Car_classification.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Optimize MobileNetV2 hyperparameters. L2 regularization is added to the Dense layer with a coefficient of 0.01"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from typing import Tuple, Callable\n",
    "import matplotlib.pyplot as plt\n",
    "import keras\n",
    "from keras import layers\n",
    "from keras import models\n",
    "from keras import regularizers\n",
    "from keras.applications import MobileNetV2\n",
    "from keras.engine.training import Model\n",
    "import functools\n",
    "import os\n",
    "import numpy as np\n",
    "import pickle\n",
    "from keras.preprocessing.image import ImageDataGenerator"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_cell_guid": "79c7e3d0-c299-4dcb-8224-4455121ee9b0",
    "_uuid": "d629ff2d2480ee46fbb7e2d37f6b5fab8052498a"
   },
   "source": [
    "### Model Defining:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_mobilenetv2_full_tune_model_alpha_1_4_concatenated_regularised(_num_classes) \\\n",
    "        -> Tuple[Model, Model, int, Callable]:\n",
    "    \"\"\"Additional training of all MobileNetV2 layers, alpha = 1.4. MobileNetV2 layers are directly included. Model is regularized.\"\"\"\n",
    "    # If imagenet weights are being loaded, alpha can be one of `0.35`, `0.50`, `0.75`, `1.0`, `1.3` or `1.4` only.\n",
    "    image_size = 224\n",
    "    channels_count = 3\n",
    "    initial_model: Model = MobileNetV2(weights='imagenet', alpha=1.4, include_top=False,\n",
    "                                       input_shape=(image_size, image_size, channels_count))\n",
    "\n",
    "    \n",
    "    initial_model.trainable = True\n",
    "    for i, layer in enumerate(initial_model.layers):\n",
    "        layer.trainable = True\n",
    "\n",
    "    # We include MobileNetV2 layers directly in the final model.\n",
    "    initial_model_output = initial_model.output\n",
    "    x = layers.GlobalAveragePooling2D()(initial_model_output)\n",
    "    x = layers.Dense(1024, activation='relu', kernel_regularizer=regularizers.l2(0.01))(x)\n",
    "    predictions = layers.Dense(_num_classes, activation='softmax')(x)\n",
    "\n",
    "    model = Model(initial_model.input, predictions)\n",
    "\n",
    "    return model, initial_model, image_size, keras.applications.mobilenet_v2.preprocess_input\n",
    "\n",
    "# ----------------------------------------------------------------------------------------------------------------------\n",
    "def get_callbacks_list(_early_stopping_patience, _reduce_lr_on_plateau_factor, _reduce_lr_on_plateau_patience):\n",
    "    \"\"\"Get callbacks for a model\"\"\"\n",
    "    return [\n",
    "        keras.callbacks.EarlyStopping(\n",
    "            monitor='val_acc',\n",
    "            patience=_early_stopping_patience\n",
    "        ),\n",
    "        keras.callbacks.ModelCheckpoint(\n",
    "            verbose=1,\n",
    "            filepath='best_model.h5',\n",
    "            monitor='val_loss',\n",
    "            save_best_only=True\n",
    "        ),\n",
    "        keras.callbacks.ReduceLROnPlateau(\n",
    "            verbose=1,\n",
    "            monitor='val_loss',\n",
    "            factor=_reduce_lr_on_plateau_factor,\n",
    "            patience=_reduce_lr_on_plateau_patience\n",
    "        ),\n",
    "    ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_1 (InputLayer)            (None, 224, 224, 3)  0                                            \n",
      "__________________________________________________________________________________________________\n",
      "Conv1_pad (ZeroPadding2D)       (None, 225, 225, 3)  0           input_1[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "Conv1 (Conv2D)                  (None, 112, 112, 48) 1296        Conv1_pad[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "bn_Conv1 (BatchNormalization)   (None, 112, 112, 48) 192         Conv1[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "Conv1_relu (ReLU)               (None, 112, 112, 48) 0           bn_Conv1[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "expanded_conv_depthwise (Depthw (None, 112, 112, 48) 432         Conv1_relu[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "expanded_conv_depthwise_BN (Bat (None, 112, 112, 48) 192         expanded_conv_depthwise[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "expanded_conv_depthwise_relu (R (None, 112, 112, 48) 0           expanded_conv_depthwise_BN[0][0] \n",
      "__________________________________________________________________________________________________\n",
      "expanded_conv_project (Conv2D)  (None, 112, 112, 24) 1152        expanded_conv_depthwise_relu[0][0\n",
      "__________________________________________________________________________________________________\n",
      "expanded_conv_project_BN (Batch (None, 112, 112, 24) 96          expanded_conv_project[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "block_1_expand (Conv2D)         (None, 112, 112, 144 3456        expanded_conv_project_BN[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "block_1_expand_BN (BatchNormali (None, 112, 112, 144 576         block_1_expand[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "block_1_expand_relu (ReLU)      (None, 112, 112, 144 0           block_1_expand_BN[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "block_1_pad (ZeroPadding2D)     (None, 113, 113, 144 0           block_1_expand_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "block_1_depthwise (DepthwiseCon (None, 56, 56, 144)  1296        block_1_pad[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "block_1_depthwise_BN (BatchNorm (None, 56, 56, 144)  576         block_1_depthwise[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "block_1_depthwise_relu (ReLU)   (None, 56, 56, 144)  0           block_1_depthwise_BN[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "block_1_project (Conv2D)        (None, 56, 56, 32)   4608        block_1_depthwise_relu[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "block_1_project_BN (BatchNormal (None, 56, 56, 32)   128         block_1_project[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "block_2_expand (Conv2D)         (None, 56, 56, 192)  6144        block_1_project_BN[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_2_expand_BN (BatchNormali (None, 56, 56, 192)  768         block_2_expand[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "block_2_expand_relu (ReLU)      (None, 56, 56, 192)  0           block_2_expand_BN[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "block_2_depthwise (DepthwiseCon (None, 56, 56, 192)  1728        block_2_expand_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "block_2_depthwise_BN (BatchNorm (None, 56, 56, 192)  768         block_2_depthwise[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "block_2_depthwise_relu (ReLU)   (None, 56, 56, 192)  0           block_2_depthwise_BN[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "block_2_project (Conv2D)        (None, 56, 56, 32)   6144        block_2_depthwise_relu[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "block_2_project_BN (BatchNormal (None, 56, 56, 32)   128         block_2_project[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "block_2_add (Add)               (None, 56, 56, 32)   0           block_1_project_BN[0][0]         \n",
      "                                                                 block_2_project_BN[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_3_expand (Conv2D)         (None, 56, 56, 192)  6144        block_2_add[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "block_3_expand_BN (BatchNormali (None, 56, 56, 192)  768         block_3_expand[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "block_3_expand_relu (ReLU)      (None, 56, 56, 192)  0           block_3_expand_BN[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "block_3_pad (ZeroPadding2D)     (None, 57, 57, 192)  0           block_3_expand_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "block_3_depthwise (DepthwiseCon (None, 28, 28, 192)  1728        block_3_pad[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "block_3_depthwise_BN (BatchNorm (None, 28, 28, 192)  768         block_3_depthwise[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "block_3_depthwise_relu (ReLU)   (None, 28, 28, 192)  0           block_3_depthwise_BN[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "block_3_project (Conv2D)        (None, 28, 28, 48)   9216        block_3_depthwise_relu[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "block_3_project_BN (BatchNormal (None, 28, 28, 48)   192         block_3_project[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "block_4_expand (Conv2D)         (None, 28, 28, 288)  13824       block_3_project_BN[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_4_expand_BN (BatchNormali (None, 28, 28, 288)  1152        block_4_expand[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "block_4_expand_relu (ReLU)      (None, 28, 28, 288)  0           block_4_expand_BN[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "block_4_depthwise (DepthwiseCon (None, 28, 28, 288)  2592        block_4_expand_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "block_4_depthwise_BN (BatchNorm (None, 28, 28, 288)  1152        block_4_depthwise[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "block_4_depthwise_relu (ReLU)   (None, 28, 28, 288)  0           block_4_depthwise_BN[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "block_4_project (Conv2D)        (None, 28, 28, 48)   13824       block_4_depthwise_relu[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "block_4_project_BN (BatchNormal (None, 28, 28, 48)   192         block_4_project[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "block_4_add (Add)               (None, 28, 28, 48)   0           block_3_project_BN[0][0]         \n",
      "                                                                 block_4_project_BN[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_5_expand (Conv2D)         (None, 28, 28, 288)  13824       block_4_add[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "block_5_expand_BN (BatchNormali (None, 28, 28, 288)  1152        block_5_expand[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "block_5_expand_relu (ReLU)      (None, 28, 28, 288)  0           block_5_expand_BN[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "block_5_depthwise (DepthwiseCon (None, 28, 28, 288)  2592        block_5_expand_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "block_5_depthwise_BN (BatchNorm (None, 28, 28, 288)  1152        block_5_depthwise[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "block_5_depthwise_relu (ReLU)   (None, 28, 28, 288)  0           block_5_depthwise_BN[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "block_5_project (Conv2D)        (None, 28, 28, 48)   13824       block_5_depthwise_relu[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "block_5_project_BN (BatchNormal (None, 28, 28, 48)   192         block_5_project[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "block_5_add (Add)               (None, 28, 28, 48)   0           block_4_add[0][0]                \n",
      "                                                                 block_5_project_BN[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_6_expand (Conv2D)         (None, 28, 28, 288)  13824       block_5_add[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "block_6_expand_BN (BatchNormali (None, 28, 28, 288)  1152        block_6_expand[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "block_6_expand_relu (ReLU)      (None, 28, 28, 288)  0           block_6_expand_BN[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "block_6_pad (ZeroPadding2D)     (None, 29, 29, 288)  0           block_6_expand_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "block_6_depthwise (DepthwiseCon (None, 14, 14, 288)  2592        block_6_pad[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "block_6_depthwise_BN (BatchNorm (None, 14, 14, 288)  1152        block_6_depthwise[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "block_6_depthwise_relu (ReLU)   (None, 14, 14, 288)  0           block_6_depthwise_BN[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "block_6_project (Conv2D)        (None, 14, 14, 88)   25344       block_6_depthwise_relu[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "block_6_project_BN (BatchNormal (None, 14, 14, 88)   352         block_6_project[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "block_7_expand (Conv2D)         (None, 14, 14, 528)  46464       block_6_project_BN[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_7_expand_BN (BatchNormali (None, 14, 14, 528)  2112        block_7_expand[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "block_7_expand_relu (ReLU)      (None, 14, 14, 528)  0           block_7_expand_BN[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "block_7_depthwise (DepthwiseCon (None, 14, 14, 528)  4752        block_7_expand_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "block_7_depthwise_BN (BatchNorm (None, 14, 14, 528)  2112        block_7_depthwise[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "block_7_depthwise_relu (ReLU)   (None, 14, 14, 528)  0           block_7_depthwise_BN[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "block_7_project (Conv2D)        (None, 14, 14, 88)   46464       block_7_depthwise_relu[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "block_7_project_BN (BatchNormal (None, 14, 14, 88)   352         block_7_project[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "block_7_add (Add)               (None, 14, 14, 88)   0           block_6_project_BN[0][0]         \n",
      "                                                                 block_7_project_BN[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_8_expand (Conv2D)         (None, 14, 14, 528)  46464       block_7_add[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "block_8_expand_BN (BatchNormali (None, 14, 14, 528)  2112        block_8_expand[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "block_8_expand_relu (ReLU)      (None, 14, 14, 528)  0           block_8_expand_BN[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "block_8_depthwise (DepthwiseCon (None, 14, 14, 528)  4752        block_8_expand_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "block_8_depthwise_BN (BatchNorm (None, 14, 14, 528)  2112        block_8_depthwise[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "block_8_depthwise_relu (ReLU)   (None, 14, 14, 528)  0           block_8_depthwise_BN[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "block_8_project (Conv2D)        (None, 14, 14, 88)   46464       block_8_depthwise_relu[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "block_8_project_BN (BatchNormal (None, 14, 14, 88)   352         block_8_project[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "block_8_add (Add)               (None, 14, 14, 88)   0           block_7_add[0][0]                \n",
      "                                                                 block_8_project_BN[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_9_expand (Conv2D)         (None, 14, 14, 528)  46464       block_8_add[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "block_9_expand_BN (BatchNormali (None, 14, 14, 528)  2112        block_9_expand[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "block_9_expand_relu (ReLU)      (None, 14, 14, 528)  0           block_9_expand_BN[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "block_9_depthwise (DepthwiseCon (None, 14, 14, 528)  4752        block_9_expand_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "block_9_depthwise_BN (BatchNorm (None, 14, 14, 528)  2112        block_9_depthwise[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "block_9_depthwise_relu (ReLU)   (None, 14, 14, 528)  0           block_9_depthwise_BN[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "block_9_project (Conv2D)        (None, 14, 14, 88)   46464       block_9_depthwise_relu[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "block_9_project_BN (BatchNormal (None, 14, 14, 88)   352         block_9_project[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "block_9_add (Add)               (None, 14, 14, 88)   0           block_8_add[0][0]                \n",
      "                                                                 block_9_project_BN[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_10_expand (Conv2D)        (None, 14, 14, 528)  46464       block_9_add[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "block_10_expand_BN (BatchNormal (None, 14, 14, 528)  2112        block_10_expand[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "block_10_expand_relu (ReLU)     (None, 14, 14, 528)  0           block_10_expand_BN[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_10_depthwise (DepthwiseCo (None, 14, 14, 528)  4752        block_10_expand_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "block_10_depthwise_BN (BatchNor (None, 14, 14, 528)  2112        block_10_depthwise[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_10_depthwise_relu (ReLU)  (None, 14, 14, 528)  0           block_10_depthwise_BN[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "block_10_project (Conv2D)       (None, 14, 14, 136)  71808       block_10_depthwise_relu[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "block_10_project_BN (BatchNorma (None, 14, 14, 136)  544         block_10_project[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "block_11_expand (Conv2D)        (None, 14, 14, 816)  110976      block_10_project_BN[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "block_11_expand_BN (BatchNormal (None, 14, 14, 816)  3264        block_11_expand[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "block_11_expand_relu (ReLU)     (None, 14, 14, 816)  0           block_11_expand_BN[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_11_depthwise (DepthwiseCo (None, 14, 14, 816)  7344        block_11_expand_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "block_11_depthwise_BN (BatchNor (None, 14, 14, 816)  3264        block_11_depthwise[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_11_depthwise_relu (ReLU)  (None, 14, 14, 816)  0           block_11_depthwise_BN[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "block_11_project (Conv2D)       (None, 14, 14, 136)  110976      block_11_depthwise_relu[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "block_11_project_BN (BatchNorma (None, 14, 14, 136)  544         block_11_project[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "block_11_add (Add)              (None, 14, 14, 136)  0           block_10_project_BN[0][0]        \n",
      "                                                                 block_11_project_BN[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "block_12_expand (Conv2D)        (None, 14, 14, 816)  110976      block_11_add[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "block_12_expand_BN (BatchNormal (None, 14, 14, 816)  3264        block_12_expand[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "block_12_expand_relu (ReLU)     (None, 14, 14, 816)  0           block_12_expand_BN[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_12_depthwise (DepthwiseCo (None, 14, 14, 816)  7344        block_12_expand_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "block_12_depthwise_BN (BatchNor (None, 14, 14, 816)  3264        block_12_depthwise[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_12_depthwise_relu (ReLU)  (None, 14, 14, 816)  0           block_12_depthwise_BN[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "block_12_project (Conv2D)       (None, 14, 14, 136)  110976      block_12_depthwise_relu[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "block_12_project_BN (BatchNorma (None, 14, 14, 136)  544         block_12_project[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "block_12_add (Add)              (None, 14, 14, 136)  0           block_11_add[0][0]               \n",
      "                                                                 block_12_project_BN[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "block_13_expand (Conv2D)        (None, 14, 14, 816)  110976      block_12_add[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "block_13_expand_BN (BatchNormal (None, 14, 14, 816)  3264        block_13_expand[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "block_13_expand_relu (ReLU)     (None, 14, 14, 816)  0           block_13_expand_BN[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_13_pad (ZeroPadding2D)    (None, 15, 15, 816)  0           block_13_expand_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "block_13_depthwise (DepthwiseCo (None, 7, 7, 816)    7344        block_13_pad[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "block_13_depthwise_BN (BatchNor (None, 7, 7, 816)    3264        block_13_depthwise[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_13_depthwise_relu (ReLU)  (None, 7, 7, 816)    0           block_13_depthwise_BN[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "block_13_project (Conv2D)       (None, 7, 7, 224)    182784      block_13_depthwise_relu[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "block_13_project_BN (BatchNorma (None, 7, 7, 224)    896         block_13_project[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "block_14_expand (Conv2D)        (None, 7, 7, 1344)   301056      block_13_project_BN[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "block_14_expand_BN (BatchNormal (None, 7, 7, 1344)   5376        block_14_expand[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "block_14_expand_relu (ReLU)     (None, 7, 7, 1344)   0           block_14_expand_BN[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_14_depthwise (DepthwiseCo (None, 7, 7, 1344)   12096       block_14_expand_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "block_14_depthwise_BN (BatchNor (None, 7, 7, 1344)   5376        block_14_depthwise[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_14_depthwise_relu (ReLU)  (None, 7, 7, 1344)   0           block_14_depthwise_BN[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "block_14_project (Conv2D)       (None, 7, 7, 224)    301056      block_14_depthwise_relu[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "block_14_project_BN (BatchNorma (None, 7, 7, 224)    896         block_14_project[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "block_14_add (Add)              (None, 7, 7, 224)    0           block_13_project_BN[0][0]        \n",
      "                                                                 block_14_project_BN[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "block_15_expand (Conv2D)        (None, 7, 7, 1344)   301056      block_14_add[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "block_15_expand_BN (BatchNormal (None, 7, 7, 1344)   5376        block_15_expand[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "block_15_expand_relu (ReLU)     (None, 7, 7, 1344)   0           block_15_expand_BN[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_15_depthwise (DepthwiseCo (None, 7, 7, 1344)   12096       block_15_expand_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "block_15_depthwise_BN (BatchNor (None, 7, 7, 1344)   5376        block_15_depthwise[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_15_depthwise_relu (ReLU)  (None, 7, 7, 1344)   0           block_15_depthwise_BN[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "block_15_project (Conv2D)       (None, 7, 7, 224)    301056      block_15_depthwise_relu[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "block_15_project_BN (BatchNorma (None, 7, 7, 224)    896         block_15_project[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "block_15_add (Add)              (None, 7, 7, 224)    0           block_14_add[0][0]               \n",
      "                                                                 block_15_project_BN[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "block_16_expand (Conv2D)        (None, 7, 7, 1344)   301056      block_15_add[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "block_16_expand_BN (BatchNormal (None, 7, 7, 1344)   5376        block_16_expand[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "block_16_expand_relu (ReLU)     (None, 7, 7, 1344)   0           block_16_expand_BN[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_16_depthwise (DepthwiseCo (None, 7, 7, 1344)   12096       block_16_expand_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "block_16_depthwise_BN (BatchNor (None, 7, 7, 1344)   5376        block_16_depthwise[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "block_16_depthwise_relu (ReLU)  (None, 7, 7, 1344)   0           block_16_depthwise_BN[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "block_16_project (Conv2D)       (None, 7, 7, 448)    602112      block_16_depthwise_relu[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "block_16_project_BN (BatchNorma (None, 7, 7, 448)    1792        block_16_project[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "Conv_1 (Conv2D)                 (None, 7, 7, 1792)   802816      block_16_project_BN[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Conv_1_bn (BatchNormalization)  (None, 7, 7, 1792)   7168        Conv_1[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "out_relu (ReLU)                 (None, 7, 7, 1792)   0           Conv_1_bn[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling2d_1 (Glo (None, 1792)         0           out_relu[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "dense_1 (Dense)                 (None, 1024)         1836032     global_average_pooling2d_1[0][0] \n",
      "__________________________________________________________________________________________________\n",
      "dense_2 (Dense)                 (None, 196)          200900      dense_1[0][0]                    \n",
      "==================================================================================================\n",
      "Total params: 6,400,644\n",
      "Trainable params: 6,352,708\n",
      "Non-trainable params: 47,936\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "batch_size = 32\n",
    "init_lr = 0.01\n",
    "momentum = 0.9\n",
    "epochs = 80\n",
    "optimazer = keras.optimizers.SGD(lr=init_lr, momentum=momentum)\n",
    "# optimazer = keras.optimizers.Adam(lr=init_lr)\n",
    "# Data directories:\n",
    "train_dir = 'data_preprocessing/Cropped_images/'\n",
    "# ----------------------------------------------------------------------------------------------------------------------\n",
    "# Find out the number of task classes by the number of directories in the training folder.:\n",
    "num_classes = len(os.listdir(train_dir))\n",
    "\n",
    "# Callback Settings:\n",
    "early_stopping_patience = 10\n",
    "reduce_lr_on_plateau_factor = 0.2\n",
    "reduce_lr_on_plateau_patience = 3\n",
    "\n",
    "# Selected model:\n",
    "model_function = get_mobilenetv2_full_tune_model_alpha_1_4_concatenated_regularised\n",
    "# We obtain the complete model, convolutional basis, image size and the function of preprocessing input images:\n",
    "model, conv_base, image_size, preprocess_function = model_function(num_classes)\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# ----------------------------------------------------------------------------------------------------------------------\n",
    "# Metric for top 5 accuracy\n",
    "top10_acc = functools.partial(keras.metrics.top_k_categorical_accuracy, k=10)\n",
    "top10_acc.__name__ = 'top10_acc'\n",
    "\n",
    "# Compile the model:\n",
    "model.compile(loss='categorical_crossentropy', optimizer=optimazer, metrics=['accuracy', top10_acc])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5"
   },
   "source": [
    "### Data Preprocessing:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 10816 images belonging to 196 classes.\n",
      "Found 1112 images belonging to 196 classes.\n"
     ]
    }
   ],
   "source": [
    "# ----------------------------------------------------------------------------------------------------------------------\n",
    "# Create Generators:\n",
    "# Expand training data.\n",
    "# Use only preprocess_function, without any other preprocessing of pixel values ​​and channels.\n",
    "image_datagen = ImageDataGenerator(\n",
    "    preprocessing_function=preprocess_function,\n",
    "    rotation_range=5,\n",
    "    width_shift_range=0.1,\n",
    "    height_shift_range=0.1,\n",
    "    shear_range=0.1,\n",
    "    zoom_range=0.1,\n",
    "    horizontal_flip=True,\n",
    "    validation_split=0.1,\n",
    "    fill_mode='nearest')\n",
    "\n",
    "train_generator = image_datagen.flow_from_directory(\n",
    "    train_dir,\n",
    "    target_size=(image_size, image_size),\n",
    "    batch_size=batch_size,\n",
    "    class_mode='categorical',\n",
    "    subset='training'\n",
    ")\n",
    "train_images_count = len(train_generator.filenames)\n",
    "\n",
    "# validation data is not expandable.\n",
    "validation_generator = image_datagen.flow_from_directory(\n",
    "    train_dir,\n",
    "    target_size=(image_size, image_size),\n",
    "    batch_size=batch_size,\n",
    "    class_mode='categorical',\n",
    "    subset='validation'\n",
    ")\n",
    "validation_images_count = len(validation_generator.filenames)\n",
    "\n",
    "\n",
    "with open('classes_name.pickle', 'wb') as handle:\n",
    "    pickle.dump(validation_generator.class_indices, handle, protocol=pickle.HIGHEST_PROTOCOL)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/80\n",
      "338/338 [==============================] - 362s 1s/step - loss: 10.7328 - acc: 0.3092 - top10_acc: 0.6092 - val_loss: 7.7384 - val_acc: 0.1857 - val_top10_acc: 0.5515\n",
      "\n",
      "Epoch 00001: val_loss improved from inf to 7.73839, saving model to best_model.h5\n",
      "Epoch 2/80\n",
      "338/338 [==============================] - 389s 1s/step - loss: 3.3388 - acc: 0.6955 - top10_acc: 0.9528 - val_loss: 5.2767 - val_acc: 0.1944 - val_top10_acc: 0.6444\n",
      "\n",
      "Epoch 00002: val_loss improved from 7.73839 to 5.27673, saving model to best_model.h5\n",
      "Epoch 3/80\n",
      "338/338 [==============================] - 378s 1s/step - loss: 1.4627 - acc: 0.8163 - top10_acc: 0.9839 - val_loss: 3.2625 - val_acc: 0.3519 - val_top10_acc: 0.7870\n",
      "\n",
      "Epoch 00003: val_loss improved from 5.27673 to 3.26252, saving model to best_model.h5\n",
      "Epoch 4/80\n",
      "338/338 [==============================] - 362s 1s/step - loss: 0.8659 - acc: 0.8736 - top10_acc: 0.9925 - val_loss: 2.1198 - val_acc: 0.5472 - val_top10_acc: 0.9065\n",
      "\n",
      "Epoch 00004: val_loss improved from 3.26252 to 2.11985, saving model to best_model.h5\n",
      "Epoch 5/80\n",
      "338/338 [==============================] - 365s 1s/step - loss: 0.6538 - acc: 0.8957 - top10_acc: 0.9966 - val_loss: 2.7048 - val_acc: 0.4306 - val_top10_acc: 0.8037\n",
      "\n",
      "Epoch 00005: val_loss did not improve from 2.11985\n",
      "Epoch 6/80\n",
      "338/338 [==============================] - 363s 1s/step - loss: 0.5137 - acc: 0.9234 - top10_acc: 0.9981 - val_loss: 1.8480 - val_acc: 0.6120 - val_top10_acc: 0.9204\n",
      "\n",
      "Epoch 00006: val_loss improved from 2.11985 to 1.84802, saving model to best_model.h5\n",
      "Epoch 7/80\n",
      "338/338 [==============================] - 357s 1s/step - loss: 0.4392 - acc: 0.9344 - top10_acc: 0.9996 - val_loss: 1.6195 - val_acc: 0.6676 - val_top10_acc: 0.9324\n",
      "\n",
      "Epoch 00007: val_loss improved from 1.84802 to 1.61949, saving model to best_model.h5\n",
      "Epoch 8/80\n",
      "338/338 [==============================] - 368s 1s/step - loss: 0.3552 - acc: 0.9529 - top10_acc: 0.9996 - val_loss: 1.1941 - val_acc: 0.7519 - val_top10_acc: 0.9667\n",
      "\n",
      "Epoch 00008: val_loss improved from 1.61949 to 1.19409, saving model to best_model.h5\n",
      "Epoch 9/80\n",
      "338/338 [==============================] - 360s 1s/step - loss: 0.2916 - acc: 0.9642 - top10_acc: 0.9998 - val_loss: 0.9307 - val_acc: 0.8037 - val_top10_acc: 0.9769\n",
      "\n",
      "Epoch 00009: val_loss improved from 1.19409 to 0.93067, saving model to best_model.h5\n",
      "Epoch 10/80\n",
      "338/338 [==============================] - 370s 1s/step - loss: 0.2680 - acc: 0.9690 - top10_acc: 0.9998 - val_loss: 0.8666 - val_acc: 0.8250 - val_top10_acc: 0.9667\n",
      "\n",
      "Epoch 00010: val_loss improved from 0.93067 to 0.86656, saving model to best_model.h5\n",
      "Epoch 11/80\n",
      "338/338 [==============================] - 353s 1s/step - loss: 0.2369 - acc: 0.9751 - top10_acc: 0.9999 - val_loss: 0.8278 - val_acc: 0.8111 - val_top10_acc: 0.9815\n",
      "\n",
      "Epoch 00011: val_loss improved from 0.86656 to 0.82782, saving model to best_model.h5\n",
      "Epoch 12/80\n",
      "338/338 [==============================] - 348s 1s/step - loss: 0.2221 - acc: 0.9734 - top10_acc: 1.0000 - val_loss: 0.8223 - val_acc: 0.8259 - val_top10_acc: 0.9769\n",
      "\n",
      "Epoch 00012: val_loss improved from 0.82782 to 0.82233, saving model to best_model.h5\n",
      "Epoch 13/80\n",
      "338/338 [==============================] - 341s 1s/step - loss: 0.2108 - acc: 0.9783 - top10_acc: 1.0000 - val_loss: 0.8153 - val_acc: 0.8259 - val_top10_acc: 0.9731\n",
      "\n",
      "Epoch 00013: val_loss improved from 0.82233 to 0.81526, saving model to best_model.h5\n",
      "Epoch 14/80\n",
      "338/338 [==============================] - 336s 994ms/step - loss: 0.1864 - acc: 0.9810 - top10_acc: 1.0000 - val_loss: 0.6951 - val_acc: 0.8491 - val_top10_acc: 0.9787\n",
      "\n",
      "Epoch 00014: val_loss improved from 0.81526 to 0.69512, saving model to best_model.h5\n",
      "Epoch 15/80\n",
      "338/338 [==============================] - 343s 1s/step - loss: 0.1825 - acc: 0.9831 - top10_acc: 1.0000 - val_loss: 0.7664 - val_acc: 0.8472 - val_top10_acc: 0.9657\n",
      "\n",
      "Epoch 00015: val_loss did not improve from 0.69512\n",
      "Epoch 16/80\n",
      "338/338 [==============================] - 334s 987ms/step - loss: 0.1554 - acc: 0.9871 - top10_acc: 1.0000 - val_loss: 0.6271 - val_acc: 0.8657 - val_top10_acc: 0.9796\n",
      "\n",
      "Epoch 00016: val_loss improved from 0.69512 to 0.62710, saving model to best_model.h5\n",
      "Epoch 17/80\n",
      "338/338 [==============================] - 336s 993ms/step - loss: 0.1452 - acc: 0.9884 - top10_acc: 1.0000 - val_loss: 0.6182 - val_acc: 0.8713 - val_top10_acc: 0.9796\n",
      "\n",
      "Epoch 00017: val_loss improved from 0.62710 to 0.61822, saving model to best_model.h5\n",
      "Epoch 18/80\n",
      "338/338 [==============================] - 333s 985ms/step - loss: 0.1425 - acc: 0.9871 - top10_acc: 0.9999 - val_loss: 0.6033 - val_acc: 0.8657 - val_top10_acc: 0.9880\n",
      "\n",
      "Epoch 00018: val_loss improved from 0.61822 to 0.60332, saving model to best_model.h5\n",
      "Epoch 19/80\n",
      "338/338 [==============================] - 335s 990ms/step - loss: 0.1270 - acc: 0.9920 - top10_acc: 0.9998 - val_loss: 0.5468 - val_acc: 0.8778 - val_top10_acc: 0.9898\n",
      "\n",
      "Epoch 00019: val_loss improved from 0.60332 to 0.54679, saving model to best_model.h5\n",
      "Epoch 20/80\n",
      "338/338 [==============================] - 332s 982ms/step - loss: 0.1141 - acc: 0.9924 - top10_acc: 1.0000 - val_loss: 0.5082 - val_acc: 0.8843 - val_top10_acc: 0.9861\n",
      "\n",
      "Epoch 00020: val_loss improved from 0.54679 to 0.50822, saving model to best_model.h5\n",
      "Epoch 21/80\n",
      "338/338 [==============================] - 337s 996ms/step - loss: 0.1043 - acc: 0.9944 - top10_acc: 1.0000 - val_loss: 0.4984 - val_acc: 0.8907 - val_top10_acc: 0.9833\n",
      "\n",
      "Epoch 00021: val_loss improved from 0.50822 to 0.49842, saving model to best_model.h5\n",
      "Epoch 22/80\n",
      "338/338 [==============================] - 337s 996ms/step - loss: 0.0999 - acc: 0.9937 - top10_acc: 1.0000 - val_loss: 0.4548 - val_acc: 0.8926 - val_top10_acc: 0.9880\n",
      "\n",
      "Epoch 00022: val_loss improved from 0.49842 to 0.45478, saving model to best_model.h5\n",
      "Epoch 23/80\n",
      "338/338 [==============================] - 338s 999ms/step - loss: 0.1019 - acc: 0.9948 - top10_acc: 0.9999 - val_loss: 0.4988 - val_acc: 0.8926 - val_top10_acc: 0.9843\n",
      "\n",
      "Epoch 00023: val_loss did not improve from 0.45478\n",
      "Epoch 24/80\n",
      "338/338 [==============================] - 335s 990ms/step - loss: 0.1031 - acc: 0.9929 - top10_acc: 1.0000 - val_loss: 0.4814 - val_acc: 0.8880 - val_top10_acc: 0.9870\n",
      "\n",
      "Epoch 00024: val_loss did not improve from 0.45478\n",
      "Epoch 25/80\n",
      "338/338 [==============================] - 331s 980ms/step - loss: 0.0933 - acc: 0.9949 - top10_acc: 1.0000 - val_loss: 0.4718 - val_acc: 0.8843 - val_top10_acc: 0.9917\n",
      "\n",
      "Epoch 00025: val_loss did not improve from 0.45478\n",
      "\n",
      "Epoch 00025: ReduceLROnPlateau reducing learning rate to 0.0019999999552965165.\n",
      "Epoch 26/80\n",
      "338/338 [==============================] - 337s 996ms/step - loss: 0.0800 - acc: 0.9973 - top10_acc: 1.0000 - val_loss: 0.3622 - val_acc: 0.9222 - val_top10_acc: 0.9963\n",
      "\n",
      "Epoch 00026: val_loss improved from 0.45478 to 0.36219, saving model to best_model.h5\n",
      "Epoch 27/80\n",
      "338/338 [==============================] - 336s 993ms/step - loss: 0.0702 - acc: 0.9980 - top10_acc: 1.0000 - val_loss: 0.3498 - val_acc: 0.9241 - val_top10_acc: 0.9889\n",
      "\n",
      "Epoch 00027: val_loss improved from 0.36219 to 0.34982, saving model to best_model.h5\n",
      "Epoch 28/80\n",
      "338/338 [==============================] - 338s 999ms/step - loss: 0.0671 - acc: 0.9983 - top10_acc: 1.0000 - val_loss: 0.3343 - val_acc: 0.9148 - val_top10_acc: 0.9944\n",
      "\n",
      "Epoch 00028: val_loss improved from 0.34982 to 0.33432, saving model to best_model.h5\n",
      "Epoch 29/80\n",
      "338/338 [==============================] - 335s 991ms/step - loss: 0.0657 - acc: 0.9977 - top10_acc: 1.0000 - val_loss: 0.3562 - val_acc: 0.9074 - val_top10_acc: 0.9972\n",
      "\n",
      "Epoch 00029: val_loss did not improve from 0.33432\n",
      "Epoch 30/80\n",
      "338/338 [==============================] - 335s 991ms/step - loss: 0.0637 - acc: 0.9983 - top10_acc: 1.0000 - val_loss: 0.3505 - val_acc: 0.9222 - val_top10_acc: 0.9935\n",
      "\n",
      "Epoch 00030: val_loss did not improve from 0.33432\n",
      "Epoch 31/80\n",
      "338/338 [==============================] - 334s 989ms/step - loss: 0.0633 - acc: 0.9979 - top10_acc: 1.0000 - val_loss: 0.3367 - val_acc: 0.9250 - val_top10_acc: 0.9935\n",
      "\n",
      "Epoch 00031: val_loss did not improve from 0.33432\n",
      "\n",
      "Epoch 00031: ReduceLROnPlateau reducing learning rate to 0.0003999999724328518.\n",
      "Epoch 32/80\n",
      "338/338 [==============================] - 337s 998ms/step - loss: 0.0620 - acc: 0.9983 - top10_acc: 1.0000 - val_loss: 0.3438 - val_acc: 0.9241 - val_top10_acc: 0.9954\n",
      "\n",
      "Epoch 00032: val_loss did not improve from 0.33432\n",
      "Epoch 33/80\n",
      "338/338 [==============================] - 338s 1s/step - loss: 0.0609 - acc: 0.9991 - top10_acc: 1.0000 - val_loss: 0.3262 - val_acc: 0.9250 - val_top10_acc: 0.9954\n",
      "\n",
      "Epoch 00033: val_loss improved from 0.33432 to 0.32615, saving model to best_model.h5\n",
      "Epoch 34/80\n",
      "338/338 [==============================] - 333s 986ms/step - loss: 0.0618 - acc: 0.9982 - top10_acc: 1.0000 - val_loss: 0.3435 - val_acc: 0.9204 - val_top10_acc: 0.9944\n",
      "\n",
      "Epoch 00034: val_loss did not improve from 0.32615\n",
      "Epoch 35/80\n",
      "338/338 [==============================] - 332s 984ms/step - loss: 0.0612 - acc: 0.9982 - top10_acc: 1.0000 - val_loss: 0.3176 - val_acc: 0.9259 - val_top10_acc: 0.9954\n",
      "\n",
      "Epoch 00035: val_loss improved from 0.32615 to 0.31760, saving model to best_model.h5\n",
      "Epoch 36/80\n",
      "338/338 [==============================] - 333s 985ms/step - loss: 0.0605 - acc: 0.9990 - top10_acc: 1.0000 - val_loss: 0.3370 - val_acc: 0.9173 - val_top10_acc: 0.9954\n",
      "\n",
      "Epoch 00036: val_loss did not improve from 0.31760\n",
      "Epoch 37/80\n",
      "338/338 [==============================] - 336s 993ms/step - loss: 0.0601 - acc: 0.9986 - top10_acc: 1.0000 - val_loss: 0.3088 - val_acc: 0.9278 - val_top10_acc: 0.9944\n",
      "\n",
      "Epoch 00037: val_loss improved from 0.31760 to 0.30884, saving model to best_model.h5\n",
      "Epoch 38/80\n",
      "338/338 [==============================] - 331s 980ms/step - loss: 0.0601 - acc: 0.9984 - top10_acc: 1.0000 - val_loss: 0.3493 - val_acc: 0.9250 - val_top10_acc: 0.9926\n",
      "\n",
      "Epoch 00038: val_loss did not improve from 0.30884\n",
      "Epoch 39/80\n",
      "338/338 [==============================] - 332s 983ms/step - loss: 0.0598 - acc: 0.9986 - top10_acc: 1.0000 - val_loss: 0.3369 - val_acc: 0.9241 - val_top10_acc: 0.9954\n",
      "\n",
      "Epoch 00039: val_loss did not improve from 0.30884\n",
      "Epoch 40/80\n",
      "338/338 [==============================] - 336s 993ms/step - loss: 0.0602 - acc: 0.9984 - top10_acc: 1.0000 - val_loss: 0.3125 - val_acc: 0.9278 - val_top10_acc: 0.9935\n",
      "\n",
      "Epoch 00040: val_loss did not improve from 0.30884\n",
      "\n",
      "Epoch 00040: ReduceLROnPlateau reducing learning rate to 7.999999215826393e-05.\n",
      "Epoch 41/80\n",
      "338/338 [==============================] - 337s 996ms/step - loss: 0.0599 - acc: 0.9980 - top10_acc: 1.0000 - val_loss: 0.3157 - val_acc: 0.9259 - val_top10_acc: 0.9935\n",
      "\n",
      "Epoch 00041: val_loss did not improve from 0.30884\n",
      "Epoch 42/80\n",
      "338/338 [==============================] - 338s 1s/step - loss: 0.0592 - acc: 0.9989 - top10_acc: 1.0000 - val_loss: 0.3537 - val_acc: 0.9185 - val_top10_acc: 0.9972\n",
      "\n",
      "Epoch 00042: val_loss did not improve from 0.30884\n",
      "Epoch 43/80\n",
      "338/338 [==============================] - 337s 996ms/step - loss: 0.0603 - acc: 0.9984 - top10_acc: 1.0000 - val_loss: 0.3006 - val_acc: 0.9352 - val_top10_acc: 0.9944\n",
      "\n",
      "Epoch 00043: val_loss improved from 0.30884 to 0.30057, saving model to best_model.h5\n",
      "Epoch 44/80\n",
      "338/338 [==============================] - 335s 990ms/step - loss: 0.0593 - acc: 0.9990 - top10_acc: 1.0000 - val_loss: 0.3436 - val_acc: 0.9241 - val_top10_acc: 0.9926\n",
      "\n",
      "Epoch 00044: val_loss did not improve from 0.30057\n",
      "Epoch 45/80\n",
      "338/338 [==============================] - 335s 990ms/step - loss: 0.0599 - acc: 0.9985 - top10_acc: 1.0000 - val_loss: 0.3459 - val_acc: 0.9278 - val_top10_acc: 0.9935\n",
      "\n",
      "Epoch 00045: val_loss did not improve from 0.30057\n",
      "Epoch 46/80\n",
      "338/338 [==============================] - 346s 1s/step - loss: 0.0593 - acc: 0.9983 - top10_acc: 1.0000 - val_loss: 0.3531 - val_acc: 0.9093 - val_top10_acc: 0.9944\n",
      "\n",
      "Epoch 00046: val_loss did not improve from 0.30057\n",
      "\n",
      "Epoch 00046: ReduceLROnPlateau reducing learning rate to 1.599999814061448e-05.\n",
      "Epoch 47/80\n",
      "338/338 [==============================] - 330s 978ms/step - loss: 0.0594 - acc: 0.9985 - top10_acc: 1.0000 - val_loss: 0.3329 - val_acc: 0.9269 - val_top10_acc: 0.9926\n",
      "\n",
      "Epoch 00047: val_loss did not improve from 0.30057\n",
      "Epoch 48/80\n",
      "338/338 [==============================] - 341s 1s/step - loss: 0.0591 - acc: 0.9988 - top10_acc: 1.0000 - val_loss: 0.3219 - val_acc: 0.9204 - val_top10_acc: 0.9954\n",
      "\n",
      "Epoch 00048: val_loss did not improve from 0.30057\n",
      "Epoch 49/80\n",
      "338/338 [==============================] - 341s 1s/step - loss: 0.0598 - acc: 0.9982 - top10_acc: 1.0000 - val_loss: 0.3421 - val_acc: 0.9194 - val_top10_acc: 0.9954\n",
      "\n",
      "Epoch 00049: val_loss did not improve from 0.30057\n",
      "\n",
      "Epoch 00049: ReduceLROnPlateau reducing learning rate to 3.199999628122896e-06.\n",
      "Epoch 50/80\n",
      "338/338 [==============================] - 342s 1s/step - loss: 0.0595 - acc: 0.9985 - top10_acc: 1.0000 - val_loss: 0.3211 - val_acc: 0.9287 - val_top10_acc: 0.9935\n",
      "\n",
      "Epoch 00050: val_loss did not improve from 0.30057\n",
      "Epoch 51/80\n",
      "338/338 [==============================] - 343s 1s/step - loss: 0.0587 - acc: 0.9989 - top10_acc: 1.0000 - val_loss: 0.3134 - val_acc: 0.9250 - val_top10_acc: 0.9944\n",
      "\n",
      "Epoch 00051: val_loss did not improve from 0.30057\n",
      "Epoch 52/80\n",
      "338/338 [==============================] - 335s 990ms/step - loss: 0.0596 - acc: 0.9992 - top10_acc: 1.0000 - val_loss: 0.3457 - val_acc: 0.9204 - val_top10_acc: 0.9944\n",
      "\n",
      "Epoch 00052: val_loss did not improve from 0.30057\n",
      "\n",
      "Epoch 00052: ReduceLROnPlateau reducing learning rate to 6.399999165296323e-07.\n",
      "Epoch 53/80\n",
      "338/338 [==============================] - 343s 1s/step - loss: 0.0593 - acc: 0.9987 - top10_acc: 1.0000 - val_loss: 0.3446 - val_acc: 0.9176 - val_top10_acc: 0.9954\n",
      "\n",
      "Epoch 00053: val_loss did not improve from 0.30057\n"
     ]
    }
   ],
   "source": [
    "# ----------------------------------------------------------------------------------------------------------------------\n",
    "# Model Training:\n",
    "train_steps = len(train_generator.filenames) // batch_size\n",
    "validation_steps = len(validation_generator.filenames) // batch_size\n",
    "history = model.fit_generator(\n",
    "    train_generator,\n",
    "    steps_per_epoch=train_steps,\n",
    "    epochs=epochs,\n",
    "    validation_data=validation_generator,\n",
    "    validation_steps=validation_steps,\n",
    "    callbacks=get_callbacks_list(early_stopping_patience, reduce_lr_on_plateau_factor, reduce_lr_on_plateau_patience)\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Validation loss:  0.2837304408903475\n",
      "Validation acc:   0.9351851851851852\n",
      "Validation top 10 score:   0.9916666666666667\n"
     ]
    }
   ],
   "source": [
    "validation_score = model.evaluate_generator(validation_generator, steps=validation_steps)\n",
    "print('Validation loss: ', validation_score[0])\n",
    "print('Validation acc:  ', validation_score[1])\n",
    "print('Validation top 10 score:  ', validation_score[2])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def draw_training_info_plots(_history):\n",
    "    \"\"\"Draw loss graphs at the training and validation stage\"\"\"\n",
    "    acc = _history.history['acc']\n",
    "    val_acc = _history.history['val_acc']\n",
    "    loss = _history.history['loss']\n",
    "    val_loss = _history.history['val_loss']\n",
    "\n",
    "    epochs_plot = range(1, len(acc) + 1)\n",
    "    plt.plot(epochs_plot, acc, 'bo', label='Training acc')\n",
    "    plt.plot(epochs_plot, val_acc, 'b', label='Validation acc')\n",
    "    plt.title('Training and validation accuracy')\n",
    "    plt.xlabel('epoch')\n",
    "    plt.ylabel('accuracy')\n",
    "    plt.legend()\n",
    "    plt.figure()\n",
    "\n",
    "    plt.plot(epochs_plot, loss, 'bo', label='Training loss')\n",
    "    plt.plot(epochs_plot, val_loss, 'b', label='Validation loss')\n",
    "    plt.title('Training and validation loss')\n",
    "    plt.xlabel('epoch')\n",
    "    plt.ylabel('loss')\n",
    "    plt.legend()\n",
    "    plt.show()\n",
    "\n",
    "    if 'lr' in _history.history:\n",
    "        learning_rate = _history.history['lr']\n",
    "        plt.plot(epochs_plot, learning_rate, 'b', label='Learning rate')\n",
    "        plt.title('Learning rate')\n",
    "        plt.xlabel('epoch')\n",
    "        plt.ylabel('learning rate')\n",
    "        plt.legend()\n",
    "        plt.show()\n",
    "    return\n",
    "\n",
    "draw_training_info_plots(history)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# model image saving\n",
    "from keras.utils import plot_model\n",
    "# Draw an image with a printout of the model structure:\n",
    "plot_model(conv_base, show_shapes=True, to_file='conv_base.png')\n",
    "plot_model(model, show_shapes=True, to_file='model.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Testing on Image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from keras.preprocessing import image\n",
    "from keras.models import load_model\n",
    "import functools\n",
    "import pickle\n",
    "import keras\n",
    "import numpy as np\n",
    "# ----------------------------------------------------------------------------------------------------------------------\n",
    "# Metric for top 5 accuracy\n",
    "top10_acc = functools.partial(keras.metrics.top_k_categorical_accuracy, k=10)\n",
    "top10_acc.__name__ = 'top10_acc'\n",
    "\n",
    "model = load_model('model/best_model_top10.h5',custom_objects={'top10_acc': top10_acc})\n",
    "\n",
    "with open('classes_name.pickle', 'rb') as handle:\n",
    "    classes_name = pickle.load(handle)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[['Volvo 240 Sedan 1993' '72.694984']\n",
      " ['AM General Hummer SUV 2000' '12.358813']\n",
      " ['Volkswagen Golf Hatchback 1991' '1.0848004']\n",
      " ['Lamborghini Diablo Coupe 2001' '1.0669779']\n",
      " ['Ford GT Coupe 2006' '0.9957726']]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<PIL.JpegImagePlugin.JpegImageFile image mode=RGB size=224x224 at 0x20978DDAA90>"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Volvo 240 Sedan 1993/03227.jpg\n",
    "#top 5 acc\n",
    "#Volkswagen Golf Hatchback 1991/00786.jpg\n",
    "#Nissan NV Passenger Van 2012/07160.jpg\n",
    "#Nissan NV Passenger Van 2012/02358.jpg\n",
    "img_path='data_preprocessing/Cropped_images/Volvo 240 Sedan 1993/03227.jpg'\n",
    "image_size=224\n",
    "img = image.load_img(img_path, target_size=(image_size, image_size))\n",
    "\n",
    "x = image.img_to_array(img)\n",
    "prediction=model.predict(x.reshape(1,image_size,image_size,3))\n",
    "\n",
    "prediction_index_5=prediction.reshape(-1,).argsort()[-5:][::-1]\n",
    "print (np.c_[np.array(list((classes_name.keys())))[prediction_index_5],prediction.reshape(-1,)[prediction_index_5]*100])\n",
    "img"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "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": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}