final_tl_mobilenetv2_dataset6000_sidangready.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "final_tl_mobilenetv2_dataset6000_sidangready.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyOlKrNw4wYyMxkotASfta2v",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/WiryawanCEO/IndonesianBankNotes-ImageClassificationModel/blob/main/final_tl_mobilenetv2_dataset6000_sidangready.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5AfJ0aAmXNQi"
      },
      "source": [
        "Import library dan depedencies yang akan digunakan"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ai4TMwb4WrVh"
      },
      "source": [
        "import tensorflow as tf\n",
        "import os\n",
        "import sys\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "import itertools\n",
        "import math\n",
        "\n",
        "from google.colab import drive\n",
        "from pathlib import Path\n",
        "from sklearn.metrics import classification_report, confusion_matrix\n",
        "from tensorflow.keras.optimizers import Adam\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vT7-nU2NXUTN"
      },
      "source": [
        "Buat koneksi dengan Google Drive"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wQK4YVAXXMOD",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 121
        },
        "outputId": "7843ffa7-c61c-4cd7-f9f9-08305fb13b2d"
      },
      "source": [
        "drive.mount('/content/drive')\n",
        "base = Path('/content/drive/My Drive/')\n",
        "sys.path.append(str(base))\n",
        "\n",
        "zip_path = base/'Dataset73.zip'\n",
        "!cp '{zip_path}' .\n",
        "\n",
        "!unzip -q Dataset73.zip\n",
        "!rm Dataset73.zip"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3aietf%3awg%3aoauth%3a2.0%3aoob&response_type=code&scope=email%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdocs.test%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive.photos.readonly%20https%3a%2f%2fwww.googleapis.com%2fauth%2fpeopleapi.readonly\n",
            "\n",
            "Enter your authorization code:\n",
            "··········\n",
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jtuv8OOqX12h"
      },
      "source": [
        "Definisikan direktori dataset di Google Drive, serta setup datagenerator dataset training & validasi"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Kb0uI0scXj9D"
      },
      "source": [
        "train_dir = 'Dataset73/train'\n",
        "validation_dir = 'Dataset73/valid'\n",
        "\n",
        "train_datagen = ImageDataGenerator(rescale=1./255)\n",
        "val_datagen = ImageDataGenerator(rescale=1./255)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qmkOWVqGYaoh"
      },
      "source": [
        "Masukan dataset kedalam generator berdasarkan batch, kemudian resize ke ukuran 224x224 dengan mode kelas berdasarkan kategorikal"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6dDiZuytYNiE",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 50
        },
        "outputId": "2a47376a-6988-4e63-c976-44b4e983974b"
      },
      "source": [
        "train_generator = train_datagen.flow_from_directory(\n",
        "        train_dir,\n",
        "        target_size=(224, 224),\n",
        "        batch_size=100,\n",
        "        class_mode='categorical')\n",
        "\n",
        "validation_generator = val_datagen.flow_from_directory(\n",
        "        validation_dir,\n",
        "        target_size=(224, 224),\n",
        "        batch_size=100,\n",
        "        shuffle=False,\n",
        "        class_mode='categorical')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 4200 images belonging to 6 classes.\n",
            "Found 1800 images belonging to 6 classes.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mjw8QCGBa60H"
      },
      "source": [
        "Buat file labels.txt untuk referensi label kelas di implementasi nantinya sekaligus digunakan pada tahap evaluasi dengan confusion matrix dan classification report"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "97WY4RaBaita",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "d2d65dca-eb05-40ca-b5d6-5f79aac7d25f"
      },
      "source": [
        "print (train_generator.class_indices)\n",
        "\n",
        "labels = '\\n'.join(sorted(train_generator.class_indices.keys()))\n",
        "\n",
        "with open('labels.txt', 'w') as f:\n",
        "  f.write(labels)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "{'100rb': 0, '10rb': 1, '20rb': 2, '2rb': 3, '50rb': 4, '5rb': 5}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MyFRAzeWZlJf"
      },
      "source": [
        "Import model pre-trained MobileNet V2 dari keras.applications(), tentukan input shapenya menjadi 224x224, instruksikan untuk tidak menggunakan lapisan akhir model, dan pilih wightsnya ImageNet"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6xYvELFhZF-k",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 50
        },
        "outputId": "5144ed45-8061-4e1a-ee13-53e03c1e4be8"
      },
      "source": [
        "IMG_SHAPE = (224, 224, 3)\n",
        "\n",
        "base_model = tf.keras.applications.MobileNetV2(input_shape=IMG_SHAPE,\n",
        "                                               include_top=False,\n",
        "                                               alpha=1.0,\n",
        "                                               weights='imagenet')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "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"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "I2f5jOxaZnwI"
      },
      "source": [
        "Freeze lapisan-lapisan yang ada di model pre-trained MobileNet v2"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9pm2YEJ4ZPZu"
      },
      "source": [
        "base_model.trainable = False"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bitI-YTaZ_bY"
      },
      "source": [
        "Buat dan gabungkan lapisan fully connected yang dibuat dengan model pre-trained yang sudah di bekukan"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "neBaDX5IZPGu",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 252
        },
        "outputId": "bdfedf64-fbec-4eb1-8c49-3439f724f917"
      },
      "source": [
        "model = tf.keras.models.Sequential()\n",
        "model.add(base_model)\n",
        "model.add(tf.keras.layers.Flatten())\n",
        "model.add(tf.keras.layers.Dense(6, activation='softmax'))\n",
        "\n",
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "mobilenetv2_1.00_224 (Model) (None, 7, 7, 1280)        2257984   \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 62720)             0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 6)                 376326    \n",
            "=================================================================\n",
            "Total params: 2,634,310\n",
            "Trainable params: 376,326\n",
            "Non-trainable params: 2,257,984\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LwQgTgMIbUN8"
      },
      "source": [
        "Konfigurasi parameter pelatihan dan mulai proses training-nya"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "m2jS5pP9aOje",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 689
        },
        "outputId": "b08221a5-b4ac-4586-afb0-8f84e04a977b"
      },
      "source": [
        "model.compile(loss='categorical_crossentropy',\n",
        "              optimizer=Adam(lr=0.0001),\n",
        "              metrics=['categorical_accuracy'])\n",
        "\n",
        "history = model.fit(\n",
        "      train_generator,\n",
        "      steps_per_epoch=42,  # 4200 images = batch_size * steps\n",
        "      epochs=20,\n",
        "      validation_data=validation_generator,\n",
        "      validation_steps=18,  # 1800 images = batch_size * steps\n",
        "      verbose=1)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/20\n",
            "42/42 [==============================] - 64s 2s/step - loss: 1.3628 - categorical_accuracy: 0.5090 - val_loss: 0.7428 - val_categorical_accuracy: 0.7272\n",
            "Epoch 2/20\n",
            "42/42 [==============================] - 63s 2s/step - loss: 0.3456 - categorical_accuracy: 0.9038 - val_loss: 0.4543 - val_categorical_accuracy: 0.8478\n",
            "Epoch 3/20\n",
            "42/42 [==============================] - 63s 2s/step - loss: 0.1584 - categorical_accuracy: 0.9817 - val_loss: 0.3756 - val_categorical_accuracy: 0.8794\n",
            "Epoch 4/20\n",
            "42/42 [==============================] - 63s 2s/step - loss: 0.0899 - categorical_accuracy: 0.9969 - val_loss: 0.3245 - val_categorical_accuracy: 0.8989\n",
            "Epoch 5/20\n",
            "42/42 [==============================] - 63s 2s/step - loss: 0.0587 - categorical_accuracy: 0.9998 - val_loss: 0.2904 - val_categorical_accuracy: 0.9106\n",
            "Epoch 6/20\n",
            "42/42 [==============================] - 63s 1s/step - loss: 0.0435 - categorical_accuracy: 1.0000 - val_loss: 0.2707 - val_categorical_accuracy: 0.9150\n",
            "Epoch 7/20\n",
            "42/42 [==============================] - 63s 2s/step - loss: 0.0334 - categorical_accuracy: 1.0000 - val_loss: 0.2553 - val_categorical_accuracy: 0.9228\n",
            "Epoch 8/20\n",
            "42/42 [==============================] - 63s 1s/step - loss: 0.0269 - categorical_accuracy: 1.0000 - val_loss: 0.2429 - val_categorical_accuracy: 0.9233\n",
            "Epoch 9/20\n",
            "42/42 [==============================] - 63s 1s/step - loss: 0.0221 - categorical_accuracy: 1.0000 - val_loss: 0.2337 - val_categorical_accuracy: 0.9272\n",
            "Epoch 10/20\n",
            "42/42 [==============================] - 62s 1s/step - loss: 0.0188 - categorical_accuracy: 1.0000 - val_loss: 0.2286 - val_categorical_accuracy: 0.9267\n",
            "Epoch 11/20\n",
            "42/42 [==============================] - 62s 1s/step - loss: 0.0161 - categorical_accuracy: 1.0000 - val_loss: 0.2209 - val_categorical_accuracy: 0.9289\n",
            "Epoch 12/20\n",
            "42/42 [==============================] - 62s 1s/step - loss: 0.0140 - categorical_accuracy: 1.0000 - val_loss: 0.2140 - val_categorical_accuracy: 0.9328\n",
            "Epoch 13/20\n",
            "42/42 [==============================] - 62s 1s/step - loss: 0.0124 - categorical_accuracy: 1.0000 - val_loss: 0.2093 - val_categorical_accuracy: 0.9306\n",
            "Epoch 14/20\n",
            "42/42 [==============================] - 62s 1s/step - loss: 0.0109 - categorical_accuracy: 1.0000 - val_loss: 0.2034 - val_categorical_accuracy: 0.9356\n",
            "Epoch 15/20\n",
            "42/42 [==============================] - 63s 1s/step - loss: 0.0098 - categorical_accuracy: 1.0000 - val_loss: 0.2003 - val_categorical_accuracy: 0.9328\n",
            "Epoch 16/20\n",
            "42/42 [==============================] - 63s 1s/step - loss: 0.0089 - categorical_accuracy: 1.0000 - val_loss: 0.1957 - val_categorical_accuracy: 0.9350\n",
            "Epoch 17/20\n",
            "42/42 [==============================] - 62s 1s/step - loss: 0.0081 - categorical_accuracy: 1.0000 - val_loss: 0.1942 - val_categorical_accuracy: 0.9356\n",
            "Epoch 18/20\n",
            "42/42 [==============================] - 62s 1s/step - loss: 0.0074 - categorical_accuracy: 1.0000 - val_loss: 0.1905 - val_categorical_accuracy: 0.9367\n",
            "Epoch 19/20\n",
            "42/42 [==============================] - 62s 1s/step - loss: 0.0067 - categorical_accuracy: 1.0000 - val_loss: 0.1873 - val_categorical_accuracy: 0.9394\n",
            "Epoch 20/20\n",
            "42/42 [==============================] - 63s 1s/step - loss: 0.0062 - categorical_accuracy: 1.0000 - val_loss: 0.1860 - val_categorical_accuracy: 0.9361\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XcGQcp_9cKjM"
      },
      "source": [
        "Evaluasi hasil pelatihan dengan grafik visualisasi proses training"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uWCl8K0Ib8cE",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 562
        },
        "outputId": "59a59192-6982-4a6a-ba19-126aa53b098f"
      },
      "source": [
        "# Retrieve a list of accuracy results on training and validation data\n",
        "# sets for each training epoch\n",
        "acc = history.history['categorical_accuracy']\n",
        "val_acc = history.history['val_categorical_accuracy']\n",
        "\n",
        "# Retrieve a list of list results on training and validation data\n",
        "# sets for each training epoch\n",
        "loss = history.history['loss']\n",
        "val_loss = history.history['val_loss']\n",
        "\n",
        "# Get number of epochs\n",
        "epochs = range(len(acc))\n",
        "\n",
        "# Plot training and validation accuracy per epoch\n",
        "plt.plot(epochs, acc, label = 'Training Accuracy')\n",
        "plt.plot(epochs, val_acc, label = 'Validation Accuracy')\n",
        "plt.legend(loc='lower right')\n",
        "plt.title('Training and validation accuracy')\n",
        "\n",
        "plt.figure()\n",
        "\n",
        "# Plot training and validation loss per epoch\n",
        "plt.plot(epochs, loss, label = 'Training Loss')\n",
        "plt.plot(epochs, val_loss, label = 'Validation Loss')\n",
        "plt.legend(loc='upper right')\n",
        "plt.title('Training and validation loss')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 1.0, 'Training and validation loss')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        },
        {
          "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": "BP4Et9txb2eK"
      },
      "source": [
        "Evaluasi akurasi akhir model pada kumpulan dataset validasi"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FmpV05w8bi2P",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 50
        },
        "outputId": "e369b445-689b-449f-a37a-7c42aea8f919"
      },
      "source": [
        "validation_generator.reset()\n",
        "STEP_SIZE_VALID=validation_generator.n//validation_generator.batch_size\n",
        "model.evaluate(validation_generator, steps=STEP_SIZE_VALID)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "18/18 [==============================] - 18s 1s/step - loss: 0.1860 - categorical_accuracy: 0.9361\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.18595214188098907, 0.9361110925674438]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iE4vdNqRedCu"
      },
      "source": [
        "Definisikan plotting confusion matrix terlebih dahulu"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DeAg7D2Bdh3N"
      },
      "source": [
        "def plot_confusion_matrix(cm, classes,\n",
        "                          normalize=False,\n",
        "                          title='Confusion matrix',\n",
        "                          cmap=plt.cm.Blues):\n",
        "    \"\"\"\n",
        "    This function prints and plots the confusion matrix.\n",
        "    Normalization can be applied by setting `normalize=True`.\n",
        "    \"\"\"\n",
        "    if normalize:\n",
        "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
        "        print(\"Normalized confusion matrix\")\n",
        "    else:\n",
        "        print('Confusion matrix, without normalization')\n",
        "\n",
        "    print(cm)\n",
        "\n",
        "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
        "    plt.title(title)\n",
        "    plt.colorbar()\n",
        "    tick_marks = np.arange(len(classes))\n",
        "    plt.xticks(tick_marks, classes, rotation=45)\n",
        "    plt.yticks(tick_marks, classes)\n",
        "\n",
        "    fmt = '.2f' if normalize else 'd'\n",
        "    thresh = cm.max() / 2.\n",
        "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
        "        plt.text(j, i, format(cm[i, j], fmt),\n",
        "                 horizontalalignment=\"center\",\n",
        "                 color=\"white\" if cm[i, j] > thresh else \"black\")\n",
        "\n",
        "    plt.ylabel('True label')\n",
        "    plt.xlabel('Predicted label')\n",
        "    plt.tight_layout()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YAtTFzLKej3_"
      },
      "source": [
        "Evaluasi klasifikasi dengan menggunakan confusion matrix"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Npc8dSLJdqg5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 812
        },
        "outputId": "a2a16446-a4f3-46a0-e2e8-7c8c31ed95c3"
      },
      "source": [
        "validation_generator.reset()\n",
        "Y_pred = model.predict(validation_generator, steps=STEP_SIZE_VALID)\n",
        "y_pred = np.argmax(Y_pred, axis=1)\n",
        "target_names = !cat labels.txt\n",
        "\n",
        "cm = confusion_matrix(validation_generator.classes, y_pred)\n",
        "\n",
        "# Compute confusion matrix\n",
        "cnf_matrix = cm\n",
        "np.set_printoptions(precision=2)\n",
        "\n",
        "# Plot non-normalized confusion matrix\n",
        "plt.figure()\n",
        "plot_confusion_matrix(cnf_matrix, classes=target_names, title='Confusion matrix, without normalization')\n",
        "\n",
        "# Plot normalized confusion matrix\n",
        "plt.figure()\n",
        "plot_confusion_matrix(cnf_matrix, classes=target_names, normalize=True, title='Normalized confusion matrix')\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Confusion matrix, without normalization\n",
            "[[284   3   3   0   0  10]\n",
            " [  0 270   6   3  13   8]\n",
            " [  1   5 286   2   4   2]\n",
            " [  1   7   1 283   3   5]\n",
            " [  1   7   1   1 289   1]\n",
            " [  4  10   2  10   1 273]]\n",
            "Normalized confusion matrix\n",
            "[[0.95 0.01 0.01 0.   0.   0.03]\n",
            " [0.   0.9  0.02 0.01 0.04 0.03]\n",
            " [0.   0.02 0.95 0.01 0.01 0.01]\n",
            " [0.   0.02 0.   0.94 0.01 0.02]\n",
            " [0.   0.02 0.   0.   0.96 0.  ]\n",
            " [0.01 0.03 0.01 0.03 0.   0.91]]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "v2Cr2Th6epsR"
      },
      "source": [
        "Evaluasi klasifikasi dengan laporan klasifikasi (classification report)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "I4lHYM-HdXBY",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 252
        },
        "outputId": "c174d31f-0b7c-4c76-c4bc-ff236c7ed1f4"
      },
      "source": [
        "print('Classification Report')\n",
        "target_names = !cat labels.txt\n",
        "print(classification_report(validation_generator.classes, y_pred, target_names=target_names))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Classification Report\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "       100rb       0.98      0.95      0.96       300\n",
            "        10rb       0.89      0.90      0.90       300\n",
            "        20rb       0.96      0.95      0.95       300\n",
            "         2rb       0.95      0.94      0.94       300\n",
            "        50rb       0.93      0.96      0.95       300\n",
            "         5rb       0.91      0.91      0.91       300\n",
            "\n",
            "    accuracy                           0.94      1800\n",
            "   macro avg       0.94      0.94      0.94      1800\n",
            "weighted avg       0.94      0.94      0.94      1800\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vMwHBsEye9hR"
      },
      "source": [
        "Simpan model dan konversikan kedalam format TensorFlow Lite"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XAhAVd_Me8nO",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "803c9ed7-248e-45d1-e786-8af4179973c8"
      },
      "source": [
        "converter = tf.lite.TFLiteConverter.from_keras_model(model)\n",
        "tflite_model = converter.convert()\n",
        "open (\"final_tl_mobilenetv2_dataset73_simpleLayer.tflite\" , \"wb\") .write(tflite_model)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "10352724"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 15
        }
      ]
    }
  ]
}