4.Make predictions_CNN.ipynb

{
  "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",
      "source": [
        "pip install tensorflow_addons"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qrk1OjrZmJoq",
        "outputId": "4fee6fcf-c278-43e3-b761-2437bc0af568"
      },
      "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[K     |████████████████████████████████| 1.1 MB 5.5 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",
      "execution_count": 2,
      "metadata": {
        "id": "XbyZT9iGxPdK"
      },
      "outputs": [],
      "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()))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "id": "z9qaEwiW5_KT"
      },
      "outputs": [],
      "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"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sTQAgmxDAuhZ"
      },
      "source": [
        "## Load Dataset"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "lOCgLHS3AuB7"
      },
      "outputs": [],
      "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://users.uoa.gr/~nalygizakis/BS/Predicted%20spatial%20distribution%20dataset.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')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "k3fKEBZS0BaQ",
        "outputId": "2a70e416-b6e1-4fa9-8f55-3463415fe1b9"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'/content/BlackSea/train'"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ],
      "source": [
        "dataset_train"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "c7bUOEw30BaR",
        "outputId": "0f7496f8-dfeb-4e7a-f6d4-674adeb9abe3"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'/content/BlackSea/evaluation'"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ],
      "source": [
        "dataset_evaluation"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "F1EmdPK7A5Tj"
      },
      "source": [
        "## Train, Validation and Test Generators"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "id": "GOzbLuAW2AG2"
      },
      "outputs": [],
      "source": [
        "batch_size = 32\n",
        "img_height = 150\n",
        "img_width = 150"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "id": "lzdCiOx4B9vh"
      },
      "outputs": [],
      "source": [
        "datagen = image.ImageDataGenerator(\n",
        "    rescale=1./255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    validation_split=0.2)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Yf1__QmcA4qa",
        "outputId": "7cb581b4-71c7-4363-a2fe-4454acbac8f9"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 1126 images belonging to 3 classes.\n"
          ]
        }
      ],
      "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')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "48IOOsS84H0R",
        "outputId": "bf2ead73-7e92-4567-a047-992cdfaf6281"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 1406 files belonging to 3 classes.\n",
            "Using 1125 files for training.\n"
          ]
        }
      ],
      "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)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 353
        },
        "id": "cBmbVCBi7VxS",
        "outputId": "de67a633-97c6-4358-bccf-c2719b3ba845"
      },
      "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"
          }
        }
      ],
      "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\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "L-i9sojx0BaU",
        "outputId": "46b1d062-9ab1-4f9a-d26f-9d7d64b4f3db"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 280 images belonging to 3 classes.\n"
          ]
        }
      ],
      "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')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WWhKPEQH4MyV",
        "outputId": "e1391efc-47d0-4f7c-face-9a1c98b56933"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 1406 files belonging to 3 classes.\n",
            "Using 281 files for validation.\n"
          ]
        }
      ],
      "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)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "metadata": {
        "id": "PGS9G1yY7lTF"
      },
      "outputs": [],
      "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)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 15,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EFlYlQW-0BaU",
        "outputId": "5367dad5-7f0e-4c57-b172-0c911bf740ac"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 224 images belonging to 3 classes.\n"
          ]
        }
      ],
      "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')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "U1R-mmwD4ahk",
        "outputId": "d6b91b8e-73b0-4f04-c9d7-9770918cf0b3"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 224 files belonging to 3 classes.\n"
          ]
        }
      ],
      "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)"
      ]
    },
    {
      "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",
      "execution_count": 17,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KCOzteRGBUAd",
        "outputId": "17a2a93b-b62e-41cd-bc4d-b4852354e9e0"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "7\n"
          ]
        }
      ],
      "source": [
        "test_steps_per_epoch = numpy.math.ceil(test_generator.samples / test_generator.batch_size)\n",
        "print(test_steps_per_epoch)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "metadata": {
        "id": "AN3P4Glv7-14"
      },
      "outputs": [],
      "source": [
        "num_classes = 3"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0egjiA5w0BaZ"
      },
      "source": [
        "# 1.Basic tensorflow CNN"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "metadata": {
        "id": "WEyY0lrz0Bab"
      },
      "outputs": [],
      "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",
        "])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 22,
      "metadata": {
        "id": "-iRqtbGO0Bab"
      },
      "outputs": [],
      "source": [
        "model_CNN1.compile(optimizer='adam',\n",
        "              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
        "               metrics=['acc', f1_m, precision_m, recall_m])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 23,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "emVI8BQU0Bab",
        "outputId": "060568ff-e2d6-462c-8677-8736d06ff747"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\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 (Flatten)           (None, 43808)             0         \n",
            "                                                                 \n",
            " dense (Dense)               (None, 128)               5607552   \n",
            "                                                                 \n",
            " dense_1 (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"
          ]
        }
      ],
      "source": [
        "model_CNN1.summary()"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "callback = tf.keras.callbacks.EarlyStopping(monitor='f1_m', patience=10)\n",
        "epochs = 100"
      ],
      "metadata": {
        "id": "U-GlfAW_mksk"
      },
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": 25,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2Iyobyk_0Bac",
        "outputId": "19b7bfaa-8f57-41c2-a604-2475395dd04c"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "36/36 [==============================] - 27s 695ms/step - loss: 0.7731 - acc: 0.8516 - f1_m: 0.7035 - precision_m: 0.5469 - recall_m: 1.2629 - val_loss: 0.0437 - val_acc: 0.9893 - val_f1_m: 0.6415 - val_precision_m: 0.4454 - val_recall_m: 1.1716\n",
            "Epoch 2/100\n",
            "36/36 [==============================] - 27s 730ms/step - loss: 0.0632 - acc: 0.9813 - f1_m: 0.7927 - precision_m: 0.5693 - recall_m: 1.3403 - val_loss: 0.0602 - val_acc: 0.9893 - val_f1_m: 0.7156 - val_precision_m: 0.4892 - val_recall_m: 1.3591\n",
            "Epoch 3/100\n",
            "36/36 [==============================] - 27s 732ms/step - loss: 0.0725 - acc: 0.9769 - f1_m: 0.8274 - precision_m: 0.5956 - recall_m: 1.3923 - val_loss: 0.0248 - val_acc: 0.9929 - val_f1_m: 0.7882 - val_precision_m: 0.5609 - val_recall_m: 1.3657\n",
            "Epoch 4/100\n",
            "36/36 [==============================] - 24s 663ms/step - loss: 0.0413 - acc: 0.9858 - f1_m: 0.8497 - precision_m: 0.5986 - recall_m: 1.5068 - val_loss: 0.0313 - val_acc: 0.9893 - val_f1_m: 0.8062 - val_precision_m: 0.5731 - val_recall_m: 1.3912\n",
            "Epoch 5/100\n",
            "36/36 [==============================] - 24s 658ms/step - loss: 0.0450 - acc: 0.9876 - f1_m: 0.8615 - precision_m: 0.6099 - recall_m: 1.5257 - val_loss: 0.0356 - val_acc: 0.9858 - val_f1_m: 0.8051 - val_precision_m: 0.5753 - val_recall_m: 1.3550\n",
            "Epoch 6/100\n",
            "36/36 [==============================] - 24s 649ms/step - loss: 0.0423 - acc: 0.9884 - f1_m: 0.8419 - precision_m: 0.6064 - recall_m: 1.3951 - val_loss: 0.0301 - val_acc: 0.9929 - val_f1_m: 0.7790 - val_precision_m: 0.5589 - val_recall_m: 1.3103\n",
            "Epoch 7/100\n",
            "36/36 [==============================] - 24s 652ms/step - loss: 0.0555 - acc: 0.9796 - f1_m: 0.8264 - precision_m: 0.6023 - recall_m: 1.3424 - val_loss: 0.0314 - val_acc: 0.9929 - val_f1_m: 0.8089 - val_precision_m: 0.5876 - val_recall_m: 1.3220\n",
            "Epoch 8/100\n",
            "36/36 [==============================] - 24s 659ms/step - loss: 0.0344 - acc: 0.9902 - f1_m: 0.8508 - precision_m: 0.6214 - recall_m: 1.3631 - val_loss: 0.0329 - val_acc: 0.9929 - val_f1_m: 0.7808 - val_precision_m: 0.5644 - val_recall_m: 1.3190\n",
            "Epoch 9/100\n",
            "36/36 [==============================] - 25s 665ms/step - loss: 0.0371 - acc: 0.9867 - f1_m: 0.8103 - precision_m: 0.6020 - recall_m: 1.2573 - val_loss: 0.0370 - val_acc: 0.9893 - val_f1_m: 0.8160 - val_precision_m: 0.6042 - val_recall_m: 1.2853\n",
            "Epoch 10/100\n",
            "36/36 [==============================] - 24s 660ms/step - loss: 0.0366 - acc: 0.9876 - f1_m: 0.7781 - precision_m: 0.5883 - recall_m: 1.1752 - val_loss: 0.0427 - val_acc: 0.9893 - val_f1_m: 0.7426 - val_precision_m: 0.5452 - val_recall_m: 1.1834\n",
            "Epoch 11/100\n",
            "36/36 [==============================] - 24s 657ms/step - loss: 0.0358 - acc: 0.9884 - f1_m: 0.7859 - precision_m: 0.5936 - recall_m: 1.1929 - val_loss: 0.0316 - val_acc: 0.9929 - val_f1_m: 0.7984 - val_precision_m: 0.5894 - val_recall_m: 1.2720\n"
          ]
        }
      ],
      "source": [
        "history_CNN1 = model_CNN1.fit(train_ds,\n",
        "                    validation_data=val_ds,\n",
        "                    callbacks=[callback],\n",
        "                    epochs=epochs)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 26,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vr39Ffrf0goB",
        "outputId": "f802a7f6-3c62-4d0b-e00e-6f162df9085e"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Test loss: 0.02286449447274208\n",
            "Test accuracy: 1.0\n",
            "INFO:tensorflow:Assets written to: ./model_CNN1/assets\n",
            "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",
            "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",
            "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"
          ]
        }
      ],
      "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_CNN1, filepath)\n",
        "\n",
        "# Load the model\n",
        "model = load_model(filepath, compile = False)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 27,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 499
        },
        "id": "sOt2_T1F-vAJ",
        "outputId": "074690e9-d99b-4342-bf5b-86dc1efa991c"
      },
      "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"
          }
        }
      ],
      "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()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 28,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9QnrNrEr0Bac",
        "outputId": "6de26d74-8b8e-4a39-c114-be301d90a20f"
      },
      "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])"
            ]
          },
          "metadata": {},
          "execution_count": 28
        }
      ],
      "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"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 29,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "j5X_d4knWrSe",
        "outputId": "91b9a181-164c-47b3-b4ea-832fdb6eb202"
      },
      "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": 29
        }
      ],
      "source": [
        "true_classes = test_generator.classes #true_classes = test_ds.class_names\n",
        "true_classes"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 30,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PxViYn_CWrSe",
        "outputId": "104aeb0c-79b4-446b-e73b-39abca96ee2a"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['Danube', 'Dnieper', 'UnknownOrigin']"
            ]
          },
          "metadata": {},
          "execution_count": 30
        }
      ],
      "source": [
        "class_labels = list(test_generator.class_indices.keys())\n",
        "class_labels"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 31,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 639
        },
        "id": "AJAcOrW9WrSe",
        "outputId": "339adea3-f214-40e9-a01d-bb6a94eb9ce7"
      },
      "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      1.00      1.00        78\n",
            "UnknownOrigin       1.00      1.00      1.00        37\n",
            "\n",
            "     accuracy                           1.00       224\n",
            "    macro avg       1.00      1.00      1.00       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"
          }
        }
      ],
      "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()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "C2QlIy1xWrSe"
      },
      "source": [
        "# Predictions"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "dataset_for_prediction = os.path.join(dataset_file_path, 'prediction')\n",
        "dataset_for_prediction"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "id": "aoG2q52ykO8J",
        "outputId": "bf87221a-b845-4044-a56d-d19c48487461"
      },
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'/content/BlackSea/prediction'"
            ]
          },
          "metadata": {},
          "execution_count": 47
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "dataset_for_prediction = datagen.flow_from_directory(\n",
        "    dataset_for_prediction, \n",
        "    target_size=(img_height, img_height), \n",
        "    batch_size=batch_size, \n",
        "    shuffle = False,\n",
        "    class_mode='categorical')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XS7OVuQXjY9_",
        "outputId": "f0df5473-88e1-4258-d502-485834e05f4c"
      },
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 30489 images belonging to 1 classes.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "dataset_for_prediction = tf.keras.preprocessing.image_dataset_from_directory(\n",
        "  'BlackSea/prediction',\n",
        "  seed=123,\n",
        "  image_size=(img_height, img_width),\n",
        "  shuffle = False,\n",
        "  batch_size=batch_size)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pmZ72T_blibO",
        "outputId": "e1047ada-5747-4a88-c2dd-30ff7e2abb66"
      },
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 30489 files belonging to 1 classes.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 50,
      "metadata": {
        "id": "Qihdefi0WrSe"
      },
      "outputs": [],
      "source": [
        "predictions = model_CNN1.predict(dataset_for_prediction)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "len(predictions)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iDziKH2J14xA",
        "outputId": "fcc8f449-6d88-4c9c-fd2e-69a2cee626bb"
      },
      "execution_count": 51,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "30489"
            ]
          },
          "metadata": {},
          "execution_count": 51
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "predictions[1:10]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4Dz-Xg5-3xib",
        "outputId": "f76d98ab-bb1a-4208-9df1-667987e62108"
      },
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[ 5.3017187,  1.0853667, -7.0040984],\n",
              "       [-4.7547708,  7.3037767, -1.4437723],\n",
              "       [ 5.006926 , -1.0553998, -2.2026606],\n",
              "       [ 9.155664 , -2.329165 , -5.8794293],\n",
              "       [-4.678572 ,  7.4452076, -1.6684153],\n",
              "       [-3.7419634,  7.634418 , -2.8910692],\n",
              "       [-4.737501 ,  7.3450694, -1.502234 ],\n",
              "       [-3.993742 ,  4.8222804,  0.8692965],\n",
              "       [-3.5704474,  7.642258 , -3.1189241]], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 52
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Get most likely class\n",
        "predicted_classes = numpy.argmax(predictions, axis=1)\n",
        "predicted_classes\n",
        "len(predicted_classes)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qmydMt2f0nYW",
        "outputId": "00dfbee4-4e4e-401c-9de5-ddf75b7974a0"
      },
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "30489"
            ]
          },
          "metadata": {},
          "execution_count": 43
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "files_for_prediction = dataset_for_prediction.file_paths"
      ],
      "metadata": {
        "id": "nadc49QA5eCC"
      },
      "execution_count": 54,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "df = pd.DataFrame({'x':files_for_prediction, 'y':predicted_classes})"
      ],
      "metadata": {
        "id": "rzAYiPuk5kod"
      },
      "execution_count": 56,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df.to_csv(\"results.csv\")"
      ],
      "metadata": {
        "id": "l4zGh0pj1_a9"
      },
      "execution_count": 57,
      "outputs": []
    }
  ],
  "metadata": {
    "colab": {
      "collapsed_sections": [
        "sTQAgmxDAuhZ",
        "F1EmdPK7A5Tj"
      ],
      "name": "Make_predictions_CNN.ipynb",
      "provenance": []
    },
    "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"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}