transfer_learning.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "from tensorflow.keras.models import Model\n",
    "from tensorflow.keras.applications.mobilenet_v2 import MobileNetV2, preprocess_input\n",
    "from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten, Dense, Dropout, GlobalAveragePooling2D\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
    "import numpy as np\n",
    "from tensorflow.keras.optimizers import RMSprop"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Failed to limit GPU memory\n"
     ]
    }
   ],
   "source": [
    "os.environ['CUDA_DEVICE_ORDER']=\"PCI_BUS_ID\"\n",
    "os.environ['CUDA_VISIBLE_DEVICES']=\"0\"\n",
    "\n",
    "import tensorflow as tf\n",
    "physical_devices= tf.config.list_physical_devices('GPU')\n",
    "try: \n",
    "    tf.config.experimental.set_memory_growth(physical_devices[0], True)\n",
    "except:\n",
    "    print('Failed to limit GPU memory')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 4188 images belonging to 13 classes.\n",
      "Found 1794 images belonging to 13 classes.\n"
     ]
    }
   ],
   "source": [
    "data_dir = 'D:/DatasetMuseumFix/'\n",
    "\n",
    "# Image data generator with rescaling\n",
    "datagen = ImageDataGenerator(rescale=1./255, validation_split=0.3)\n",
    "\n",
    "# Training dataset\n",
    "train_generator = datagen.flow_from_directory(\n",
    "    data_dir,\n",
    "    target_size=(224, 224),\n",
    "    batch_size=32,\n",
    "    class_mode='categorical',  \n",
    "    subset='training' \n",
    ")\n",
    "\n",
    "# Validation dataset\n",
    "validation_generator = datagen.flow_from_directory(\n",
    "    data_dir,\n",
    "    target_size=(224, 224),\n",
    "    batch_size=32,\n",
    "    class_mode='categorical',  \n",
    "    subset='validation'\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:`input_shape` is undefined or non-square, or `rows` is not in [96, 128, 160, 192, 224]. Weights for input shape (224, 224) will be loaded as the default.\n"
     ]
    }
   ],
   "source": [
    "base_model = MobileNetV2(weights='imagenet', include_top=False)\n",
    "x = base_model.output\n",
    "x = GlobalAveragePooling2D()(x)\n",
    "x = Dense(512, activation='relu')(x)\n",
    "x = Dense(256, activation='relu')(x)\n",
    "x = Dense(128, activation='relu')(x)\n",
    "preds = Dense(13, activation='softmax')(x)  \n",
    "\n",
    "model = Model(inputs=base_model.input, outputs=preds)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "__________________________________________________________________________________________________\n",
      " Layer (type)                   Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      " input_1 (InputLayer)           [(None, None, None,  0           []                               \n",
      "                                 3)]                                                              \n",
      "                                                                                                  \n",
      " Conv1 (Conv2D)                 (None, None, None,   864         ['input_1[0][0]']                \n",
      "                                32)                                                               \n",
      "                                                                                                  \n",
      " bn_Conv1 (BatchNormalization)  (None, None, None,   128         ['Conv1[0][0]']                  \n",
      "                                32)                                                               \n",
      "                                                                                                  \n",
      " Conv1_relu (ReLU)              (None, None, None,   0           ['bn_Conv1[0][0]']               \n",
      "                                32)                                                               \n",
      "                                                                                                  \n",
      " expanded_conv_depthwise (Depth  (None, None, None,   288        ['Conv1_relu[0][0]']             \n",
      " wiseConv2D)                    32)                                                               \n",
      "                                                                                                  \n",
      " expanded_conv_depthwise_BN (Ba  (None, None, None,   128        ['expanded_conv_depthwise[0][0]']\n",
      " tchNormalization)              32)                                                               \n",
      "                                                                                                  \n",
      " expanded_conv_depthwise_relu (  (None, None, None,   0          ['expanded_conv_depthwise_BN[0][0\n",
      " ReLU)                          32)                              ]']                              \n",
      "                                                                                                  \n",
      " expanded_conv_project (Conv2D)  (None, None, None,   512        ['expanded_conv_depthwise_relu[0]\n",
      "                                16)                              [0]']                            \n",
      "                                                                                                  \n",
      " expanded_conv_project_BN (Batc  (None, None, None,   64         ['expanded_conv_project[0][0]']  \n",
      " hNormalization)                16)                                                               \n",
      "                                                                                                  \n",
      " block_1_expand (Conv2D)        (None, None, None,   1536        ['expanded_conv_project_BN[0][0]'\n",
      "                                96)                              ]                                \n",
      "                                                                                                  \n",
      " block_1_expand_BN (BatchNormal  (None, None, None,   384        ['block_1_expand[0][0]']         \n",
      " ization)                       96)                                                               \n",
      "                                                                                                  \n",
      " block_1_expand_relu (ReLU)     (None, None, None,   0           ['block_1_expand_BN[0][0]']      \n",
      "                                96)                                                               \n",
      "                                                                                                  \n",
      " block_1_pad (ZeroPadding2D)    (None, None, None,   0           ['block_1_expand_relu[0][0]']    \n",
      "                                96)                                                               \n",
      "                                                                                                  \n",
      " block_1_depthwise (DepthwiseCo  (None, None, None,   864        ['block_1_pad[0][0]']            \n",
      " nv2D)                          96)                                                               \n",
      "                                                                                                  \n",
      " block_1_depthwise_BN (BatchNor  (None, None, None,   384        ['block_1_depthwise[0][0]']      \n",
      " malization)                    96)                                                               \n",
      "                                                                                                  \n",
      " block_1_depthwise_relu (ReLU)  (None, None, None,   0           ['block_1_depthwise_BN[0][0]']   \n",
      "                                96)                                                               \n",
      "                                                                                                  \n",
      " block_1_project (Conv2D)       (None, None, None,   2304        ['block_1_depthwise_relu[0][0]'] \n",
      "                                24)                                                               \n",
      "                                                                                                  \n",
      " block_1_project_BN (BatchNorma  (None, None, None,   96         ['block_1_project[0][0]']        \n",
      " lization)                      24)                                                               \n",
      "                                                                                                  \n",
      " block_2_expand (Conv2D)        (None, None, None,   3456        ['block_1_project_BN[0][0]']     \n",
      "                                144)                                                              \n",
      "                                                                                                  \n",
      " block_2_expand_BN (BatchNormal  (None, None, None,   576        ['block_2_expand[0][0]']         \n",
      " ization)                       144)                                                              \n",
      "                                                                                                  \n",
      " block_2_expand_relu (ReLU)     (None, None, None,   0           ['block_2_expand_BN[0][0]']      \n",
      "                                144)                                                              \n",
      "                                                                                                  \n",
      " block_2_depthwise (DepthwiseCo  (None, None, None,   1296       ['block_2_expand_relu[0][0]']    \n",
      " nv2D)                          144)                                                              \n",
      "                                                                                                  \n",
      " block_2_depthwise_BN (BatchNor  (None, None, None,   576        ['block_2_depthwise[0][0]']      \n",
      " malization)                    144)                                                              \n",
      "                                                                                                  \n",
      " block_2_depthwise_relu (ReLU)  (None, None, None,   0           ['block_2_depthwise_BN[0][0]']   \n",
      "                                144)                                                              \n",
      "                                                                                                  \n",
      " block_2_project (Conv2D)       (None, None, None,   3456        ['block_2_depthwise_relu[0][0]'] \n",
      "                                24)                                                               \n",
      "                                                                                                  \n",
      " block_2_project_BN (BatchNorma  (None, None, None,   96         ['block_2_project[0][0]']        \n",
      " lization)                      24)                                                               \n",
      "                                                                                                  \n",
      " block_2_add (Add)              (None, None, None,   0           ['block_1_project_BN[0][0]',     \n",
      "                                24)                               'block_2_project_BN[0][0]']     \n",
      "                                                                                                  \n",
      " block_3_expand (Conv2D)        (None, None, None,   3456        ['block_2_add[0][0]']            \n",
      "                                144)                                                              \n",
      "                                                                                                  \n",
      " block_3_expand_BN (BatchNormal  (None, None, None,   576        ['block_3_expand[0][0]']         \n",
      " ization)                       144)                                                              \n",
      "                                                                                                  \n",
      " block_3_expand_relu (ReLU)     (None, None, None,   0           ['block_3_expand_BN[0][0]']      \n",
      "                                144)                                                              \n",
      "                                                                                                  \n",
      " block_3_pad (ZeroPadding2D)    (None, None, None,   0           ['block_3_expand_relu[0][0]']    \n",
      "                                144)                                                              \n",
      "                                                                                                  \n",
      " block_3_depthwise (DepthwiseCo  (None, None, None,   1296       ['block_3_pad[0][0]']            \n",
      " nv2D)                          144)                                                              \n",
      "                                                                                                  \n",
      " block_3_depthwise_BN (BatchNor  (None, None, None,   576        ['block_3_depthwise[0][0]']      \n",
      " malization)                    144)                                                              \n",
      "                                                                                                  \n",
      " block_3_depthwise_relu (ReLU)  (None, None, None,   0           ['block_3_depthwise_BN[0][0]']   \n",
      "                                144)                                                              \n",
      "                                                                                                  \n",
      " block_3_project (Conv2D)       (None, None, None,   4608        ['block_3_depthwise_relu[0][0]'] \n",
      "                                32)                                                               \n",
      "                                                                                                  \n",
      " block_3_project_BN (BatchNorma  (None, None, None,   128        ['block_3_project[0][0]']        \n",
      " lization)                      32)                                                               \n",
      "                                                                                                  \n",
      " block_4_expand (Conv2D)        (None, None, None,   6144        ['block_3_project_BN[0][0]']     \n",
      "                                192)                                                              \n",
      "                                                                                                  \n",
      " block_4_expand_BN (BatchNormal  (None, None, None,   768        ['block_4_expand[0][0]']         \n",
      " ization)                       192)                                                              \n",
      "                                                                                                  \n",
      " block_4_expand_relu (ReLU)     (None, None, None,   0           ['block_4_expand_BN[0][0]']      \n",
      "                                192)                                                              \n",
      "                                                                                                  \n",
      " block_4_depthwise (DepthwiseCo  (None, None, None,   1728       ['block_4_expand_relu[0][0]']    \n",
      " nv2D)                          192)                                                              \n",
      "                                                                                                  \n",
      " block_4_depthwise_BN (BatchNor  (None, None, None,   768        ['block_4_depthwise[0][0]']      \n",
      " malization)                    192)                                                              \n",
      "                                                                                                  \n",
      " block_4_depthwise_relu (ReLU)  (None, None, None,   0           ['block_4_depthwise_BN[0][0]']   \n",
      "                                192)                                                              \n",
      "                                                                                                  \n",
      " block_4_project (Conv2D)       (None, None, None,   6144        ['block_4_depthwise_relu[0][0]'] \n",
      "                                32)                                                               \n",
      "                                                                                                  \n",
      " block_4_project_BN (BatchNorma  (None, None, None,   128        ['block_4_project[0][0]']        \n",
      " lization)                      32)                                                               \n",
      "                                                                                                  \n",
      " block_4_add (Add)              (None, None, None,   0           ['block_3_project_BN[0][0]',     \n",
      "                                32)                               'block_4_project_BN[0][0]']     \n",
      "                                                                                                  \n",
      " block_5_expand (Conv2D)        (None, None, None,   6144        ['block_4_add[0][0]']            \n",
      "                                192)                                                              \n",
      "                                                                                                  \n",
      " block_5_expand_BN (BatchNormal  (None, None, None,   768        ['block_5_expand[0][0]']         \n",
      " ization)                       192)                                                              \n",
      "                                                                                                  \n",
      " block_5_expand_relu (ReLU)     (None, None, None,   0           ['block_5_expand_BN[0][0]']      \n",
      "                                192)                                                              \n",
      "                                                                                                  \n",
      " block_5_depthwise (DepthwiseCo  (None, None, None,   1728       ['block_5_expand_relu[0][0]']    \n",
      " nv2D)                          192)                                                              \n",
      "                                                                                                  \n",
      " block_5_depthwise_BN (BatchNor  (None, None, None,   768        ['block_5_depthwise[0][0]']      \n",
      " malization)                    192)                                                              \n",
      "                                                                                                  \n",
      " block_5_depthwise_relu (ReLU)  (None, None, None,   0           ['block_5_depthwise_BN[0][0]']   \n",
      "                                192)                                                              \n",
      "                                                                                                  \n",
      " block_5_project (Conv2D)       (None, None, None,   6144        ['block_5_depthwise_relu[0][0]'] \n",
      "                                32)                                                               \n",
      "                                                                                                  \n",
      " block_5_project_BN (BatchNorma  (None, None, None,   128        ['block_5_project[0][0]']        \n",
      " lization)                      32)                                                               \n",
      "                                                                                                  \n",
      " block_5_add (Add)              (None, None, None,   0           ['block_4_add[0][0]',            \n",
      "                                32)                               'block_5_project_BN[0][0]']     \n",
      "                                                                                                  \n",
      " block_6_expand (Conv2D)        (None, None, None,   6144        ['block_5_add[0][0]']            \n",
      "                                192)                                                              \n",
      "                                                                                                  \n",
      " block_6_expand_BN (BatchNormal  (None, None, None,   768        ['block_6_expand[0][0]']         \n",
      " ization)                       192)                                                              \n",
      "                                                                                                  \n",
      " block_6_expand_relu (ReLU)     (None, None, None,   0           ['block_6_expand_BN[0][0]']      \n",
      "                                192)                                                              \n",
      "                                                                                                  \n",
      " block_6_pad (ZeroPadding2D)    (None, None, None,   0           ['block_6_expand_relu[0][0]']    \n",
      "                                192)                                                              \n",
      "                                                                                                  \n",
      " block_6_depthwise (DepthwiseCo  (None, None, None,   1728       ['block_6_pad[0][0]']            \n",
      " nv2D)                          192)                                                              \n",
      "                                                                                                  \n",
      " block_6_depthwise_BN (BatchNor  (None, None, None,   768        ['block_6_depthwise[0][0]']      \n",
      " malization)                    192)                                                              \n",
      "                                                                                                  \n",
      " block_6_depthwise_relu (ReLU)  (None, None, None,   0           ['block_6_depthwise_BN[0][0]']   \n",
      "                                192)                                                              \n",
      "                                                                                                  \n",
      " block_6_project (Conv2D)       (None, None, None,   12288       ['block_6_depthwise_relu[0][0]'] \n",
      "                                64)                                                               \n",
      "                                                                                                  \n",
      " block_6_project_BN (BatchNorma  (None, None, None,   256        ['block_6_project[0][0]']        \n",
      " lization)                      64)                                                               \n",
      "                                                                                                  \n",
      " block_7_expand (Conv2D)        (None, None, None,   24576       ['block_6_project_BN[0][0]']     \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_7_expand_BN (BatchNormal  (None, None, None,   1536       ['block_7_expand[0][0]']         \n",
      " ization)                       384)                                                              \n",
      "                                                                                                  \n",
      " block_7_expand_relu (ReLU)     (None, None, None,   0           ['block_7_expand_BN[0][0]']      \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_7_depthwise (DepthwiseCo  (None, None, None,   3456       ['block_7_expand_relu[0][0]']    \n",
      " nv2D)                          384)                                                              \n",
      "                                                                                                  \n",
      " block_7_depthwise_BN (BatchNor  (None, None, None,   1536       ['block_7_depthwise[0][0]']      \n",
      " malization)                    384)                                                              \n",
      "                                                                                                  \n",
      " block_7_depthwise_relu (ReLU)  (None, None, None,   0           ['block_7_depthwise_BN[0][0]']   \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_7_project (Conv2D)       (None, None, None,   24576       ['block_7_depthwise_relu[0][0]'] \n",
      "                                64)                                                               \n",
      "                                                                                                  \n",
      " block_7_project_BN (BatchNorma  (None, None, None,   256        ['block_7_project[0][0]']        \n",
      " lization)                      64)                                                               \n",
      "                                                                                                  \n",
      " block_7_add (Add)              (None, None, None,   0           ['block_6_project_BN[0][0]',     \n",
      "                                64)                               'block_7_project_BN[0][0]']     \n",
      "                                                                                                  \n",
      " block_8_expand (Conv2D)        (None, None, None,   24576       ['block_7_add[0][0]']            \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_8_expand_BN (BatchNormal  (None, None, None,   1536       ['block_8_expand[0][0]']         \n",
      " ization)                       384)                                                              \n",
      "                                                                                                  \n",
      " block_8_expand_relu (ReLU)     (None, None, None,   0           ['block_8_expand_BN[0][0]']      \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_8_depthwise (DepthwiseCo  (None, None, None,   3456       ['block_8_expand_relu[0][0]']    \n",
      " nv2D)                          384)                                                              \n",
      "                                                                                                  \n",
      " block_8_depthwise_BN (BatchNor  (None, None, None,   1536       ['block_8_depthwise[0][0]']      \n",
      " malization)                    384)                                                              \n",
      "                                                                                                  \n",
      " block_8_depthwise_relu (ReLU)  (None, None, None,   0           ['block_8_depthwise_BN[0][0]']   \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_8_project (Conv2D)       (None, None, None,   24576       ['block_8_depthwise_relu[0][0]'] \n",
      "                                64)                                                               \n",
      "                                                                                                  \n",
      " block_8_project_BN (BatchNorma  (None, None, None,   256        ['block_8_project[0][0]']        \n",
      " lization)                      64)                                                               \n",
      "                                                                                                  \n",
      " block_8_add (Add)              (None, None, None,   0           ['block_7_add[0][0]',            \n",
      "                                64)                               'block_8_project_BN[0][0]']     \n",
      "                                                                                                  \n",
      " block_9_expand (Conv2D)        (None, None, None,   24576       ['block_8_add[0][0]']            \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_9_expand_BN (BatchNormal  (None, None, None,   1536       ['block_9_expand[0][0]']         \n",
      " ization)                       384)                                                              \n",
      "                                                                                                  \n",
      " block_9_expand_relu (ReLU)     (None, None, None,   0           ['block_9_expand_BN[0][0]']      \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_9_depthwise (DepthwiseCo  (None, None, None,   3456       ['block_9_expand_relu[0][0]']    \n",
      " nv2D)                          384)                                                              \n",
      "                                                                                                  \n",
      " block_9_depthwise_BN (BatchNor  (None, None, None,   1536       ['block_9_depthwise[0][0]']      \n",
      " malization)                    384)                                                              \n",
      "                                                                                                  \n",
      " block_9_depthwise_relu (ReLU)  (None, None, None,   0           ['block_9_depthwise_BN[0][0]']   \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_9_project (Conv2D)       (None, None, None,   24576       ['block_9_depthwise_relu[0][0]'] \n",
      "                                64)                                                               \n",
      "                                                                                                  \n",
      " block_9_project_BN (BatchNorma  (None, None, None,   256        ['block_9_project[0][0]']        \n",
      " lization)                      64)                                                               \n",
      "                                                                                                  \n",
      " block_9_add (Add)              (None, None, None,   0           ['block_8_add[0][0]',            \n",
      "                                64)                               'block_9_project_BN[0][0]']     \n",
      "                                                                                                  \n",
      " block_10_expand (Conv2D)       (None, None, None,   24576       ['block_9_add[0][0]']            \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_10_expand_BN (BatchNorma  (None, None, None,   1536       ['block_10_expand[0][0]']        \n",
      " lization)                      384)                                                              \n",
      "                                                                                                  \n",
      " block_10_expand_relu (ReLU)    (None, None, None,   0           ['block_10_expand_BN[0][0]']     \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_10_depthwise (DepthwiseC  (None, None, None,   3456       ['block_10_expand_relu[0][0]']   \n",
      " onv2D)                         384)                                                              \n",
      "                                                                                                  \n",
      " block_10_depthwise_BN (BatchNo  (None, None, None,   1536       ['block_10_depthwise[0][0]']     \n",
      " rmalization)                   384)                                                              \n",
      "                                                                                                  \n",
      " block_10_depthwise_relu (ReLU)  (None, None, None,   0          ['block_10_depthwise_BN[0][0]']  \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_10_project (Conv2D)      (None, None, None,   36864       ['block_10_depthwise_relu[0][0]']\n",
      "                                96)                                                               \n",
      "                                                                                                  \n",
      " block_10_project_BN (BatchNorm  (None, None, None,   384        ['block_10_project[0][0]']       \n",
      " alization)                     96)                                                               \n",
      "                                                                                                  \n",
      " block_11_expand (Conv2D)       (None, None, None,   55296       ['block_10_project_BN[0][0]']    \n",
      "                                576)                                                              \n",
      "                                                                                                  \n",
      " block_11_expand_BN (BatchNorma  (None, None, None,   2304       ['block_11_expand[0][0]']        \n",
      " lization)                      576)                                                              \n",
      "                                                                                                  \n",
      " block_11_expand_relu (ReLU)    (None, None, None,   0           ['block_11_expand_BN[0][0]']     \n",
      "                                576)                                                              \n",
      "                                                                                                  \n",
      " block_11_depthwise (DepthwiseC  (None, None, None,   5184       ['block_11_expand_relu[0][0]']   \n",
      " onv2D)                         576)                                                              \n",
      "                                                                                                  \n",
      " block_11_depthwise_BN (BatchNo  (None, None, None,   2304       ['block_11_depthwise[0][0]']     \n",
      " rmalization)                   576)                                                              \n",
      "                                                                                                  \n",
      " block_11_depthwise_relu (ReLU)  (None, None, None,   0          ['block_11_depthwise_BN[0][0]']  \n",
      "                                576)                                                              \n",
      "                                                                                                  \n",
      " block_11_project (Conv2D)      (None, None, None,   55296       ['block_11_depthwise_relu[0][0]']\n",
      "                                96)                                                               \n",
      "                                                                                                  \n",
      " block_11_project_BN (BatchNorm  (None, None, None,   384        ['block_11_project[0][0]']       \n",
      " alization)                     96)                                                               \n",
      "                                                                                                  \n",
      " block_11_add (Add)             (None, None, None,   0           ['block_10_project_BN[0][0]',    \n",
      "                                96)                               'block_11_project_BN[0][0]']    \n",
      "                                                                                                  \n",
      " block_12_expand (Conv2D)       (None, None, None,   55296       ['block_11_add[0][0]']           \n",
      "                                576)                                                              \n",
      "                                                                                                  \n",
      " block_12_expand_BN (BatchNorma  (None, None, None,   2304       ['block_12_expand[0][0]']        \n",
      " lization)                      576)                                                              \n",
      "                                                                                                  \n",
      " block_12_expand_relu (ReLU)    (None, None, None,   0           ['block_12_expand_BN[0][0]']     \n",
      "                                576)                                                              \n",
      "                                                                                                  \n",
      " block_12_depthwise (DepthwiseC  (None, None, None,   5184       ['block_12_expand_relu[0][0]']   \n",
      " onv2D)                         576)                                                              \n",
      "                                                                                                  \n",
      " block_12_depthwise_BN (BatchNo  (None, None, None,   2304       ['block_12_depthwise[0][0]']     \n",
      " rmalization)                   576)                                                              \n",
      "                                                                                                  \n",
      " block_12_depthwise_relu (ReLU)  (None, None, None,   0          ['block_12_depthwise_BN[0][0]']  \n",
      "                                576)                                                              \n",
      "                                                                                                  \n",
      " block_12_project (Conv2D)      (None, None, None,   55296       ['block_12_depthwise_relu[0][0]']\n",
      "                                96)                                                               \n",
      "                                                                                                  \n",
      " block_12_project_BN (BatchNorm  (None, None, None,   384        ['block_12_project[0][0]']       \n",
      " alization)                     96)                                                               \n",
      "                                                                                                  \n",
      " block_12_add (Add)             (None, None, None,   0           ['block_11_add[0][0]',           \n",
      "                                96)                               'block_12_project_BN[0][0]']    \n",
      "                                                                                                  \n",
      " block_13_expand (Conv2D)       (None, None, None,   55296       ['block_12_add[0][0]']           \n",
      "                                576)                                                              \n",
      "                                                                                                  \n",
      " block_13_expand_BN (BatchNorma  (None, None, None,   2304       ['block_13_expand[0][0]']        \n",
      " lization)                      576)                                                              \n",
      "                                                                                                  \n",
      " block_13_expand_relu (ReLU)    (None, None, None,   0           ['block_13_expand_BN[0][0]']     \n",
      "                                576)                                                              \n",
      "                                                                                                  \n",
      " block_13_pad (ZeroPadding2D)   (None, None, None,   0           ['block_13_expand_relu[0][0]']   \n",
      "                                576)                                                              \n",
      "                                                                                                  \n",
      " block_13_depthwise (DepthwiseC  (None, None, None,   5184       ['block_13_pad[0][0]']           \n",
      " onv2D)                         576)                                                              \n",
      "                                                                                                  \n",
      " block_13_depthwise_BN (BatchNo  (None, None, None,   2304       ['block_13_depthwise[0][0]']     \n",
      " rmalization)                   576)                                                              \n",
      "                                                                                                  \n",
      " block_13_depthwise_relu (ReLU)  (None, None, None,   0          ['block_13_depthwise_BN[0][0]']  \n",
      "                                576)                                                              \n",
      "                                                                                                  \n",
      " block_13_project (Conv2D)      (None, None, None,   92160       ['block_13_depthwise_relu[0][0]']\n",
      "                                160)                                                              \n",
      "                                                                                                  \n",
      " block_13_project_BN (BatchNorm  (None, None, None,   640        ['block_13_project[0][0]']       \n",
      " alization)                     160)                                                              \n",
      "                                                                                                  \n",
      " block_14_expand (Conv2D)       (None, None, None,   153600      ['block_13_project_BN[0][0]']    \n",
      "                                960)                                                              \n",
      "                                                                                                  \n",
      " block_14_expand_BN (BatchNorma  (None, None, None,   3840       ['block_14_expand[0][0]']        \n",
      " lization)                      960)                                                              \n",
      "                                                                                                  \n",
      " block_14_expand_relu (ReLU)    (None, None, None,   0           ['block_14_expand_BN[0][0]']     \n",
      "                                960)                                                              \n",
      "                                                                                                  \n",
      " block_14_depthwise (DepthwiseC  (None, None, None,   8640       ['block_14_expand_relu[0][0]']   \n",
      " onv2D)                         960)                                                              \n",
      "                                                                                                  \n",
      " block_14_depthwise_BN (BatchNo  (None, None, None,   3840       ['block_14_depthwise[0][0]']     \n",
      " rmalization)                   960)                                                              \n",
      "                                                                                                  \n",
      " block_14_depthwise_relu (ReLU)  (None, None, None,   0          ['block_14_depthwise_BN[0][0]']  \n",
      "                                960)                                                              \n",
      "                                                                                                  \n",
      " block_14_project (Conv2D)      (None, None, None,   153600      ['block_14_depthwise_relu[0][0]']\n",
      "                                160)                                                              \n",
      "                                                                                                  \n",
      " block_14_project_BN (BatchNorm  (None, None, None,   640        ['block_14_project[0][0]']       \n",
      " alization)                     160)                                                              \n",
      "                                                                                                  \n",
      " block_14_add (Add)             (None, None, None,   0           ['block_13_project_BN[0][0]',    \n",
      "                                160)                              'block_14_project_BN[0][0]']    \n",
      "                                                                                                  \n",
      " block_15_expand (Conv2D)       (None, None, None,   153600      ['block_14_add[0][0]']           \n",
      "                                960)                                                              \n",
      "                                                                                                  \n",
      " block_15_expand_BN (BatchNorma  (None, None, None,   3840       ['block_15_expand[0][0]']        \n",
      " lization)                      960)                                                              \n",
      "                                                                                                  \n",
      " block_15_expand_relu (ReLU)    (None, None, None,   0           ['block_15_expand_BN[0][0]']     \n",
      "                                960)                                                              \n",
      "                                                                                                  \n",
      " block_15_depthwise (DepthwiseC  (None, None, None,   8640       ['block_15_expand_relu[0][0]']   \n",
      " onv2D)                         960)                                                              \n",
      "                                                                                                  \n",
      " block_15_depthwise_BN (BatchNo  (None, None, None,   3840       ['block_15_depthwise[0][0]']     \n",
      " rmalization)                   960)                                                              \n",
      "                                                                                                  \n",
      " block_15_depthwise_relu (ReLU)  (None, None, None,   0          ['block_15_depthwise_BN[0][0]']  \n",
      "                                960)                                                              \n",
      "                                                                                                  \n",
      " block_15_project (Conv2D)      (None, None, None,   153600      ['block_15_depthwise_relu[0][0]']\n",
      "                                160)                                                              \n",
      "                                                                                                  \n",
      " block_15_project_BN (BatchNorm  (None, None, None,   640        ['block_15_project[0][0]']       \n",
      " alization)                     160)                                                              \n",
      "                                                                                                  \n",
      " block_15_add (Add)             (None, None, None,   0           ['block_14_add[0][0]',           \n",
      "                                160)                              'block_15_project_BN[0][0]']    \n",
      "                                                                                                  \n",
      " block_16_expand (Conv2D)       (None, None, None,   153600      ['block_15_add[0][0]']           \n",
      "                                960)                                                              \n",
      "                                                                                                  \n",
      " block_16_expand_BN (BatchNorma  (None, None, None,   3840       ['block_16_expand[0][0]']        \n",
      " lization)                      960)                                                              \n",
      "                                                                                                  \n",
      " block_16_expand_relu (ReLU)    (None, None, None,   0           ['block_16_expand_BN[0][0]']     \n",
      "                                960)                                                              \n",
      "                                                                                                  \n",
      " block_16_depthwise (DepthwiseC  (None, None, None,   8640       ['block_16_expand_relu[0][0]']   \n",
      " onv2D)                         960)                                                              \n",
      "                                                                                                  \n",
      " block_16_depthwise_BN (BatchNo  (None, None, None,   3840       ['block_16_depthwise[0][0]']     \n",
      " rmalization)                   960)                                                              \n",
      "                                                                                                  \n",
      " block_16_depthwise_relu (ReLU)  (None, None, None,   0          ['block_16_depthwise_BN[0][0]']  \n",
      "                                960)                                                              \n",
      "                                                                                                  \n",
      " block_16_project (Conv2D)      (None, None, None,   307200      ['block_16_depthwise_relu[0][0]']\n",
      "                                320)                                                              \n",
      "                                                                                                  \n",
      " block_16_project_BN (BatchNorm  (None, None, None,   1280       ['block_16_project[0][0]']       \n",
      " alization)                     320)                                                              \n",
      "                                                                                                  \n",
      " Conv_1 (Conv2D)                (None, None, None,   409600      ['block_16_project_BN[0][0]']    \n",
      "                                1280)                                                             \n",
      "                                                                                                  \n",
      " Conv_1_bn (BatchNormalization)  (None, None, None,   5120       ['Conv_1[0][0]']                 \n",
      "                                1280)                                                             \n",
      "                                                                                                  \n",
      " out_relu (ReLU)                (None, None, None,   0           ['Conv_1_bn[0][0]']              \n",
      "                                1280)                                                             \n",
      "                                                                                                  \n",
      " global_average_pooling2d (Glob  (None, 1280)        0           ['out_relu[0][0]']               \n",
      " alAveragePooling2D)                                                                              \n",
      "                                                                                                  \n",
      " dense (Dense)                  (None, 512)          655872      ['global_average_pooling2d[0][0]'\n",
      "                                                                 ]                                \n",
      "                                                                                                  \n",
      " dense_1 (Dense)                (None, 256)          131328      ['dense[0][0]']                  \n",
      "                                                                                                  \n",
      " dense_2 (Dense)                (None, 128)          32896       ['dense_1[0][0]']                \n",
      "                                                                                                  \n",
      " dense_3 (Dense)                (None, 13)           1677        ['dense_2[0][0]']                \n",
      "                                                                                                  \n",
      "==================================================================================================\n",
      "Total params: 3,079,757\n",
      "Trainable params: 3,045,645\n",
      "Non-trainable params: 34,112\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "for layers in model.layers[:-5]:\n",
    "    layers.trainable = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_3\"\n",
      "__________________________________________________________________________________________________\n",
      " Layer (type)                   Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      " input_4 (InputLayer)           [(None, None, None,  0           []                               \n",
      "                                 3)]                                                              \n",
      "                                                                                                  \n",
      " Conv1 (Conv2D)                 (None, None, None,   864         ['input_4[0][0]']                \n",
      "                                32)                                                               \n",
      "                                                                                                  \n",
      " bn_Conv1 (BatchNormalization)  (None, None, None,   128         ['Conv1[0][0]']                  \n",
      "                                32)                                                               \n",
      "                                                                                                  \n",
      " Conv1_relu (ReLU)              (None, None, None,   0           ['bn_Conv1[0][0]']               \n",
      "                                32)                                                               \n",
      "                                                                                                  \n",
      " expanded_conv_depthwise (Depth  (None, None, None,   288        ['Conv1_relu[0][0]']             \n",
      " wiseConv2D)                    32)                                                               \n",
      "                                                                                                  \n",
      " expanded_conv_depthwise_BN (Ba  (None, None, None,   128        ['expanded_conv_depthwise[0][0]']\n",
      " tchNormalization)              32)                                                               \n",
      "                                                                                                  \n",
      " expanded_conv_depthwise_relu (  (None, None, None,   0          ['expanded_conv_depthwise_BN[0][0\n",
      " ReLU)                          32)                              ]']                              \n",
      "                                                                                                  \n",
      " expanded_conv_project (Conv2D)  (None, None, None,   512        ['expanded_conv_depthwise_relu[0]\n",
      "                                16)                              [0]']                            \n",
      "                                                                                                  \n",
      " expanded_conv_project_BN (Batc  (None, None, None,   64         ['expanded_conv_project[0][0]']  \n",
      " hNormalization)                16)                                                               \n",
      "                                                                                                  \n",
      " block_1_expand (Conv2D)        (None, None, None,   1536        ['expanded_conv_project_BN[0][0]'\n",
      "                                96)                              ]                                \n",
      "                                                                                                  \n",
      " block_1_expand_BN (BatchNormal  (None, None, None,   384        ['block_1_expand[0][0]']         \n",
      " ization)                       96)                                                               \n",
      "                                                                                                  \n",
      " block_1_expand_relu (ReLU)     (None, None, None,   0           ['block_1_expand_BN[0][0]']      \n",
      "                                96)                                                               \n",
      "                                                                                                  \n",
      " block_1_pad (ZeroPadding2D)    (None, None, None,   0           ['block_1_expand_relu[0][0]']    \n",
      "                                96)                                                               \n",
      "                                                                                                  \n",
      " block_1_depthwise (DepthwiseCo  (None, None, None,   864        ['block_1_pad[0][0]']            \n",
      " nv2D)                          96)                                                               \n",
      "                                                                                                  \n",
      " block_1_depthwise_BN (BatchNor  (None, None, None,   384        ['block_1_depthwise[0][0]']      \n",
      " malization)                    96)                                                               \n",
      "                                                                                                  \n",
      " block_1_depthwise_relu (ReLU)  (None, None, None,   0           ['block_1_depthwise_BN[0][0]']   \n",
      "                                96)                                                               \n",
      "                                                                                                  \n",
      " block_1_project (Conv2D)       (None, None, None,   2304        ['block_1_depthwise_relu[0][0]'] \n",
      "                                24)                                                               \n",
      "                                                                                                  \n",
      " block_1_project_BN (BatchNorma  (None, None, None,   96         ['block_1_project[0][0]']        \n",
      " lization)                      24)                                                               \n",
      "                                                                                                  \n",
      " block_2_expand (Conv2D)        (None, None, None,   3456        ['block_1_project_BN[0][0]']     \n",
      "                                144)                                                              \n",
      "                                                                                                  \n",
      " block_2_expand_BN (BatchNormal  (None, None, None,   576        ['block_2_expand[0][0]']         \n",
      " ization)                       144)                                                              \n",
      "                                                                                                  \n",
      " block_2_expand_relu (ReLU)     (None, None, None,   0           ['block_2_expand_BN[0][0]']      \n",
      "                                144)                                                              \n",
      "                                                                                                  \n",
      " block_2_depthwise (DepthwiseCo  (None, None, None,   1296       ['block_2_expand_relu[0][0]']    \n",
      " nv2D)                          144)                                                              \n",
      "                                                                                                  \n",
      " block_2_depthwise_BN (BatchNor  (None, None, None,   576        ['block_2_depthwise[0][0]']      \n",
      " malization)                    144)                                                              \n",
      "                                                                                                  \n",
      " block_2_depthwise_relu (ReLU)  (None, None, None,   0           ['block_2_depthwise_BN[0][0]']   \n",
      "                                144)                                                              \n",
      "                                                                                                  \n",
      " block_2_project (Conv2D)       (None, None, None,   3456        ['block_2_depthwise_relu[0][0]'] \n",
      "                                24)                                                               \n",
      "                                                                                                  \n",
      " block_2_project_BN (BatchNorma  (None, None, None,   96         ['block_2_project[0][0]']        \n",
      " lization)                      24)                                                               \n",
      "                                                                                                  \n",
      " block_2_add (Add)              (None, None, None,   0           ['block_1_project_BN[0][0]',     \n",
      "                                24)                               'block_2_project_BN[0][0]']     \n",
      "                                                                                                  \n",
      " block_3_expand (Conv2D)        (None, None, None,   3456        ['block_2_add[0][0]']            \n",
      "                                144)                                                              \n",
      "                                                                                                  \n",
      " block_3_expand_BN (BatchNormal  (None, None, None,   576        ['block_3_expand[0][0]']         \n",
      " ization)                       144)                                                              \n",
      "                                                                                                  \n",
      " block_3_expand_relu (ReLU)     (None, None, None,   0           ['block_3_expand_BN[0][0]']      \n",
      "                                144)                                                              \n",
      "                                                                                                  \n",
      " block_3_pad (ZeroPadding2D)    (None, None, None,   0           ['block_3_expand_relu[0][0]']    \n",
      "                                144)                                                              \n",
      "                                                                                                  \n",
      " block_3_depthwise (DepthwiseCo  (None, None, None,   1296       ['block_3_pad[0][0]']            \n",
      " nv2D)                          144)                                                              \n",
      "                                                                                                  \n",
      " block_3_depthwise_BN (BatchNor  (None, None, None,   576        ['block_3_depthwise[0][0]']      \n",
      " malization)                    144)                                                              \n",
      "                                                                                                  \n",
      " block_3_depthwise_relu (ReLU)  (None, None, None,   0           ['block_3_depthwise_BN[0][0]']   \n",
      "                                144)                                                              \n",
      "                                                                                                  \n",
      " block_3_project (Conv2D)       (None, None, None,   4608        ['block_3_depthwise_relu[0][0]'] \n",
      "                                32)                                                               \n",
      "                                                                                                  \n",
      " block_3_project_BN (BatchNorma  (None, None, None,   128        ['block_3_project[0][0]']        \n",
      " lization)                      32)                                                               \n",
      "                                                                                                  \n",
      " block_4_expand (Conv2D)        (None, None, None,   6144        ['block_3_project_BN[0][0]']     \n",
      "                                192)                                                              \n",
      "                                                                                                  \n",
      " block_4_expand_BN (BatchNormal  (None, None, None,   768        ['block_4_expand[0][0]']         \n",
      " ization)                       192)                                                              \n",
      "                                                                                                  \n",
      " block_4_expand_relu (ReLU)     (None, None, None,   0           ['block_4_expand_BN[0][0]']      \n",
      "                                192)                                                              \n",
      "                                                                                                  \n",
      " block_4_depthwise (DepthwiseCo  (None, None, None,   1728       ['block_4_expand_relu[0][0]']    \n",
      " nv2D)                          192)                                                              \n",
      "                                                                                                  \n",
      " block_4_depthwise_BN (BatchNor  (None, None, None,   768        ['block_4_depthwise[0][0]']      \n",
      " malization)                    192)                                                              \n",
      "                                                                                                  \n",
      " block_4_depthwise_relu (ReLU)  (None, None, None,   0           ['block_4_depthwise_BN[0][0]']   \n",
      "                                192)                                                              \n",
      "                                                                                                  \n",
      " block_4_project (Conv2D)       (None, None, None,   6144        ['block_4_depthwise_relu[0][0]'] \n",
      "                                32)                                                               \n",
      "                                                                                                  \n",
      " block_4_project_BN (BatchNorma  (None, None, None,   128        ['block_4_project[0][0]']        \n",
      " lization)                      32)                                                               \n",
      "                                                                                                  \n",
      " block_4_add (Add)              (None, None, None,   0           ['block_3_project_BN[0][0]',     \n",
      "                                32)                               'block_4_project_BN[0][0]']     \n",
      "                                                                                                  \n",
      " block_5_expand (Conv2D)        (None, None, None,   6144        ['block_4_add[0][0]']            \n",
      "                                192)                                                              \n",
      "                                                                                                  \n",
      " block_5_expand_BN (BatchNormal  (None, None, None,   768        ['block_5_expand[0][0]']         \n",
      " ization)                       192)                                                              \n",
      "                                                                                                  \n",
      " block_5_expand_relu (ReLU)     (None, None, None,   0           ['block_5_expand_BN[0][0]']      \n",
      "                                192)                                                              \n",
      "                                                                                                  \n",
      " block_5_depthwise (DepthwiseCo  (None, None, None,   1728       ['block_5_expand_relu[0][0]']    \n",
      " nv2D)                          192)                                                              \n",
      "                                                                                                  \n",
      " block_5_depthwise_BN (BatchNor  (None, None, None,   768        ['block_5_depthwise[0][0]']      \n",
      " malization)                    192)                                                              \n",
      "                                                                                                  \n",
      " block_5_depthwise_relu (ReLU)  (None, None, None,   0           ['block_5_depthwise_BN[0][0]']   \n",
      "                                192)                                                              \n",
      "                                                                                                  \n",
      " block_5_project (Conv2D)       (None, None, None,   6144        ['block_5_depthwise_relu[0][0]'] \n",
      "                                32)                                                               \n",
      "                                                                                                  \n",
      " block_5_project_BN (BatchNorma  (None, None, None,   128        ['block_5_project[0][0]']        \n",
      " lization)                      32)                                                               \n",
      "                                                                                                  \n",
      " block_5_add (Add)              (None, None, None,   0           ['block_4_add[0][0]',            \n",
      "                                32)                               'block_5_project_BN[0][0]']     \n",
      "                                                                                                  \n",
      " block_6_expand (Conv2D)        (None, None, None,   6144        ['block_5_add[0][0]']            \n",
      "                                192)                                                              \n",
      "                                                                                                  \n",
      " block_6_expand_BN (BatchNormal  (None, None, None,   768        ['block_6_expand[0][0]']         \n",
      " ization)                       192)                                                              \n",
      "                                                                                                  \n",
      " block_6_expand_relu (ReLU)     (None, None, None,   0           ['block_6_expand_BN[0][0]']      \n",
      "                                192)                                                              \n",
      "                                                                                                  \n",
      " block_6_pad (ZeroPadding2D)    (None, None, None,   0           ['block_6_expand_relu[0][0]']    \n",
      "                                192)                                                              \n",
      "                                                                                                  \n",
      " block_6_depthwise (DepthwiseCo  (None, None, None,   1728       ['block_6_pad[0][0]']            \n",
      " nv2D)                          192)                                                              \n",
      "                                                                                                  \n",
      " block_6_depthwise_BN (BatchNor  (None, None, None,   768        ['block_6_depthwise[0][0]']      \n",
      " malization)                    192)                                                              \n",
      "                                                                                                  \n",
      " block_6_depthwise_relu (ReLU)  (None, None, None,   0           ['block_6_depthwise_BN[0][0]']   \n",
      "                                192)                                                              \n",
      "                                                                                                  \n",
      " block_6_project (Conv2D)       (None, None, None,   12288       ['block_6_depthwise_relu[0][0]'] \n",
      "                                64)                                                               \n",
      "                                                                                                  \n",
      " block_6_project_BN (BatchNorma  (None, None, None,   256        ['block_6_project[0][0]']        \n",
      " lization)                      64)                                                               \n",
      "                                                                                                  \n",
      " block_7_expand (Conv2D)        (None, None, None,   24576       ['block_6_project_BN[0][0]']     \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_7_expand_BN (BatchNormal  (None, None, None,   1536       ['block_7_expand[0][0]']         \n",
      " ization)                       384)                                                              \n",
      "                                                                                                  \n",
      " block_7_expand_relu (ReLU)     (None, None, None,   0           ['block_7_expand_BN[0][0]']      \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_7_depthwise (DepthwiseCo  (None, None, None,   3456       ['block_7_expand_relu[0][0]']    \n",
      " nv2D)                          384)                                                              \n",
      "                                                                                                  \n",
      " block_7_depthwise_BN (BatchNor  (None, None, None,   1536       ['block_7_depthwise[0][0]']      \n",
      " malization)                    384)                                                              \n",
      "                                                                                                  \n",
      " block_7_depthwise_relu (ReLU)  (None, None, None,   0           ['block_7_depthwise_BN[0][0]']   \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_7_project (Conv2D)       (None, None, None,   24576       ['block_7_depthwise_relu[0][0]'] \n",
      "                                64)                                                               \n",
      "                                                                                                  \n",
      " block_7_project_BN (BatchNorma  (None, None, None,   256        ['block_7_project[0][0]']        \n",
      " lization)                      64)                                                               \n",
      "                                                                                                  \n",
      " block_7_add (Add)              (None, None, None,   0           ['block_6_project_BN[0][0]',     \n",
      "                                64)                               'block_7_project_BN[0][0]']     \n",
      "                                                                                                  \n",
      " block_8_expand (Conv2D)        (None, None, None,   24576       ['block_7_add[0][0]']            \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_8_expand_BN (BatchNormal  (None, None, None,   1536       ['block_8_expand[0][0]']         \n",
      " ization)                       384)                                                              \n",
      "                                                                                                  \n",
      " block_8_expand_relu (ReLU)     (None, None, None,   0           ['block_8_expand_BN[0][0]']      \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_8_depthwise (DepthwiseCo  (None, None, None,   3456       ['block_8_expand_relu[0][0]']    \n",
      " nv2D)                          384)                                                              \n",
      "                                                                                                  \n",
      " block_8_depthwise_BN (BatchNor  (None, None, None,   1536       ['block_8_depthwise[0][0]']      \n",
      " malization)                    384)                                                              \n",
      "                                                                                                  \n",
      " block_8_depthwise_relu (ReLU)  (None, None, None,   0           ['block_8_depthwise_BN[0][0]']   \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_8_project (Conv2D)       (None, None, None,   24576       ['block_8_depthwise_relu[0][0]'] \n",
      "                                64)                                                               \n",
      "                                                                                                  \n",
      " block_8_project_BN (BatchNorma  (None, None, None,   256        ['block_8_project[0][0]']        \n",
      " lization)                      64)                                                               \n",
      "                                                                                                  \n",
      " block_8_add (Add)              (None, None, None,   0           ['block_7_add[0][0]',            \n",
      "                                64)                               'block_8_project_BN[0][0]']     \n",
      "                                                                                                  \n",
      " block_9_expand (Conv2D)        (None, None, None,   24576       ['block_8_add[0][0]']            \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_9_expand_BN (BatchNormal  (None, None, None,   1536       ['block_9_expand[0][0]']         \n",
      " ization)                       384)                                                              \n",
      "                                                                                                  \n",
      " block_9_expand_relu (ReLU)     (None, None, None,   0           ['block_9_expand_BN[0][0]']      \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_9_depthwise (DepthwiseCo  (None, None, None,   3456       ['block_9_expand_relu[0][0]']    \n",
      " nv2D)                          384)                                                              \n",
      "                                                                                                  \n",
      " block_9_depthwise_BN (BatchNor  (None, None, None,   1536       ['block_9_depthwise[0][0]']      \n",
      " malization)                    384)                                                              \n",
      "                                                                                                  \n",
      " block_9_depthwise_relu (ReLU)  (None, None, None,   0           ['block_9_depthwise_BN[0][0]']   \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_9_project (Conv2D)       (None, None, None,   24576       ['block_9_depthwise_relu[0][0]'] \n",
      "                                64)                                                               \n",
      "                                                                                                  \n",
      " block_9_project_BN (BatchNorma  (None, None, None,   256        ['block_9_project[0][0]']        \n",
      " lization)                      64)                                                               \n",
      "                                                                                                  \n",
      " block_9_add (Add)              (None, None, None,   0           ['block_8_add[0][0]',            \n",
      "                                64)                               'block_9_project_BN[0][0]']     \n",
      "                                                                                                  \n",
      " block_10_expand (Conv2D)       (None, None, None,   24576       ['block_9_add[0][0]']            \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_10_expand_BN (BatchNorma  (None, None, None,   1536       ['block_10_expand[0][0]']        \n",
      " lization)                      384)                                                              \n",
      "                                                                                                  \n",
      " block_10_expand_relu (ReLU)    (None, None, None,   0           ['block_10_expand_BN[0][0]']     \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_10_depthwise (DepthwiseC  (None, None, None,   3456       ['block_10_expand_relu[0][0]']   \n",
      " onv2D)                         384)                                                              \n",
      "                                                                                                  \n",
      " block_10_depthwise_BN (BatchNo  (None, None, None,   1536       ['block_10_depthwise[0][0]']     \n",
      " rmalization)                   384)                                                              \n",
      "                                                                                                  \n",
      " block_10_depthwise_relu (ReLU)  (None, None, None,   0          ['block_10_depthwise_BN[0][0]']  \n",
      "                                384)                                                              \n",
      "                                                                                                  \n",
      " block_10_project (Conv2D)      (None, None, None,   36864       ['block_10_depthwise_relu[0][0]']\n",
      "                                96)                                                               \n",
      "                                                                                                  \n",
      " block_10_project_BN (BatchNorm  (None, None, None,   384        ['block_10_project[0][0]']       \n",
      " alization)                     96)                                                               \n",
      "                                                                                                  \n",
      " block_11_expand (Conv2D)       (None, None, None,   55296       ['block_10_project_BN[0][0]']    \n",
      "                                576)                                                              \n",
      "                                                                                                  \n",
      " block_11_expand_BN (BatchNorma  (None, None, None,   2304       ['block_11_expand[0][0]']        \n",
      " lization)                      576)                                                              \n",
      "                                                                                                  \n",
      " block_11_expand_relu (ReLU)    (None, None, None,   0           ['block_11_expand_BN[0][0]']     \n",
      "                                576)                                                              \n",
      "                                                                                                  \n",
      " block_11_depthwise (DepthwiseC  (None, None, None,   5184       ['block_11_expand_relu[0][0]']   \n",
      " onv2D)                         576)                                                              \n",
      "                                                                                                  \n",
      " block_11_depthwise_BN (BatchNo  (None, None, None,   2304       ['block_11_depthwise[0][0]']     \n",
      " rmalization)                   576)                                                              \n",
      "                                                                                                  \n",
      " block_11_depthwise_relu (ReLU)  (None, None, None,   0          ['block_11_depthwise_BN[0][0]']  \n",
      "                                576)                                                              \n",
      "                                                                                                  \n",
      " block_11_project (Conv2D)      (None, None, None,   55296       ['block_11_depthwise_relu[0][0]']\n",
      "                                96)                                                               \n",
      "                                                                                                  \n",
      " block_11_project_BN (BatchNorm  (None, None, None,   384        ['block_11_project[0][0]']       \n",
      " alization)                     96)                                                               \n",
      "                                                                                                  \n",
      " block_11_add (Add)             (None, None, None,   0           ['block_10_project_BN[0][0]',    \n",
      "                                96)                               'block_11_project_BN[0][0]']    \n",
      "                                                                                                  \n",
      " block_12_expand (Conv2D)       (None, None, None,   55296       ['block_11_add[0][0]']           \n",
      "                                576)                                                              \n",
      "                                                                                                  \n",
      " block_12_expand_BN (BatchNorma  (None, None, None,   2304       ['block_12_expand[0][0]']        \n",
      " lization)                      576)                                                              \n",
      "                                                                                                  \n",
      " block_12_expand_relu (ReLU)    (None, None, None,   0           ['block_12_expand_BN[0][0]']     \n",
      "                                576)                                                              \n",
      "                                                                                                  \n",
      " block_12_depthwise (DepthwiseC  (None, None, None,   5184       ['block_12_expand_relu[0][0]']   \n",
      " onv2D)                         576)                                                              \n",
      "                                                                                                  \n",
      " block_12_depthwise_BN (BatchNo  (None, None, None,   2304       ['block_12_depthwise[0][0]']     \n",
      " rmalization)                   576)                                                              \n",
      "                                                                                                  \n",
      " block_12_depthwise_relu (ReLU)  (None, None, None,   0          ['block_12_depthwise_BN[0][0]']  \n",
      "                                576)                                                              \n",
      "                                                                                                  \n",
      " block_12_project (Conv2D)      (None, None, None,   55296       ['block_12_depthwise_relu[0][0]']\n",
      "                                96)                                                               \n",
      "                                                                                                  \n",
      " block_12_project_BN (BatchNorm  (None, None, None,   384        ['block_12_project[0][0]']       \n",
      " alization)                     96)                                                               \n",
      "                                                                                                  \n",
      " block_12_add (Add)             (None, None, None,   0           ['block_11_add[0][0]',           \n",
      "                                96)                               'block_12_project_BN[0][0]']    \n",
      "                                                                                                  \n",
      " block_13_expand (Conv2D)       (None, None, None,   55296       ['block_12_add[0][0]']           \n",
      "                                576)                                                              \n",
      "                                                                                                  \n",
      " block_13_expand_BN (BatchNorma  (None, None, None,   2304       ['block_13_expand[0][0]']        \n",
      " lization)                      576)                                                              \n",
      "                                                                                                  \n",
      " block_13_expand_relu (ReLU)    (None, None, None,   0           ['block_13_expand_BN[0][0]']     \n",
      "                                576)                                                              \n",
      "                                                                                                  \n",
      " block_13_pad (ZeroPadding2D)   (None, None, None,   0           ['block_13_expand_relu[0][0]']   \n",
      "                                576)                                                              \n",
      "                                                                                                  \n",
      " block_13_depthwise (DepthwiseC  (None, None, None,   5184       ['block_13_pad[0][0]']           \n",
      " onv2D)                         576)                                                              \n",
      "                                                                                                  \n",
      " block_13_depthwise_BN (BatchNo  (None, None, None,   2304       ['block_13_depthwise[0][0]']     \n",
      " rmalization)                   576)                                                              \n",
      "                                                                                                  \n",
      " block_13_depthwise_relu (ReLU)  (None, None, None,   0          ['block_13_depthwise_BN[0][0]']  \n",
      "                                576)                                                              \n",
      "                                                                                                  \n",
      " block_13_project (Conv2D)      (None, None, None,   92160       ['block_13_depthwise_relu[0][0]']\n",
      "                                160)                                                              \n",
      "                                                                                                  \n",
      " block_13_project_BN (BatchNorm  (None, None, None,   640        ['block_13_project[0][0]']       \n",
      " alization)                     160)                                                              \n",
      "                                                                                                  \n",
      " block_14_expand (Conv2D)       (None, None, None,   153600      ['block_13_project_BN[0][0]']    \n",
      "                                960)                                                              \n",
      "                                                                                                  \n",
      " block_14_expand_BN (BatchNorma  (None, None, None,   3840       ['block_14_expand[0][0]']        \n",
      " lization)                      960)                                                              \n",
      "                                                                                                  \n",
      " block_14_expand_relu (ReLU)    (None, None, None,   0           ['block_14_expand_BN[0][0]']     \n",
      "                                960)                                                              \n",
      "                                                                                                  \n",
      " block_14_depthwise (DepthwiseC  (None, None, None,   8640       ['block_14_expand_relu[0][0]']   \n",
      " onv2D)                         960)                                                              \n",
      "                                                                                                  \n",
      " block_14_depthwise_BN (BatchNo  (None, None, None,   3840       ['block_14_depthwise[0][0]']     \n",
      " rmalization)                   960)                                                              \n",
      "                                                                                                  \n",
      " block_14_depthwise_relu (ReLU)  (None, None, None,   0          ['block_14_depthwise_BN[0][0]']  \n",
      "                                960)                                                              \n",
      "                                                                                                  \n",
      " block_14_project (Conv2D)      (None, None, None,   153600      ['block_14_depthwise_relu[0][0]']\n",
      "                                160)                                                              \n",
      "                                                                                                  \n",
      " block_14_project_BN (BatchNorm  (None, None, None,   640        ['block_14_project[0][0]']       \n",
      " alization)                     160)                                                              \n",
      "                                                                                                  \n",
      " block_14_add (Add)             (None, None, None,   0           ['block_13_project_BN[0][0]',    \n",
      "                                160)                              'block_14_project_BN[0][0]']    \n",
      "                                                                                                  \n",
      " block_15_expand (Conv2D)       (None, None, None,   153600      ['block_14_add[0][0]']           \n",
      "                                960)                                                              \n",
      "                                                                                                  \n",
      " block_15_expand_BN (BatchNorma  (None, None, None,   3840       ['block_15_expand[0][0]']        \n",
      " lization)                      960)                                                              \n",
      "                                                                                                  \n",
      " block_15_expand_relu (ReLU)    (None, None, None,   0           ['block_15_expand_BN[0][0]']     \n",
      "                                960)                                                              \n",
      "                                                                                                  \n",
      " block_15_depthwise (DepthwiseC  (None, None, None,   8640       ['block_15_expand_relu[0][0]']   \n",
      " onv2D)                         960)                                                              \n",
      "                                                                                                  \n",
      " block_15_depthwise_BN (BatchNo  (None, None, None,   3840       ['block_15_depthwise[0][0]']     \n",
      " rmalization)                   960)                                                              \n",
      "                                                                                                  \n",
      " block_15_depthwise_relu (ReLU)  (None, None, None,   0          ['block_15_depthwise_BN[0][0]']  \n",
      "                                960)                                                              \n",
      "                                                                                                  \n",
      " block_15_project (Conv2D)      (None, None, None,   153600      ['block_15_depthwise_relu[0][0]']\n",
      "                                160)                                                              \n",
      "                                                                                                  \n",
      " block_15_project_BN (BatchNorm  (None, None, None,   640        ['block_15_project[0][0]']       \n",
      " alization)                     160)                                                              \n",
      "                                                                                                  \n",
      " block_15_add (Add)             (None, None, None,   0           ['block_14_add[0][0]',           \n",
      "                                160)                              'block_15_project_BN[0][0]']    \n",
      "                                                                                                  \n",
      " block_16_expand (Conv2D)       (None, None, None,   153600      ['block_15_add[0][0]']           \n",
      "                                960)                                                              \n",
      "                                                                                                  \n",
      " block_16_expand_BN (BatchNorma  (None, None, None,   3840       ['block_16_expand[0][0]']        \n",
      " lization)                      960)                                                              \n",
      "                                                                                                  \n",
      " block_16_expand_relu (ReLU)    (None, None, None,   0           ['block_16_expand_BN[0][0]']     \n",
      "                                960)                                                              \n",
      "                                                                                                  \n",
      " block_16_depthwise (DepthwiseC  (None, None, None,   8640       ['block_16_expand_relu[0][0]']   \n",
      " onv2D)                         960)                                                              \n",
      "                                                                                                  \n",
      " block_16_depthwise_BN (BatchNo  (None, None, None,   3840       ['block_16_depthwise[0][0]']     \n",
      " rmalization)                   960)                                                              \n",
      "                                                                                                  \n",
      " block_16_depthwise_relu (ReLU)  (None, None, None,   0          ['block_16_depthwise_BN[0][0]']  \n",
      "                                960)                                                              \n",
      "                                                                                                  \n",
      " block_16_project (Conv2D)      (None, None, None,   307200      ['block_16_depthwise_relu[0][0]']\n",
      "                                320)                                                              \n",
      "                                                                                                  \n",
      " block_16_project_BN (BatchNorm  (None, None, None,   1280       ['block_16_project[0][0]']       \n",
      " alization)                     320)                                                              \n",
      "                                                                                                  \n",
      " Conv_1 (Conv2D)                (None, None, None,   409600      ['block_16_project_BN[0][0]']    \n",
      "                                1280)                                                             \n",
      "                                                                                                  \n",
      " Conv_1_bn (BatchNormalization)  (None, None, None,   5120       ['Conv_1[0][0]']                 \n",
      "                                1280)                                                             \n",
      "                                                                                                  \n",
      " out_relu (ReLU)                (None, None, None,   0           ['Conv_1_bn[0][0]']              \n",
      "                                1280)                                                             \n",
      "                                                                                                  \n",
      " global_average_pooling2d_3 (Gl  (None, 1280)        0           ['out_relu[0][0]']               \n",
      " obalAveragePooling2D)                                                                            \n",
      "                                                                                                  \n",
      " dense_12 (Dense)               (None, 512)          655872      ['global_average_pooling2d_3[0][0\n",
      "                                                                 ]']                              \n",
      "                                                                                                  \n",
      " dense_13 (Dense)               (None, 256)          131328      ['dense_12[0][0]']               \n",
      "                                                                                                  \n",
      " dense_14 (Dense)               (None, 128)          32896       ['dense_13[0][0]']               \n",
      "                                                                                                  \n",
      " dense_15 (Dense)               (None, 13)           1677        ['dense_14[0][0]']               \n",
      "                                                                                                  \n",
      "==================================================================================================\n",
      "Total params: 3,079,757\n",
      "Trainable params: 821,773\n",
      "Non-trainable params: 2,257,984\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/20\n",
      "131/131 [==============================] - 320s 2s/step - loss: 0.4904 - accuracy: 0.8782 - val_loss: 0.1930 - val_accuracy: 0.9509\n",
      "Epoch 2/20\n",
      "131/131 [==============================] - 214s 2s/step - loss: 0.0117 - accuracy: 0.9995 - val_loss: 0.1449 - val_accuracy: 0.9588\n",
      "Epoch 3/20\n",
      "131/131 [==============================] - 223s 2s/step - loss: 0.0041 - accuracy: 1.0000 - val_loss: 0.1398 - val_accuracy: 0.9610\n",
      "Epoch 4/20\n",
      "131/131 [==============================] - 211s 2s/step - loss: 0.0019 - accuracy: 1.0000 - val_loss: 0.1286 - val_accuracy: 0.9632\n",
      "Epoch 5/20\n",
      "131/131 [==============================] - 217s 2s/step - loss: 0.0012 - accuracy: 1.0000 - val_loss: 0.1269 - val_accuracy: 0.9649\n",
      "Epoch 6/20\n",
      "131/131 [==============================] - 212s 2s/step - loss: 7.7909e-04 - accuracy: 1.0000 - val_loss: 0.1424 - val_accuracy: 0.9627\n",
      "Epoch 7/20\n",
      "131/131 [==============================] - 212s 2s/step - loss: 5.5970e-04 - accuracy: 1.0000 - val_loss: 0.1223 - val_accuracy: 0.9660\n",
      "Epoch 8/20\n",
      "131/131 [==============================] - 198s 2s/step - loss: 4.1916e-04 - accuracy: 1.0000 - val_loss: 0.1341 - val_accuracy: 0.9654\n",
      "Epoch 9/20\n",
      "131/131 [==============================] - 216s 2s/step - loss: 3.2592e-04 - accuracy: 1.0000 - val_loss: 0.1308 - val_accuracy: 0.9671\n",
      "Epoch 10/20\n",
      "131/131 [==============================] - 214s 2s/step - loss: 2.5702e-04 - accuracy: 1.0000 - val_loss: 0.1351 - val_accuracy: 0.9654\n",
      "Epoch 11/20\n",
      "131/131 [==============================] - 206s 2s/step - loss: 2.0999e-04 - accuracy: 1.0000 - val_loss: 0.1326 - val_accuracy: 0.9671\n",
      "Epoch 12/20\n",
      "131/131 [==============================] - 211s 2s/step - loss: 1.7318e-04 - accuracy: 1.0000 - val_loss: 0.1384 - val_accuracy: 0.9654\n",
      "Epoch 13/20\n",
      "131/131 [==============================] - 213s 2s/step - loss: 1.4362e-04 - accuracy: 1.0000 - val_loss: 0.1353 - val_accuracy: 0.9671\n",
      "Epoch 14/20\n",
      "131/131 [==============================] - 208s 2s/step - loss: 1.2167e-04 - accuracy: 1.0000 - val_loss: 0.1341 - val_accuracy: 0.9677\n",
      "Epoch 15/20\n",
      "131/131 [==============================] - 214s 2s/step - loss: 1.0356e-04 - accuracy: 1.0000 - val_loss: 0.1389 - val_accuracy: 0.9671\n",
      "Epoch 16/20\n",
      "131/131 [==============================] - 211s 2s/step - loss: 8.9166e-05 - accuracy: 1.0000 - val_loss: 0.1339 - val_accuracy: 0.9677\n",
      "Epoch 17/20\n",
      "131/131 [==============================] - 211s 2s/step - loss: 7.7339e-05 - accuracy: 1.0000 - val_loss: 0.1354 - val_accuracy: 0.9671\n",
      "Epoch 18/20\n",
      "131/131 [==============================] - 217s 2s/step - loss: 6.7238e-05 - accuracy: 1.0000 - val_loss: 0.1361 - val_accuracy: 0.9677\n",
      "Epoch 19/20\n",
      "131/131 [==============================] - 223s 2s/step - loss: 5.8980e-05 - accuracy: 1.0000 - val_loss: 0.1391 - val_accuracy: 0.9671\n",
      "Epoch 20/20\n",
      "131/131 [==============================] - 219s 2s/step - loss: 5.1959e-05 - accuracy: 1.0000 - val_loss: 0.1383 - val_accuracy: 0.9671\n"
     ]
    }
   ],
   "source": [
    "opt = Adam(learning_rate=0.0001)\n",
    "model.compile(optimizer=opt, loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "\n",
    "history = model.fit(\n",
    "    train_generator,\n",
    "    validation_data=validation_generator,\n",
    "    epochs=20\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.plot(history.history['accuracy'])\n",
    "plt.plot(history.history['val_accuracy'])\n",
    "plt.title('Model accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.legend(['Train', 'Validation'], loc='upper left')\n",
    "# plt.ylim([0, 0.1])\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.title('Model loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend(['Train', 'Validation'], loc='upper left')\n",
    "# plt.ylim([0, 5])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "D:/DatasetMuseum/PaintingIGustiNgurahRai\n",
      "1/1 [==============================] - 0s 194ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 208ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 114ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 101ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 108ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 174ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 101ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 104ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 92ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 133ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 176ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 188ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 158ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 100ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 166ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 110ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 90ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 104ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 98ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 110ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 111ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 109ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 85ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 86ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 119ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 151ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 114ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 98ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 97ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 92ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 87ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 85ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 107ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 114ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 120ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 102ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 82ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 81ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 90ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 90ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 93ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 113ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 98ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 102ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 98ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 133ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 147ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 120ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 127ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 83ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 123ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 103ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 72ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 119ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 73ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 139ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 139ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 97ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 112ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 74ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 115ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 78ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 90ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 68ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 71ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 65ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 71ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 69ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 70ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 85ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 75ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 67ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 68ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 76ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 78ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 75ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 105ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 72ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 71ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 66ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 73ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 70ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 149ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 104ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 106ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 107ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 102ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 80ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 87ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 144ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 86ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 120ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 71ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 73ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 72ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 81ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 70ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 75ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 67ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 77ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 82ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 73ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 75ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 68ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 142ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 114ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 85ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 86ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 100ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 112ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 99ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 84ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 82ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 105ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 129ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 123ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 110ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 104ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 119ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 113ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 78ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 86ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 90ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 78ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 131ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 93ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 79ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 82ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 118ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 133ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 77ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 81ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 82ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 83ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 84ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 84ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 85ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 93ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 81ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 84ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 92ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 82ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 86ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 84ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 92ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 111ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 82ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 80ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 103ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 82ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 80ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 81ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 98ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 76ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 87ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 122ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 99ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 79ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 86ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 82ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 84ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 107ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 125ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 134ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 86ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 92ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 173ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 97ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 98ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 104ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 92ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 85ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 90ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 84ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 103ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 97ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 97ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 110ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 139ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 131ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 120ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 102ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 103ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 87ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 107ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 121ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 124ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 83ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 100ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 86ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 85ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 99ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 103ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 87ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 87ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 97ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 85ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 90ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 116ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 106ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 97ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 75ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 82ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 86ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 87ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 82ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 112ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 86ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 92ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 87ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 207ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 208ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 244ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 182ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 195ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 179ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 100ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 104ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 101ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 100ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 213ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 121ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 107ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 113ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 105ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 98ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 122ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 97ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 98ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 114ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 103ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 101ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 129ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 147ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 106ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 105ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 99ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 84ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 101ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 81ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 100ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 101ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 90ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 147ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 172ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 206ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 133ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 86ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 97ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 90ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 101ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 102ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 107ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 122ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 129ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 106ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 109ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 110ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 101ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 86ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 94ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 103ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 94ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 118ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 97ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 94ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 107ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 94ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 105ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 125ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 159ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 199ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 126ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 136ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 149ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 166ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 134ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 82ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 114ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 97ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 83ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 84ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 113ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 97ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 105ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 100ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 86ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 98ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 102ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 83ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 93ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 102ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 87ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 94ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 85ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 101ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 158ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 86ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 84ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 131ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 109ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 107ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 108ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 84ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 110ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 97ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 99ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 105ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 93ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 80ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 106ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 106ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 85ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 100ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 111ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 93ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 92ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 124ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 92ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 104ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 116ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 105ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 115ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 110ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 127ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 118ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 115ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 125ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 135ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 124ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 116ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 113ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 110ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 112ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 108ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 109ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 101ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 84ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 98ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 86ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 114ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 83ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 99ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 94ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 103ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 86ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 104ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 87ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 92ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 85ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 97ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 98ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 102ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 83ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 86ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 100ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 110ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 78ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 93ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 105ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 118ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 106ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 92ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 90ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 90ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 113ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 85ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 87ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 107ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 142ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 101ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 117ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 87ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 85ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 94ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 100ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 98ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 103ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 100ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 87ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 103ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 99ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 177ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 86ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 110ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 90ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 87ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 103ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 107ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 108ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 101ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 100ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 85ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 123ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 119ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 100ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 99ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 84ms/step\n",
      "Predicted Class: 0\n",
      "RandomData\n",
      "1/1 [==============================] - 0s 92ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 103ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 90ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 90ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 163ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 94ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 103ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 97ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 128ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 173ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 120ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 99ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 125ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 93ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 108ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 98ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 102ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 87ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 87ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 109ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 93ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 101ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 133ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 100ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 98ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 93ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 112ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 84ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 107ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 121ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 104ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 87ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 116ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 101ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 98ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 92ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 134ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 106ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 111ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 105ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 104ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 124ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 93ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 107ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 104ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 102ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 119ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 103ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 128ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 105ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 101ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 100ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 109ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 118ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 111ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 105ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 99ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 103ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 100ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 104ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 97ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 94ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 98ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 99ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 145ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 102ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 110ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 92ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 170ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 93ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 85ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 93ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 112ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 93ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 101ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 86ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 109ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 105ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 98ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 97ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 124ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 112ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 102ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 102ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 102ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 132ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 99ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 104ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 101ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 93ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 86ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 98ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 101ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 98ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 101ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 113ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 94ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 96ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 99ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 136ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 92ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 98ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 102ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 101ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 199ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 104ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 92ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 103ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 106ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 85ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 85ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 93ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 108ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 84ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 87ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 115ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 102ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 87ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 107ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 117ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 92ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 90ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 102ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 106ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 89ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 83ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 112ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 91ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 94ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 90ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 103ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 88ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 78ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 85ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 87ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 93ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 95ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n",
      "1/1 [==============================] - 0s 105ms/step\n",
      "Predicted Class: 10\n",
      "Painting I Gusti Ngurah Rai\n"
     ]
    }
   ],
   "source": [
    "from PIL import Image\n",
    "\n",
    "folder = 'D:/DatasetMuseum/PaintingIGustiNgurahRai'\n",
    "dirs = os.listdir(folder)\n",
    "print(folder)\n",
    "\n",
    "for item in dirs:\n",
    "    fullpath = os.path.join(folder, item)\n",
    "    if os.path.isfile(fullpath):\n",
    "        img = Image.open(fullpath).convert('RGB')\n",
    "        img = img.resize((224, 224))\n",
    "        img_array = np.array(img) / 255.0 \n",
    "        img_array = np.expand_dims(img_array, axis=0)\n",
    "        predictions = model.predict(img_array)\n",
    "        predicted_class = np.argmax(predictions)\n",
    "        print(\"Predicted Class:\", predicted_class)\n",
    "\n",
    "        if predicted_class==1:\n",
    "            print(\"Bali Following Independence\")\n",
    "        elif predicted_class==2:\n",
    "            print(\"Dang Hyang Nirartha\")\n",
    "        elif predicted_class==3:\n",
    "            print(\"Sagung Wah In Preparation To Attack The Dutch\")\n",
    "        elif predicted_class==4:\n",
    "            print(\"Spreading The News of The Proclamation of Indonesia\")\n",
    "        elif predicted_class==5:\n",
    "            print(\"The Battle of Jagaraga\")\n",
    "        elif predicted_class==6:\n",
    "            print(\"The Crowning Of Sri Kresna Kepakisan\")\n",
    "        elif predicted_class==7:\n",
    "            print(\"The Establishment Of Indonesian People Struggling Board Of Sunda Kecil\")\n",
    "        elif predicted_class==8:\n",
    "            print(\"The Naval Battle In Bali Strait\")\n",
    "        elif predicted_class==9:\n",
    "            print(\"The Rise Of Youth Organization\")\n",
    "        elif predicted_class==10:\n",
    "            print(\"Painting I Gusti Ngurah Rai\")\n",
    "        elif predicted_class==11:\n",
    "            print(\"Painting Menyambut Proklamasi Kemerdekaan RI\")\n",
    "        elif predicted_class==12:\n",
    "            print(\"Painting Ngurah Rai Mengatur Siasat Tempur\")\n",
    "        else:\n",
    "            print(\"RandomData\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "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 53). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: C:/Users/user/miniconda3/CapstoneProject\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: C:/Users/user/miniconda3/CapstoneProject\\assets\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "12146656"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from tensorflow.keras.models import load_model\n",
    "import pathlib\n",
    "\n",
    "model.save('model_fix.h5')\n",
    "loaded_model = load_model('model_fix.h5')\n",
    "\n",
    "\n",
    "export_dir = 'C:/Users/user/miniconda3/CapstoneProject'\n",
    "tf.saved_model.save(model, export_dir)\n",
    "\n",
    "converter = tf.lite.TFLiteConverter.from_saved_model(export_dir)\n",
    "tflite_model = converter.convert()\n",
    "\n",
    "tflite_model_file = pathlib.Path('C:/Users/user/miniconda3/CapstoneProject/model_fix.tflite')\n",
    "tflite_model_file.write_bytes(tflite_model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}