tflite_batik_TA_mobilenet_without_TL.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/ihsan02tempest/Batik-CNN-MobileNet-transfer-learning/blob/master/tflite_batik_TA_mobilenet_without_TL.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "chRXZoZqWJgU"
      },
      "source": [
        "TRAINING DATASET MOTIF BATIK MENGGUNAKAN PENDEKATAN MOBILENET DAN FRAMEWORK TENSORFLOWLITE\n",
        "\n",
        "Sample yang dibutuhkan =\n",
        "1. Motif Batik Corak Insang\n",
        "2. Motif Batik Dayak\n",
        "3. Motif Batik Megamendung\n",
        "4. Motif Batik Ikat Celup\n",
        "\n",
        "Source = https://www.kaggle.com/datasets/ihsan10/motif-batik-ta"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "djQDPyXWSctS",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "efc6adfa-e2f9-40ad-96ce-777265ef79f9"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ]
        }
      ],
      "source": [
        "#@title MOUNT DATA DARI GOOGLE DRIVE\n",
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "P9rJFwxFFKqv"
      },
      "outputs": [],
      "source": [
        "#@title ALOKASI AKSES KE FILE DATASET KAGGLE\n",
        "! chmod 600 /content/drive/MyDrive/Colab-Notebooks/KLASIFIKASI-BATIK/kaggle.json"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cWmhGO8vgGkx",
        "outputId": "08522007-6955-48c5-9497-e1186c57e656"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading motif-batik-ta.zip to /content\n",
            "100% 35.8M/35.8M [00:03<00:00, 16.0MB/s]\n",
            "100% 35.8M/35.8M [00:03<00:00, 11.8MB/s]\n"
          ]
        }
      ],
      "source": [
        "#@title AKSES PENYIMPANAN GOOGLE DRIVE UNTUK MENGIMPORT DATASET\n",
        "! KAGGLE_CONFIG_DIR=/content/drive/MyDrive/Colab-Notebooks/KLASIFIKASI-BATIK/ kaggle datasets download -d ihsan10/motif-batik-ta"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "lrGNkDj6gV_P"
      },
      "outputs": [],
      "source": [
        "#@title IMPORT DAN EKSTRAK ZIP YANG DIBUTUHKAN\n",
        "import zipfile\n",
        "zip_file = zipfile.ZipFile('/content/drive/MyDrive/Colab-Notebooks/KLASIFIKASI-BATIK/gambar motif batik.zip')\n",
        "zip_file.extractall('/content/drive/MyDrive/Colab-Notebooks/KLASIFIKASI-BATIK/')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "qlSWLBt4gcw2"
      },
      "outputs": [],
      "source": [
        "#@title IMPORT LIBRARY YANG DIBUTUHKAN\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Conv2D, GlobalAveragePooling2D, Activation, Dropout, Flatten, Dense, InputLayer, ZeroPadding2D, BatchNormalization, DepthwiseConv2D\n",
        "from keras.preprocessing import image\n",
        "from keras.optimizers import Adam\n",
        "from keras import regularizers\n",
        "from tensorflow import keras\n",
        "from tensorflow import lite\n",
        "from sklearn.metrics import confusion_matrix\n",
        "import tensorflow as tf\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "import sklearn.metrics as metrics\n",
        "import os\n",
        "os.environ[\"TF_XLA_FLAGS\"] = \"--tf_xla_cpu_global_jit\""
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "QIqLqX43hh9Y"
      },
      "outputs": [],
      "source": [
        "#@title PENENTUAN PARAMETER PELATIHAN DATA\n",
        "img_width, img_height = 128, 128\n",
        "\n",
        "train_data_dir = '/content/drive/MyDrive/Colab-Notebooks/KLASIFIKASI-BATIK/gambar motif batik/TRAIN'\n",
        "validation_data_dir = '/content/drive/MyDrive/Colab-Notebooks/KLASIFIKASI-BATIK/gambar motif batik/VALIDATION'\n",
        "test_data_dir = '/content/drive/MyDrive/Colab-Notebooks/KLASIFIKASI-BATIK/gambar motif batik/TEST'\n",
        "\n",
        "label_class = [\"Batik Corak Insang\", \"Batik Dayak\", \"Batik Ikat Celup\", \"Batik Megamendung\"]\n",
        "image_size = 128\n",
        "num_train = 100\n",
        "num_test = 20\n",
        "num_val = 10"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "06zG7zZR8lZ0",
        "outputId": "c7408dc4-7331-4e94-999e-a0af3a014b7a"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 440 images belonging to 4 classes.\n",
            "Found 80 images belonging to 4 classes.\n",
            "Found 120 images belonging to 4 classes.\n"
          ]
        }
      ],
      "source": [
        "#@title PRE-PROCESSING\n",
        "train_datagen = ImageDataGenerator( rescale = 1/255,\n",
        "                                    horizontal_flip=True,\n",
        "                                    vertical_flip=True,\n",
        "                                    fill_mode='nearest')\n",
        "\n",
        "validation_datagen = ImageDataGenerator( rescale = 1/255 )\n",
        "\n",
        "test_datagen = ImageDataGenerator( rescale = 1/255 )\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(train_data_dir, \n",
        "                                                    classes=label_class,\n",
        "                                                    batch_size=32,\n",
        "                                                    class_mode='categorical',\n",
        "                                                    target_size = (128, 128))\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(validation_data_dir, \n",
        "                                                    batch_size=16,\n",
        "                                                    classes=label_class,\n",
        "                                                    class_mode='categorical',\n",
        "                                                    target_size = (128, 128),\n",
        "                                                    shuffle=False)\n",
        "\n",
        "test_generator = test_datagen.flow_from_directory(test_data_dir, \n",
        "                                                    batch_size=16,\n",
        "                                                    classes=label_class,\n",
        "                                                    class_mode='categorical',\n",
        "                                                    target_size = (128, 128),\n",
        "                                                    shuffle=False)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "OL_WbLti7HAS"
      },
      "outputs": [],
      "source": [
        "#@title MODEL ARSITEKTUR MOBILENET (CNN DEPTHWISE SEPARABLE CONVOLUTION)\n",
        "model = Sequential()\n",
        "model.add(InputLayer(input_shape=(128,128,3), name='input_layer'))\n",
        "\n",
        "#layer 1\n",
        "model.add(ZeroPadding2D(((0,1),(0,1)), name='Conv2D_1_Pad'))\n",
        "model.add(Conv2D(32, kernel_size=(3,3), strides=(2,2), kernel_regularizer=regularizers.l1(0.01), name='Conv2D_1'))\n",
        "model.add(BatchNormalization(axis=1, name='Conv2D_1_Bn'))\n",
        "\n",
        "model.add(Activation(('relu'), name='Conv2D_1_Relu'))\n",
        "model.add(DepthwiseConv2D(kernel_size=(3,3),strides=(1,1), \n",
        "                          kernel_regularizer=regularizers.l1(0.01), padding='same', depth_multiplier=1, use_bias=False, name='ConvDw2D_1'))\n",
        "model.add(BatchNormalization(axis=1, name='ConvDw2D_1_Bn'))\n",
        "model.add(Activation(('relu'), name='ConvDw2D_1_Relu'))\n",
        "model.add(Conv2D(64, kernel_size=(1,1), strides=(1,1), name='ConvPw2D_1' ))\n",
        "model.add(BatchNormalization(axis=1, name='ConvPw2D_1_Bn'))\n",
        "model.add(Activation(('relu'), name='ConvPw2D_1_Relu'))\n",
        "\n",
        "#Layer 2\n",
        "model.add(ZeroPadding2D(((0,1),(0,1)), name='Conv2D_2_Pad'))\n",
        "model.add(DepthwiseConv2D(kernel_size=(3,3), padding='same', \n",
        "                          kernel_regularizer=regularizers.l1(0.01), strides=(2,2), depth_multiplier=1, use_bias=False, name='ConvDw2D_2'))\n",
        "model.add(BatchNormalization(axis=1, name='ConvDw2D_2_Bn'))\n",
        "model.add(Activation(('relu'), name='ConvDw2D_2_Relu'))\n",
        "model.add(Conv2D(128, kernel_size=(1,1), strides=(1,1), name='ConvPw2D_2' ))\n",
        "model.add(BatchNormalization(axis=1, name='ConvPw2D_2_Bn'))\n",
        "model.add(Activation(('relu'), name='ConvPw2D_2_Relu'))\n",
        "\n",
        "#Layer 3\n",
        "model.add(ZeroPadding2D(((0,1),(0,1)), name='Conv2D_3_Pad'))\n",
        "model.add(DepthwiseConv2D(kernel_size=(3,3), padding='same', \n",
        "                          kernel_regularizer=regularizers.l1(0.01), strides=(1,1), depth_multiplier=1, use_bias=False, name='ConvDw2D_3'))\n",
        "model.add(BatchNormalization(axis=1, name='ConvDw2D_3_Bn'))\n",
        "model.add(Activation(('relu'), name='ConvDw2D_3_Relu'))\n",
        "model.add(Conv2D(128, kernel_size=(1,1), strides=(1,1), name='ConvPw2D_3' ))\n",
        "model.add(BatchNormalization(axis=1, name='ConvPw2D_3_Bn'))\n",
        "model.add(Activation(('relu'), name='ConvPw2D_3_Relu'))\n",
        "\n",
        "#Average Pooling\n",
        "model.add(GlobalAveragePooling2D())\n",
        "\n",
        "#Fully-Connected\n",
        "model.add(Dropout(0.25))\n",
        "predictions = model.add(Dense(4, activation = 'softmax'))"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#@title CEK SUMMARY TRAINING MODEL\n",
        "model.summary()"
      ],
      "metadata": {
        "id": "h2jA7xuPAAY6",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c65e95c0-3dc2-40e0-b8ae-3f4d3f9e2079"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " Conv2D_1_Pad (ZeroPadding2D  (None, 129, 129, 3)      0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " Conv2D_1 (Conv2D)           (None, 64, 64, 32)        896       \n",
            "                                                                 \n",
            " Conv2D_1_Bn (BatchNormaliza  (None, 64, 64, 32)       256       \n",
            " tion)                                                           \n",
            "                                                                 \n",
            " Conv2D_1_Relu (Activation)  (None, 64, 64, 32)        0         \n",
            "                                                                 \n",
            " ConvDw2D_1 (DepthwiseConv2D  (None, 64, 64, 32)       288       \n",
            " )                                                               \n",
            "                                                                 \n",
            " ConvDw2D_1_Bn (BatchNormali  (None, 64, 64, 32)       256       \n",
            " zation)                                                         \n",
            "                                                                 \n",
            " ConvDw2D_1_Relu (Activation  (None, 64, 64, 32)       0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " ConvPw2D_1 (Conv2D)         (None, 64, 64, 64)        2112      \n",
            "                                                                 \n",
            " ConvPw2D_1_Bn (BatchNormali  (None, 64, 64, 64)       256       \n",
            " zation)                                                         \n",
            "                                                                 \n",
            " ConvPw2D_1_Relu (Activation  (None, 64, 64, 64)       0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " Conv2D_2_Pad (ZeroPadding2D  (None, 65, 65, 64)       0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " ConvDw2D_2 (DepthwiseConv2D  (None, 33, 33, 64)       576       \n",
            " )                                                               \n",
            "                                                                 \n",
            " ConvDw2D_2_Bn (BatchNormali  (None, 33, 33, 64)       132       \n",
            " zation)                                                         \n",
            "                                                                 \n",
            " ConvDw2D_2_Relu (Activation  (None, 33, 33, 64)       0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " ConvPw2D_2 (Conv2D)         (None, 33, 33, 128)       8320      \n",
            "                                                                 \n",
            " ConvPw2D_2_Bn (BatchNormali  (None, 33, 33, 128)      132       \n",
            " zation)                                                         \n",
            "                                                                 \n",
            " ConvPw2D_2_Relu (Activation  (None, 33, 33, 128)      0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " Conv2D_3_Pad (ZeroPadding2D  (None, 34, 34, 128)      0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " ConvDw2D_3 (DepthwiseConv2D  (None, 34, 34, 128)      1152      \n",
            " )                                                               \n",
            "                                                                 \n",
            " ConvDw2D_3_Bn (BatchNormali  (None, 34, 34, 128)      136       \n",
            " zation)                                                         \n",
            "                                                                 \n",
            " ConvDw2D_3_Relu (Activation  (None, 34, 34, 128)      0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " ConvPw2D_3 (Conv2D)         (None, 34, 34, 128)       16512     \n",
            "                                                                 \n",
            " ConvPw2D_3_Bn (BatchNormali  (None, 34, 34, 128)      136       \n",
            " zation)                                                         \n",
            "                                                                 \n",
            " ConvPw2D_3_Relu (Activation  (None, 34, 34, 128)      0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " global_average_pooling2d (G  (None, 128)              0         \n",
            " lobalAveragePooling2D)                                          \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 128)               0         \n",
            "                                                                 \n",
            " dense (Dense)               (None, 4)                 516       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 31,676\n",
            "Trainable params: 31,024\n",
            "Non-trainable params: 652\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "zr72diWj7VBW"
      },
      "outputs": [],
      "source": [
        "#@title COMPILE MASUKAN\n",
        "opt = keras.optimizers.Adam(learning_rate=0.001,beta_1=0.9, beta_2=0.999, amsgrad=True)\n",
        "model.compile(optimizer=opt, loss='categorical_crossentropy', metrics=['accuracy'])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "IfwNtFU98PNK"
      },
      "outputs": [],
      "source": [
        "#@title CHECKPOINT PENYIMPANAN MODEL\n",
        "#import ModelCheckpoint\n",
        "from keras.callbacks import ModelCheckpoint\n",
        "\n",
        "#menentukan path folder menyimpan model\n",
        "filepath = '/content/drive/MyDrive/Colab-Notebooks/KLASIFIKASI-BATIK/my-saved-model/model8/model-batik.h5'"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "RWLEAWzy8P2Q"
      },
      "outputs": [],
      "source": [
        "#@title MENYIMPAN MODEL DENGAN KRITERIA TERBAIK\n",
        "checkpoint = ModelCheckpoint(filepath,\n",
        "                             save_weights_only=True,\n",
        "                             save_best_only=True,\n",
        "                             monitor='val_loss',\n",
        "                             mode='min')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "D3hWjmii7fPr",
        "outputId": "45d04583-1b65-4fac-ba0b-7c7731b66a7c"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "10/10 [==============================] - 46s 3s/step - loss: 2.0952 - accuracy: 0.2500 - val_loss: 1.9401 - val_accuracy: 0.2500\n",
            "Epoch 2/100\n",
            "10/10 [==============================] - 36s 3s/step - loss: 1.8583 - accuracy: 0.3654 - val_loss: 1.8899 - val_accuracy: 0.2500\n",
            "Epoch 3/100\n",
            "10/10 [==============================] - 39s 3s/step - loss: 1.7770 - accuracy: 0.4031 - val_loss: 1.8507 - val_accuracy: 0.2500\n",
            "Epoch 4/100\n",
            "10/10 [==============================] - 33s 3s/step - loss: 1.6333 - accuracy: 0.4750 - val_loss: 1.8208 - val_accuracy: 0.2500\n",
            "Epoch 5/100\n",
            "10/10 [==============================] - 31s 2s/step - loss: 1.5589 - accuracy: 0.4968 - val_loss: 1.7982 - val_accuracy: 0.2500\n",
            "Epoch 6/100\n",
            "10/10 [==============================] - 34s 3s/step - loss: 1.4488 - accuracy: 0.5417 - val_loss: 1.7804 - val_accuracy: 0.2500\n",
            "Epoch 7/100\n",
            "10/10 [==============================] - 35s 3s/step - loss: 1.4149 - accuracy: 0.5353 - val_loss: 1.7638 - val_accuracy: 0.2500\n",
            "Epoch 8/100\n",
            "10/10 [==============================] - 35s 3s/step - loss: 1.3152 - accuracy: 0.5994 - val_loss: 1.7492 - val_accuracy: 0.2500\n",
            "Epoch 9/100\n",
            "10/10 [==============================] - 34s 3s/step - loss: 1.2852 - accuracy: 0.6375 - val_loss: 1.7358 - val_accuracy: 0.2500\n",
            "Epoch 10/100\n",
            "10/10 [==============================] - 34s 3s/step - loss: 1.2371 - accuracy: 0.6186 - val_loss: 1.7242 - val_accuracy: 0.2500\n",
            "Epoch 11/100\n",
            "10/10 [==============================] - 31s 2s/step - loss: 1.2039 - accuracy: 0.6506 - val_loss: 1.7142 - val_accuracy: 0.2500\n",
            "Epoch 12/100\n",
            "10/10 [==============================] - 36s 3s/step - loss: 1.1518 - accuracy: 0.6442 - val_loss: 1.7048 - val_accuracy: 0.2500\n",
            "Epoch 13/100\n",
            "10/10 [==============================] - 34s 3s/step - loss: 1.1753 - accuracy: 0.6667 - val_loss: 1.6967 - val_accuracy: 0.2500\n",
            "Epoch 14/100\n",
            "10/10 [==============================] - 35s 3s/step - loss: 1.0703 - accuracy: 0.6827 - val_loss: 1.6891 - val_accuracy: 0.2500\n",
            "Epoch 15/100\n",
            "10/10 [==============================] - 34s 3s/step - loss: 1.1154 - accuracy: 0.6154 - val_loss: 1.6833 - val_accuracy: 0.2500\n",
            "Epoch 16/100\n",
            "10/10 [==============================] - 38s 3s/step - loss: 1.0310 - accuracy: 0.7094 - val_loss: 1.6781 - val_accuracy: 0.2500\n",
            "Epoch 17/100\n",
            "10/10 [==============================] - 37s 3s/step - loss: 1.0378 - accuracy: 0.6763 - val_loss: 1.6755 - val_accuracy: 0.2500\n",
            "Epoch 18/100\n",
            "10/10 [==============================] - 36s 3s/step - loss: 1.0038 - accuracy: 0.7125 - val_loss: 1.6731 - val_accuracy: 0.2500\n",
            "Epoch 19/100\n",
            "10/10 [==============================] - 31s 2s/step - loss: 1.0082 - accuracy: 0.6923 - val_loss: 1.6705 - val_accuracy: 0.2500\n",
            "Epoch 20/100\n",
            "10/10 [==============================] - 36s 3s/step - loss: 0.9611 - accuracy: 0.7212 - val_loss: 1.6713 - val_accuracy: 0.2500\n",
            "Epoch 21/100\n",
            "10/10 [==============================] - 33s 3s/step - loss: 0.9850 - accuracy: 0.7115 - val_loss: 1.6712 - val_accuracy: 0.2500\n",
            "Epoch 22/100\n",
            "10/10 [==============================] - 35s 3s/step - loss: 0.9349 - accuracy: 0.7188 - val_loss: 1.6746 - val_accuracy: 0.2500\n",
            "Epoch 23/100\n",
            "10/10 [==============================] - 34s 3s/step - loss: 0.9645 - accuracy: 0.7340 - val_loss: 1.6733 - val_accuracy: 0.2875\n",
            "Epoch 24/100\n",
            "10/10 [==============================] - 31s 3s/step - loss: 1.0115 - accuracy: 0.6923 - val_loss: 1.6769 - val_accuracy: 0.2500\n",
            "Epoch 25/100\n",
            "10/10 [==============================] - 34s 3s/step - loss: 1.0979 - accuracy: 0.6625 - val_loss: 1.6763 - val_accuracy: 0.2750\n",
            "Epoch 26/100\n",
            "10/10 [==============================] - 34s 3s/step - loss: 0.9859 - accuracy: 0.6969 - val_loss: 1.6769 - val_accuracy: 0.3125\n",
            "Epoch 27/100\n",
            "10/10 [==============================] - 33s 3s/step - loss: 0.9466 - accuracy: 0.6987 - val_loss: 1.6769 - val_accuracy: 0.3250\n",
            "Epoch 28/100\n",
            "10/10 [==============================] - 32s 2s/step - loss: 0.9562 - accuracy: 0.6859 - val_loss: 1.6673 - val_accuracy: 0.2500\n",
            "Epoch 29/100\n",
            "10/10 [==============================] - 38s 3s/step - loss: 1.0026 - accuracy: 0.7083 - val_loss: 1.6675 - val_accuracy: 0.2750\n",
            "Epoch 30/100\n",
            "10/10 [==============================] - 31s 3s/step - loss: 0.9472 - accuracy: 0.7179 - val_loss: 1.6689 - val_accuracy: 0.2625\n",
            "Epoch 31/100\n",
            "10/10 [==============================] - 34s 3s/step - loss: 0.8668 - accuracy: 0.7372 - val_loss: 1.6630 - val_accuracy: 0.2750\n",
            "Epoch 32/100\n",
            "10/10 [==============================] - 35s 3s/step - loss: 0.8630 - accuracy: 0.7436 - val_loss: 1.6616 - val_accuracy: 0.2750\n",
            "Epoch 33/100\n",
            "10/10 [==============================] - 36s 3s/step - loss: 0.8402 - accuracy: 0.7532 - val_loss: 1.6444 - val_accuracy: 0.2750\n",
            "Epoch 34/100\n",
            "10/10 [==============================] - 38s 3s/step - loss: 0.8321 - accuracy: 0.7596 - val_loss: 1.6355 - val_accuracy: 0.2750\n",
            "Epoch 35/100\n",
            "10/10 [==============================] - 37s 3s/step - loss: 0.8284 - accuracy: 0.7532 - val_loss: 1.6111 - val_accuracy: 0.2875\n",
            "Epoch 36/100\n",
            "10/10 [==============================] - 35s 3s/step - loss: 0.8499 - accuracy: 0.7469 - val_loss: 1.5774 - val_accuracy: 0.2875\n",
            "Epoch 37/100\n",
            "10/10 [==============================] - 32s 3s/step - loss: 0.8031 - accuracy: 0.7628 - val_loss: 1.6075 - val_accuracy: 0.3250\n",
            "Epoch 38/100\n",
            "10/10 [==============================] - 34s 3s/step - loss: 0.8062 - accuracy: 0.7596 - val_loss: 1.5845 - val_accuracy: 0.3250\n",
            "Epoch 39/100\n",
            "10/10 [==============================] - 31s 2s/step - loss: 0.8311 - accuracy: 0.7853 - val_loss: 1.4900 - val_accuracy: 0.4000\n",
            "Epoch 40/100\n",
            "10/10 [==============================] - 34s 3s/step - loss: 0.8004 - accuracy: 0.7724 - val_loss: 1.3914 - val_accuracy: 0.3875\n",
            "Epoch 41/100\n",
            "10/10 [==============================] - 31s 2s/step - loss: 0.7427 - accuracy: 0.7781 - val_loss: 1.3672 - val_accuracy: 0.3625\n",
            "Epoch 42/100\n",
            "10/10 [==============================] - 39s 3s/step - loss: 0.7350 - accuracy: 0.7756 - val_loss: 1.2572 - val_accuracy: 0.5500\n",
            "Epoch 43/100\n",
            "10/10 [==============================] - 35s 3s/step - loss: 0.7890 - accuracy: 0.7875 - val_loss: 1.2465 - val_accuracy: 0.5000\n",
            "Epoch 44/100\n",
            "10/10 [==============================] - 31s 3s/step - loss: 0.7505 - accuracy: 0.7981 - val_loss: 1.1885 - val_accuracy: 0.5875\n",
            "Epoch 45/100\n",
            "10/10 [==============================] - 36s 3s/step - loss: 0.7096 - accuracy: 0.7885 - val_loss: 1.1319 - val_accuracy: 0.6375\n",
            "Epoch 46/100\n",
            "10/10 [==============================] - 34s 3s/step - loss: 0.7412 - accuracy: 0.7906 - val_loss: 1.1666 - val_accuracy: 0.5875\n",
            "Epoch 47/100\n",
            "10/10 [==============================] - 31s 2s/step - loss: 0.7176 - accuracy: 0.8045 - val_loss: 1.0887 - val_accuracy: 0.5875\n",
            "Epoch 48/100\n",
            "10/10 [==============================] - 36s 3s/step - loss: 0.7127 - accuracy: 0.8000 - val_loss: 1.0018 - val_accuracy: 0.6375\n",
            "Epoch 49/100\n",
            "10/10 [==============================] - 36s 3s/step - loss: 0.7455 - accuracy: 0.7853 - val_loss: 0.8903 - val_accuracy: 0.7250\n",
            "Epoch 50/100\n",
            "10/10 [==============================] - 34s 3s/step - loss: 0.7145 - accuracy: 0.8141 - val_loss: 0.9257 - val_accuracy: 0.7000\n",
            "Epoch 51/100\n",
            "10/10 [==============================] - 37s 3s/step - loss: 0.7189 - accuracy: 0.8156 - val_loss: 0.8696 - val_accuracy: 0.7625\n",
            "Epoch 52/100\n",
            "10/10 [==============================] - 31s 3s/step - loss: 0.8302 - accuracy: 0.7500 - val_loss: 0.8649 - val_accuracy: 0.8000\n",
            "Epoch 53/100\n",
            "10/10 [==============================] - 32s 2s/step - loss: 0.8425 - accuracy: 0.7404 - val_loss: 0.8146 - val_accuracy: 0.7875\n",
            "Epoch 54/100\n",
            "10/10 [==============================] - 37s 3s/step - loss: 0.7590 - accuracy: 0.8109 - val_loss: 0.7899 - val_accuracy: 0.7875\n",
            "Epoch 55/100\n",
            "10/10 [==============================] - 31s 3s/step - loss: 0.7241 - accuracy: 0.8077 - val_loss: 0.6722 - val_accuracy: 0.8250\n",
            "Epoch 56/100\n",
            "10/10 [==============================] - 33s 3s/step - loss: 0.6648 - accuracy: 0.8173 - val_loss: 0.6745 - val_accuracy: 0.8375\n",
            "Epoch 57/100\n",
            "10/10 [==============================] - 31s 3s/step - loss: 0.6669 - accuracy: 0.8205 - val_loss: 0.7345 - val_accuracy: 0.8125\n",
            "Epoch 58/100\n",
            "10/10 [==============================] - 31s 3s/step - loss: 0.6702 - accuracy: 0.8013 - val_loss: 0.6813 - val_accuracy: 0.8750\n",
            "Epoch 59/100\n",
            "10/10 [==============================] - 32s 3s/step - loss: 0.6196 - accuracy: 0.8469 - val_loss: 0.6789 - val_accuracy: 0.7875\n",
            "Epoch 60/100\n",
            "10/10 [==============================] - 38s 3s/step - loss: 0.6558 - accuracy: 0.8156 - val_loss: 0.5905 - val_accuracy: 0.8750\n",
            "Epoch 61/100\n",
            "10/10 [==============================] - 38s 3s/step - loss: 0.6966 - accuracy: 0.8031 - val_loss: 0.6385 - val_accuracy: 0.7875\n",
            "Epoch 62/100\n",
            "10/10 [==============================] - 36s 3s/step - loss: 0.6301 - accuracy: 0.8281 - val_loss: 0.6505 - val_accuracy: 0.7875\n",
            "Epoch 63/100\n",
            "10/10 [==============================] - 32s 3s/step - loss: 0.6336 - accuracy: 0.8109 - val_loss: 0.5767 - val_accuracy: 0.8750\n",
            "Epoch 64/100\n",
            "10/10 [==============================] - 31s 2s/step - loss: 0.5920 - accuracy: 0.8686 - val_loss: 0.5759 - val_accuracy: 0.8875\n",
            "Epoch 65/100\n",
            "10/10 [==============================] - 36s 3s/step - loss: 0.6799 - accuracy: 0.7853 - val_loss: 0.5457 - val_accuracy: 0.8750\n",
            "Epoch 66/100\n",
            "10/10 [==============================] - 33s 3s/step - loss: 0.7178 - accuracy: 0.7821 - val_loss: 0.6683 - val_accuracy: 0.7750\n",
            "Epoch 67/100\n",
            "10/10 [==============================] - 39s 3s/step - loss: 0.6810 - accuracy: 0.8156 - val_loss: 0.5922 - val_accuracy: 0.8375\n",
            "Epoch 68/100\n",
            "10/10 [==============================] - 37s 3s/step - loss: 0.6519 - accuracy: 0.8333 - val_loss: 0.5355 - val_accuracy: 0.8750\n",
            "Epoch 69/100\n",
            "10/10 [==============================] - 35s 3s/step - loss: 0.6400 - accuracy: 0.8156 - val_loss: 0.5407 - val_accuracy: 0.9000\n",
            "Epoch 70/100\n",
            "10/10 [==============================] - 32s 3s/step - loss: 0.5369 - accuracy: 0.8718 - val_loss: 0.5638 - val_accuracy: 0.9125\n",
            "Epoch 71/100\n",
            "10/10 [==============================] - 35s 3s/step - loss: 0.5859 - accuracy: 0.8429 - val_loss: 0.5222 - val_accuracy: 0.8875\n",
            "Epoch 72/100\n",
            "10/10 [==============================] - 32s 2s/step - loss: 0.6187 - accuracy: 0.8281 - val_loss: 0.5387 - val_accuracy: 0.8375\n",
            "Epoch 73/100\n",
            "10/10 [==============================] - 35s 3s/step - loss: 0.6265 - accuracy: 0.8397 - val_loss: 0.5344 - val_accuracy: 0.8500\n",
            "Epoch 74/100\n",
            "10/10 [==============================] - 36s 3s/step - loss: 0.6955 - accuracy: 0.8077 - val_loss: 0.6585 - val_accuracy: 0.8250\n",
            "Epoch 75/100\n",
            "10/10 [==============================] - 35s 3s/step - loss: 0.7670 - accuracy: 0.7628 - val_loss: 0.5068 - val_accuracy: 0.9000\n",
            "Epoch 76/100\n",
            "10/10 [==============================] - 34s 3s/step - loss: 0.6800 - accuracy: 0.8109 - val_loss: 0.5325 - val_accuracy: 0.8625\n",
            "Epoch 77/100\n",
            "10/10 [==============================] - 32s 2s/step - loss: 0.5800 - accuracy: 0.8500 - val_loss: 0.5278 - val_accuracy: 0.9000\n",
            "Epoch 78/100\n",
            "10/10 [==============================] - 32s 2s/step - loss: 0.5906 - accuracy: 0.8281 - val_loss: 0.5079 - val_accuracy: 0.8875\n",
            "Epoch 79/100\n",
            "10/10 [==============================] - 37s 3s/step - loss: 0.5446 - accuracy: 0.8814 - val_loss: 0.5014 - val_accuracy: 0.8875\n",
            "Epoch 80/100\n",
            "10/10 [==============================] - 35s 3s/step - loss: 0.5267 - accuracy: 0.8590 - val_loss: 0.4746 - val_accuracy: 0.9000\n",
            "Epoch 81/100\n",
            "10/10 [==============================] - 36s 3s/step - loss: 0.5182 - accuracy: 0.8782 - val_loss: 0.5095 - val_accuracy: 0.9250\n",
            "Epoch 82/100\n",
            "10/10 [==============================] - 35s 3s/step - loss: 0.4871 - accuracy: 0.8782 - val_loss: 0.4862 - val_accuracy: 0.9000\n",
            "Epoch 83/100\n",
            "10/10 [==============================] - 37s 3s/step - loss: 0.4825 - accuracy: 0.9135 - val_loss: 0.4478 - val_accuracy: 0.9125\n",
            "Epoch 84/100\n",
            "10/10 [==============================] - 35s 3s/step - loss: 0.5052 - accuracy: 0.8718 - val_loss: 0.5111 - val_accuracy: 0.8500\n",
            "Epoch 85/100\n",
            "10/10 [==============================] - 36s 3s/step - loss: 0.5184 - accuracy: 0.8590 - val_loss: 0.4968 - val_accuracy: 0.9000\n",
            "Epoch 86/100\n",
            "10/10 [==============================] - 34s 3s/step - loss: 0.5007 - accuracy: 0.8686 - val_loss: 0.4819 - val_accuracy: 0.8625\n",
            "Epoch 87/100\n",
            "10/10 [==============================] - 31s 3s/step - loss: 0.4769 - accuracy: 0.9000 - val_loss: 0.5305 - val_accuracy: 0.8750\n",
            "Epoch 88/100\n",
            "10/10 [==============================] - 34s 3s/step - loss: 0.4755 - accuracy: 0.8846 - val_loss: 0.4278 - val_accuracy: 0.9000\n",
            "Epoch 89/100\n",
            "10/10 [==============================] - 31s 3s/step - loss: 0.4830 - accuracy: 0.8782 - val_loss: 0.4860 - val_accuracy: 0.8625\n",
            "Epoch 90/100\n",
            "10/10 [==============================] - 34s 3s/step - loss: 0.5181 - accuracy: 0.8718 - val_loss: 0.5130 - val_accuracy: 0.8875\n",
            "Epoch 91/100\n",
            "10/10 [==============================] - 35s 3s/step - loss: 0.4938 - accuracy: 0.8526 - val_loss: 0.4751 - val_accuracy: 0.9125\n",
            "Epoch 92/100\n",
            "10/10 [==============================] - 36s 3s/step - loss: 0.4658 - accuracy: 0.8974 - val_loss: 0.4540 - val_accuracy: 0.9125\n",
            "Epoch 93/100\n",
            "10/10 [==============================] - 35s 3s/step - loss: 0.4854 - accuracy: 0.8910 - val_loss: 0.4898 - val_accuracy: 0.8750\n",
            "Epoch 94/100\n",
            "10/10 [==============================] - 32s 2s/step - loss: 0.4934 - accuracy: 0.8814 - val_loss: 0.4856 - val_accuracy: 0.9250\n",
            "Epoch 95/100\n",
            "10/10 [==============================] - 37s 3s/step - loss: 0.4786 - accuracy: 0.8656 - val_loss: 0.5227 - val_accuracy: 0.9375\n",
            "Epoch 96/100\n",
            "10/10 [==============================] - 35s 3s/step - loss: 0.4367 - accuracy: 0.9000 - val_loss: 0.4441 - val_accuracy: 0.8750\n",
            "Epoch 97/100\n",
            "10/10 [==============================] - 32s 2s/step - loss: 0.4547 - accuracy: 0.8782 - val_loss: 0.4104 - val_accuracy: 0.9000\n",
            "Epoch 98/100\n",
            "10/10 [==============================] - 36s 3s/step - loss: 0.4405 - accuracy: 0.8969 - val_loss: 0.4131 - val_accuracy: 0.9000\n",
            "Epoch 99/100\n",
            "10/10 [==============================] - 38s 3s/step - loss: 0.4282 - accuracy: 0.9071 - val_loss: 0.3844 - val_accuracy: 0.9125\n",
            "Epoch 100/100\n",
            "10/10 [==============================] - 32s 3s/step - loss: 0.4002 - accuracy: 0.9250 - val_loss: 0.4386 - val_accuracy: 0.9125\n"
          ]
        }
      ],
      "source": [
        "#@title TRANING MODEL\n",
        "history = model.fit(train_generator, \n",
        "                    validation_data=validation_generator,\n",
        "                    steps_per_epoch=10,\n",
        "                    epochs=100,\n",
        "                    validation_steps=10,\n",
        "                    callbacks=[checkpoint],\n",
        "                    verbose=1,\n",
        "                    use_multiprocessing=True,\n",
        "                    workers=4)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 887
        },
        "id": "uO_ThwZOv2_T",
        "outputId": "bfe964e9-9688-4d8b-87ed-3e53b941d2d8"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "#@title GRAFIK NILAI AKURASI TRAINING DAN VALIDASI\n",
        "plt.plot(history.history['accuracy'])\n",
        "plt.plot(history.history['val_accuracy'])\n",
        "plt.title('Nilai Akurasi pada tiap epoch')\n",
        "plt.legend(['training', 'validation'], loc='best')\n",
        "plt.show()\n",
        "\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.title('Nilai Loss pada tiap epoch')\n",
        "plt.legend(['training', 'validation'], loc='best')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NBjQbbIQxUHh",
        "outputId": "aec8f0d2-54fa-4805-a4fb-c0af9784c8eb"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "14/14 - 15s - loss: 0.3576 - accuracy: 0.9409 - 15s/epoch - 1s/step\n",
            "5/5 - 2s - loss: 0.3844 - accuracy: 0.9125 - 2s/epoch - 363ms/step\n"
          ]
        }
      ],
      "source": [
        "#@title EVALUASI MODEL\n",
        "model.load_weights(filepath)\n",
        "train_loss, train_accuracy = model.evaluate(train_generator, verbose=2)\n",
        "val_loss, val_accuracy = model.evaluate(validation_generator, verbose=2)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.save('/content/drive/MyDrive/Colab-Notebooks/KLASIFIKASI-BATIK/my-saved-model/model8/model-batik.h5')\n",
        "loaded_model=tf.keras.models.load_model('/content/drive/MyDrive/Colab-Notebooks/KLASIFIKASI-BATIK/my-saved-model/model8/model-batik.h5')"
      ],
      "metadata": {
        "id": "tegKPWMMwouT"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "loaded_model_json = loaded_model.to_json()\n",
        "with open('loaded_model.json', 'w') as json_file:\n",
        "    json_file.write(loaded_model_json)"
      ],
      "metadata": {
        "id": "Y3Z3cKhCwq9D"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "converter = tf.lite.TFLiteConverter.from_keras_model(loaded_model)\n",
        "converter.experimental_new_converter = True\n",
        "tflite_model = converter.convert()\n",
        "open(\"/content/drive/MyDrive/Colab-Notebooks/KLASIFIKASI-BATIK/my-saved-model/model8/converted_model.tflite\", \"wb\").write(tflite_model)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rlIEIcPKwtvs",
        "outputId": "19adeb7d-4331-40d6-a70c-45f20591c6a5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 5 of 8). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "137460"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#@title CONFUSION MATRIX DAN KLASIFIKASI REPORT\n",
        "model.load_weights(filepath)\n",
        "test_steps_per_epoch = np.math.ceil(test_generator.samples / test_generator.batch_size)\n",
        "predictions = model.predict(test_generator, steps=test_steps_per_epoch)\n",
        "predicted_classes = np.argmax(predictions, axis=1)\n",
        "class_labels = list(test_generator.class_indices.keys())\n",
        "report = metrics.classification_report(test_generator.classes, predicted_classes, target_names=class_labels)\n",
        "print('Confusion Matrix')\n",
        "print(confusion_matrix(test_generator.classes, predicted_classes))\n",
        "print('Classification Report')\n",
        "print(report)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hUwBkTvkomch",
        "outputId": "7d6ab622-5cbe-4a69-d1ae-737473c864ec"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "8/8 [==============================] - 13s 2s/step\n",
            "Confusion Matrix\n",
            "[[27  2  0  1]\n",
            " [ 0 29  0  1]\n",
            " [ 0  1 28  1]\n",
            " [ 1  1  3 25]]\n",
            "Classification Report\n",
            "                    precision    recall  f1-score   support\n",
            "\n",
            "Batik Corak Insang       0.96      0.90      0.93        30\n",
            "       Batik Dayak       0.88      0.97      0.92        30\n",
            "  Batik Ikat Celup       0.90      0.93      0.92        30\n",
            " Batik Megamendung       0.89      0.83      0.86        30\n",
            "\n",
            "          accuracy                           0.91       120\n",
            "         macro avg       0.91      0.91      0.91       120\n",
            "      weighted avg       0.91      0.91      0.91       120\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#@title HITUNG LANGSUNG AKURASI, PRESISI, RECALL, DAN F1-SCORE\n",
        "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score\n",
        "\n",
        "# hitung akurasi\n",
        "accuracy = accuracy_score(test_generator.classes, predicted_classes)\n",
        "\n",
        "# hitung presisi\n",
        "precision = precision_score(test_generator.classes, predicted_classes, average='weighted')\n",
        "\n",
        "# hitung recall\n",
        "recall = recall_score(test_generator.classes, predicted_classes, average='weighted')\n",
        "\n",
        "# hitung f1-score\n",
        "f1 = f1_score(test_generator.classes, predicted_classes, average='weighted')\n",
        "\n",
        "print(\"Accuracy:\", accuracy)\n",
        "print(\"Precision:\", precision)\n",
        "print(\"Recall:\", recall)\n",
        "print(\"F1-Score:\", f1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1aCifCkOlEfK",
        "outputId": "52a57b4a-7929-4be5-ad62-cb9e401a89fe"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Accuracy: 0.9083333333333333\n",
            "Precision: 0.9097891355955873\n",
            "Recall: 0.9083333333333333\n",
            "F1-Score: 0.9079427889490072\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#@title DISPLAY HASIL DARI EVALUASI CONFUSION MATRIX\n",
        "from sklearn.metrics import ConfusionMatrixDisplay, confusion_matrix\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "\n",
        "fig, ax = plt.subplots(figsize=(10, 10))\n",
        "cm = confusion_matrix(test_generator.classes, predicted_classes)\n",
        "\n",
        "disp = ConfusionMatrixDisplay(confusion_matrix=cm, display_labels=label_class)\n",
        "\n",
        "disp.plot(cmap=plt.cm.binary, ax=ax)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 815
        },
        "id": "OyeVXncOlIUP",
        "outputId": "2c950650-93b4-4245-b889-a7cf05598421"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x1000 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    }
  ],
  "metadata": {
    "colab": {
      "provenance": [],
      "mount_file_id": "1VTuNXKUowaDj8_Nkg4C90EHYWGg9QPR9",
      "authorship_tag": "ABX9TyOvxYG5lgza2aFqPKyNtgKQ",
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}