4_training_mobileNetV3Small.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CnJiZ1Q-K0qs"
      },
      "source": [
        "AIM: To train MobileNetV3Small model on different parameters."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "TB7SxHDsUD2w"
      },
      "outputs": [],
      "source": [
        "# Note: Comment or delete the current code block if dataset is loaded from local disk.\n",
        "# loading dataset from kaggle\n",
        "# upload kaggle.json file\n",
        "! pip install -q kaggle\n",
        "from google.colab import files\n",
        "print(\"Upload 'kaggle.json' file...\")\n",
        "files.upload()\n",
        "\n",
        "# download the dataset(zipped folder) and unzip it\n",
        "! mkdir ~/.kaggle\n",
        "! cp kaggle.json ~/.kaggle/\n",
        "! chmod 600 ~/.kaggle/kaggle.json\n",
        "! kaggle datasets download alessiocorrado99/animals10\n",
        "! unzip animals10.zip"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "OTJGpCW0UKgb"
      },
      "outputs": [],
      "source": [
        "# Note: Load the class files into current working directory.\n",
        "# importing required libraries\n",
        "from tensorflow.keras.utils import image_dataset_from_directory\n",
        "import tensorflow as tf\n",
        "\n",
        "from keras.callbacks import ModelCheckpoint\n",
        "\n",
        "from imutils import paths\n",
        "import matplotlib.pyplot as plt\n",
        "import os\n",
        "\n",
        "import mobilenetv3s # class file\n",
        "import helper_functions # class file"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "jXiUug4QUvxi"
      },
      "outputs": [],
      "source": [
        "# defining image size and batch size\n",
        "batch_size=64\n",
        "image_width=224\n",
        "image_height=224"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hOXNndS2LPoB",
        "outputId": "1187889b-87bf-42ea-c723-929b670e768b"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Class Labels: ['cane', 'cavallo', 'elefante', 'farfalla', 'gallina', 'gatto', 'mucca', 'pecora', 'ragno', 'scoiattolo']\n",
            "\n",
            "Number of Classes: 10\n",
            "\n",
            "Total Elements in Each Class : [4863, 2623, 1446, 2112, 3098, 1668, 1866, 1820, 4821, 1862]\n",
            "\n",
            "Class Weight: {0: 1.0, 1: 1.854, 2: 3.363, 3: 2.303, 4: 1.57, 5: 2.915, 6: 2.606, 7: 2.672, 8: 1.009, 9: 2.612}\n"
          ]
        }
      ],
      "source": [
        "# getting output class labels and calculating class weight\n",
        "class_labels = sorted(os.listdir('raw-img')) # get class labels\n",
        "print('\\nClass Labels: {}'.format(class_labels))\n",
        "num_classes = len(class_labels) # get number of output classes\n",
        "print('\\nNumber of Classes: {}'.format(num_classes))\n",
        "\n",
        "# calculating total elements in each class\n",
        "class_totals = [len(os.listdir('raw-img' + f\"/{label}\")) for label in class_labels]\n",
        "print(\"\\nTotal Elements in Each Class : {}\".format(class_totals))\n",
        "\n",
        "# calculate class weight\n",
        "class_weight = {i:round((max(class_totals) / class_total), 3) for i, class_total in enumerate(class_totals)}\n",
        "print('\\nClass Weight: {}'.format(class_weight))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WRecGxWFx46_",
        "outputId": "0475cd3b-73d5-4bf7-f072-2d7266dab6e1"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Processing...\n",
            "Completed...\n"
          ]
        }
      ],
      "source": [
        "# get image paths and move them into train and validation directories\n",
        "imagePaths = list(paths.list_images(\"raw-img\")) # getting path to the directory of images\n",
        "print('Processing...')\n",
        "helper_functions.move_images(imagePaths=imagePaths) # calling function\n",
        "print('Completed...')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5mnkH8bNUyNp",
        "outputId": "cb5a859b-76fd-4b16-c857-9d30bc41deb3"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 18325 files belonging to 10 classes.\n",
            "Found 7854 files belonging to 10 classes.\n"
          ]
        }
      ],
      "source": [
        "# get training dataset\n",
        "train_ds = image_dataset_from_directory(\n",
        "  \"train_dir/\", label_mode='categorical', seed=42, shuffle=True,\n",
        "  image_size=(image_height, image_width),\n",
        "  batch_size=batch_size)\n",
        "\n",
        "# get validation dataset\n",
        "val_ds = image_dataset_from_directory(\n",
        "  \"valid_dir/\", label_mode='categorical', seed=42, shuffle=True,\n",
        "  image_size=(image_height, image_width),\n",
        "  batch_size=batch_size)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "mpOT6QyjWZ_z"
      },
      "outputs": [],
      "source": [
        "# using buffered prefetching\n",
        "shuffle_size = len(train_ds) # shuffles entire train dataset at each epoch\n",
        "\n",
        "AUTOTUNE = tf.data.AUTOTUNE\n",
        "train_ds = train_ds.cache().shuffle(shuffle_size).prefetch(buffer_size=AUTOTUNE)\n",
        "val_ds = val_ds.cache().prefetch(buffer_size=AUTOTUNE)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Model 1:**\n",
        "\n",
        "      INPUT => MobileNetV3Small => GAP => DO => SOFTMAX\n",
        "\n",
        "Observations:\n",
        "\n",
        "\n",
        "\n",
        "1.   Trainable parameters: 5,770.\n",
        "2.   Even after 100 epochs validation loss is lower than training loss.\n",
        "    * Trend: Training loss is reducing slowly.\n",
        "    * Training for more epochs may results in a good model.\n",
        "    * Model is taking more time to learn might be due to the fact that it is not too deep or densed, i.e., fewer parameters.\n",
        "\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "QK_oLgKRPdbD"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "lx5igNZaWaoV",
        "outputId": "59752e3b-ca84-48d1-e542-e0e4da17f5f4"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/mobilenet_v3/weights_mobilenet_v3_small_224_1.0_float_no_top_v2.h5\n",
            "4334752/4334752 [==============================] - 0s 0us/step\n",
            "Model Summary...\n",
            "Model: \"MobileNetV3Small\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_2 (InputLayer)        [(None, 224, 224, 3)]     0         \n",
            "                                                                 \n",
            " MobilenetV3small (Functiona  (None, 7, 7, 576)        939120    \n",
            " l)                                                              \n",
            "                                                                 \n",
            " global_average_pooling2d (G  (None, 576)              0         \n",
            " lobalAveragePooling2D)                                          \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 576)               0         \n",
            "                                                                 \n",
            " dense (Dense)               (None, 10)                5770      \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 944,890\n",
            "Trainable params: 5,770\n",
            "Non-trainable params: 939,120\n",
            "_________________________________________________________________\n",
            "\n",
            "\n",
            "Model Training...\n",
            "Epoch 1/100\n",
            "287/287 [==============================] - 46s 74ms/step - loss: 1.2466 - accuracy: 0.7975 - val_loss: 0.2854 - val_accuracy: 0.9170\n",
            "Epoch 2/100\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.5210 - accuracy: 0.9169 - val_loss: 0.2324 - val_accuracy: 0.9281\n",
            "Epoch 3/100\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.4289 - accuracy: 0.9287 - val_loss: 0.2068 - val_accuracy: 0.9351\n",
            "Epoch 4/100\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.3914 - accuracy: 0.9329 - val_loss: 0.1998 - val_accuracy: 0.9372\n",
            "Epoch 5/100\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.3704 - accuracy: 0.9371 - val_loss: 0.1874 - val_accuracy: 0.9423\n",
            "Epoch 6/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.3451 - accuracy: 0.9405 - val_loss: 0.1937 - val_accuracy: 0.9388\n",
            "Epoch 7/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.3374 - accuracy: 0.9434 - val_loss: 0.1837 - val_accuracy: 0.9437\n",
            "Epoch 8/100\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.3247 - accuracy: 0.9428 - val_loss: 0.1867 - val_accuracy: 0.9417\n",
            "Epoch 9/100\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.3221 - accuracy: 0.9447 - val_loss: 0.1817 - val_accuracy: 0.9431\n",
            "Epoch 10/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.3131 - accuracy: 0.9446 - val_loss: 0.1851 - val_accuracy: 0.9427\n",
            "Epoch 11/100\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.3002 - accuracy: 0.9485 - val_loss: 0.1786 - val_accuracy: 0.9447\n",
            "Epoch 12/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2993 - accuracy: 0.9473 - val_loss: 0.1780 - val_accuracy: 0.9451\n",
            "Epoch 13/100\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.2952 - accuracy: 0.9488 - val_loss: 0.1894 - val_accuracy: 0.9398\n",
            "Epoch 14/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2888 - accuracy: 0.9484 - val_loss: 0.1838 - val_accuracy: 0.9416\n",
            "Epoch 15/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2841 - accuracy: 0.9488 - val_loss: 0.1751 - val_accuracy: 0.9464\n",
            "Epoch 16/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2851 - accuracy: 0.9478 - val_loss: 0.1799 - val_accuracy: 0.9439\n",
            "Epoch 17/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2826 - accuracy: 0.9497 - val_loss: 0.1773 - val_accuracy: 0.9459\n",
            "Epoch 18/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2760 - accuracy: 0.9509 - val_loss: 0.1751 - val_accuracy: 0.9470\n",
            "Epoch 19/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2764 - accuracy: 0.9501 - val_loss: 0.1833 - val_accuracy: 0.9445\n",
            "Epoch 20/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2714 - accuracy: 0.9507 - val_loss: 0.1795 - val_accuracy: 0.9445\n",
            "Epoch 21/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2717 - accuracy: 0.9507 - val_loss: 0.1852 - val_accuracy: 0.9437\n",
            "Epoch 22/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2716 - accuracy: 0.9501 - val_loss: 0.1823 - val_accuracy: 0.9456\n",
            "Epoch 23/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2654 - accuracy: 0.9514 - val_loss: 0.1801 - val_accuracy: 0.9463\n",
            "Epoch 24/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2660 - accuracy: 0.9496 - val_loss: 0.1786 - val_accuracy: 0.9456\n",
            "Epoch 25/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2688 - accuracy: 0.9506 - val_loss: 0.1867 - val_accuracy: 0.9422\n",
            "Epoch 26/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2669 - accuracy: 0.9514 - val_loss: 0.1823 - val_accuracy: 0.9453\n",
            "Epoch 27/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2642 - accuracy: 0.9512 - val_loss: 0.1831 - val_accuracy: 0.9464\n",
            "Epoch 28/100\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.2582 - accuracy: 0.9530 - val_loss: 0.1877 - val_accuracy: 0.9436\n",
            "Epoch 29/100\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.2584 - accuracy: 0.9536 - val_loss: 0.1872 - val_accuracy: 0.9440\n",
            "Epoch 30/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2621 - accuracy: 0.9522 - val_loss: 0.1842 - val_accuracy: 0.9446\n",
            "Epoch 31/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2604 - accuracy: 0.9524 - val_loss: 0.1858 - val_accuracy: 0.9426\n",
            "Epoch 32/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2551 - accuracy: 0.9519 - val_loss: 0.1837 - val_accuracy: 0.9458\n",
            "Epoch 33/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2544 - accuracy: 0.9513 - val_loss: 0.1859 - val_accuracy: 0.9454\n",
            "Epoch 34/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2592 - accuracy: 0.9514 - val_loss: 0.1867 - val_accuracy: 0.9439\n",
            "Epoch 35/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2561 - accuracy: 0.9521 - val_loss: 0.1822 - val_accuracy: 0.9465\n",
            "Epoch 36/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2588 - accuracy: 0.9517 - val_loss: 0.1878 - val_accuracy: 0.9446\n",
            "Epoch 37/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2553 - accuracy: 0.9521 - val_loss: 0.1861 - val_accuracy: 0.9451\n",
            "Epoch 38/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2545 - accuracy: 0.9520 - val_loss: 0.1874 - val_accuracy: 0.9449\n",
            "Epoch 39/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2559 - accuracy: 0.9529 - val_loss: 0.1951 - val_accuracy: 0.9423\n",
            "Epoch 40/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2572 - accuracy: 0.9537 - val_loss: 0.1884 - val_accuracy: 0.9445\n",
            "Epoch 41/100\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.2504 - accuracy: 0.9544 - val_loss: 0.1865 - val_accuracy: 0.9459\n",
            "Epoch 42/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2503 - accuracy: 0.9521 - val_loss: 0.1879 - val_accuracy: 0.9453\n",
            "Epoch 43/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2606 - accuracy: 0.9523 - val_loss: 0.1963 - val_accuracy: 0.9423\n",
            "Epoch 44/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2630 - accuracy: 0.9510 - val_loss: 0.1898 - val_accuracy: 0.9436\n",
            "Epoch 45/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2468 - accuracy: 0.9542 - val_loss: 0.1902 - val_accuracy: 0.9455\n",
            "Epoch 46/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2536 - accuracy: 0.9532 - val_loss: 0.1971 - val_accuracy: 0.9404\n",
            "Epoch 47/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2536 - accuracy: 0.9517 - val_loss: 0.1851 - val_accuracy: 0.9468\n",
            "Epoch 48/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2518 - accuracy: 0.9529 - val_loss: 0.1919 - val_accuracy: 0.9451\n",
            "Epoch 49/100\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.2421 - accuracy: 0.9564 - val_loss: 0.1855 - val_accuracy: 0.9470\n",
            "Epoch 50/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2558 - accuracy: 0.9522 - val_loss: 0.1922 - val_accuracy: 0.9460\n",
            "Epoch 51/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2554 - accuracy: 0.9519 - val_loss: 0.1895 - val_accuracy: 0.9460\n",
            "Epoch 52/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2512 - accuracy: 0.9526 - val_loss: 0.1900 - val_accuracy: 0.9456\n",
            "Epoch 53/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2532 - accuracy: 0.9521 - val_loss: 0.2039 - val_accuracy: 0.9388\n",
            "Epoch 54/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2531 - accuracy: 0.9523 - val_loss: 0.1979 - val_accuracy: 0.9409\n",
            "Epoch 55/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2571 - accuracy: 0.9539 - val_loss: 0.1934 - val_accuracy: 0.9464\n",
            "Epoch 56/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2586 - accuracy: 0.9506 - val_loss: 0.1901 - val_accuracy: 0.9451\n",
            "Epoch 57/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2493 - accuracy: 0.9509 - val_loss: 0.1888 - val_accuracy: 0.9465\n",
            "Epoch 58/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2529 - accuracy: 0.9521 - val_loss: 0.1950 - val_accuracy: 0.9445\n",
            "Epoch 59/100\n",
            "287/287 [==============================] - 16s 55ms/step - loss: 0.2563 - accuracy: 0.9523 - val_loss: 0.1893 - val_accuracy: 0.9464\n",
            "Epoch 60/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2423 - accuracy: 0.9536 - val_loss: 0.1914 - val_accuracy: 0.9455\n",
            "Epoch 61/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2588 - accuracy: 0.9519 - val_loss: 0.1976 - val_accuracy: 0.9426\n",
            "Epoch 62/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2498 - accuracy: 0.9536 - val_loss: 0.1951 - val_accuracy: 0.9450\n",
            "Epoch 63/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2502 - accuracy: 0.9533 - val_loss: 0.1973 - val_accuracy: 0.9431\n",
            "Epoch 64/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2439 - accuracy: 0.9543 - val_loss: 0.1955 - val_accuracy: 0.9439\n",
            "Epoch 65/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2514 - accuracy: 0.9538 - val_loss: 0.2012 - val_accuracy: 0.9404\n",
            "Epoch 66/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2535 - accuracy: 0.9521 - val_loss: 0.1951 - val_accuracy: 0.9442\n",
            "Epoch 67/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2496 - accuracy: 0.9527 - val_loss: 0.1984 - val_accuracy: 0.9439\n",
            "Epoch 68/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2456 - accuracy: 0.9524 - val_loss: 0.1953 - val_accuracy: 0.9447\n",
            "Epoch 69/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2398 - accuracy: 0.9540 - val_loss: 0.1920 - val_accuracy: 0.9461\n",
            "Epoch 70/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2432 - accuracy: 0.9553 - val_loss: 0.1973 - val_accuracy: 0.9461\n",
            "Epoch 71/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2572 - accuracy: 0.9515 - val_loss: 0.2003 - val_accuracy: 0.9426\n",
            "Epoch 72/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2464 - accuracy: 0.9538 - val_loss: 0.1927 - val_accuracy: 0.9473\n",
            "Epoch 73/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2478 - accuracy: 0.9523 - val_loss: 0.1934 - val_accuracy: 0.9461\n",
            "Epoch 74/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2566 - accuracy: 0.9508 - val_loss: 0.1970 - val_accuracy: 0.9427\n",
            "Epoch 75/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2495 - accuracy: 0.9522 - val_loss: 0.1934 - val_accuracy: 0.9468\n",
            "Epoch 76/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2470 - accuracy: 0.9537 - val_loss: 0.2010 - val_accuracy: 0.9424\n",
            "Epoch 77/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2430 - accuracy: 0.9537 - val_loss: 0.2014 - val_accuracy: 0.9437\n",
            "Epoch 78/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2442 - accuracy: 0.9533 - val_loss: 0.1966 - val_accuracy: 0.9469\n",
            "Epoch 79/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2380 - accuracy: 0.9558 - val_loss: 0.1994 - val_accuracy: 0.9444\n",
            "Epoch 80/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2458 - accuracy: 0.9523 - val_loss: 0.1931 - val_accuracy: 0.9454\n",
            "Epoch 81/100\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.2468 - accuracy: 0.9538 - val_loss: 0.1965 - val_accuracy: 0.9442\n",
            "Epoch 82/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2518 - accuracy: 0.9541 - val_loss: 0.1970 - val_accuracy: 0.9439\n",
            "Epoch 83/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2525 - accuracy: 0.9515 - val_loss: 0.1980 - val_accuracy: 0.9459\n",
            "Epoch 84/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2408 - accuracy: 0.9541 - val_loss: 0.1989 - val_accuracy: 0.9440\n",
            "Epoch 85/100\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.2445 - accuracy: 0.9547 - val_loss: 0.1940 - val_accuracy: 0.9456\n",
            "Epoch 86/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2393 - accuracy: 0.9559 - val_loss: 0.2037 - val_accuracy: 0.9413\n",
            "Epoch 87/100\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.2545 - accuracy: 0.9499 - val_loss: 0.2026 - val_accuracy: 0.9449\n",
            "Epoch 88/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2459 - accuracy: 0.9537 - val_loss: 0.2001 - val_accuracy: 0.9440\n",
            "Epoch 89/100\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.2431 - accuracy: 0.9531 - val_loss: 0.1946 - val_accuracy: 0.9477\n",
            "Epoch 90/100\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.2507 - accuracy: 0.9510 - val_loss: 0.1968 - val_accuracy: 0.9461\n",
            "Epoch 91/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2443 - accuracy: 0.9542 - val_loss: 0.2037 - val_accuracy: 0.9426\n",
            "Epoch 92/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2520 - accuracy: 0.9537 - val_loss: 0.1967 - val_accuracy: 0.9453\n",
            "Epoch 93/100\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.2427 - accuracy: 0.9556 - val_loss: 0.1973 - val_accuracy: 0.9444\n",
            "Epoch 94/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2443 - accuracy: 0.9531 - val_loss: 0.1986 - val_accuracy: 0.9441\n",
            "Epoch 95/100\n",
            "287/287 [==============================] - 16s 55ms/step - loss: 0.2456 - accuracy: 0.9529 - val_loss: 0.1984 - val_accuracy: 0.9446\n",
            "Epoch 96/100\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.2504 - accuracy: 0.9524 - val_loss: 0.1988 - val_accuracy: 0.9428\n",
            "Epoch 97/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2475 - accuracy: 0.9536 - val_loss: 0.1970 - val_accuracy: 0.9455\n",
            "Epoch 98/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2384 - accuracy: 0.9551 - val_loss: 0.2008 - val_accuracy: 0.9430\n",
            "Epoch 99/100\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.2498 - accuracy: 0.9528 - val_loss: 0.1997 - val_accuracy: 0.9450\n",
            "Epoch 100/100\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2443 - accuracy: 0.9535 - val_loss: 0.2088 - val_accuracy: 0.9422\n",
            "\n",
            "\n",
            "Model Graphs...\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 576x576 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# defining model architecture, subsequent training, and plotting graphs\n",
        "# building the model\n",
        "model = mobilenetv3s.MobileNetV3S.build(width=image_width, height=image_height, depth=3,\n",
        "                                        classes=num_classes, data_aug=False, dense_layer=False)\n",
        "\n",
        "# compiling and summarizing the model\n",
        "model.compile(optimizer=tf.keras.optimizers.Adam(learning_rate=0.001),\n",
        "              loss=tf.keras.losses.CategoricalCrossentropy(from_logits=False),\n",
        "              metrics=['accuracy'])\n",
        "print('Model Summary...')\n",
        "model.summary()\n",
        "\n",
        "# defining callback - checkpoint\n",
        "checkpoint = ModelCheckpoint('Models/MobileNetV3S_Basic/mobilenetv3s_basic_{epoch:02d}_{val_loss:.3f}.h5', monitor='val_loss', \n",
        "                             mode='min', verbose=0, save_freq='epoch')\n",
        "# total epochs\n",
        "epochs = 100\n",
        "\n",
        "# training the model on defined parameters\n",
        "print('\\n\\nModel Training...')\n",
        "history = model.fit(\n",
        "  train_ds, validation_data=val_ds, class_weight=class_weight, epochs=epochs, verbose=1,\n",
        "  batch_size=batch_size, callbacks=[checkpoint] \n",
        ")\n",
        "\n",
        "# plot accuracy and loss graphs of the trained model\n",
        "print('\\n\\nModel Graphs...')\n",
        "fig = helper_functions.plot_graphs(history=history)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Model 2:**\n",
        "\n",
        "      INPUT => MobileNetV3Small => GAP => FC => DO => SOFTMAX\n",
        "\n",
        "Observations:\n",
        "\n",
        "\n",
        "\n",
        "1.   Trainable parameters: 75,146.\n",
        "2.   Best epoch: 7\n",
        "    *   loss: 0.1769 - accuracy: 0.9687 - val_loss: 0.1780 - val_accuracy: 0.9459\n",
        "3.   Overfitting starts from epoch 8.\n",
        "\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "CCyy2f18RCo4"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "J57W-cXaQIef",
        "outputId": "30a4e995-a0f2-401c-9ebe-f565f69fd6c0"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model Summary...\n",
            "Model: \"MobileNetV3Small\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_4 (InputLayer)        [(None, 224, 224, 3)]     0         \n",
            "                                                                 \n",
            " MobilenetV3small (Functiona  (None, 7, 7, 576)        939120    \n",
            " l)                                                              \n",
            "                                                                 \n",
            " global_average_pooling2d_1   (None, 576)              0         \n",
            " (GlobalAveragePooling2D)                                        \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 128)               73856     \n",
            "                                                                 \n",
            " dropout_1 (Dropout)         (None, 128)               0         \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 10)                1290      \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 1,014,266\n",
            "Trainable params: 75,146\n",
            "Non-trainable params: 939,120\n",
            "_________________________________________________________________\n",
            "\n",
            "\n",
            "Model Training...\n",
            "Epoch 1/30\n",
            "287/287 [==============================] - 19s 57ms/step - loss: 0.7041 - accuracy: 0.8795 - val_loss: 0.2117 - val_accuracy: 0.9315\n",
            "Epoch 2/30\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.3610 - accuracy: 0.9389 - val_loss: 0.1905 - val_accuracy: 0.9404\n",
            "Epoch 3/30\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.3069 - accuracy: 0.9473 - val_loss: 0.1844 - val_accuracy: 0.9403\n",
            "Epoch 4/30\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2648 - accuracy: 0.9540 - val_loss: 0.1845 - val_accuracy: 0.9432\n",
            "Epoch 5/30\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.2297 - accuracy: 0.9592 - val_loss: 0.1780 - val_accuracy: 0.9441\n",
            "Epoch 6/30\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.2003 - accuracy: 0.9625 - val_loss: 0.1682 - val_accuracy: 0.9470\n",
            "Epoch 7/30\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.1769 - accuracy: 0.9687 - val_loss: 0.1780 - val_accuracy: 0.9459\n",
            "Epoch 8/30\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.1509 - accuracy: 0.9720 - val_loss: 0.1727 - val_accuracy: 0.9501\n",
            "Epoch 9/30\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.1394 - accuracy: 0.9742 - val_loss: 0.1794 - val_accuracy: 0.9482\n",
            "Epoch 10/30\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.1289 - accuracy: 0.9762 - val_loss: 0.1776 - val_accuracy: 0.9475\n",
            "Epoch 11/30\n",
            "287/287 [==============================] - 16s 55ms/step - loss: 0.1149 - accuracy: 0.9790 - val_loss: 0.1868 - val_accuracy: 0.9488\n",
            "Epoch 12/30\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.0971 - accuracy: 0.9824 - val_loss: 0.1934 - val_accuracy: 0.9487\n",
            "Epoch 13/30\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.0927 - accuracy: 0.9832 - val_loss: 0.1826 - val_accuracy: 0.9501\n",
            "Epoch 14/30\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.0845 - accuracy: 0.9844 - val_loss: 0.1979 - val_accuracy: 0.9511\n",
            "Epoch 15/30\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.0774 - accuracy: 0.9856 - val_loss: 0.1981 - val_accuracy: 0.9496\n",
            "Epoch 16/30\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.0674 - accuracy: 0.9877 - val_loss: 0.2042 - val_accuracy: 0.9507\n",
            "Epoch 17/30\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.0591 - accuracy: 0.9894 - val_loss: 0.2113 - val_accuracy: 0.9489\n",
            "Epoch 18/30\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.0647 - accuracy: 0.9883 - val_loss: 0.2255 - val_accuracy: 0.9484\n",
            "Epoch 19/30\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.0607 - accuracy: 0.9878 - val_loss: 0.2269 - val_accuracy: 0.9446\n",
            "Epoch 20/30\n",
            "287/287 [==============================] - 16s 54ms/step - loss: 0.0492 - accuracy: 0.9902 - val_loss: 0.2166 - val_accuracy: 0.9496\n",
            "Epoch 21/30\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.0488 - accuracy: 0.9909 - val_loss: 0.2279 - val_accuracy: 0.9493\n",
            "Epoch 22/30\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.0373 - accuracy: 0.9930 - val_loss: 0.2335 - val_accuracy: 0.9473\n",
            "Epoch 23/30\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.0499 - accuracy: 0.9903 - val_loss: 0.2319 - val_accuracy: 0.9489\n",
            "Epoch 24/30\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.0348 - accuracy: 0.9936 - val_loss: 0.2407 - val_accuracy: 0.9503\n",
            "Epoch 25/30\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.0369 - accuracy: 0.9932 - val_loss: 0.2400 - val_accuracy: 0.9515\n",
            "Epoch 26/30\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.0377 - accuracy: 0.9927 - val_loss: 0.2482 - val_accuracy: 0.9488\n",
            "Epoch 27/30\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.0364 - accuracy: 0.9932 - val_loss: 0.2597 - val_accuracy: 0.9503\n",
            "Epoch 28/30\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.0261 - accuracy: 0.9952 - val_loss: 0.2654 - val_accuracy: 0.9488\n",
            "Epoch 29/30\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.0362 - accuracy: 0.9933 - val_loss: 0.2786 - val_accuracy: 0.9447\n",
            "Epoch 30/30\n",
            "287/287 [==============================] - 15s 54ms/step - loss: 0.0322 - accuracy: 0.9942 - val_loss: 0.2713 - val_accuracy: 0.9481\n",
            "\n",
            "\n",
            "Model Graphs...\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 576x576 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# defining model architecture, subsequent training, and plotting graphs\n",
        "# building the model\n",
        "model = mobilenetv3s.MobileNetV3S.build(width=image_width, height=image_height, depth=3,\n",
        "                                        classes=num_classes, data_aug=False, dense_layer=True)\n",
        "\n",
        "# compiling and summarizing the model\n",
        "model.compile(optimizer=tf.keras.optimizers.Adam(learning_rate=0.001),\n",
        "              loss=tf.keras.losses.CategoricalCrossentropy(from_logits=False),\n",
        "              metrics=['accuracy'])\n",
        "print('Model Summary...')\n",
        "model.summary()\n",
        "\n",
        "# defining callback - checkpoint\n",
        "checkpoint = ModelCheckpoint('Models/MobileNetV3S_Dense/mobilenetv3s_dense_{epoch:02d}_{val_loss:.3f}.h5', monitor='val_loss', \n",
        "                             mode='min', verbose=0, save_freq='epoch')\n",
        "# total epochs\n",
        "epochs = 30\n",
        "\n",
        "# training the model on defined parameters\n",
        "print('\\n\\nModel Training...')\n",
        "history = model.fit(\n",
        "  train_ds, validation_data=val_ds, class_weight=class_weight, epochs=epochs, verbose=1,\n",
        "  batch_size=batch_size, callbacks=[checkpoint] \n",
        ")\n",
        "\n",
        "# plot accuracy and loss graphs of the trained model\n",
        "print('\\n\\nModel Graphs...')\n",
        "fig = helper_functions.plot_graphs(history=history)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Model 3:**\n",
        "\n",
        "      INPUT => AUGMENTATION => MobileNetV3Small => GAP => DO => SOFTMAX\n",
        "\n",
        "Observations:\n",
        "\n",
        "\n",
        "\n",
        "1.   Trainable parameters: 5,770.\n",
        "2.   Validation loss is always lower than training loss.\n",
        "    *   Probable Reason: Due to the use of data augmentation technique, the model might be seeing harder examples while training. \n",
        "\n",
        "\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "jIlx1jwVRwDW"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "5H6igs8VQMan",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "6e8028d7-0f1d-4551-addd-545dfe442d9b"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model Summary...\n",
            "Model: \"MobileNetV3Small\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_6 (InputLayer)        [(None, 224, 224, 3)]     0         \n",
            "                                                                 \n",
            " sequential_2 (Sequential)   (None, 224, 224, 3)       0         \n",
            "                                                                 \n",
            " MobilenetV3small (Functiona  (None, 7, 7, 576)        939120    \n",
            " l)                                                              \n",
            "                                                                 \n",
            " global_average_pooling2d_2   (None, 576)              0         \n",
            " (GlobalAveragePooling2D)                                        \n",
            "                                                                 \n",
            " dropout_2 (Dropout)         (None, 576)               0         \n",
            "                                                                 \n",
            " dense_3 (Dense)             (None, 10)                5770      \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 944,890\n",
            "Trainable params: 5,770\n",
            "Non-trainable params: 939,120\n",
            "_________________________________________________________________\n",
            "\n",
            "\n",
            "Model Training...\n",
            "Epoch 1/30\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:tensorflow:Using a while_loop for converting RngReadAndSkip\n",
            "WARNING:tensorflow:Using a while_loop for converting Bitcast\n",
            "WARNING:tensorflow:Using a while_loop for converting Bitcast\n",
            "WARNING:tensorflow:Using a while_loop for converting StatelessRandomUniformV2\n",
            "WARNING:tensorflow:Using a while_loop for converting RngReadAndSkip\n",
            "WARNING:tensorflow:Using a while_loop for converting Bitcast\n",
            "WARNING:tensorflow:Using a while_loop for converting Bitcast\n",
            "WARNING:tensorflow:Using a while_loop for converting StatelessRandomUniformV2\n",
            "WARNING:tensorflow:Using a while_loop for converting ImageProjectiveTransformV3\n",
            "WARNING:tensorflow:Using a while_loop for converting RngReadAndSkip\n",
            "WARNING:tensorflow:Using a while_loop for converting Bitcast\n",
            "WARNING:tensorflow:Using a while_loop for converting Bitcast\n",
            "WARNING:tensorflow:Using a while_loop for converting StatelessRandomUniformV2\n",
            "WARNING:tensorflow:Using a while_loop for converting RngReadAndSkip\n",
            "WARNING:tensorflow:Using a while_loop for converting Bitcast\n",
            "WARNING:tensorflow:Using a while_loop for converting Bitcast\n",
            "WARNING:tensorflow:Using a while_loop for converting StatelessRandomUniformV2\n",
            "WARNING:tensorflow:Using a while_loop for converting ImageProjectiveTransformV3\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "287/287 [==============================] - 95s 313ms/step - loss: 1.4805 - accuracy: 0.7483 - val_loss: 0.2826 - val_accuracy: 0.9147\n",
            "Epoch 2/30\n",
            "287/287 [==============================] - 90s 314ms/step - loss: 0.7606 - accuracy: 0.8719 - val_loss: 0.2332 - val_accuracy: 0.9273\n",
            "Epoch 3/30\n",
            "287/287 [==============================] - 89s 310ms/step - loss: 0.6628 - accuracy: 0.8869 - val_loss: 0.2181 - val_accuracy: 0.9298\n",
            "Epoch 4/30\n",
            "287/287 [==============================] - 90s 315ms/step - loss: 0.6345 - accuracy: 0.8886 - val_loss: 0.2020 - val_accuracy: 0.9357\n",
            "Epoch 5/30\n",
            "287/287 [==============================] - 89s 311ms/step - loss: 0.6150 - accuracy: 0.8916 - val_loss: 0.2027 - val_accuracy: 0.9358\n",
            "Epoch 6/30\n",
            "287/287 [==============================] - 90s 313ms/step - loss: 0.5738 - accuracy: 0.8992 - val_loss: 0.2065 - val_accuracy: 0.9323\n",
            "Epoch 7/30\n",
            "287/287 [==============================] - 89s 309ms/step - loss: 0.5666 - accuracy: 0.8995 - val_loss: 0.1895 - val_accuracy: 0.9407\n",
            "Epoch 8/30\n",
            "287/287 [==============================] - 88s 307ms/step - loss: 0.5523 - accuracy: 0.8996 - val_loss: 0.1872 - val_accuracy: 0.9421\n",
            "Epoch 9/30\n",
            "287/287 [==============================] - 88s 307ms/step - loss: 0.5509 - accuracy: 0.8983 - val_loss: 0.1902 - val_accuracy: 0.9388\n",
            "Epoch 10/30\n",
            "287/287 [==============================] - 88s 308ms/step - loss: 0.5376 - accuracy: 0.9020 - val_loss: 0.1821 - val_accuracy: 0.9419\n",
            "Epoch 11/30\n",
            "287/287 [==============================] - 88s 307ms/step - loss: 0.5379 - accuracy: 0.9027 - val_loss: 0.1859 - val_accuracy: 0.9404\n",
            "Epoch 12/30\n",
            "287/287 [==============================] - 88s 307ms/step - loss: 0.5320 - accuracy: 0.9068 - val_loss: 0.2017 - val_accuracy: 0.9344\n",
            "Epoch 13/30\n",
            "287/287 [==============================] - 89s 310ms/step - loss: 0.5258 - accuracy: 0.9048 - val_loss: 0.1808 - val_accuracy: 0.9433\n",
            "Epoch 14/30\n",
            "287/287 [==============================] - 90s 312ms/step - loss: 0.5270 - accuracy: 0.9058 - val_loss: 0.1801 - val_accuracy: 0.9455\n",
            "Epoch 15/30\n",
            "287/287 [==============================] - 89s 311ms/step - loss: 0.5107 - accuracy: 0.9068 - val_loss: 0.1841 - val_accuracy: 0.9427\n",
            "Epoch 16/30\n",
            "287/287 [==============================] - 89s 311ms/step - loss: 0.5287 - accuracy: 0.9040 - val_loss: 0.1804 - val_accuracy: 0.9421\n",
            "Epoch 17/30\n",
            "287/287 [==============================] - 88s 307ms/step - loss: 0.5144 - accuracy: 0.9094 - val_loss: 0.1823 - val_accuracy: 0.9423\n",
            "Epoch 18/30\n",
            "287/287 [==============================] - 88s 307ms/step - loss: 0.5221 - accuracy: 0.9037 - val_loss: 0.1912 - val_accuracy: 0.9408\n",
            "Epoch 19/30\n",
            "287/287 [==============================] - 88s 307ms/step - loss: 0.5072 - accuracy: 0.9102 - val_loss: 0.1892 - val_accuracy: 0.9398\n",
            "Epoch 20/30\n",
            "287/287 [==============================] - 88s 307ms/step - loss: 0.5200 - accuracy: 0.9072 - val_loss: 0.1795 - val_accuracy: 0.9439\n",
            "Epoch 21/30\n",
            "287/287 [==============================] - 88s 307ms/step - loss: 0.5111 - accuracy: 0.9079 - val_loss: 0.1848 - val_accuracy: 0.9410\n",
            "Epoch 22/30\n",
            "287/287 [==============================] - 89s 309ms/step - loss: 0.5150 - accuracy: 0.9062 - val_loss: 0.1759 - val_accuracy: 0.9460\n",
            "Epoch 23/30\n",
            "287/287 [==============================] - 88s 307ms/step - loss: 0.5183 - accuracy: 0.9061 - val_loss: 0.1794 - val_accuracy: 0.9447\n",
            "Epoch 24/30\n",
            "287/287 [==============================] - 88s 306ms/step - loss: 0.5212 - accuracy: 0.9080 - val_loss: 0.1844 - val_accuracy: 0.9428\n",
            "Epoch 25/30\n",
            "287/287 [==============================] - 88s 307ms/step - loss: 0.5178 - accuracy: 0.9070 - val_loss: 0.1829 - val_accuracy: 0.9431\n",
            "Epoch 26/30\n",
            "287/287 [==============================] - 88s 306ms/step - loss: 0.5097 - accuracy: 0.9072 - val_loss: 0.1920 - val_accuracy: 0.9381\n",
            "Epoch 27/30\n",
            "287/287 [==============================] - 88s 307ms/step - loss: 0.5273 - accuracy: 0.9062 - val_loss: 0.1826 - val_accuracy: 0.9413\n",
            "Epoch 28/30\n",
            "287/287 [==============================] - 88s 308ms/step - loss: 0.5139 - accuracy: 0.9076 - val_loss: 0.1894 - val_accuracy: 0.9410\n",
            "Epoch 29/30\n",
            "287/287 [==============================] - 88s 308ms/step - loss: 0.5209 - accuracy: 0.9066 - val_loss: 0.1853 - val_accuracy: 0.9417\n",
            "Epoch 30/30\n",
            "287/287 [==============================] - 88s 307ms/step - loss: 0.5012 - accuracy: 0.9085 - val_loss: 0.1794 - val_accuracy: 0.9439\n",
            "\n",
            "\n",
            "Model Graphs...\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 576x576 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# defining model architecture, subsequent training, and plotting graphs\n",
        "# building the model\n",
        "model = mobilenetv3s.MobileNetV3S.build(width=image_width, height=image_height, depth=3,\n",
        "                                        classes=num_classes, data_aug=True, dense_layer=False)\n",
        "\n",
        "# compiling and summarizing the model\n",
        "model.compile(optimizer=tf.keras.optimizers.Adam(learning_rate=0.001),\n",
        "              loss=tf.keras.losses.CategoricalCrossentropy(from_logits=False),\n",
        "              metrics=['accuracy'])\n",
        "print('Model Summary...')\n",
        "model.summary()\n",
        "\n",
        "# defining callback - checkpoint\n",
        "checkpoint = ModelCheckpoint('Models/MobileNetV3S_Aug/mobilenetv3s_aug_{epoch:02d}_{val_loss:.3f}.h5', monitor='val_loss', \n",
        "                             mode='min', verbose=0, save_freq='epoch')\n",
        "# total epochs\n",
        "epochs = 30\n",
        "\n",
        "# training the model on defined parameters\n",
        "print('\\n\\nModel Training...')\n",
        "history = model.fit(\n",
        "  train_ds, validation_data=val_ds, class_weight=class_weight, epochs=epochs, verbose=1,\n",
        "  batch_size=batch_size, callbacks=[checkpoint] \n",
        ")\n",
        "\n",
        "# plot accuracy and loss graphs of the trained model\n",
        "print('\\n\\nModel Graphs...')\n",
        "fig = helper_functions.plot_graphs(history=history)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Model 4:**\n",
        "\n",
        "      INPUT => AUGMENTATION => MobileNetV3Small => GAP => FC => DO => SOFTMAX\n",
        "\n",
        "Observations:\n",
        "\n",
        "\n",
        "\n",
        "1.   Trainable parameters: 75,146.\n",
        "2.   Validation loss is lower than training loss.\n",
        "    * Trend: Training loss is reducing slowly.\n",
        "    * Training for more epochs may results in a good model.\n",
        "    * Model is taking more time to learn might be due to the fact that it is not too deep or densed, i.e., fewer parameters. \n",
        "\n",
        "\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "Lway27LdSGKO"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# defining model architecture, subsequent training, and plotting graphs\n",
        "# building the model\n",
        "model = mobilenetv3s.MobileNetV3S.build(width=image_width, height=image_height, depth=3,\n",
        "                                        classes=num_classes, data_aug=True, dense_layer=True)\n",
        "\n",
        "# compiling and summarizing the model\n",
        "model.compile(optimizer=tf.keras.optimizers.Adam(learning_rate=0.001),\n",
        "              loss=tf.keras.losses.CategoricalCrossentropy(from_logits=False),\n",
        "              metrics=['accuracy'])\n",
        "print('Model Summary...')\n",
        "model.summary()\n",
        "\n",
        "# defining callback - checkpoint\n",
        "checkpoint = ModelCheckpoint('Models/MobileNetV3S_Aug_Dense/mobilenetv3s_aug_dense_{epoch:02d}_{val_loss:.3f}.h5', monitor='val_loss', \n",
        "                             mode='min', verbose=0, save_freq='epoch')\n",
        "# total epochs\n",
        "epochs = 25\n",
        "\n",
        "# training the model on defined parameters\n",
        "print('\\n\\nModel Training...')\n",
        "history = model.fit(\n",
        "  train_ds, validation_data=val_ds, class_weight=class_weight, epochs=epochs, verbose=1,\n",
        "  batch_size=batch_size, callbacks=[checkpoint] \n",
        ")\n",
        "\n",
        "# plot accuracy and loss graphs of the trained model\n",
        "print('\\n\\nModel Graphs...')\n",
        "fig = helper_functions.plot_graphs(history=history)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "_Sj_cqs1Rq57",
        "outputId": "7c190524-33c5-45b7-ecb2-94b0bc940964"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model Summary...\n",
            "Model: \"MobileNetV3Small\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_10 (InputLayer)       [(None, 224, 224, 3)]     0         \n",
            "                                                                 \n",
            " sequential_4 (Sequential)   (None, 224, 224, 3)       0         \n",
            "                                                                 \n",
            " MobilenetV3small (Functiona  (None, 7, 7, 576)        939120    \n",
            " l)                                                              \n",
            "                                                                 \n",
            " global_average_pooling2d_4   (None, 576)              0         \n",
            " (GlobalAveragePooling2D)                                        \n",
            "                                                                 \n",
            " dense_6 (Dense)             (None, 128)               73856     \n",
            "                                                                 \n",
            " dropout_4 (Dropout)         (None, 128)               0         \n",
            "                                                                 \n",
            " dense_7 (Dense)             (None, 10)                1290      \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 1,014,266\n",
            "Trainable params: 75,146\n",
            "Non-trainable params: 939,120\n",
            "_________________________________________________________________\n",
            "\n",
            "\n",
            "Model Training...\n",
            "Epoch 1/25\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:tensorflow:Using a while_loop for converting RngReadAndSkip\n",
            "WARNING:tensorflow:Using a while_loop for converting Bitcast\n",
            "WARNING:tensorflow:Using a while_loop for converting Bitcast\n",
            "WARNING:tensorflow:Using a while_loop for converting StatelessRandomUniformV2\n",
            "WARNING:tensorflow:Using a while_loop for converting RngReadAndSkip\n",
            "WARNING:tensorflow:Using a while_loop for converting Bitcast\n",
            "WARNING:tensorflow:Using a while_loop for converting Bitcast\n",
            "WARNING:tensorflow:Using a while_loop for converting StatelessRandomUniformV2\n",
            "WARNING:tensorflow:Using a while_loop for converting ImageProjectiveTransformV3\n",
            "WARNING:tensorflow:Using a while_loop for converting RngReadAndSkip\n",
            "WARNING:tensorflow:Using a while_loop for converting Bitcast\n",
            "WARNING:tensorflow:Using a while_loop for converting Bitcast\n",
            "WARNING:tensorflow:Using a while_loop for converting StatelessRandomUniformV2\n",
            "WARNING:tensorflow:Using a while_loop for converting RngReadAndSkip\n",
            "WARNING:tensorflow:Using a while_loop for converting Bitcast\n",
            "WARNING:tensorflow:Using a while_loop for converting Bitcast\n",
            "WARNING:tensorflow:Using a while_loop for converting StatelessRandomUniformV2\n",
            "WARNING:tensorflow:Using a while_loop for converting ImageProjectiveTransformV3\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "287/287 [==============================] - 95s 315ms/step - loss: 0.9782 - accuracy: 0.8291 - val_loss: 0.2297 - val_accuracy: 0.9241\n",
            "Epoch 2/25\n",
            "287/287 [==============================] - 90s 312ms/step - loss: 0.6113 - accuracy: 0.8938 - val_loss: 0.1937 - val_accuracy: 0.9371\n",
            "Epoch 3/25\n",
            "287/287 [==============================] - 89s 312ms/step - loss: 0.5397 - accuracy: 0.9059 - val_loss: 0.1821 - val_accuracy: 0.9424\n",
            "Epoch 4/25\n",
            "287/287 [==============================] - 89s 311ms/step - loss: 0.5188 - accuracy: 0.9060 - val_loss: 0.1759 - val_accuracy: 0.9436\n",
            "Epoch 5/25\n",
            "287/287 [==============================] - 90s 314ms/step - loss: 0.4839 - accuracy: 0.9130 - val_loss: 0.1969 - val_accuracy: 0.9363\n",
            "Epoch 6/25\n",
            "287/287 [==============================] - 89s 311ms/step - loss: 0.4639 - accuracy: 0.9164 - val_loss: 0.1895 - val_accuracy: 0.9388\n",
            "Epoch 7/25\n",
            "287/287 [==============================] - 89s 310ms/step - loss: 0.4403 - accuracy: 0.9205 - val_loss: 0.1893 - val_accuracy: 0.9395\n",
            "Epoch 8/25\n",
            "287/287 [==============================] - 89s 311ms/step - loss: 0.4297 - accuracy: 0.9222 - val_loss: 0.1741 - val_accuracy: 0.9445\n",
            "Epoch 9/25\n",
            "287/287 [==============================] - 90s 314ms/step - loss: 0.4093 - accuracy: 0.9256 - val_loss: 0.1695 - val_accuracy: 0.9454\n",
            "Epoch 10/25\n",
            "287/287 [==============================] - 89s 311ms/step - loss: 0.3987 - accuracy: 0.9274 - val_loss: 0.1757 - val_accuracy: 0.9439\n",
            "Epoch 11/25\n",
            "287/287 [==============================] - 89s 311ms/step - loss: 0.3868 - accuracy: 0.9274 - val_loss: 0.1672 - val_accuracy: 0.9481\n",
            "Epoch 12/25\n",
            "287/287 [==============================] - 89s 311ms/step - loss: 0.3726 - accuracy: 0.9297 - val_loss: 0.1773 - val_accuracy: 0.9441\n",
            "Epoch 13/25\n",
            "287/287 [==============================] - 89s 311ms/step - loss: 0.3783 - accuracy: 0.9299 - val_loss: 0.1751 - val_accuracy: 0.9459\n",
            "Epoch 14/25\n",
            "287/287 [==============================] - 89s 311ms/step - loss: 0.3629 - accuracy: 0.9311 - val_loss: 0.1747 - val_accuracy: 0.9451\n",
            "Epoch 15/25\n",
            "287/287 [==============================] - 89s 310ms/step - loss: 0.3611 - accuracy: 0.9334 - val_loss: 0.1736 - val_accuracy: 0.9464\n",
            "Epoch 16/25\n",
            "287/287 [==============================] - 90s 313ms/step - loss: 0.3498 - accuracy: 0.9365 - val_loss: 0.1803 - val_accuracy: 0.9467\n",
            "Epoch 17/25\n",
            "287/287 [==============================] - 89s 311ms/step - loss: 0.3479 - accuracy: 0.9362 - val_loss: 0.1681 - val_accuracy: 0.9501\n",
            "Epoch 18/25\n",
            "287/287 [==============================] - 89s 310ms/step - loss: 0.3297 - accuracy: 0.9380 - val_loss: 0.1721 - val_accuracy: 0.9493\n",
            "Epoch 19/25\n",
            "287/287 [==============================] - 89s 310ms/step - loss: 0.3315 - accuracy: 0.9386 - val_loss: 0.1655 - val_accuracy: 0.9497\n",
            "Epoch 20/25\n",
            "287/287 [==============================] - 90s 312ms/step - loss: 0.3178 - accuracy: 0.9401 - val_loss: 0.1657 - val_accuracy: 0.9506\n",
            "Epoch 21/25\n",
            "287/287 [==============================] - 90s 313ms/step - loss: 0.3153 - accuracy: 0.9408 - val_loss: 0.1641 - val_accuracy: 0.9507\n",
            "Epoch 22/25\n",
            "287/287 [==============================] - 89s 311ms/step - loss: 0.3178 - accuracy: 0.9422 - val_loss: 0.1781 - val_accuracy: 0.9449\n",
            "Epoch 23/25\n",
            "287/287 [==============================] - 90s 313ms/step - loss: 0.3062 - accuracy: 0.9426 - val_loss: 0.1708 - val_accuracy: 0.9488\n",
            "Epoch 24/25\n",
            "287/287 [==============================] - 90s 313ms/step - loss: 0.3090 - accuracy: 0.9412 - val_loss: 0.1694 - val_accuracy: 0.9486\n",
            "Epoch 25/25\n",
            "287/287 [==============================] - 90s 313ms/step - loss: 0.3027 - accuracy: 0.9422 - val_loss: 0.1660 - val_accuracy: 0.9514\n",
            "\n",
            "\n",
            "Model Graphs...\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 576x576 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Conclusion:**\n",
        "\n",
        "\n",
        "\n",
        "1.   Epoch 7 of Model 2 gives the best results. However, the results are not as good as given by best model from MobileNetV3Large.\n",
        "2.   If model is taking more time to learn, then it might be due to the fact that it is not too deep or densed, i.e., fewer parameters."
      ],
      "metadata": {
        "id": "ZKdKKD2-Ssvy"
      }
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "machine_shape": "hm",
      "provenance": []
    },
    "gpuClass": "standard",
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}