Sistem Cerdas Final Project.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Tugas Siscer.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "h3Z5isQ0b5Av"
      },
      "source": [
        "# Klasifikasi Tumor Otak pada Gambar MRI\r\n",
        "### Fahmi Jabbar (1801389)\r\n",
        "---\r\n",
        "**Informasi Dataset :**\r\n",
        "![Dataset](https://image.prntscr.com/image/3D0qQ1ZySkuarGolLUndOg.png)\r\n",
        "\r\n",
        "Sumber : https://www.kaggle.com/sartajbhuvaji/brain-tumor-classification-mri\r\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FCw5ojCGMFlP"
      },
      "source": [
        "## Import Library"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-QtvQSoKarlp"
      },
      "source": [
        "import matplotlib.pyplot as plt\r\n",
        "import matplotlib.image as mpimg\r\n",
        "import os, random, time\r\n",
        "import pandas as pd\r\n",
        "import numpy as np\r\n",
        "import cv2\r\n",
        "import tensorflow as tf\r\n",
        "import seaborn as sns\r\n",
        "from sklearn.metrics import classification_report, confusion_matrix"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RUvCoByedqCd",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "675f26b3-2d2c-4f22-faa5-a0c59732612c"
      },
      "source": [
        "!pip install split-folders tqdm\r\n",
        "import splitfolders"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting split-folders\n",
            "  Downloading https://files.pythonhosted.org/packages/b8/5f/3c2b2f7ea5e047c8cdc3bb00ae582c5438fcdbbedcc23b3cc1c2c7aae642/split_folders-0.4.3-py3-none-any.whl\n",
            "Requirement already satisfied: tqdm in /usr/local/lib/python3.6/dist-packages (4.41.1)\n",
            "Installing collected packages: split-folders\n",
            "Successfully installed split-folders-0.4.3\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "M8XzZPNgMI-8"
      },
      "source": [
        "## Cek Hardware"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2zcNcjNhMYPC"
      },
      "source": [
        "### Ketersediaan Hardware Penunjang"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "izRhuQu9tsDb",
        "outputId": "451fe6ba-32ec-47c7-c1f1-0b88a88a53ac"
      },
      "source": [
        "tf.config.list_physical_devices()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[PhysicalDevice(name='/physical_device:CPU:0', device_type='CPU'),\n",
              " PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4wGwBjsUMeXi"
      },
      "source": [
        "### Detail GPU"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "83WSvbQCup9C",
        "outputId": "3d325930-f193-445e-900d-c9fa432118b3"
      },
      "source": [
        "!nvidia-smi"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Sun Jan  3 12:57:58 2021       \n",
            "+-----------------------------------------------------------------------------+\n",
            "| NVIDIA-SMI 460.27.04    Driver Version: 418.67       CUDA Version: 10.1     |\n",
            "|-------------------------------+----------------------+----------------------+\n",
            "| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\n",
            "| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\n",
            "|                               |                      |               MIG M. |\n",
            "|===============================+======================+======================|\n",
            "|   0  Tesla T4            Off  | 00000000:00:04.0 Off |                    0 |\n",
            "| N/A   66C    P8    11W /  70W |     10MiB / 15079MiB |      0%      Default |\n",
            "|                               |                      |                 ERR! |\n",
            "+-------------------------------+----------------------+----------------------+\n",
            "                                                                               \n",
            "+-----------------------------------------------------------------------------+\n",
            "| Processes:                                                                  |\n",
            "|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |\n",
            "|        ID   ID                                                   Usage      |\n",
            "|=============================================================================|\n",
            "|  No running processes found                                                 |\n",
            "+-----------------------------------------------------------------------------+\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dPCXddHbMNIL"
      },
      "source": [
        "## Persiapan Dataset"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "f-gopimlMhjX"
      },
      "source": [
        "### Download dataset dengan kaggle API"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ivXcTS2bbnMc"
      },
      "source": [
        "import os\r\n",
        "os.environ['KAGGLE_USERNAME'] = \"fahmijabbar\"\r\n",
        "os.environ['KAGGLE_KEY'] = \"38196642493f1560965029fefcee361d\""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nOUXxTIWdE9J",
        "outputId": "c530ff2b-20c6-43e3-cf15-80c96431e398"
      },
      "source": [
        "!kaggle datasets download -d sartajbhuvaji/brain-tumor-classification-mri"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading customer-segmentation-tutorial-in-python.zip to /content\n",
            "\r  0% 0.00/1.55k [00:00<?, ?B/s]\n",
            "\r100% 1.55k/1.55k [00:00<00:00, 2.58MB/s]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oi4ICyzRMnqA"
      },
      "source": [
        "### Unzip Dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Z5Z1gQDCdfcY"
      },
      "source": [
        "!unzip -q /content/brain-tumor-classification-mri.zip -d /content/dataset/"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GGtaoIYqMsLZ"
      },
      "source": [
        "### Rename dan copy dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3Zy9WYYPnhTr"
      },
      "source": [
        "path = '/content/dataset'\r\n",
        "folder = os.listdir(path)\r\n",
        "for subfolder in folder:\r\n",
        "  subpath = path+\"/\"+subfolder \r\n",
        "  files = os.listdir(subpath)\r\n",
        "  for subs in files:\r\n",
        "    subs = path+\"/\"+subfolder+\"/\"+subs\r\n",
        "    files = os.listdir(subs)\r\n",
        "    for index, file in enumerate(files):\r\n",
        "      os.rename(os.path.join(subs, file), os.path.join(subs, ''.join([subfolder,'-',str(index), '.jpg'])))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E31tYi4KmEsh"
      },
      "source": [
        "!mv /content/dataset/Training/glioma_tumor/ /content/dataset/glioma_tumor\r\n",
        "!mv /content/dataset/Training/meningioma_tumor/ /content/dataset/meningioma_tumor\r\n",
        "!mv /content/dataset/Training/no_tumor/ /content/dataset/no_tumor\r\n",
        "!mv /content/dataset/Training/pituitary_tumor/ /content/dataset/pituitary_tumor\r\n",
        "!mv /content/dataset/Testing/glioma_tumor/* /content/dataset/glioma_tumor/\r\n",
        "!mv /content/dataset/Testing/meningioma_tumor/* /content/dataset/meningioma_tumor/\r\n",
        "!mv /content/dataset/Testing/no_tumor/* /content/dataset/no_tumor/\r\n",
        "!mv /content/dataset/Testing/pituitary_tumor/* /content/dataset/pituitary_tumor/\r\n",
        "!rm -rf /content/dataset/Training/\r\n",
        "!rm -rf /content/dataset/Testing/"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3QIkIFJuMxb1"
      },
      "source": [
        "## Eksplorasi dan Analisis Dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PhmlueNKf2yB"
      },
      "source": [
        "base_dir = '/content/dataset/'\r\n",
        "labels = os.listdir(base_dir)\r\n",
        "countedlabel = []"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7VTUeLMLf9GA",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b5a70cce-f61f-445a-bb43-e7d3eb544f3e"
      },
      "source": [
        "# Count total images each classes\r\n",
        "file_count = []\r\n",
        "for label in labels:\r\n",
        "  dir = base_dir + label\r\n",
        "  path, dirs, files = next(os.walk(dir))\r\n",
        "  file_count.append(len(files))\r\n",
        "print(\"Consist of \" + str(len(file_count)) + \" Classes\")\r\n",
        "\r\n",
        "# Count total images\r\n",
        "total = 0\r\n",
        "for num in file_count:\r\n",
        "  total += num\r\n",
        "print(\"And \" + str(total) + \" Total Images\")\r\n",
        "\r\n",
        "# Append total image in classes\r\n",
        "x = 0\r\n",
        "while x < len(labels):\r\n",
        "  countedlabel.append(labels[x] + \" (\" + str(file_count[x]) + \")\")\r\n",
        "  x += 1"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Consist of 4 Classes\n",
            "And 3264 Total Images\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1-9d4vXKgBwP",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 785
        },
        "outputId": "8365c587-16ac-43ad-e4ee-f699f8b1f784"
      },
      "source": [
        "plt.figure(figsize=(8,13))\r\n",
        "plt.barh(np.arange(len(labels)), file_count)\r\n",
        "plt.yticks(np.arange(len(labels)), countedlabel)\r\n",
        "plt.title('Dataset Information')\r\n",
        "plt.xlabel('Total')\r\n",
        "plt.ylabel('Labels')\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x936 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IftNiwQ-M5Jq"
      },
      "source": [
        "## Pembagian dataset\r\n",
        "![Dataset Split](https://developers.google.com/machine-learning/crash-course/images/PartitionThreeSets.svg)\r\n",
        "\r\n",
        "Referensi : https://developers.google.com/machine-learning/crash-course/validation/another-partition"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4Cx4wmCVeiuh",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6c78b04e-9a7b-467e-a0ab-cb9bb36e2c1f"
      },
      "source": [
        "splitfolders.ratio(base_dir, output=\"/content/splitted/\", seed=1337, ratio=(.7, .2, .1))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Copying files: 3264 files [00:00, 7728.17 files/s]\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "akJ0Y-XrNijo"
      },
      "source": [
        "## Augmentasi Dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PN474_brhcB8"
      },
      "source": [
        "train_datagen = tf.keras.preprocessing.image.ImageDataGenerator(\r\n",
        "    rescale = 1./255, #Normalization\r\n",
        ")\r\n",
        "\r\n",
        "validation_datagen = tf.keras.preprocessing.image.ImageDataGenerator(\r\n",
        "    rescale=1./255, #Normalization\r\n",
        ")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BLWAr-WBNnt5"
      },
      "source": [
        "## Feed gambar dataset kedalam generator"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1h-y7XYzhj9F",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "3f3cd8f0-7e4d-44a8-8550-6ef71fc36983"
      },
      "source": [
        "img_shape = (150, 150, 3)\r\n",
        "train_generator = train_datagen.flow_from_directory(\r\n",
        "    directory=\"/content/splitted/train\",\r\n",
        "    target_size= img_shape[:2],\r\n",
        "    class_mode='categorical',\r\n",
        "    color_mode='rgb',\r\n",
        "    batch_size=64)\r\n",
        "\r\n",
        "validation_generator = validation_datagen.flow_from_directory(\r\n",
        "    directory=\"/content/splitted/val\",\r\n",
        "    target_size= img_shape[:2],\r\n",
        "    class_mode='categorical',\r\n",
        "    color_mode='rgb',\r\n",
        "    batch_size=8)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 2283 images belonging to 4 classes.\n",
            "Found 652 images belonging to 4 classes.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RHhNNgvlNvDJ"
      },
      "source": [
        "### Helper Function :"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1o0J7xZeQukf"
      },
      "source": [
        "transparent = (255,255,255)\r\n",
        "yellow = (245, 180, 0)\r\n",
        "green = (0, 170, 0)\r\n",
        "red = (220, 0, 0)\r\n",
        "\r\n",
        "def print_in_color(txt_msg,fore_tupple,back_tupple):\r\n",
        "    rf,gf,bf=fore_tupple\r\n",
        "    rb,gb,bb=back_tupple\r\n",
        "    if(rb == 255 and gb == 255 and bb == 255):\r\n",
        "      msg='{0}' + txt_msg\r\n",
        "      mat='\\33[38;2;' + str(rf) +';' + str(gf) + ';' + str(bf) + ';48;2'+ 'm' \r\n",
        "      print(msg .format(mat))\r\n",
        "      print('\\33[0m')\r\n",
        "    else:\r\n",
        "      msg='{0}' + txt_msg\r\n",
        "      mat='\\33[38;2;' + str(rf) +';' + str(gf) + ';' + str(bf) + ';48;2;' + str(rb) + ';' +str(gb) + ';' + str(bb) +'m' \r\n",
        "      print(msg .format(mat))\r\n",
        "      print('\\33[0m')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "I-WAx5iCNzdM"
      },
      "source": [
        "## Arsitektur Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_8FWUm8rtgZG"
      },
      "source": [
        "tf.keras.backend.clear_session()\r\n",
        "model = tf.keras.models.Sequential([\r\n",
        "          tf.keras.layers.Conv2D(64, (5,5), activation='relu', padding='same', input_shape=(150, 150, 3)),\r\n",
        "          tf.keras.layers.MaxPooling2D(pool_size=(2,2)),\r\n",
        "          tf.keras.layers.Dropout(0.25),\r\n",
        "          tf.keras.layers.Conv2D(128, (3,3), padding='same', activation='relu'),\r\n",
        "          tf.keras.layers.MaxPooling2D(pool_size=(2,2), strides=(2,2)),\r\n",
        "          tf.keras.layers.Dropout(0.25),\r\n",
        "          tf.keras.layers.Conv2D(128, (3,3), padding='same', activation='relu'),\r\n",
        "          tf.keras.layers.MaxPooling2D(pool_size=(2,2), strides=(2,2)),\r\n",
        "          tf.keras.layers.Dropout(0.3),\r\n",
        "          tf.keras.layers.Conv2D(128, (2,2), padding='same', activation='relu'),\r\n",
        "          tf.keras.layers.MaxPooling2D(pool_size=(2,2), strides=(2,2)),\r\n",
        "          tf.keras.layers.Dropout(0.3),\r\n",
        "          tf.keras.layers.Conv2D(256, (2,2), padding='same', activation='relu'),\r\n",
        "          tf.keras.layers.MaxPooling2D(pool_size=(2,2), strides=(2,2)),\r\n",
        "          tf.keras.layers.Dropout(0.3),\r\n",
        "          tf.keras.layers.Flatten(),\r\n",
        "          tf.keras.layers.Dense(1024, activation='relu'),\r\n",
        "          tf.keras.layers.Dropout(0.5),\r\n",
        "          tf.keras.layers.Dense(4, activation='softmax')\r\n",
        "])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KCs8OmWptiXf",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "54c3c1a9-e8e9-4cac-e253-ab774751c166"
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d (Conv2D)              (None, 150, 150, 64)      4864      \n",
            "_________________________________________________________________\n",
            "max_pooling2d (MaxPooling2D) (None, 75, 75, 64)        0         \n",
            "_________________________________________________________________\n",
            "dropout (Dropout)            (None, 75, 75, 64)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_1 (Conv2D)            (None, 75, 75, 128)       73856     \n",
            "_________________________________________________________________\n",
            "max_pooling2d_1 (MaxPooling2 (None, 37, 37, 128)       0         \n",
            "_________________________________________________________________\n",
            "dropout_1 (Dropout)          (None, 37, 37, 128)       0         \n",
            "_________________________________________________________________\n",
            "conv2d_2 (Conv2D)            (None, 37, 37, 128)       147584    \n",
            "_________________________________________________________________\n",
            "max_pooling2d_2 (MaxPooling2 (None, 18, 18, 128)       0         \n",
            "_________________________________________________________________\n",
            "dropout_2 (Dropout)          (None, 18, 18, 128)       0         \n",
            "_________________________________________________________________\n",
            "conv2d_3 (Conv2D)            (None, 18, 18, 128)       65664     \n",
            "_________________________________________________________________\n",
            "max_pooling2d_3 (MaxPooling2 (None, 9, 9, 128)         0         \n",
            "_________________________________________________________________\n",
            "dropout_3 (Dropout)          (None, 9, 9, 128)         0         \n",
            "_________________________________________________________________\n",
            "conv2d_4 (Conv2D)            (None, 9, 9, 256)         131328    \n",
            "_________________________________________________________________\n",
            "max_pooling2d_4 (MaxPooling2 (None, 4, 4, 256)         0         \n",
            "_________________________________________________________________\n",
            "dropout_4 (Dropout)          (None, 4, 4, 256)         0         \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 4096)              0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 1024)              4195328   \n",
            "_________________________________________________________________\n",
            "dropout_5 (Dropout)          (None, 1024)              0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 4)                 4100      \n",
            "=================================================================\n",
            "Total params: 4,622,724\n",
            "Trainable params: 4,622,724\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "QFBCYclBdzOW",
        "outputId": "0f22f10a-5b95-4b42-a04f-956034af566b"
      },
      "source": [
        "tf.keras.utils.plot_model(model)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 20
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NBXmNvqMN3N8"
      },
      "source": [
        "### Kompilasi model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CANYKwPztnSI"
      },
      "source": [
        "model.compile(loss='categorical_crossentropy',\r\n",
        "              optimizer=tf.keras.optimizers.Adam(lr=0.001, beta_1=0.9, beta_2=0.999),\r\n",
        "              metrics=['accuracy'])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "taFbPfuPN7ya"
      },
      "source": [
        "### Penerapan teknik callback"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ru0FcYfsdhY9"
      },
      "source": [
        "class completeCallback(tf.keras.callbacks.Callback):\r\n",
        "  best_weights=model.get_weights()\r\n",
        "  def __init__(self, patience=2, threshold=.85, factor=.75):\r\n",
        "    super(LRA, self).__init__()\r\n",
        "    self.patience=patience\r\n",
        "    self.threshold=threshold\r\n",
        "    self.factor=factor\r\n",
        "    self.lr=float(tf.keras.backend.get_value(model.optimizer.lr))\r\n",
        "    self.highest_tracc=0.0 \r\n",
        "    self.lowest_vloss=np.inf\r\n",
        "    self.count=0\r\n",
        "    msg='Training Model with Self Adjusting Learning Rate\\n==============================================='\r\n",
        "    print_in_color (msg, (0, 0, 230), (255,255,255))\r\n",
        "      \r\n",
        "  def on_epoch_end(self, epoch, logs=None):\r\n",
        "    if(logs.get('accuracy') > 0.92 and logs.get('val_accuracy') >= 0.92):\r\n",
        "      msg = \"\\n\\n[i] Reached Accuracy 92% & Val Accuracy 92%, Training is stoped\\n\"\r\n",
        "      print_in_color(msg, (0, 0, 230), transparent)\r\n",
        "      self.model.stop_training = True\r\n",
        "    else:\r\n",
        "      lr=float(tf.keras.backend.get_value(self.model.optimizer.lr))\r\n",
        "      v_loss=logs.get('val_loss')\r\n",
        "      acc=logs.get('accuracy')\r\n",
        "      if acc < self.threshold:\r\n",
        "        if acc>self.highest_tracc: \r\n",
        "          msg= f'\\n[i] training accuracy improved from  {self.highest_tracc:.4f} to {acc:.4f} learning rate held at {lr:.6f}'\r\n",
        "          print_in_color(msg, green, transparent)\r\n",
        "          self.highest_tracc=acc\r\n",
        "          LRA.best_weights=model.get_weights()\r\n",
        "          count=0\r\n",
        "          if v_loss<self.lowest_vloss:\r\n",
        "            self.lowest_vloss=v_loss                    \r\n",
        "        else:  \r\n",
        "          if self.count>=self.patience-1:\r\n",
        "            self.lr= lr * self.factor\r\n",
        "            tf.keras.backend.set_value(model.optimizer.lr, self.lr)\r\n",
        "            self.count=0\r\n",
        "            if v_loss<self.lowest_vloss:\r\n",
        "              self.lowest_vloss=v_loss\r\n",
        "            msg=f'\\n[i] for epoch {epoch+1} training accuracy did not improve for {self.patience} consecutive epochs, learning rate adjusted to {lr:.6f}'\r\n",
        "            print_in_color(msg, red, transparent)\r\n",
        "          else:\r\n",
        "            self.count=self.count+1\r\n",
        "            msg=f'\\n[i] try number {self.count}, for  epoch {epoch+1} training accuracy did not improve'\r\n",
        "            print_in_color(msg, yellow, transparent)\r\n",
        "      else:\r\n",
        "        if v_loss< self.lowest_vloss:\r\n",
        "          msg=f'\\n[i] validation loss improved from {self.lowest_vloss:.4f} to {v_loss:.4}, saving best weights'\r\n",
        "          print_in_color(msg, green, transparent)\r\n",
        "          self.lowest_vloss=v_loss                \r\n",
        "          LRA.best_weights=model.get_weights()\r\n",
        "          self.count=0      \r\n",
        "        else:\r\n",
        "          if self.count>=self.patience-1:\r\n",
        "            self.lr=self.lr * self.factor\r\n",
        "            msg=f'\\n[i] for epoch {epoch+1} validation loss failed to improve for {self.patience} consecutive epochs, learning rate adjusted to {self.lr:9.6f}'\r\n",
        "            self.count=0\r\n",
        "            print_in_color(msg, red, transparent)\r\n",
        "            tf.keras.backend.set_value(model.optimizer.lr, self.lr)\r\n",
        "          else: \r\n",
        "            self.count =self.count +1 \r\n",
        "            msg=f'\\n[i] try number {self.count}, for epoch {epoch+1} validation loss did not improve'\r\n",
        "            print_in_color(msg, yellow, transparent)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JFwPeObtOAiG"
      },
      "source": [
        "## Training Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TP6c-RXDuLrj",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "cd41ef8c-6233-495f-d13d-e6cbb78db781"
      },
      "source": [
        "with tf.device('/device:GPU:0'):\r\n",
        "  history = model.fit(train_generator, \r\n",
        "                      epochs = 50, \r\n",
        "                      validation_data = validation_generator, \r\n",
        "                      verbose = 1,\r\n",
        "                      callbacks=[completeCallback()])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\u001b[38;2;0;0;230;48;2mTraining Model with Self Adjusting Learning Rate\n",
            "===============================================\n",
            "\u001b[0m\n",
            "Epoch 1/50\n",
            "36/36 [==============================] - 13s 339ms/step - loss: 1.4129 - accuracy: 0.3047 - val_loss: 1.2857 - val_accuracy: 0.4617\n",
            "\u001b[38;2;0;170;0;48;2m\n",
            "[i] training accuracy improved from  0.0000 to 0.3530 learning rate held at 0.001000\n",
            "\u001b[0m\n",
            "Epoch 2/50\n",
            "36/36 [==============================] - 12s 326ms/step - loss: 1.2188 - accuracy: 0.4582 - val_loss: 1.2433 - val_accuracy: 0.4831\n",
            "\u001b[38;2;0;170;0;48;2m\n",
            "[i] training accuracy improved from  0.3530 to 0.4788 learning rate held at 0.001000\n",
            "\u001b[0m\n",
            "Epoch 3/50\n",
            "36/36 [==============================] - 12s 329ms/step - loss: 1.0522 - accuracy: 0.5488 - val_loss: 1.1619 - val_accuracy: 0.4601\n",
            "\u001b[38;2;0;170;0;48;2m\n",
            "[i] training accuracy improved from  0.4788 to 0.5629 learning rate held at 0.001000\n",
            "\u001b[0m\n",
            "Epoch 4/50\n",
            "36/36 [==============================] - 12s 330ms/step - loss: 0.9529 - accuracy: 0.5931 - val_loss: 0.8952 - val_accuracy: 0.5844\n",
            "\u001b[38;2;0;170;0;48;2m\n",
            "[i] training accuracy improved from  0.5629 to 0.6093 learning rate held at 0.001000\n",
            "\u001b[0m\n",
            "Epoch 5/50\n",
            "36/36 [==============================] - 12s 330ms/step - loss: 0.8721 - accuracy: 0.6271 - val_loss: 0.9266 - val_accuracy: 0.5644\n",
            "\u001b[38;2;0;170;0;48;2m\n",
            "[i] training accuracy improved from  0.6093 to 0.6496 learning rate held at 0.001000\n",
            "\u001b[0m\n",
            "Epoch 6/50\n",
            "36/36 [==============================] - 12s 330ms/step - loss: 0.7735 - accuracy: 0.6697 - val_loss: 0.7676 - val_accuracy: 0.6227\n",
            "\u001b[38;2;0;170;0;48;2m\n",
            "[i] training accuracy improved from  0.6496 to 0.6881 learning rate held at 0.001000\n",
            "\u001b[0m\n",
            "Epoch 7/50\n",
            "36/36 [==============================] - 12s 329ms/step - loss: 0.6402 - accuracy: 0.7613 - val_loss: 0.6321 - val_accuracy: 0.7239\n",
            "\u001b[38;2;0;170;0;48;2m\n",
            "[i] training accuracy improved from  0.6881 to 0.7560 learning rate held at 0.001000\n",
            "\u001b[0m\n",
            "Epoch 8/50\n",
            "36/36 [==============================] - 12s 329ms/step - loss: 0.5525 - accuracy: 0.7776 - val_loss: 0.6857 - val_accuracy: 0.6887\n",
            "\u001b[38;2;0;170;0;48;2m\n",
            "[i] training accuracy improved from  0.7560 to 0.7841 learning rate held at 0.001000\n",
            "\u001b[0m\n",
            "Epoch 9/50\n",
            "36/36 [==============================] - 12s 332ms/step - loss: 0.5430 - accuracy: 0.7825 - val_loss: 0.5929 - val_accuracy: 0.7500\n",
            "\u001b[38;2;0;170;0;48;2m\n",
            "[i] training accuracy improved from  0.7841 to 0.7941 learning rate held at 0.001000\n",
            "\u001b[0m\n",
            "Epoch 10/50\n",
            "36/36 [==============================] - 12s 335ms/step - loss: 0.4715 - accuracy: 0.8140 - val_loss: 0.6942 - val_accuracy: 0.7163\n",
            "\u001b[38;2;0;170;0;48;2m\n",
            "[i] training accuracy improved from  0.7941 to 0.8152 learning rate held at 0.001000\n",
            "\u001b[0m\n",
            "Epoch 11/50\n",
            "36/36 [==============================] - 12s 334ms/step - loss: 0.4701 - accuracy: 0.8081 - val_loss: 0.4450 - val_accuracy: 0.8113\n",
            "\u001b[38;2;245;180;0;48;2m\n",
            "[i] try number 1, for  epoch 11 training accuracy did not improve\n",
            "\u001b[0m\n",
            "Epoch 12/50\n",
            "36/36 [==============================] - 12s 328ms/step - loss: 0.4331 - accuracy: 0.8298 - val_loss: 0.5180 - val_accuracy: 0.7883\n",
            "\u001b[38;2;0;170;0;48;2m\n",
            "[i] training accuracy improved from  0.8152 to 0.8292 learning rate held at 0.001000\n",
            "\u001b[0m\n",
            "Epoch 13/50\n",
            "36/36 [==============================] - 12s 331ms/step - loss: 0.3651 - accuracy: 0.8570 - val_loss: 0.4458 - val_accuracy: 0.8328\n",
            "\u001b[38;2;0;170;0;48;2m\n",
            "[i] validation loss improved from 0.5180 to 0.4458, saving best weights\n",
            "\u001b[0m\n",
            "Epoch 14/50\n",
            "36/36 [==============================] - 12s 327ms/step - loss: 0.3121 - accuracy: 0.8799 - val_loss: 0.4028 - val_accuracy: 0.8558\n",
            "\u001b[38;2;0;170;0;48;2m\n",
            "[i] validation loss improved from 0.4458 to 0.4028, saving best weights\n",
            "\u001b[0m\n",
            "Epoch 15/50\n",
            "36/36 [==============================] - 12s 326ms/step - loss: 0.2710 - accuracy: 0.8956 - val_loss: 0.5168 - val_accuracy: 0.7991\n",
            "\u001b[38;2;245;180;0;48;2m\n",
            "[i] try number 1, for epoch 15 validation loss did not improve\n",
            "\u001b[0m\n",
            "Epoch 16/50\n",
            "36/36 [==============================] - 12s 331ms/step - loss: 0.2813 - accuracy: 0.8981 - val_loss: 0.4140 - val_accuracy: 0.8635\n",
            "\u001b[38;2;220;0;0;48;2m\n",
            "[i] for epoch 16 validation loss failed to improve for 2 consecutive epochs, learning rate adjusted to  0.000750\n",
            "\u001b[0m\n",
            "Epoch 17/50\n",
            "36/36 [==============================] - 12s 326ms/step - loss: 0.2558 - accuracy: 0.9076 - val_loss: 0.3862 - val_accuracy: 0.8604\n",
            "\u001b[38;2;0;170;0;48;2m\n",
            "[i] validation loss improved from 0.4028 to 0.3862, saving best weights\n",
            "\u001b[0m\n",
            "Epoch 18/50\n",
            "36/36 [==============================] - 12s 331ms/step - loss: 0.1937 - accuracy: 0.9256 - val_loss: 0.4239 - val_accuracy: 0.8528\n",
            "\u001b[38;2;245;180;0;48;2m\n",
            "[i] try number 1, for epoch 18 validation loss did not improve\n",
            "\u001b[0m\n",
            "Epoch 19/50\n",
            "36/36 [==============================] - 12s 323ms/step - loss: 0.1922 - accuracy: 0.9211 - val_loss: 0.3519 - val_accuracy: 0.8850\n",
            "\u001b[38;2;0;170;0;48;2m\n",
            "[i] validation loss improved from 0.3862 to 0.3519, saving best weights\n",
            "\u001b[0m\n",
            "Epoch 20/50\n",
            "36/36 [==============================] - 12s 324ms/step - loss: 0.1759 - accuracy: 0.9385 - val_loss: 0.3245 - val_accuracy: 0.8819\n",
            "\u001b[38;2;0;170;0;48;2m\n",
            "[i] validation loss improved from 0.3519 to 0.3245, saving best weights\n",
            "\u001b[0m\n",
            "Epoch 21/50\n",
            "36/36 [==============================] - 12s 325ms/step - loss: 0.1461 - accuracy: 0.9498 - val_loss: 0.3143 - val_accuracy: 0.8988\n",
            "\u001b[38;2;0;170;0;48;2m\n",
            "[i] validation loss improved from 0.3245 to 0.3143, saving best weights\n",
            "\u001b[0m\n",
            "Epoch 22/50\n",
            "36/36 [==============================] - 12s 325ms/step - loss: 0.1607 - accuracy: 0.9367 - val_loss: 0.3469 - val_accuracy: 0.8880\n",
            "\u001b[38;2;245;180;0;48;2m\n",
            "[i] try number 1, for epoch 22 validation loss did not improve\n",
            "\u001b[0m\n",
            "Epoch 23/50\n",
            "36/36 [==============================] - 12s 324ms/step - loss: 0.1340 - accuracy: 0.9500 - val_loss: 0.3987 - val_accuracy: 0.8742\n",
            "\u001b[38;2;220;0;0;48;2m\n",
            "[i] for epoch 23 validation loss failed to improve for 2 consecutive epochs, learning rate adjusted to  0.000563\n",
            "\u001b[0m\n",
            "Epoch 24/50\n",
            "36/36 [==============================] - 12s 329ms/step - loss: 0.1439 - accuracy: 0.9420 - val_loss: 0.3439 - val_accuracy: 0.9049\n",
            "\u001b[38;2;245;180;0;48;2m\n",
            "[i] try number 1, for epoch 24 validation loss did not improve\n",
            "\u001b[0m\n",
            "Epoch 25/50\n",
            "36/36 [==============================] - 12s 331ms/step - loss: 0.0930 - accuracy: 0.9666 - val_loss: 0.4038 - val_accuracy: 0.8926\n",
            "\u001b[38;2;220;0;0;48;2m\n",
            "[i] for epoch 25 validation loss failed to improve for 2 consecutive epochs, learning rate adjusted to  0.000422\n",
            "\u001b[0m\n",
            "Epoch 26/50\n",
            "36/36 [==============================] - 12s 322ms/step - loss: 0.0938 - accuracy: 0.9659 - val_loss: 0.3095 - val_accuracy: 0.9110\n",
            "\u001b[38;2;0;170;0;48;2m\n",
            "[i] validation loss improved from 0.3143 to 0.3095, saving best weights\n",
            "\u001b[0m\n",
            "Epoch 27/50\n",
            "36/36 [==============================] - 12s 323ms/step - loss: 0.0911 - accuracy: 0.9692 - val_loss: 0.3453 - val_accuracy: 0.9049\n",
            "\u001b[38;2;245;180;0;48;2m\n",
            "[i] try number 1, for epoch 27 validation loss did not improve\n",
            "\u001b[0m\n",
            "Epoch 28/50\n",
            "36/36 [==============================] - 12s 320ms/step - loss: 0.0643 - accuracy: 0.9817 - val_loss: 0.3188 - val_accuracy: 0.9248\n",
            "\u001b[38;2;0;0;230;48;2m\n",
            "\n",
            "[i] Reached Accuracy 92% & Val Accuracy 92%, Training is stoped\n",
            "\n",
            "\u001b[0m\n",
            "\u001b[38;2;220;0;0;48;2m\n",
            "[i] for epoch 28 validation loss failed to improve for 2 consecutive epochs, learning rate adjusted to  0.000316\n",
            "\u001b[0m\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ahnbPM3kOD2D"
      },
      "source": [
        "## Hasil Training Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DW-4vEDZmsZR"
      },
      "source": [
        "# Bila model yang di training belum cukup bagus\r\n",
        "# gunakan model yang sebelumnya pernah dibuat\r\n",
        "model = tf.keras.models.load_model('/content/model0201.h5')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aeu2oJS2OKFW"
      },
      "source": [
        "### Evaluasi Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "p2nxHeVQOWmB"
      },
      "source": [
        "test_datagen = tf.keras.preprocessing.image.ImageDataGenerator(rescale=1. / 255)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "F-4phc3fPAbM",
        "outputId": "a97bfbd0-9bd1-4894-81e3-a31f3e1a4012"
      },
      "source": [
        "test_generator = test_datagen.flow_from_directory(\r\n",
        "  directory='/content/splitted/test',\r\n",
        "  target_size=(150, 150),\r\n",
        "  shuffle=False,\r\n",
        "  batch_size=2,\r\n",
        "  class_mode='categorical')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 329 images belonging to 4 classes.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SnxkRwnYPHTL",
        "outputId": "d14add50-c5e1-41f7-b5b6-f0342cb03e45"
      },
      "source": [
        "scores = model.evaluate(test_generator)\r\n",
        "print(\"\\n%s: %.2f%%\" % (model.metrics_names[1], scores[1]*100))\r\n",
        "print(\"%s: %.2f%%\" % (model.metrics_names[0], scores[0]*100))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "165/165 [==============================] - 9s 11ms/step - loss: 0.2882 - accuracy: 0.9179\n",
            "\n",
            "accuracy: 91.79%\n",
            "loss: 28.82%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2_6amt9-OGs8"
      },
      "source": [
        "### Grafik akurasi dan Loss"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M67WEp4ossCp",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 545
        },
        "outputId": "1cf5b595-5efb-4caa-cf70-be293de4c0d0"
      },
      "source": [
        "acc = history.history['accuracy']\r\n",
        "val_acc = history.history['val_accuracy']\r\n",
        "loss = history.history['loss']\r\n",
        "val_loss = history.history['val_loss']\r\n",
        "\r\n",
        "epochs = range(len(acc))\r\n",
        "\r\n",
        "plt.plot(epochs, acc, 'r', label='Training accuracy')\r\n",
        "plt.plot(epochs, val_acc, 'b', label='Validation accuracy')\r\n",
        "plt.title('Training and validation accuracy')\r\n",
        "plt.legend()\r\n",
        "plt.figure()\r\n",
        "\r\n",
        "plt.plot(epochs, loss, 'r', label='Training Loss')\r\n",
        "plt.plot(epochs, val_loss, 'b', label='Validation Loss')\r\n",
        "plt.title('Training and validation loss')\r\n",
        "plt.legend()\r\n",
        "\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5IaMcCqgRAe0"
      },
      "source": [
        "### Confussion Matrix"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 594
        },
        "id": "IGrRf7kqRC-P",
        "outputId": "bd24acba-540e-4f40-a7b8-e3bcd34c5ca7"
      },
      "source": [
        "y_pred = model.predict(test_generator)\r\n",
        "y_pred = np.argmax(y_pred, axis=1)\r\n",
        "df_cm = pd.DataFrame(confusion_matrix(test_generator.classes, y_pred), columns=labels, index= labels)\r\n",
        "df_cm.index.name = 'Actual'\r\n",
        "df_cm.columns.name = 'Predicted'\r\n",
        "plt.figure(figsize = (7,7))\r\n",
        "sns.set(font_scale=1.4)#for label size\r\n",
        "sns.heatmap(df_cm, cmap=\"YlGnBu\", annot=True,annot_kws={\"size\": 17})\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 504x504 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ebTkL5gLOOTn"
      },
      "source": [
        "### Prediksi Model Secara acak dan Tampilan heatmap menggunakan teknik Guided GradCam"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 519
        },
        "id": "a1NKRnkIZBqb",
        "outputId": "d4d5996d-7761-449e-9234-b5ba08789c33"
      },
      "source": [
        "#Memilih gambar secara acak\r\n",
        "folders = random.choice(os.listdir(\"/content/splitted/test/\"))\r\n",
        "files = random.choice(os.listdir(\"/content/splitted/test/\"+folders+\"/\"))\r\n",
        "\r\n",
        "image = np.array(tf.keras.preprocessing.image.load_img(\"/content/splitted/test/\"+folders+\"/\"+files, target_size=(150, 150, 3)))\r\n",
        "x = tf.keras.preprocessing.image.img_to_array(image)\r\n",
        "x /= 255\r\n",
        "x = np.expand_dims(x, axis=0)\r\n",
        "x = np.vstack([x])\r\n",
        "classes = model.predict(x, batch_size=1)\r\n",
        "plt.imshow(image)\r\n",
        "plt.title(folders)\r\n",
        "plt.axis(\"off\")\r\n",
        "plt.show()\r\n",
        "last_conv_layer = model.get_layer(\"conv2d_4\")\r\n",
        "last_conv_layer_model = tf.keras.Model(model.inputs, last_conv_layer.output)\r\n",
        "classifier_input = tf.keras.Input(shape=last_conv_layer.output.shape[1:])\r\n",
        "x = classifier_input\r\n",
        "for layer_name in [\"max_pooling2d_4\",\"dropout_4\",\"flatten\"]:\r\n",
        "  x = model.get_layer(layer_name)(x)\r\n",
        "classifier_model = tf.keras.Model(classifier_input, x)\r\n",
        "with tf.GradientTape() as tape:\r\n",
        "  last_conv_layer_output = last_conv_layer_model(image[np.newaxis, ...])\r\n",
        "  tape.watch(last_conv_layer_output)\r\n",
        "  preds = classifier_model(last_conv_layer_output)\r\n",
        "  top_pred_index = tf.argmax(preds[0])\r\n",
        "  top_class_channel = preds[:, top_pred_index]\r\n",
        "grads = tape.gradient(top_class_channel, last_conv_layer_output)[0]\r\n",
        "last_conv_layer_output = last_conv_layer_output[0]\r\n",
        "guided_grads = (\r\n",
        "    tf.cast(last_conv_layer_output > 0, \"float32\")\r\n",
        "    * tf.cast(grads > 0, \"float32\")\r\n",
        "    * grads\r\n",
        ")\r\n",
        "pooled_guided_grads = tf.reduce_mean(guided_grads, axis=(0, 1))\r\n",
        "guided_gradcam = np.ones(last_conv_layer_output.shape[:2], dtype=np.float32)\r\n",
        "for i, w in enumerate(pooled_guided_grads):\r\n",
        "    guided_gradcam += w * last_conv_layer_output[:, :, i]\r\n",
        "guided_gradcam = cv2.resize(guided_gradcam.numpy(), (150, 150))\r\n",
        "guided_gradcam = np.clip(guided_gradcam, 0, np.max(guided_gradcam))\r\n",
        "guided_gradcam = (guided_gradcam - guided_gradcam.min()) / (\r\n",
        "    guided_gradcam.max() - guided_gradcam.min()\r\n",
        ")\r\n",
        "plt.imshow(image)\r\n",
        "labels = test_generator.class_indices\r\n",
        "labels = {value:key for key, value in labels.items()}\r\n",
        "plt.title(\"Predicted: \"+str(labels[np.argmax(classes)]))\r\n",
        "plt.imshow(guided_gradcam, alpha=0.5,cmap='afmhot')\r\n",
        "plt.axis(\"off\")\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EfTOBLaWXAMt"
      },
      "source": [
        "## Simpan Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "L8MfwdmUKXKv"
      },
      "source": [
        "model.save('/content/model0201.h5')"
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}