_16 (BatchNo (None, None, None, 9 288         conv2d_16[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_13 (Activation)      (None, None, None, 4 0           batch_normalization_13[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_16 (Activation)      (None, None, None, 9 0           batch_normalization_16[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "average_pooling2d_1 (AveragePoo (None, None, None, 2 0           mixed0[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_12 (Conv2D)              (None, None, None, 6 16384       mixed0[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_14 (Conv2D)              (None, None, None, 6 76800       activation_13[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_17 (Conv2D)              (None, None, None, 9 82944       activation_16[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_18 (Conv2D)              (None, None, None, 6 16384       average_pooling2d_1[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_12 (BatchNo (None, None, None, 6 192         conv2d_12[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_14 (BatchNo (None, None, None, 6 192         conv2d_14[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_17 (BatchNo (None, None, None, 9 288         conv2d_17[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_18 (BatchNo (None, None, None, 6 192         conv2d_18[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_12 (Activation)      (None, None, None, 6 0           batch_normalization_12[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_14 (Activation)      (None, None, None, 6 0           batch_normalization_14[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_17 (Activation)      (None, None, None, 9 0           batch_normalization_17[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_18 (Activation)      (None, None, None, 6 0           batch_normalization_18[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "mixed1 (Concatenate)            (None, None, None, 2 0           activation_12[0][0]              \n",
            "                                                                 activation_14[0][0]              \n",
            "                                                                 activation_17[0][0]              \n",
            "                                                                 activation_18[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_22 (Conv2D)              (None, None, None, 6 18432       mixed1[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_22 (BatchNo (None, None, None, 6 192         conv2d_22[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_22 (Activation)      (None, None, None, 6 0           batch_normalization_22[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_20 (Conv2D)              (None, None, None, 4 13824       mixed1[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_23 (Conv2D)              (None, None, None, 9 55296       activation_22[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_20 (BatchNo (None, None, None, 4 144         conv2d_20[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_23 (BatchNo (None, None, None, 9 288         conv2d_23[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_20 (Activation)      (None, None, None, 4 0           batch_normalization_20[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_23 (Activation)      (None, None, None, 9 0           batch_normalization_23[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "average_pooling2d_2 (AveragePoo (None, None, None, 2 0           mixed1[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_19 (Conv2D)              (None, None, None, 6 18432       mixed1[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_21 (Conv2D)              (None, None, None, 6 76800       activation_20[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_24 (Conv2D)              (None, None, None, 9 82944       activation_23[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_25 (Conv2D)              (None, None, None, 6 18432       average_pooling2d_2[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_19 (BatchNo (None, None, None, 6 192         conv2d_19[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_21 (BatchNo (None, None, None, 6 192         conv2d_21[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_24 (BatchNo (None, None, None, 9 288         conv2d_24[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_25 (BatchNo (None, None, None, 6 192         conv2d_25[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_19 (Activation)      (None, None, None, 6 0           batch_normalization_19[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_21 (Activation)      (None, None, None, 6 0           batch_normalization_21[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_24 (Activation)      (None, None, None, 9 0           batch_normalization_24[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_25 (Activation)      (None, None, None, 6 0           batch_normalization_25[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "mixed2 (Concatenate)            (None, None, None, 2 0           activation_19[0][0]              \n",
            "                                                                 activation_21[0][0]              \n",
            "                                                                 activation_24[0][0]              \n",
            "                                                                 activation_25[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_27 (Conv2D)              (None, None, None, 6 18432       mixed2[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_27 (BatchNo (None, None, None, 6 192         conv2d_27[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_27 (Activation)      (None, None, None, 6 0           batch_normalization_27[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_28 (Conv2D)              (None, None, None, 9 55296       activation_27[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_28 (BatchNo (None, None, None, 9 288         conv2d_28[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_28 (Activation)      (None, None, None, 9 0           batch_normalization_28[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_26 (Conv2D)              (None, None, None, 3 995328      mixed2[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_29 (Conv2D)              (None, None, None, 9 82944       activation_28[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_26 (BatchNo (None, None, None, 3 1152        conv2d_26[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_29 (BatchNo (None, None, None, 9 288         conv2d_29[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_26 (Activation)      (None, None, None, 3 0           batch_normalization_26[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_29 (Activation)      (None, None, None, 9 0           batch_normalization_29[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling2d_2 (MaxPooling2D)  (None, None, None, 2 0           mixed2[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "mixed3 (Concatenate)            (None, None, None, 7 0           activation_26[0][0]              \n",
            "                                                                 activation_29[0][0]              \n",
            "                                                                 max_pooling2d_2[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_34 (Conv2D)              (None, None, None, 1 98304       mixed3[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_34 (BatchNo (None, None, None, 1 384         conv2d_34[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_34 (Activation)      (None, None, None, 1 0           batch_normalization_34[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_35 (Conv2D)              (None, None, None, 1 114688      activation_34[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_35 (BatchNo (None, None, None, 1 384         conv2d_35[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_35 (Activation)      (None, None, None, 1 0           batch_normalization_35[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_31 (Conv2D)              (None, None, None, 1 98304       mixed3[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_36 (Conv2D)              (None, None, None, 1 114688      activation_35[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_31 (BatchNo (None, None, None, 1 384         conv2d_31[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_36 (BatchNo (None, None, None, 1 384         conv2d_36[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_31 (Activation)      (None, None, None, 1 0           batch_normalization_31[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_36 (Activation)      (None, None, None, 1 0           batch_normalization_36[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_32 (Conv2D)              (None, None, None, 1 114688      activation_31[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_37 (Conv2D)              (None, None, None, 1 114688      activation_36[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_32 (BatchNo (None, None, None, 1 384         conv2d_32[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_37 (BatchNo (None, None, None, 1 384         conv2d_37[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_32 (Activation)      (None, None, None, 1 0           batch_normalization_32[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_37 (Activation)      (None, None, None, 1 0           batch_normalization_37[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "average_pooling2d_3 (AveragePoo (None, None, None, 7 0           mixed3[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_30 (Conv2D)              (None, None, None, 1 147456      mixed3[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_33 (Conv2D)              (None, None, None, 1 172032      activation_32[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_38 (Conv2D)              (None, None, None, 1 172032      activation_37[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_39 (Conv2D)              (None, None, None, 1 147456      average_pooling2d_3[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_30 (BatchNo (None, None, None, 1 576         conv2d_30[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_33 (BatchNo (None, None, None, 1 576         conv2d_33[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_38 (BatchNo (None, None, None, 1 576         conv2d_38[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_39 (BatchNo (None, None, None, 1 576         conv2d_39[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_30 (Activation)      (None, None, None, 1 0           batch_normalization_30[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_33 (Activation)      (None, None, None, 1 0           batch_normalization_33[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_38 (Activation)      (None, None, None, 1 0           batch_normalization_38[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_39 (Activation)      (None, None, None, 1 0           batch_normalization_39[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "mixed4 (Concatenate)            (None, None, None, 7 0           activation_30[0][0]              \n",
            "                                                                 activation_33[0][0]              \n",
            "                                                                 activation_38[0][0]              \n",
            "                                                                 activation_39[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_44 (Conv2D)              (None, None, None, 1 122880      mixed4[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_44 (BatchNo (None, None, None, 1 480         conv2d_44[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_44 (Activation)      (None, None, None, 1 0           batch_normalization_44[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_45 (Conv2D)              (None, None, None, 1 179200      activation_44[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_45 (BatchNo (None, None, None, 1 480         conv2d_45[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_45 (Activation)      (None, None, None, 1 0           batch_normalization_45[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_41 (Conv2D)              (None, None, None, 1 122880      mixed4[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_46 (Conv2D)              (None, None, None, 1 179200      activation_45[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_41 (BatchNo (None, None, None, 1 480         conv2d_41[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_46 (BatchNo (None, None, None, 1 480         conv2d_46[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_41 (Activation)      (None, None, None, 1 0           batch_normalization_41[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_46 (Activation)      (None, None, None, 1 0           batch_normalization_46[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_42 (Conv2D)              (None, None, None, 1 179200      activation_41[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_47 (Conv2D)              (None, None, None, 1 179200      activation_46[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_42 (BatchNo (None, None, None, 1 480         conv2d_42[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_47 (BatchNo (None, None, None, 1 480         conv2d_47[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_42 (Activation)      (None, None, None, 1 0           batch_normalization_42[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_47 (Activation)      (None, None, None, 1 0           batch_normalization_47[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "average_pooling2d_4 (AveragePoo (None, None, None, 7 0           mixed4[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_40 (Conv2D)              (None, None, None, 1 147456      mixed4[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_43 (Conv2D)              (None, None, None, 1 215040      activation_42[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_48 (Conv2D)              (None, None, None, 1 215040      activation_47[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_49 (Conv2D)              (None, None, None, 1 147456      average_pooling2d_4[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_40 (BatchNo (None, None, None, 1 576         conv2d_40[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_43 (BatchNo (None, None, None, 1 576         conv2d_43[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_48 (BatchNo (None, None, None, 1 576         conv2d_48[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_49 (BatchNo (None, None, None, 1 576         conv2d_49[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_40 (Activation)      (None, None, None, 1 0           batch_normalization_40[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_43 (Activation)      (None, None, None, 1 0           batch_normalization_43[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_48 (Activation)      (None, None, None, 1 0           batch_normalization_48[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_49 (Activation)      (None, None, None, 1 0           batch_normalization_49[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "mixed5 (Concatenate)            (None, None, None, 7 0           activation_40[0][0]              \n",
            "                                                                 activation_43[0][0]              \n",
            "                                                                 activation_48[0][0]              \n",
            "                                                                 activation_49[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_54 (Conv2D)              (None, None, None, 1 122880      mixed5[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_54 (BatchNo (None, None, None, 1 480         conv2d_54[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_54 (Activation)      (None, None, None, 1 0           batch_normalization_54[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_55 (Conv2D)              (None, None, None, 1 179200      activation_54[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_55 (BatchNo (None, None, None, 1 480         conv2d_55[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_55 (Activation)      (None, None, None, 1 0           batch_normalization_55[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_51 (Conv2D)              (None, None, None, 1 122880      mixed5[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_56 (Conv2D)              (None, None, None, 1 179200      activation_55[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_51 (BatchNo (None, None, None, 1 480         conv2d_51[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_56 (BatchNo (None, None, None, 1 480         conv2d_56[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_51 (Activation)      (None, None, None, 1 0           batch_normalization_51[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_56 (Activation)      (None, None, None, 1 0           batch_normalization_56[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_52 (Conv2D)              (None, None, None, 1 179200      activation_51[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_57 (Conv2D)              (None, None, None, 1 179200      activation_56[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_52 (BatchNo (None, None, None, 1 480         conv2d_52[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_57 (BatchNo (None, None, None, 1 480         conv2d_57[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_52 (Activation)      (None, None, None, 1 0           batch_normalization_52[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_57 (Activation)      (None, None, None, 1 0           batch_normalization_57[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "average_pooling2d_5 (AveragePoo (None, None, None, 7 0           mixed5[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_50 (Conv2D)              (None, None, None, 1 147456      mixed5[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_53 (Conv2D)              (None, None, None, 1 215040      activation_52[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_58 (Conv2D)              (None, None, None, 1 215040      activation_57[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_59 (Conv2D)              (None, None, None, 1 147456      average_pooling2d_5[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_50 (BatchNo (None, None, None, 1 576         conv2d_50[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_53 (BatchNo (None, None, None, 1 576         conv2d_53[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_58 (BatchNo (None, None, None, 1 576         conv2d_58[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_59 (BatchNo (None, None, None, 1 576         conv2d_59[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_50 (Activation)      (None, None, None, 1 0           batch_normalization_50[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_53 (Activation)      (None, None, None, 1 0           batch_normalization_53[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_58 (Activation)      (None, None, None, 1 0           batch_normalization_58[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_59 (Activation)      (None, None, None, 1 0           batch_normalization_59[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "mixed6 (Concatenate)            (None, None, None, 7 0           activation_50[0][0]              \n",
            "                                                                 activation_53[0][0]              \n",
            "                                                                 activation_58[0][0]              \n",
            "                                                                 activation_59[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_64 (Conv2D)              (None, None, None, 1 147456      mixed6[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_64 (BatchNo (None, None, None, 1 576         conv2d_64[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_64 (Activation)      (None, None, None, 1 0           batch_normalization_64[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_65 (Conv2D)              (None, None, None, 1 258048      activation_64[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_65 (BatchNo (None, None, None, 1 576         conv2d_65[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_65 (Activation)      (None, None, None, 1 0           batch_normalization_65[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_61 (Conv2D)              (None, None, None, 1 147456      mixed6[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_66 (Conv2D)              (None, None, None, 1 258048      activation_65[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_61 (BatchNo (None, None, None, 1 576         conv2d_61[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_66 (BatchNo (None, None, None, 1 576         conv2d_66[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_61 (Activation)      (None, None, None, 1 0           batch_normalization_61[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_66 (Activation)      (None, None, None, 1 0           batch_normalization_66[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_62 (Conv2D)              (None, None, None, 1 258048      activation_61[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_67 (Conv2D)              (None, None, None, 1 258048      activation_66[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_62 (BatchNo (None, None, None, 1 576         conv2d_62[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_67 (BatchNo (None, None, None, 1 576         conv2d_67[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_62 (Activation)      (None, None, None, 1 0           batch_normalization_62[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_67 (Activation)      (None, None, None, 1 0           batch_normalization_67[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "average_pooling2d_6 (AveragePoo (None, None, None, 7 0           mixed6[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_60 (Conv2D)              (None, None, None, 1 147456      mixed6[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_63 (Conv2D)              (None, None, None, 1 258048      activation_62[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_68 (Conv2D)              (None, None, None, 1 258048      activation_67[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_69 (Conv2D)              (None, None, None, 1 147456      average_pooling2d_6[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_60 (BatchNo (None, None, None, 1 576         conv2d_60[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_63 (BatchNo (None, None, None, 1 576         conv2d_63[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_68 (BatchNo (None, None, None, 1 576         conv2d_68[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_69 (BatchNo (None, None, None, 1 576         conv2d_69[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_60 (Activation)      (None, None, None, 1 0           batch_normalization_60[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_63 (Activation)      (None, None, None, 1 0           batch_normalization_63[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_68 (Activation)      (None, None, None, 1 0           batch_normalization_68[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_69 (Activation)      (None, None, None, 1 0           batch_normalization_69[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "mixed7 (Concatenate)            (None, None, None, 7 0           activation_60[0][0]              \n",
            "                                                                 activation_63[0][0]              \n",
            "                                                                 activation_68[0][0]              \n",
            "                                                                 activation_69[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_72 (Conv2D)              (None, None, None, 1 147456      mixed7[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_72 (BatchNo (None, None, None, 1 576         conv2d_72[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_72 (Activation)      (None, None, None, 1 0           batch_normalization_72[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_73 (Conv2D)              (None, None, None, 1 258048      activation_72[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_73 (BatchNo (None, None, None, 1 576         conv2d_73[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_73 (Activation)      (None, None, None, 1 0           batch_normalization_73[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_70 (Conv2D)              (None, None, None, 1 147456      mixed7[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_74 (Conv2D)              (None, None, None, 1 258048      activation_73[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_70 (BatchNo (None, None, None, 1 576         conv2d_70[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_74 (BatchNo (None, None, None, 1 576         conv2d_74[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_70 (Activation)      (None, None, None, 1 0           batch_normalization_70[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_74 (Activation)      (None, None, None, 1 0           batch_normalization_74[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_71 (Conv2D)              (None, None, None, 3 552960      activation_70[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_75 (Conv2D)              (None, None, None, 1 331776      activation_74[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_71 (BatchNo (None, None, None, 3 960         conv2d_71[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_75 (BatchNo (None, None, None, 1 576         conv2d_75[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_71 (Activation)      (None, None, None, 3 0           batch_normalization_71[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_75 (Activation)      (None, None, None, 1 0           batch_normalization_75[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling2d_3 (MaxPooling2D)  (None, None, None, 7 0           mixed7[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "mixed8 (Concatenate)            (None, None, None, 1 0           activation_71[0][0]              \n",
            "                                                                 activation_75[0][0]              \n",
            "                                                                 max_pooling2d_3[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_80 (Conv2D)              (None, None, None, 4 573440      mixed8[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_80 (BatchNo (None, None, None, 4 1344        conv2d_80[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_80 (Activation)      (None, None, None, 4 0           batch_normalization_80[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_77 (Conv2D)              (None, None, None, 3 491520      mixed8[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_81 (Conv2D)              (None, None, None, 3 1548288     activation_80[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_77 (BatchNo (None, None, None, 3 1152        conv2d_77[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_81 (BatchNo (None, None, None, 3 1152        conv2d_81[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_77 (Activation)      (None, None, None, 3 0           batch_normalization_77[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_81 (Activation)      (None, None, None, 3 0           batch_normalization_81[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_78 (Conv2D)              (None, None, None, 3 442368      activation_77[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_79 (Conv2D)              (None, None, None, 3 442368      activation_77[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_82 (Conv2D)              (None, None, None, 3 442368      activation_81[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_83 (Conv2D)              (None, None, None, 3 442368      activation_81[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "average_pooling2d_7 (AveragePoo (None, None, None, 1 0           mixed8[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_76 (Conv2D)              (None, None, None, 3 409600      mixed8[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_78 (BatchNo (None, None, None, 3 1152        conv2d_78[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_79 (BatchNo (None, None, None, 3 1152        conv2d_79[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_82 (BatchNo (None, None, None, 3 1152        conv2d_82[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_83 (BatchNo (None, None, None, 3 1152        conv2d_83[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_84 (Conv2D)              (None, None, None, 1 245760      average_pooling2d_7[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_76 (BatchNo (None, None, None, 3 960         conv2d_76[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_78 (Activation)      (None, None, None, 3 0           batch_normalization_78[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_79 (Activation)      (None, None, None, 3 0           batch_normalization_79[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_82 (Activation)      (None, None, None, 3 0           batch_normalization_82[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_83 (Activation)      (None, None, None, 3 0           batch_normalization_83[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_84 (BatchNo (None, None, None, 1 576         conv2d_84[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_76 (Activation)      (None, None, None, 3 0           batch_normalization_76[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "mixed9_0 (Concatenate)          (None, None, None, 7 0           activation_78[0][0]              \n",
            "                                                                 activation_79[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, None, None, 7 0           activation_82[0][0]              \n",
            "                                                                 activation_83[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "activation_84 (Activation)      (None, None, None, 1 0           batch_normalization_84[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "mixed9 (Concatenate)            (None, None, None, 2 0           activation_76[0][0]              \n",
            "                                                                 mixed9_0[0][0]                   \n",
            "                                                                 concatenate[0][0]                \n",
            "                                                                 activation_84[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_89 (Conv2D)              (None, None, None, 4 917504      mixed9[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_89 (BatchNo (None, None, None, 4 1344        conv2d_89[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_89 (Activation)      (None, None, None, 4 0           batch_normalization_89[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_86 (Conv2D)              (None, None, None, 3 786432      mixed9[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_90 (Conv2D)              (None, None, None, 3 1548288     activation_89[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_86 (BatchNo (None, None, None, 3 1152        conv2d_86[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_90 (BatchNo (None, None, None, 3 1152        conv2d_90[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_86 (Activation)      (None, None, None, 3 0           batch_normalization_86[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_90 (Activation)      (None, None, None, 3 0           batch_normalization_90[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_87 (Conv2D)              (None, None, None, 3 442368      activation_86[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_88 (Conv2D)              (None, None, None, 3 442368      activation_86[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_91 (Conv2D)              (None, None, None, 3 442368      activation_90[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_92 (Conv2D)              (None, None, None, 3 442368      activation_90[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "average_pooling2d_8 (AveragePoo (None, None, None, 2 0           mixed9[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_85 (Conv2D)              (None, None, None, 3 655360      mixed9[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_87 (BatchNo (None, None, None, 3 1152        conv2d_87[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_88 (BatchNo (None, None, None, 3 1152        conv2d_88[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_91 (BatchNo (None, None, None, 3 1152        conv2d_91[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_92 (BatchNo (None, None, None, 3 1152        conv2d_92[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_93 (Conv2D)              (None, None, None, 1 393216      average_pooling2d_8[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_85 (BatchNo (None, None, None, 3 960         conv2d_85[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_87 (Activation)      (None, None, None, 3 0           batch_normalization_87[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_88 (Activation)      (None, None, None, 3 0           batch_normalization_88[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_91 (Activation)      (None, None, None, 3 0           batch_normalization_91[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_92 (Activation)      (None, None, None, 3 0           batch_normalization_92[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_93 (BatchNo (None, None, None, 1 576         conv2d_93[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_85 (Activation)      (None, None, None, 3 0           batch_normalization_85[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "mixed9_1 (Concatenate)          (None, None, None, 7 0           activation_87[0][0]              \n",
            "                                                                 activation_88[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "concatenate_1 (Concatenate)     (None, None, None, 7 0           activation_91[0][0]              \n",
            "                                                                 activation_92[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "activation_93 (Activation)      (None, None, None, 1 0           batch_normalization_93[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "mixed10 (Concatenate)           (None, None, None, 2 0           activation_85[0][0]              \n",
            "                                                                 mixed9_1[0][0]                   \n",
            "                                                                 concatenate_1[0][0]              \n",
            "                                                                 activation_93[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "global_average_pooling2d_1 (Glo (None, 2048)         0           mixed10[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "dense_2 (Dense)                 (None, 1024)         2098176     global_average_pooling2d_1[0][0] \n",
            "__________________________________________________________________________________________________\n",
            "dropout_1 (Dropout)             (None, 1024)         0           dense_2[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "dense_3 (Dense)                 (None, 3)            3075        dropout_1[0][0]                  \n",
            "==================================================================================================\n",
            "Total params: 23,904,035\n",
            "Trainable params: 2,101,251\n",
            "Non-trainable params: 21,802,784\n",
            "__________________________________________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tFaIWvWICpYM"
      },
      "source": [
        "### Model Training"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9f8WNPlZzqCQ",
        "outputId": "31b31708-3959-496d-f60d-449eb3911b2a"
      },
      "source": [
        "model.fit(\n",
        "    train_generator,\n",
        "    validation_data=validation_generator,\n",
        "    steps_per_epoch=1126//32,\n",
        "    epochs=5,\n",
        "    validation_steps=280//32)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/5\n",
            "35/35 [==============================] - 31s 771ms/step - loss: 1.1153 - accuracy: 0.6746 - val_loss: 0.3353 - val_accuracy: 0.8945\n",
            "Epoch 2/5\n",
            "35/35 [==============================] - 26s 733ms/step - loss: 0.2571 - accuracy: 0.9333 - val_loss: 0.1917 - val_accuracy: 0.9414\n",
            "Epoch 3/5\n",
            "35/35 [==============================] - 26s 738ms/step - loss: 0.2115 - accuracy: 0.9287 - val_loss: 0.1302 - val_accuracy: 0.9570\n",
            "Epoch 4/5\n",
            "35/35 [==============================] - 25s 729ms/step - loss: 0.1600 - accuracy: 0.9543 - val_loss: 0.1688 - val_accuracy: 0.9492\n",
            "Epoch 5/5\n",
            "35/35 [==============================] - 26s 733ms/step - loss: 0.1605 - accuracy: 0.9506 - val_loss: 0.1190 - val_accuracy: 0.9570\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "<tensorflow.python.keras.callbacks.History at 0x7f618406e590>"
            ]
          },
          "execution_count": 13,
          "metadata": {
            "tags": []
          },
          "output_type": "execute_result"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KvzI7YknJbJC"
      },
      "source": [
        "### Model Accuracy"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "brX7sMVUJa10",
        "outputId": "fefdaef0-194d-48e1-99a6-ccd6a4af04b3"
      },
      "source": [
        "score = model.evaluate(test_generator, verbose=0)\n",
        "print(\"Test loss:\", score[0])\n",
        "print(\"Test accuracy:\", score[1])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Test loss: 0.03664422780275345\n",
            "Test accuracy: 0.9910714030265808\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ir20KcmrCsqz"
      },
      "source": [
        "### Classification Report and Confusion Matrix Results"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 628
        },
        "id": "KZM_J_rvzrx9",
        "outputId": "729cfb42-4d64-4f96-818c-d66710c69a76"
      },
      "source": [
        "predictions = model.predict(test_generator, steps=test_steps_per_epoch)\n",
        "# Get most likely class\n",
        "predicted_classes = numpy.argmax(predictions, axis=1)\n",
        "\n",
        "true_classes = test_generator.classes\n",
        "class_labels = list(test_generator.class_indices.keys())   \n",
        "\n",
        "report = classification_report(true_classes, predicted_classes, target_names=class_labels, zero_division=0)\n",
        "print(report)\n",
        "cmrforest = confusion_matrix(true_classes, predicted_classes)\n",
        "plt.figure(figsize=(10,7))\n",
        "sns.heatmap(cmrforest ,annot=True , fmt = 'd')\n",
        "plt.xlabel('Predicted')\n",
        "plt.ylabel('Truth')\n",
        "plt.title('Confusion Report')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "               precision    recall  f1-score   support\n",
            "\n",
            "       Danube       0.98      1.00      0.99       109\n",
            "      Dnieper       1.00      0.97      0.99        78\n",
            "UnknownOrigin       1.00      1.00      1.00        37\n",
            "\n",
            "     accuracy                           0.99       224\n",
            "    macro avg       0.99      0.99      0.99       224\n",
            " weighted avg       0.99      0.99      0.99       224\n",
            "\n"
          ]
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x504 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light",
            "tags": []
          },
          "output_type": "display_data"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "b-9AUv_sCzHK"
      },
      "source": [
        "## MobileNetV2"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MBk9xGPbC3Bn"
      },
      "source": [
        "### Initialization and Summary"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Kodvkg930LBd",
        "outputId": "0d1b4a65-738e-44de-820c-59ee56cba441"
      },
      "source": [
        "base_model = MobileNetV2(weights='imagenet', include_top=False)\n",
        "base_model.trainable = False\n",
        "\n",
        "# add a global spatial average pooling layer\n",
        "x = base_model.output\n",
        "x = GlobalAveragePooling2D()(x)\n",
        "# add a fully-connected layer\n",
        "x = Dense(1024, activation='relu')(x)\n",
        "# add a dropout layer\n",
        "x = Dropout(0.2)(x)\n",
        "# and a logistic layer -- we have 3 classes\n",
        "predictions = Dense(3, activation='softmax')(x)\n",
        "\n",
        "# this is the model we will train\n",
        "model = Model(inputs=base_model.input, outputs=predictions)\n",
        "\n",
        "# compile the model (should be done *after* setting layers to non-trainable)\n",
        "model.compile(loss=\"categorical_crossentropy\", optimizer=\"adam\", metrics=[\"accuracy\"])\n",
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:`input_shape` is undefined or non-square, or `rows` is not in [96, 128, 160, 192, 224]. Weights for input shape (224, 224) will be loaded as the default.\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/mobilenet_v2/mobilenet_v2_weights_tf_dim_ordering_tf_kernels_1.0_224_no_top.h5\n",
            "9412608/9406464 [==============================] - 0s 0us/step\n",
            "Model: \"model_2\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_3 (InputLayer)            [(None, None, None,  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "Conv1 (Conv2D)                  (None, None, None, 3 864         input_3[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "bn_Conv1 (BatchNormalization)   (None, None, None, 3 128         Conv1[0][0]                      \n",
            "__________________________________________________________________________________________________\n",
            "Conv1_relu (ReLU)               (None, None, None, 3 0           bn_Conv1[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "expanded_conv_depthwise (Depthw (None, None, None, 3 288         Conv1_relu[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "expanded_conv_depthwise_BN (Bat (None, None, None, 3 128         expanded_conv_depthwise[0][0]    \n",
            "__________________________________________________________________________________________________\n",
            "expanded_conv_depthwise_relu (R (None, None, None, 3 0           expanded_conv_depthwise_BN[0][0] \n",
            "__________________________________________________________________________________________________\n",
            "expanded_conv_project (Conv2D)  (None, None, None, 1 512         expanded_conv_depthwise_relu[0][0\n",
            "__________________________________________________________________________________________________\n",
            "expanded_conv_project_BN (Batch (None, None, None, 1 64          expanded_conv_project[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "block_1_expand (Conv2D)         (None, None, None, 9 1536        expanded_conv_project_BN[0][0]   \n",
            "__________________________________________________________________________________________________\n",
            "block_1_expand_BN (BatchNormali (None, None, None, 9 384         block_1_expand[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "block_1_expand_relu (ReLU)      (None, None, None, 9 0           block_1_expand_BN[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "block_1_pad (ZeroPadding2D)     (None, None, None, 9 0           block_1_expand_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "block_1_depthwise (DepthwiseCon (None, None, None, 9 864         block_1_pad[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "block_1_depthwise_BN (BatchNorm (None, None, None, 9 384         block_1_depthwise[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "block_1_depthwise_relu (ReLU)   (None, None, None, 9 0           block_1_depthwise_BN[0][0]       \n",
            "__________________________________________________________________________________________________\n",
            "block_1_project (Conv2D)        (None, None, None, 2 2304        block_1_depthwise_relu[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "block_1_project_BN (BatchNormal (None, None, None, 2 96          block_1_project[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "block_2_expand (Conv2D)         (None, None, None, 1 3456        block_1_project_BN[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_2_expand_BN (BatchNormali (None, None, None, 1 576         block_2_expand[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "block_2_expand_relu (ReLU)      (None, None, None, 1 0           block_2_expand_BN[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "block_2_depthwise (DepthwiseCon (None, None, None, 1 1296        block_2_expand_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "block_2_depthwise_BN (BatchNorm (None, None, None, 1 576         block_2_depthwise[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "block_2_depthwise_relu (ReLU)   (None, None, None, 1 0           block_2_depthwise_BN[0][0]       \n",
            "__________________________________________________________________________________________________\n",
            "block_2_project (Conv2D)        (None, None, None, 2 3456        block_2_depthwise_relu[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "block_2_project_BN (BatchNormal (None, None, None, 2 96          block_2_project[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "block_2_add (Add)               (None, None, None, 2 0           block_1_project_BN[0][0]         \n",
            "                                                                 block_2_project_BN[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_3_expand (Conv2D)         (None, None, None, 1 3456        block_2_add[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "block_3_expand_BN (BatchNormali (None, None, None, 1 576         block_3_expand[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "block_3_expand_relu (ReLU)      (None, None, None, 1 0           block_3_expand_BN[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "block_3_pad (ZeroPadding2D)     (None, None, None, 1 0           block_3_expand_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "block_3_depthwise (DepthwiseCon (None, None, None, 1 1296        block_3_pad[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "block_3_depthwise_BN (BatchNorm (None, None, None, 1 576         block_3_depthwise[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "block_3_depthwise_relu (ReLU)   (None, None, None, 1 0           block_3_depthwise_BN[0][0]       \n",
            "__________________________________________________________________________________________________\n",
            "block_3_project (Conv2D)        (None, None, None, 3 4608        block_3_depthwise_relu[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "block_3_project_BN (BatchNormal (None, None, None, 3 128         block_3_project[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "block_4_expand (Conv2D)         (None, None, None, 1 6144        block_3_project_BN[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_4_expand_BN (BatchNormali (None, None, None, 1 768         block_4_expand[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "block_4_expand_relu (ReLU)      (None, None, None, 1 0           block_4_expand_BN[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "block_4_depthwise (DepthwiseCon (None, None, None, 1 1728        block_4_expand_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "block_4_depthwise_BN (BatchNorm (None, None, None, 1 768         block_4_depthwise[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "block_4_depthwise_relu (ReLU)   (None, None, None, 1 0           block_4_depthwise_BN[0][0]       \n",
            "__________________________________________________________________________________________________\n",
            "block_4_project (Conv2D)        (None, None, None, 3 6144        block_4_depthwise_relu[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "block_4_project_BN (BatchNormal (None, None, None, 3 128         block_4_project[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "block_4_add (Add)               (None, None, None, 3 0           block_3_project_BN[0][0]         \n",
            "                                                                 block_4_project_BN[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_5_expand (Conv2D)         (None, None, None, 1 6144        block_4_add[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "block_5_expand_BN (BatchNormali (None, None, None, 1 768         block_5_expand[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "block_5_expand_relu (ReLU)      (None, None, None, 1 0           block_5_expand_BN[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "block_5_depthwise (DepthwiseCon (None, None, None, 1 1728        block_5_expand_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "block_5_depthwise_BN (BatchNorm (None, None, None, 1 768         block_5_depthwise[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "block_5_depthwise_relu (ReLU)   (None, None, None, 1 0           block_5_depthwise_BN[0][0]       \n",
            "__________________________________________________________________________________________________\n",
            "block_5_project (Conv2D)        (None, None, None, 3 6144        block_5_depthwise_relu[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "block_5_project_BN (BatchNormal (None, None, None, 3 128         block_5_project[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "block_5_add (Add)               (None, None, None, 3 0           block_4_add[0][0]                \n",
            "                                                                 block_5_project_BN[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_6_expand (Conv2D)         (None, None, None, 1 6144        block_5_add[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "block_6_expand_BN (BatchNormali (None, None, None, 1 768         block_6_expand[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "block_6_expand_relu (ReLU)      (None, None, None, 1 0           block_6_expand_BN[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "block_6_pad (ZeroPadding2D)     (None, None, None, 1 0           block_6_expand_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "block_6_depthwise (DepthwiseCon (None, None, None, 1 1728        block_6_pad[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "block_6_depthwise_BN (BatchNorm (None, None, None, 1 768         block_6_depthwise[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "block_6_depthwise_relu (ReLU)   (None, None, None, 1 0           block_6_depthwise_BN[0][0]       \n",
            "__________________________________________________________________________________________________\n",
            "block_6_project (Conv2D)        (None, None, None, 6 12288       block_6_depthwise_relu[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "block_6_project_BN (BatchNormal (None, None, None, 6 256         block_6_project[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "block_7_expand (Conv2D)         (None, None, None, 3 24576       block_6_project_BN[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_7_expand_BN (BatchNormali (None, None, None, 3 1536        block_7_expand[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "block_7_expand_relu (ReLU)      (None, None, None, 3 0           block_7_expand_BN[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "block_7_depthwise (DepthwiseCon (None, None, None, 3 3456        block_7_expand_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "block_7_depthwise_BN (BatchNorm (None, None, None, 3 1536        block_7_depthwise[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "block_7_depthwise_relu (ReLU)   (None, None, None, 3 0           block_7_depthwise_BN[0][0]       \n",
            "__________________________________________________________________________________________________\n",
            "block_7_project (Conv2D)        (None, None, None, 6 24576       block_7_depthwise_relu[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "block_7_project_BN (BatchNormal (None, None, None, 6 256         block_7_project[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "block_7_add (Add)               (None, None, None, 6 0           block_6_project_BN[0][0]         \n",
            "                                                                 block_7_project_BN[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_8_expand (Conv2D)         (None, None, None, 3 24576       block_7_add[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "block_8_expand_BN (BatchNormali (None, None, None, 3 1536        block_8_expand[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "block_8_expand_relu (ReLU)      (None, None, None, 3 0           block_8_expand_BN[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "block_8_depthwise (DepthwiseCon (None, None, None, 3 3456        block_8_expand_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "block_8_depthwise_BN (BatchNorm (None, None, None, 3 1536        block_8_depthwise[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "block_8_depthwise_relu (ReLU)   (None, None, None, 3 0           block_8_depthwise_BN[0][0]       \n",
            "__________________________________________________________________________________________________\n",
            "block_8_project (Conv2D)        (None, None, None, 6 24576       block_8_depthwise_relu[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "block_8_project_BN (BatchNormal (None, None, None, 6 256         block_8_project[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "block_8_add (Add)               (None, None, None, 6 0           block_7_add[0][0]                \n",
            "                                                                 block_8_project_BN[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_9_expand (Conv2D)         (None, None, None, 3 24576       block_8_add[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "block_9_expand_BN (BatchNormali (None, None, None, 3 1536        block_9_expand[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "block_9_expand_relu (ReLU)      (None, None, None, 3 0           block_9_expand_BN[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "block_9_depthwise (DepthwiseCon (None, None, None, 3 3456        block_9_expand_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "block_9_depthwise_BN (BatchNorm (None, None, None, 3 1536        block_9_depthwise[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "block_9_depthwise_relu (ReLU)   (None, None, None, 3 0           block_9_depthwise_BN[0][0]       \n",
            "__________________________________________________________________________________________________\n",
            "block_9_project (Conv2D)        (None, None, None, 6 24576       block_9_depthwise_relu[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "block_9_project_BN (BatchNormal (None, None, None, 6 256         block_9_project[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "block_9_add (Add)               (None, None, None, 6 0           block_8_add[0][0]                \n",
            "                                                                 block_9_project_BN[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_10_expand (Conv2D)        (None, None, None, 3 24576       block_9_add[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "block_10_expand_BN (BatchNormal (None, None, None, 3 1536        block_10_expand[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "block_10_expand_relu (ReLU)     (None, None, None, 3 0           block_10_expand_BN[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_10_depthwise (DepthwiseCo (None, None, None, 3 3456        block_10_expand_relu[0][0]       \n",
            "__________________________________________________________________________________________________\n",
            "block_10_depthwise_BN (BatchNor (None, None, None, 3 1536        block_10_depthwise[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_10_depthwise_relu (ReLU)  (None, None, None, 3 0           block_10_depthwise_BN[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "block_10_project (Conv2D)       (None, None, None, 9 36864       block_10_depthwise_relu[0][0]    \n",
            "__________________________________________________________________________________________________\n",
            "block_10_project_BN (BatchNorma (None, None, None, 9 384         block_10_project[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "block_11_expand (Conv2D)        (None, None, None, 5 55296       block_10_project_BN[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "block_11_expand_BN (BatchNormal (None, None, None, 5 2304        block_11_expand[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "block_11_expand_relu (ReLU)     (None, None, None, 5 0           block_11_expand_BN[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_11_depthwise (DepthwiseCo (None, None, None, 5 5184        block_11_expand_relu[0][0]       \n",
            "__________________________________________________________________________________________________\n",
            "block_11_depthwise_BN (BatchNor (None, None, None, 5 2304        block_11_depthwise[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_11_depthwise_relu (ReLU)  (None, None, None, 5 0           block_11_depthwise_BN[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "block_11_project (Conv2D)       (None, None, None, 9 55296       block_11_depthwise_relu[0][0]    \n",
            "__________________________________________________________________________________________________\n",
            "block_11_project_BN (BatchNorma (None, None, None, 9 384         block_11_project[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "block_11_add (Add)              (None, None, None, 9 0           block_10_project_BN[0][0]        \n",
            "                                                                 block_11_project_BN[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "block_12_expand (Conv2D)        (None, None, None, 5 55296       block_11_add[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "block_12_expand_BN (BatchNormal (None, None, None, 5 2304        block_12_expand[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "block_12_expand_relu (ReLU)     (None, None, None, 5 0           block_12_expand_BN[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_12_depthwise (DepthwiseCo (None, None, None, 5 5184        block_12_expand_relu[0][0]       \n",
            "__________________________________________________________________________________________________\n",
            "block_12_depthwise_BN (BatchNor (None, None, None, 5 2304        block_12_depthwise[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_12_depthwise_relu (ReLU)  (None, None, None, 5 0           block_12_depthwise_BN[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "block_12_project (Conv2D)       (None, None, None, 9 55296       block_12_depthwise_relu[0][0]    \n",
            "__________________________________________________________________________________________________\n",
            "block_12_project_BN (BatchNorma (None, None, None, 9 384         block_12_project[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "block_12_add (Add)              (None, None, None, 9 0           block_11_add[0][0]               \n",
            "                                                                 block_12_project_BN[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "block_13_expand (Conv2D)        (None, None, None, 5 55296       block_12_add[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "block_13_expand_BN (BatchNormal (None, None, None, 5 2304        block_13_expand[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "block_13_expand_relu (ReLU)     (None, None, None, 5 0           block_13_expand_BN[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_13_pad (ZeroPadding2D)    (None, None, None, 5 0           block_13_expand_relu[0][0]       \n",
            "__________________________________________________________________________________________________\n",
            "block_13_depthwise (DepthwiseCo (None, None, None, 5 5184        block_13_pad[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "block_13_depthwise_BN (BatchNor (None, None, None, 5 2304        block_13_depthwise[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_13_depthwise_relu (ReLU)  (None, None, None, 5 0           block_13_depthwise_BN[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "block_13_project (Conv2D)       (None, None, None, 1 92160       block_13_depthwise_relu[0][0]    \n",
            "__________________________________________________________________________________________________\n",
            "block_13_project_BN (BatchNorma (None, None, None, 1 640         block_13_project[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "block_14_expand (Conv2D)        (None, None, None, 9 153600      block_13_project_BN[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "block_14_expand_BN (BatchNormal (None, None, None, 9 3840        block_14_expand[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "block_14_expand_relu (ReLU)     (None, None, None, 9 0           block_14_expand_BN[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_14_depthwise (DepthwiseCo (None, None, None, 9 8640        block_14_expand_relu[0][0]       \n",
            "__________________________________________________________________________________________________\n",
            "block_14_depthwise_BN (BatchNor (None, None, None, 9 3840        block_14_depthwise[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_14_depthwise_relu (ReLU)  (None, None, None, 9 0           block_14_depthwise_BN[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "block_14_project (Conv2D)       (None, None, None, 1 153600      block_14_depthwise_relu[0][0]    \n",
            "__________________________________________________________________________________________________\n",
            "block_14_project_BN (BatchNorma (None, None, None, 1 640         block_14_project[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "block_14_add (Add)              (None, None, None, 1 0           block_13_project_BN[0][0]        \n",
            "                                                                 block_14_project_BN[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "block_15_expand (Conv2D)        (None, None, None, 9 153600      block_14_add[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "block_15_expand_BN (BatchNormal (None, None, None, 9 3840        block_15_expand[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "block_15_expand_relu (ReLU)     (None, None, None, 9 0           block_15_expand_BN[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_15_depthwise (DepthwiseCo (None, None, None, 9 8640        block_15_expand_relu[0][0]       \n",
            "__________________________________________________________________________________________________\n",
            "block_15_depthwise_BN (BatchNor (None, None, None, 9 3840        block_15_depthwise[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_15_depthwise_relu (ReLU)  (None, None, None, 9 0           block_15_depthwise_BN[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "block_15_project (Conv2D)       (None, None, None, 1 153600      block_15_depthwise_relu[0][0]    \n",
            "__________________________________________________________________________________________________\n",
            "block_15_project_BN (BatchNorma (None, None, None, 1 640         block_15_project[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "block_15_add (Add)              (None, None, None, 1 0           block_14_add[0][0]               \n",
            "                                                                 block_15_project_BN[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "block_16_expand (Conv2D)        (None, None, None, 9 153600      block_15_add[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "block_16_expand_BN (BatchNormal (None, None, None, 9 3840        block_16_expand[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "block_16_expand_relu (ReLU)     (None, None, None, 9 0           block_16_expand_BN[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_16_depthwise (DepthwiseCo (None, None, None, 9 8640        block_16_expand_relu[0][0]       \n",
            "__________________________________________________________________________________________________\n",
            "block_16_depthwise_BN (BatchNor (None, None, None, 9 3840        block_16_depthwise[0][0]         \n",
            "__________________________________________________________________________________________________\n",
            "block_16_depthwise_relu (ReLU)  (None, None, None, 9 0           block_16_depthwise_BN[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "block_16_project (Conv2D)       (None, None, None, 3 307200      block_16_depthwise_relu[0][0]    \n",
            "__________________________________________________________________________________________________\n",
            "block_16_project_BN (BatchNorma (None, None, None, 3 1280        block_16_project[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "Conv_1 (Conv2D)                 (None, None, None, 1 409600      block_16_project_BN[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "Conv_1_bn (BatchNormalization)  (None, None, None, 1 5120        Conv_1[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "out_relu (ReLU)                 (None, None, None, 1 0           Conv_1_bn[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "global_average_pooling2d_2 (Glo (None, 1280)         0           out_relu[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "dense_4 (Dense)                 (None, 1024)         1311744     global_average_pooling2d_2[0][0] \n",
            "__________________________________________________________________________________________________\n",
            "dropout_2 (Dropout)             (None, 1024)         0           dense_4[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "dense_5 (Dense)                 (None, 3)            3075        dropout_2[0][0]                  \n",
            "==================================================================================================\n",
            "Total params: 3,572,803\n",
            "Trainable params: 1,314,819\n",
            "Non-trainable params: 2,257,984\n",
            "__________________________________________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "p57-N9aOC5Ex"
      },
      "source": [
        "### Model Training"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ck2TBSyM0WNK",
        "outputId": "d2841e63-c9c2-4d4e-987f-5e77907efe73"
      },
      "source": [
        "model.fit(\n",
        "    train_generator,\n",
        "    validation_data=validation_generator,\n",
        "    steps_per_epoch=1126//32,\n",
        "    epochs=9,\n",
        "    validation_steps=280//32)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/9\n",
            "35/35 [==============================] - 29s 763ms/step - loss: 0.4266 - accuracy: 0.8912 - val_loss: 0.0344 - val_accuracy: 0.9922\n",
            "Epoch 2/9\n",
            "35/35 [==============================] - 25s 724ms/step - loss: 0.1032 - accuracy: 0.9771 - val_loss: 0.1260 - val_accuracy: 0.9648\n",
            "Epoch 3/9\n",
            "35/35 [==============================] - 25s 711ms/step - loss: 0.1093 - accuracy: 0.9726 - val_loss: 0.0133 - val_accuracy: 0.9961\n",
            "Epoch 4/9\n",
            "35/35 [==============================] - 25s 713ms/step - loss: 0.0866 - accuracy: 0.9753 - val_loss: 0.2561 - val_accuracy: 0.8984\n",
            "Epoch 5/9\n",
            "35/35 [==============================] - 25s 711ms/step - loss: 0.0968 - accuracy: 0.9735 - val_loss: 0.1049 - val_accuracy: 0.9570\n",
            "Epoch 6/9\n",
            "35/35 [==============================] - 25s 713ms/step - loss: 0.0909 - accuracy: 0.9735 - val_loss: 0.0363 - val_accuracy: 0.9883\n",
            "Epoch 7/9\n",
            "35/35 [==============================] - 25s 714ms/step - loss: 0.0846 - accuracy: 0.9781 - val_loss: 0.0219 - val_accuracy: 0.9922\n",
            "Epoch 8/9\n",
            "35/35 [==============================] - 25s 714ms/step - loss: 0.0829 - accuracy: 0.9781 - val_loss: 0.0620 - val_accuracy: 0.9805\n",
            "Epoch 9/9\n",
            "35/35 [==============================] - 25s 719ms/step - loss: 0.0819 - accuracy: 0.9771 - val_loss: 0.0532 - val_accuracy: 0.9805\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "<tensorflow.python.keras.callbacks.History at 0x7f617f592050>"
            ]
          },
          "execution_count": 17,
          "metadata": {
            "tags": []
          },
          "output_type": "execute_result"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ei0q1lvFJgEo"
      },
      "source": [
        "### Model Accuracy"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WNUP1j1EJgab",
        "outputId": "6889d3d4-9958-4ee4-9a54-079a072f9ab9"
      },
      "source": [
        "score = model.evaluate(test_generator, verbose=0)\n",
        "print(\"Test loss:\", score[0])\n",
        "print(\"Test accuracy:\", score[1])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Test loss: 0.20453548431396484\n",
            "Test accuracy: 0.9285714030265808\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VWR9jkUbDAqB"
      },
      "source": [
        "### Classification Report and Confusion Matrix Results"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 628
        },
        "id": "olzLMmaB0Xkt",
        "outputId": "bc2bf172-340c-43c1-e011-9014da2a476c"
      },
      "source": [
        "predictions = model.predict(test_generator, steps=test_steps_per_epoch)\n",
        "# Get most likely class\n",
        "predicted_classes = numpy.argmax(predictions, axis=1)\n",
        "\n",
        "true_classes = test_generator.classes\n",
        "class_labels = list(test_generator.class_indices.keys())   \n",
        "\n",
        "report = classification_report(true_classes, predicted_classes, target_names=class_labels, zero_division=0)\n",
        "print(report)\n",
        "cmrforest = confusion_matrix(true_classes, predicted_classes)\n",
        "plt.figure(figsize=(10,7))\n",
        "sns.heatmap(cmrforest ,annot=True , fmt = 'd')\n",
        "plt.xlabel('Predicted')\n",
        "plt.ylabel('Truth')\n",
        "plt.title('Confusion Report')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "               precision    recall  f1-score   support\n",
            "\n",
            "       Danube       1.00      0.85      0.92       109\n",
            "      Dnieper       1.00      1.00      1.00        78\n",
            "UnknownOrigin       0.70      1.00      0.82        37\n",
            "\n",
            "     accuracy                           0.93       224\n",
            "    macro avg       0.90      0.95      0.91       224\n",
            " weighted avg       0.95      0.93      0.93       224\n",
            "\n"
          ]
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x504 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light",
            "tags": []
          },
          "output_type": "display_data"
        }
      ]
    }
  ]
}