2_1_train_age_model.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "machine_shape": "hm"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jmgbfqAKOsxR"
      },
      "source": [
        "# Plan of Action"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RxEzjKuOPr5n"
      },
      "source": [
        "**Our age prediction CNN model shall be defined and trained by**:\n",
        "1. Importing **training and test datasets** from Google Drive Input Sub-folder\n",
        "2. **Training dataset is already augmented** and has 234,000 images\n",
        "3. **Greyscaling images** instead of using RGB color images\n",
        "4. Defining our intuitively **distributed classes of age-ranges**\n",
        "5. Using **60 epochs** on our **optimized CNN Architecture**, comprising of:\n",
        "    - an input *Conv2D* layer (with 32 filters) paired with an *AveragePooling2D* layer,\n",
        "    - 3 pairs of *Conv2D* (with 64, 128 & 256 filters) and *AveragePooling2D* layers,\n",
        "    - a *GlobalAveragePooling2D* layer,\n",
        "    - 1 *Dense* layer with 132 nodes, and\n",
        "    - an output *Dense* layer with 7 nodes."
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Mount Google Drive & Imports"
      ],
      "metadata": {
        "id": "io9rnPUceyKx"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ndBDuhivUu7C",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "db0c3318-eb11-4fb3-a720-45414585010e"
      },
      "source": [
        "#@title Mount Google Drive {display-mode: \"form\"}\n",
        "\n",
        "# This code will be hidden when the notebook is loaded.\n",
        "\n",
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_kTFrNNPU7wP"
      },
      "source": [
        "# Imports\n",
        "\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "import cv2\n",
        "import os\n",
        "from zipfile import ZipFile\n",
        "import time\n",
        "from datetime import datetime\n",
        "import itertools\n",
        "\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import confusion_matrix\n",
        "\n",
        "import tensorflow as tf\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import Dense\n",
        "from tensorflow.keras.layers import Conv2D, AveragePooling2D, GlobalAveragePooling2D\n",
        "from tensorflow.keras import utils\n",
        "from tensorflow.keras.callbacks import TensorBoard, ModelCheckpoint\n",
        "\n",
        "# Setting random seeds to reduce the amount of randomness in the neural net weights and results\n",
        "# The results may still not be exactly reproducible\n",
        "\n",
        "np.random.seed(42)\n",
        "tf.random.set_seed(42)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ggkatHAQU75X",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8dbff126-b83b-4f9f-983b-0cfdbd2ab7f0",
        "cellView": "form"
      },
      "source": [
        "#@title Check for GPU\n",
        "\n",
        "# Testing to ensure GPU is being utilized\n",
        "# Ensure that the Runtime Type for this notebook is set to GPU\n",
        "# If a GPU device is not found, change the runtime type under: Runtime>> Change runtime type>> Hardware accelerator>> GPU\n",
        "# and run the notebook from the beginning again.\n",
        "\n",
        "device_name = tf.test.gpu_device_name()\n",
        "if device_name != '/device:GPU:0':\n",
        "    raise SystemError('GPU device not found')\n",
        "print('Found GPU at: {}'.format(device_name))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found GPU at: /device:GPU:0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Data Preparation"
      ],
      "metadata": {
        "id": "xmhd2RT_fIaS"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Import Dataset"
      ],
      "metadata": {
        "id": "5P0vu49KbWXH"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tCN57I6AU7_g",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d1539ac1-8bd3-4d87-c11a-37f01f066e60"
      },
      "source": [
        "# Unzipping the dataset file combined_faces.zip\n",
        "\n",
        "combined_faces_zip_path = \"/content/combined_faces (1).zip\"\n",
        "\n",
        "with ZipFile(combined_faces_zip_path, 'r') as myzip:\n",
        "    myzip.extractall()\n",
        "    print('Done unzipping combined_faces.zip')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Done unzipping combined_faces.zip\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "q61d9-wxU8Ea",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "9f8a6897-92ba-4594-96da-c3d3747264d1"
      },
      "source": [
        "# Unzipping the dataset file combined_faces.zip(augumented wala)\n",
        "\n",
        "combined_faces_zip_path = \"\"\n",
        "\n",
        "with ZipFile(combined_faces_zip_path, 'r') as myzip:\n",
        "    myzip.extractall()\n",
        "    print('Done unzipping combined_faces_train_augmented.zip')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Done unzipping combined_faces_train_augmented.zip\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Yu50xj5jU8I6"
      },
      "source": [
        "# Importing the augmented training dataset and testing dataset to create tensors of images using the filename paths.\n",
        "\n",
        "train_aug_df = pd.read_csv(\"/content/images_filenames_labels_train_augmented.csv\")\n",
        "test_df = pd.read_csv(\"/content/images_filenames_labels_test.csv\")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Know your Dataset"
      ],
      "metadata": {
        "id": "C6lsicAOcbYZ"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "80VuyS3lOZub",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "766ee1a8-2a3b-4557-9407-f866336ce1be"
      },
      "source": [
        "train_aug_df.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-c5f4d58a-7e74-49cb-8d6a-5b2c11c63276\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>filename</th>\n",
              "      <th>age</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>/content/content/combined_faces_train_augmente...</td>\n",
              "      <td>43</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>/content/content/combined_faces_train_augmente...</td>\n",
              "      <td>71</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>/content/content/combined_faces_train_augmente...</td>\n",
              "      <td>6</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>/content/content/combined_faces_train_augmente...</td>\n",
              "      <td>18</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>/content/content/combined_faces_train_augmente...</td>\n",
              "      <td>25</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-c5f4d58a-7e74-49cb-8d6a-5b2c11c63276')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-c5f4d58a-7e74-49cb-8d6a-5b2c11c63276 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-c5f4d58a-7e74-49cb-8d6a-5b2c11c63276');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ],
            "text/plain": [
              "                                            filename  age\n",
              "0  /content/content/combined_faces_train_augmente...   43\n",
              "1  /content/content/combined_faces_train_augmente...   71\n",
              "2  /content/content/combined_faces_train_augmente...    6\n",
              "3  /content/content/combined_faces_train_augmente...   18\n",
              "4  /content/content/combined_faces_train_augmente...   25"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EBcZJQchO-JH",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "7ae079e4-efd7-4078-cece-dfbf1216b579"
      },
      "source": [
        "test_df.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-a5b7bba2-520c-4cfa-a0e4-519a53084644\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>filename</th>\n",
              "      <th>age</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>/content/content/combined_faces/8_163.jpg</td>\n",
              "      <td>8</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>/content/content/combined_faces/38_66.jpg</td>\n",
              "      <td>38</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>/content/content/combined_faces/40_177.jpg</td>\n",
              "      <td>40</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>/content/content/combined_faces/36_267.jpg</td>\n",
              "      <td>36</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>/content/content/combined_faces/8_349.jpg</td>\n",
              "      <td>8</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-a5b7bba2-520c-4cfa-a0e4-519a53084644')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-a5b7bba2-520c-4cfa-a0e4-519a53084644 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-a5b7bba2-520c-4cfa-a0e4-519a53084644');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ],
            "text/plain": [
              "                                     filename  age\n",
              "0   /content/content/combined_faces/8_163.jpg    8\n",
              "1   /content/content/combined_faces/38_66.jpg   38\n",
              "2  /content/content/combined_faces/40_177.jpg   40\n",
              "3  /content/content/combined_faces/36_267.jpg   36\n",
              "4   /content/content/combined_faces/8_349.jpg    8"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5D3RBtwwU8W1",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "76b711e5-c5d4-4fe5-cded-c34d98fd56d5"
      },
      "source": [
        "train_aug_df.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(234400, 2)"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "J4Qz1-9mU8bN",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ebf922a5-4532-41e2-cdcc-2238434a9b15"
      },
      "source": [
        "test_df.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(10046, 2)"
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Define Age Classes"
      ],
      "metadata": {
        "id": "BT0E02E3cipb"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QqlcD7x6Yjcd"
      },
      "source": [
        "# Defining a function to return the class labels corresponding to the re-distributed 7 age-ranges.\n",
        "\n",
        "def class_labels_reassign(age):\n",
        "\n",
        "    if 1 <= age <= 2:\n",
        "        return 0\n",
        "    elif 3 <= age <= 9:\n",
        "        return 1\n",
        "    elif 10 <= age <= 20:\n",
        "        return 2\n",
        "    elif 21 <= age <= 27:\n",
        "        return 3\n",
        "    elif 28 <= age <= 45:\n",
        "        return 4\n",
        "    elif 46 <= age <= 65:\n",
        "        return 5\n",
        "    else:\n",
        "        return 6"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LROsGcIaYjVr"
      },
      "source": [
        "train_aug_df['target'] = train_aug_df['age'].map(class_labels_reassign)\n",
        "test_df['target'] = test_df['age'].map(class_labels_reassign)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "j51hK9VIaxo1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "91aaa882-14ab-4e4d-cdb8-2163fdba8f30"
      },
      "source": [
        "train_aug_df.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-412c0a6c-a8cc-425b-9540-9e865c95d73d\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>filename</th>\n",
              "      <th>age</th>\n",
              "      <th>target</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>/content/content/combined_faces_train_augmente...</td>\n",
              "      <td>43</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>/content/content/combined_faces_train_augmente...</td>\n",
              "      <td>71</td>\n",
              "      <td>6</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>/content/content/combined_faces_train_augmente...</td>\n",
              "      <td>6</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>/content/content/combined_faces_train_augmente...</td>\n",
              "      <td>18</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>/content/content/combined_faces_train_augmente...</td>\n",
              "      <td>25</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-412c0a6c-a8cc-425b-9540-9e865c95d73d')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-412c0a6c-a8cc-425b-9540-9e865c95d73d button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-412c0a6c-a8cc-425b-9540-9e865c95d73d');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ],
            "text/plain": [
              "                                            filename  age  target\n",
              "0  /content/content/combined_faces_train_augmente...   43       4\n",
              "1  /content/content/combined_faces_train_augmente...   71       6\n",
              "2  /content/content/combined_faces_train_augmente...    6       1\n",
              "3  /content/content/combined_faces_train_augmente...   18       2\n",
              "4  /content/content/combined_faces_train_augmente...   25       3"
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "csZb53SYaxgu",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "d97c1e7e-df0d-41a3-92be-ece9aa22715d"
      },
      "source": [
        "test_df.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-1069ea9e-e3d5-4fef-bf42-a2323f7fbc0c\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>filename</th>\n",
              "      <th>age</th>\n",
              "      <th>target</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>/content/content/combined_faces/8_163.jpg</td>\n",
              "      <td>8</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>/content/content/combined_faces/38_66.jpg</td>\n",
              "      <td>38</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>/content/content/combined_faces/40_177.jpg</td>\n",
              "      <td>40</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>/content/content/combined_faces/36_267.jpg</td>\n",
              "      <td>36</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>/content/content/combined_faces/8_349.jpg</td>\n",
              "      <td>8</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-1069ea9e-e3d5-4fef-bf42-a2323f7fbc0c')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-1069ea9e-e3d5-4fef-bf42-a2323f7fbc0c button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-1069ea9e-e3d5-4fef-bf42-a2323f7fbc0c');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ],
            "text/plain": [
              "                                     filename  age  target\n",
              "0   /content/content/combined_faces/8_163.jpg    8       1\n",
              "1   /content/content/combined_faces/38_66.jpg   38       4\n",
              "2  /content/content/combined_faces/40_177.jpg   40       4\n",
              "3  /content/content/combined_faces/36_267.jpg   36       4\n",
              "4   /content/content/combined_faces/8_349.jpg    8       1"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Organize Dataset"
      ],
      "metadata": {
        "id": "s52pb794dHhn"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sJSY34ShavL_"
      },
      "source": [
        "# Converting the filenames and target class labels into lists for augmented train and test datasets.\n",
        "\n",
        "train_aug_filenames_list = list(train_aug_df['filename'])\n",
        "train_aug_labels_list = list(train_aug_df['target'])\n",
        "\n",
        "test_filenames_list = list(test_df['filename'])\n",
        "test_labels_list = list(test_df['target'])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "S7ye8s5savP5"
      },
      "source": [
        "# Creating tensorflow constants of filenames and labels for augmented train and test datasets from the lists defined above.\n",
        "\n",
        "train_aug_filenames_tensor = tf.constant(train_aug_filenames_list)\n",
        "train_aug_labels_tensor = tf.constant(train_aug_labels_list)\n",
        "\n",
        "test_filenames_tensor = tf.constant(test_filenames_list)\n",
        "test_labels_tensor = tf.constant(test_labels_list)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Image Greyscale Function"
      ],
      "metadata": {
        "id": "FIUoR5YidZMA"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4ALPko3Xavbo"
      },
      "source": [
        "# Defining a function to read the image, decode the image from given tensor and one-hot encode the image label class.\n",
        "# Changing the channels para in tf.io.decode_jpeg from 3 to 1 changes the output images from RGB coloured to grayscale.\n",
        "\n",
        "num_classes = 7\n",
        "\n",
        "def _parse_function(filename, label):\n",
        "\n",
        "    image_string = tf.io.read_file(filename)\n",
        "    image_decoded = tf.io.decode_jpeg(image_string, channels=1)    # channels=1 to convert to grayscale, channels=3 to convert to RGB.\n",
        "    # image_resized = tf.image.resize(image_decoded, [200, 200])\n",
        "    label = tf.one_hot(label, num_classes)\n",
        "\n",
        "    return image_decoded, label"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OTyXZ2tUavZh"
      },
      "source": [
        "# Getting the dataset ready for the neural network.\n",
        "# Using the tensor vectors defined above, accessing the images in the dataset and passing them through the function defined above.\n",
        "\n",
        "train_aug_dataset = tf.data.Dataset.from_tensor_slices((train_aug_filenames_tensor, train_aug_labels_tensor))\n",
        "train_aug_dataset = train_aug_dataset.map(_parse_function)\n",
        "# train_aug_dataset = train_aug_dataset.repeat(3)\n",
        "train_aug_dataset = train_aug_dataset.batch(512)    # Same as batch_size hyperparameter in model.fit() below.\n",
        "\n",
        "test_dataset = tf.data.Dataset.from_tensor_slices((test_filenames_tensor, test_labels_tensor))\n",
        "test_dataset = test_dataset.map(_parse_function)\n",
        "# test_dataset = test_dataset.repeat(3)\n",
        "test_dataset = test_dataset.batch(512)    # Same as batch_size hyperparameter in model.fit() below."
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# CNN Architecture"
      ],
      "metadata": {
        "id": "TSLahA9adp-o"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kcW44JspavWb",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "200d9ea4-2b9d-43ed-ab91-7f649182a40e"
      },
      "source": [
        "# Defining the architecture of the sequential neural network.\n",
        "\n",
        "final_cnn = Sequential()\n",
        "\n",
        "# Input layer with 32 filters, followed by an AveragePooling2D layer.\n",
        "final_cnn.add(Conv2D(filters=32, kernel_size=3, activation='relu', input_shape=(200, 200, 1)))    # 3rd dim = 1 for grayscale images.\n",
        "final_cnn.add(AveragePooling2D(pool_size=(2,2)))\n",
        "\n",
        "# Three Conv2D layers with filters increasing by a factor of 2 for every successive Conv2D layer.\n",
        "final_cnn.add(Conv2D(filters=64, kernel_size=3, activation='relu'))\n",
        "final_cnn.add(AveragePooling2D(pool_size=(2,2)))\n",
        "\n",
        "final_cnn.add(Conv2D(filters=128, kernel_size=3, activation='relu'))\n",
        "final_cnn.add(AveragePooling2D(pool_size=(2,2)))\n",
        "\n",
        "final_cnn.add(Conv2D(filters=256, kernel_size=3, activation='relu'))\n",
        "final_cnn.add(AveragePooling2D(pool_size=(2,2)))\n",
        "\n",
        "# A GlobalAveragePooling2D layer before going into Dense layers below.\n",
        "# GlobalAveragePooling2D layer gives no. of outputs equal to no. of filters in last Conv2D layer above (256).\n",
        "final_cnn.add(GlobalAveragePooling2D())\n",
        "\n",
        "# One Dense layer with 132 nodes so as to taper down the no. of nodes from no. of outputs of GlobalAveragePooling2D layer above towards no. of nodes in output layer below (7).\n",
        "final_cnn.add(Dense(132, activation='relu'))\n",
        "\n",
        "# Output layer with 7 nodes (equal to the no. of classes).\n",
        "final_cnn.add(Dense(7, activation='softmax'))\n",
        "\n",
        "final_cnn.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " conv2d (Conv2D)             (None, 198, 198, 32)      320       \n",
            "                                                                 \n",
            " average_pooling2d (AverageP  (None, 99, 99, 32)       0         \n",
            " ooling2D)                                                       \n",
            "                                                                 \n",
            " conv2d_1 (Conv2D)           (None, 97, 97, 64)        18496     \n",
            "                                                                 \n",
            " average_pooling2d_1 (Averag  (None, 48, 48, 64)       0         \n",
            " ePooling2D)                                                     \n",
            "                                                                 \n",
            " conv2d_2 (Conv2D)           (None, 46, 46, 128)       73856     \n",
            "                                                                 \n",
            " average_pooling2d_2 (Averag  (None, 23, 23, 128)      0         \n",
            " ePooling2D)                                                     \n",
            "                                                                 \n",
            " conv2d_3 (Conv2D)           (None, 21, 21, 256)       295168    \n",
            "                                                                 \n",
            " average_pooling2d_3 (Averag  (None, 10, 10, 256)      0         \n",
            " ePooling2D)                                                     \n",
            "                                                                 \n",
            " global_average_pooling2d (G  (None, 256)              0         \n",
            " lobalAveragePooling2D)                                          \n",
            "                                                                 \n",
            " dense (Dense)               (None, 132)               33924     \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 7)                 931       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 422,695\n",
            "Trainable params: 422,695\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uBG5MY8zavTn"
      },
      "source": [
        "# Compiling the above created CNN architecture.\n",
        "\n",
        "final_cnn.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nWy7fRfHfu8L"
      },
      "source": [
        "# Creating a TensorBoard callback object and saving it at the desired location.\n",
        "\n",
        "tensorboard = TensorBoard(log_dir=f\"/content/drive/My Drive/1_LiveProjects/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs\")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3pKfqLLbS3ld"
      },
      "source": [
        "We shall also use ***ModelCheckpoint*** as a callback while training the final CNN model so as to be able to save the model as it continues training and improving in performance over 60 epochs."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "X6dQDDhmfuzo"
      },
      "source": [
        "# Creating a ModelCheckpoint callback object to save the model according to the value of val_accuracy.\n",
        "\n",
        "checkpoint = ModelCheckpoint(filepath=f\"/content/drive/My Drive/1_LiveProjects/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\",\n",
        "                             monitor='val_accuracy',\n",
        "                             save_best_only=True,\n",
        "                             save_weights_only=False,\n",
        "                             verbose=1\n",
        "                            )"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Model Fitting"
      ],
      "metadata": {
        "id": "jxVlNLLMeLWw"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hU4LJT6xYjJp",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "29700e2d-b475-4995-d6b9-25cbe193b5c7"
      },
      "source": [
        "# Fitting the above created CNN model.\n",
        "\n",
        "final_cnn_history = final_cnn.fit(train_aug_dataset,\n",
        "                                  batch_size=512,\n",
        "                                  validation_data=test_dataset,\n",
        "                                  epochs=60,\n",
        "                                  callbacks=[tensorboard, checkpoint],\n",
        "                                  shuffle=False    # shuffle=False to reduce randomness and increase reproducibility\n",
        "                                 )"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 1.8014 - accuracy: 0.3014\n",
            "Epoch 00001: val_accuracy improved from -inf to 0.35736, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 159s 320ms/step - loss: 1.8014 - accuracy: 0.3014 - val_loss: 1.6051 - val_accuracy: 0.3574\n",
            "Epoch 2/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 1.4575 - accuracy: 0.4115\n",
            "Epoch 00002: val_accuracy improved from 0.35736 to 0.45331, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 297ms/step - loss: 1.4575 - accuracy: 0.4115 - val_loss: 1.3222 - val_accuracy: 0.4533\n",
            "Epoch 3/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 1.2988 - accuracy: 0.4691\n",
            "Epoch 00003: val_accuracy did not improve from 0.45331\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 1.2988 - accuracy: 0.4691 - val_loss: 1.3976 - val_accuracy: 0.4119\n",
            "Epoch 4/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 1.2145 - accuracy: 0.4995\n",
            "Epoch 00004: val_accuracy improved from 0.45331 to 0.54649, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 1.2145 - accuracy: 0.4995 - val_loss: 1.0892 - val_accuracy: 0.5465\n",
            "Epoch 5/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 1.2101 - accuracy: 0.5033\n",
            "Epoch 00005: val_accuracy improved from 0.54649 to 0.56122, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 1.2101 - accuracy: 0.5033 - val_loss: 1.0464 - val_accuracy: 0.5612\n",
            "Epoch 6/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 1.0598 - accuracy: 0.5577\n",
            "Epoch 00006: val_accuracy did not improve from 0.56122\n",
            "458/458 [==============================] - 135s 296ms/step - loss: 1.0598 - accuracy: 0.5577 - val_loss: 1.0446 - val_accuracy: 0.5515\n",
            "Epoch 7/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 1.0022 - accuracy: 0.5803\n",
            "Epoch 00007: val_accuracy improved from 0.56122 to 0.57963, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 1.0022 - accuracy: 0.5803 - val_loss: 0.9833 - val_accuracy: 0.5796\n",
            "Epoch 8/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.9537 - accuracy: 0.5984\n",
            "Epoch 00008: val_accuracy improved from 0.57963 to 0.61139, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 297ms/step - loss: 0.9537 - accuracy: 0.5984 - val_loss: 0.9137 - val_accuracy: 0.6114\n",
            "Epoch 9/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.9180 - accuracy: 0.6126\n",
            "Epoch 00009: val_accuracy improved from 0.61139 to 0.62463, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.9180 - accuracy: 0.6126 - val_loss: 0.8835 - val_accuracy: 0.6246\n",
            "Epoch 10/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.8856 - accuracy: 0.6266\n",
            "Epoch 00010: val_accuracy improved from 0.62463 to 0.63558, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.8856 - accuracy: 0.6266 - val_loss: 0.8617 - val_accuracy: 0.6356\n",
            "Epoch 11/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.8601 - accuracy: 0.6373\n",
            "Epoch 00011: val_accuracy did not improve from 0.63558\n",
            "458/458 [==============================] - 135s 295ms/step - loss: 0.8601 - accuracy: 0.6373 - val_loss: 0.8578 - val_accuracy: 0.6353\n",
            "Epoch 12/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.8333 - accuracy: 0.6484\n",
            "Epoch 00012: val_accuracy improved from 0.63558 to 0.65489, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.8333 - accuracy: 0.6484 - val_loss: 0.8181 - val_accuracy: 0.6549\n",
            "Epoch 13/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.8082 - accuracy: 0.6586\n",
            "Epoch 00013: val_accuracy improved from 0.65489 to 0.66016, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.8082 - accuracy: 0.6586 - val_loss: 0.8037 - val_accuracy: 0.6602\n",
            "Epoch 14/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.7901 - accuracy: 0.6658\n",
            "Epoch 00014: val_accuracy improved from 0.66016 to 0.67360, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 135s 296ms/step - loss: 0.7901 - accuracy: 0.6658 - val_loss: 0.7772 - val_accuracy: 0.6736\n",
            "Epoch 15/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.7690 - accuracy: 0.6758\n",
            "Epoch 00015: val_accuracy improved from 0.67360 to 0.67370, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.7690 - accuracy: 0.6758 - val_loss: 0.7755 - val_accuracy: 0.6737\n",
            "Epoch 16/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.7440 - accuracy: 0.6859\n",
            "Epoch 00016: val_accuracy improved from 0.67370 to 0.68246, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 135s 296ms/step - loss: 0.7440 - accuracy: 0.6859 - val_loss: 0.7635 - val_accuracy: 0.6825\n",
            "Epoch 17/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.7231 - accuracy: 0.6945\n",
            "Epoch 00017: val_accuracy did not improve from 0.68246\n",
            "458/458 [==============================] - 135s 295ms/step - loss: 0.7231 - accuracy: 0.6945 - val_loss: 0.7631 - val_accuracy: 0.6813\n",
            "Epoch 18/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.7097 - accuracy: 0.7005\n",
            "Epoch 00018: val_accuracy did not improve from 0.68246\n",
            "458/458 [==============================] - 135s 295ms/step - loss: 0.7097 - accuracy: 0.7005 - val_loss: 0.7584 - val_accuracy: 0.6819\n",
            "Epoch 19/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.6917 - accuracy: 0.7083\n",
            "Epoch 00019: val_accuracy improved from 0.68246 to 0.71163, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 135s 296ms/step - loss: 0.6917 - accuracy: 0.7083 - val_loss: 0.6939 - val_accuracy: 0.7116\n",
            "Epoch 20/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.6715 - accuracy: 0.7170\n",
            "Epoch 00020: val_accuracy improved from 0.71163 to 0.71421, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.6715 - accuracy: 0.7170 - val_loss: 0.6966 - val_accuracy: 0.7142\n",
            "Epoch 21/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.6511 - accuracy: 0.7259\n",
            "Epoch 00021: val_accuracy did not improve from 0.71421\n",
            "458/458 [==============================] - 135s 295ms/step - loss: 0.6511 - accuracy: 0.7259 - val_loss: 0.6977 - val_accuracy: 0.7111\n",
            "Epoch 22/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.6357 - accuracy: 0.7327\n",
            "Epoch 00022: val_accuracy did not improve from 0.71421\n",
            "458/458 [==============================] - 135s 295ms/step - loss: 0.6357 - accuracy: 0.7327 - val_loss: 0.7044 - val_accuracy: 0.7125\n",
            "Epoch 23/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.6182 - accuracy: 0.7404\n",
            "Epoch 00023: val_accuracy improved from 0.71421 to 0.72108, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.6182 - accuracy: 0.7404 - val_loss: 0.6808 - val_accuracy: 0.7211\n",
            "Epoch 24/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.5991 - accuracy: 0.7496\n",
            "Epoch 00024: val_accuracy improved from 0.72108 to 0.73801, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.5991 - accuracy: 0.7496 - val_loss: 0.6413 - val_accuracy: 0.7380\n",
            "Epoch 25/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.5827 - accuracy: 0.7559\n",
            "Epoch 00025: val_accuracy improved from 0.73801 to 0.74487, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.5827 - accuracy: 0.7559 - val_loss: 0.6305 - val_accuracy: 0.7449\n",
            "Epoch 26/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.5710 - accuracy: 0.7611\n",
            "Epoch 00026: val_accuracy did not improve from 0.74487\n",
            "458/458 [==============================] - 135s 295ms/step - loss: 0.5710 - accuracy: 0.7611 - val_loss: 0.6665 - val_accuracy: 0.7363\n",
            "Epoch 27/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.5570 - accuracy: 0.7670\n",
            "Epoch 00027: val_accuracy did not improve from 0.74487\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.5570 - accuracy: 0.7670 - val_loss: 0.6632 - val_accuracy: 0.7364\n",
            "Epoch 28/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.5407 - accuracy: 0.7745\n",
            "Epoch 00028: val_accuracy did not improve from 0.74487\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.5407 - accuracy: 0.7745 - val_loss: 0.6422 - val_accuracy: 0.7406\n",
            "Epoch 29/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.5321 - accuracy: 0.7782\n",
            "Epoch 00029: val_accuracy improved from 0.74487 to 0.76269, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.5321 - accuracy: 0.7782 - val_loss: 0.6065 - val_accuracy: 0.7627\n",
            "Epoch 30/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.5278 - accuracy: 0.7799\n",
            "Epoch 00030: val_accuracy did not improve from 0.76269\n",
            "458/458 [==============================] - 135s 295ms/step - loss: 0.5278 - accuracy: 0.7799 - val_loss: 0.6719 - val_accuracy: 0.7292\n",
            "Epoch 31/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.5080 - accuracy: 0.7885\n",
            "Epoch 00031: val_accuracy improved from 0.76269 to 0.76568, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.5080 - accuracy: 0.7885 - val_loss: 0.5961 - val_accuracy: 0.7657\n",
            "Epoch 32/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.4972 - accuracy: 0.7938\n",
            "Epoch 00032: val_accuracy improved from 0.76568 to 0.76687, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.4972 - accuracy: 0.7938 - val_loss: 0.5852 - val_accuracy: 0.7669\n",
            "Epoch 33/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.4822 - accuracy: 0.8000\n",
            "Epoch 00033: val_accuracy did not improve from 0.76687\n",
            "458/458 [==============================] - 135s 295ms/step - loss: 0.4822 - accuracy: 0.8000 - val_loss: 0.6130 - val_accuracy: 0.7573\n",
            "Epoch 34/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.4650 - accuracy: 0.8084\n",
            "Epoch 00034: val_accuracy did not improve from 0.76687\n",
            "458/458 [==============================] - 135s 295ms/step - loss: 0.4650 - accuracy: 0.8084 - val_loss: 0.6238 - val_accuracy: 0.7581\n",
            "Epoch 35/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.4517 - accuracy: 0.8144\n",
            "Epoch 00035: val_accuracy improved from 0.76687 to 0.78379, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.4517 - accuracy: 0.8144 - val_loss: 0.5683 - val_accuracy: 0.7838\n",
            "Epoch 36/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.4387 - accuracy: 0.8203\n",
            "Epoch 00036: val_accuracy improved from 0.78379 to 0.78778, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.4387 - accuracy: 0.8203 - val_loss: 0.5556 - val_accuracy: 0.7878\n",
            "Epoch 37/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.4171 - accuracy: 0.8303\n",
            "Epoch 00037: val_accuracy improved from 0.78778 to 0.79763, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.4171 - accuracy: 0.8303 - val_loss: 0.5502 - val_accuracy: 0.7976\n",
            "Epoch 38/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.3956 - accuracy: 0.8391\n",
            "Epoch 00038: val_accuracy did not improve from 0.79763\n",
            "458/458 [==============================] - 135s 295ms/step - loss: 0.3956 - accuracy: 0.8391 - val_loss: 0.5687 - val_accuracy: 0.7948\n",
            "Epoch 39/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.3795 - accuracy: 0.8462\n",
            "Epoch 00039: val_accuracy improved from 0.79763 to 0.80102, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 297ms/step - loss: 0.3795 - accuracy: 0.8462 - val_loss: 0.5547 - val_accuracy: 0.8010\n",
            "Epoch 40/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.3855 - accuracy: 0.8429\n",
            "Epoch 00040: val_accuracy did not improve from 0.80102\n",
            "458/458 [==============================] - 135s 295ms/step - loss: 0.3855 - accuracy: 0.8429 - val_loss: 0.5962 - val_accuracy: 0.7805\n",
            "Epoch 41/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.3824 - accuracy: 0.8445\n",
            "Epoch 00041: val_accuracy did not improve from 0.80102\n",
            "458/458 [==============================] - 135s 295ms/step - loss: 0.3824 - accuracy: 0.8445 - val_loss: 0.5755 - val_accuracy: 0.7946\n",
            "Epoch 42/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.3875 - accuracy: 0.8415\n",
            "Epoch 00042: val_accuracy did not improve from 0.80102\n",
            "458/458 [==============================] - 135s 295ms/step - loss: 0.3875 - accuracy: 0.8415 - val_loss: 0.5927 - val_accuracy: 0.7878\n",
            "Epoch 43/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.3946 - accuracy: 0.8390\n",
            "Epoch 00043: val_accuracy did not improve from 0.80102\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.3946 - accuracy: 0.8390 - val_loss: 0.5975 - val_accuracy: 0.7808\n",
            "Epoch 44/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.3896 - accuracy: 0.8411\n",
            "Epoch 00044: val_accuracy did not improve from 0.80102\n",
            "458/458 [==============================] - 135s 295ms/step - loss: 0.3896 - accuracy: 0.8411 - val_loss: 0.6046 - val_accuracy: 0.7833\n",
            "Epoch 45/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.3760 - accuracy: 0.8472\n",
            "Epoch 00045: val_accuracy did not improve from 0.80102\n",
            "458/458 [==============================] - 135s 295ms/step - loss: 0.3760 - accuracy: 0.8472 - val_loss: 0.5829 - val_accuracy: 0.7962\n",
            "Epoch 46/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.3612 - accuracy: 0.8533\n",
            "Epoch 00046: val_accuracy did not improve from 0.80102\n",
            "458/458 [==============================] - 135s 296ms/step - loss: 0.3612 - accuracy: 0.8533 - val_loss: 0.6026 - val_accuracy: 0.7879\n",
            "Epoch 47/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.3469 - accuracy: 0.8590\n",
            "Epoch 00047: val_accuracy did not improve from 0.80102\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.3469 - accuracy: 0.8590 - val_loss: 0.6070 - val_accuracy: 0.7927\n",
            "Epoch 48/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.3292 - accuracy: 0.8669\n",
            "Epoch 00048: val_accuracy improved from 0.80102 to 0.80559, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.3292 - accuracy: 0.8669 - val_loss: 0.5787 - val_accuracy: 0.8056\n",
            "Epoch 49/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.3170 - accuracy: 0.8724\n",
            "Epoch 00049: val_accuracy improved from 0.80559 to 0.81485, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.3170 - accuracy: 0.8724 - val_loss: 0.5706 - val_accuracy: 0.8149\n",
            "Epoch 50/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.3369 - accuracy: 0.8635\n",
            "Epoch 00050: val_accuracy did not improve from 0.81485\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.3369 - accuracy: 0.8635 - val_loss: 0.6074 - val_accuracy: 0.7972\n",
            "Epoch 51/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.3539 - accuracy: 0.8554\n",
            "Epoch 00051: val_accuracy did not improve from 0.81485\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.3539 - accuracy: 0.8554 - val_loss: 0.5910 - val_accuracy: 0.8042\n",
            "Epoch 52/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.3344 - accuracy: 0.8644\n",
            "Epoch 00052: val_accuracy did not improve from 0.81485\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.3344 - accuracy: 0.8644 - val_loss: 0.5696 - val_accuracy: 0.8127\n",
            "Epoch 53/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.3163 - accuracy: 0.8718\n",
            "Epoch 00053: val_accuracy improved from 0.81485 to 0.82003, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 297ms/step - loss: 0.3163 - accuracy: 0.8718 - val_loss: 0.5674 - val_accuracy: 0.8200\n",
            "Epoch 54/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.2989 - accuracy: 0.8796\n",
            "Epoch 00054: val_accuracy improved from 0.82003 to 0.82481, saving model to /content/drive/My Drive/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/age_model_checkpoint.h5\n",
            "458/458 [==============================] - 136s 297ms/step - loss: 0.2989 - accuracy: 0.8796 - val_loss: 0.5584 - val_accuracy: 0.8248\n",
            "Epoch 55/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.2739 - accuracy: 0.8907\n",
            "Epoch 00055: val_accuracy did not improve from 0.82481\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.2739 - accuracy: 0.8907 - val_loss: 0.5778 - val_accuracy: 0.8223\n",
            "Epoch 56/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.2636 - accuracy: 0.8955\n",
            "Epoch 00056: val_accuracy did not improve from 0.82481\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.2636 - accuracy: 0.8955 - val_loss: 0.6150 - val_accuracy: 0.8175\n",
            "Epoch 57/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.2489 - accuracy: 0.9022\n",
            "Epoch 00057: val_accuracy did not improve from 0.82481\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.2489 - accuracy: 0.9022 - val_loss: 0.6493 - val_accuracy: 0.8071\n",
            "Epoch 58/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.2502 - accuracy: 0.9002\n",
            "Epoch 00058: val_accuracy did not improve from 0.82481\n",
            "458/458 [==============================] - 137s 299ms/step - loss: 0.2502 - accuracy: 0.9002 - val_loss: 0.6888 - val_accuracy: 0.8015\n",
            "Epoch 59/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.2570 - accuracy: 0.8969\n",
            "Epoch 00059: val_accuracy did not improve from 0.82481\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.2570 - accuracy: 0.8969 - val_loss: 0.7584 - val_accuracy: 0.7819\n",
            "Epoch 60/60\n",
            "458/458 [==============================] - ETA: 0s - loss: 0.2733 - accuracy: 0.8898\n",
            "Epoch 00060: val_accuracy did not improve from 0.82481\n",
            "458/458 [==============================] - 136s 296ms/step - loss: 0.2733 - accuracy: 0.8898 - val_loss: 0.7837 - val_accuracy: 0.7792\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Checking Model Performance"
      ],
      "metadata": {
        "id": "siwy6UeOavHB"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RP7vxY_6YjCM"
      },
      "source": [
        "# Checking the train and test loss and accuracy values from the neural network above.\n",
        "\n",
        "train_loss = final_cnn_history.history['loss']\n",
        "test_loss = final_cnn_history.history['val_loss']\n",
        "train_accuracy = final_cnn_history.history['accuracy']\n",
        "test_accuracy = final_cnn_history.history['val_accuracy']"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "COkki9JUU8kf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 470
        },
        "outputId": "fd227df4-734f-41cd-979b-36ac79375bfb"
      },
      "source": [
        "# Plotting a line chart to visualize the loss and accuracy values by epochs.\n",
        "\n",
        "fig, ax = plt.subplots(ncols=2, figsize=(15,7))\n",
        "\n",
        "ax = ax.ravel()\n",
        "\n",
        "ax[0].plot(train_loss, label='Train Loss', color='royalblue', marker='o', markersize=5)\n",
        "ax[0].plot(test_loss, label='Test Loss', color = 'orangered', marker='o', markersize=5)\n",
        "\n",
        "ax[0].set_xlabel('Epochs', fontsize=14)\n",
        "ax[0].set_ylabel('Categorical Crossentropy', fontsize=14)\n",
        "\n",
        "ax[0].legend(fontsize=14)\n",
        "ax[0].tick_params(axis='both', labelsize=12)\n",
        "\n",
        "ax[1].plot(train_accuracy, label='Train Accuracy', color='royalblue', marker='o', markersize=5)\n",
        "ax[1].plot(test_accuracy, label='Test Accuracy', color='orangered', marker='o', markersize=5)\n",
        "\n",
        "ax[1].set_xlabel('Epochs', fontsize=14)\n",
        "ax[1].set_ylabel('Accuracy', fontsize=14)\n",
        "\n",
        "ax[1].legend(fontsize=14)\n",
        "ax[1].tick_params(axis='both', labelsize=12)\n",
        "\n",
        "fig.suptitle(x=0.5, y=0.92, t=\"Lineplots showing loss and accuracy of CNN model by epochs\", fontsize=16)\n",
        "\n",
        "# Exporting plot image in PNG format.\n",
        "plt.savefig('/content/drive/My Drive/1_LiveProjects/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/final_cnn_loss_accuracy.png', bbox_inches='tight');"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x504 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "F75NJtVJU8oD",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ea038f01-1c6b-415f-cfb8-0682d98c46b1"
      },
      "source": [
        "# Evaluating the model on test dataset.\n",
        "\n",
        "final_cnn_score = final_cnn.evaluate(test_dataset, verbose=1)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "20/20 [==============================] - 2s 109ms/step - loss: 0.7837 - accuracy: 0.7792\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h5JdZYFMU8s0",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "0f1c3605-8bfa-4bdb-ff09-2d7fd847bb82"
      },
      "source": [
        "# Printing the relevant score summary.\n",
        "\n",
        "final_cnn_labels = final_cnn.metrics_names\n",
        "print(f'CNN model {final_cnn_labels[0]} \\t\\t= {round(final_cnn_score[0], 3)}')\n",
        "print(f'CNN model {final_cnn_labels[1]} \\t= {round(final_cnn_score[1], 3)}')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "CNN model loss \t\t= 0.784\n",
            "CNN model accuracy \t= 0.779\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fatkGFAkrjCF"
      },
      "source": [
        "# Saving the model as a h5 file for possible use later.\n",
        "\n",
        "final_cnn.save(f\"/content/drive/My Drive/1_LiveProjects/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/age_model_acc_{round(final_cnn_score[1], 3)}.h5\", save_format='h5')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Plotting Confusion Matrix"
      ],
      "metadata": {
        "id": "EpFNy9QFe_pN"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "R1e1uYqHU8wx"
      },
      "source": [
        "# Generating predictions from the model above.\n",
        "\n",
        "final_cnn_pred = final_cnn.predict(test_dataset)\n",
        "final_cnn_pred = final_cnn_pred.argmax(axis=-1)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "T6hAs7SlU9Zi",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "fb5f47fe-011c-44ec-ddb2-4c79c49d9e05"
      },
      "source": [
        "# Generating a confusion matrix based on above predictions.\n",
        "\n",
        "conf_mat = confusion_matrix(test_labels_list, final_cnn_pred)\n",
        "conf_mat"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[ 948,    1,    1,    1,    4,    3,    0],\n",
              "       [ 164,  518,   69,   58,   26,    5,    5],\n",
              "       [   3,    6,  566,  260,   91,   15,    0],\n",
              "       [   1,    0,   11, 1674,  308,   13,    0],\n",
              "       [   0,    0,    9,  480, 2221,   91,    6],\n",
              "       [   2,    0,    0,   43,  359, 1203,   72],\n",
              "       [   0,    0,    0,    8,   29,   74,  698]])"
            ]
          },
          "metadata": {},
          "execution_count": 33
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uACH8n6fU95F"
      },
      "source": [
        "# Defining a function to plot the confusion matrix in a grid for easier visualization.\n",
        "\n",
        "def plot_confusion_matrix(cm, classes, normalize=False, title='Confusion Matrix', export_as='confusion_matrix', cmap=plt.cm.Blues):\n",
        "    \"\"\"\n",
        "    This function prints and plots the confusion matrix.\n",
        "    Normalization can be applied by setting `normalize=True`.\n",
        "    \"\"\"\n",
        "    if normalize:\n",
        "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
        "        print(\"Normalized confusion matrix\")\n",
        "    else:\n",
        "        print('Confusion matrix, without normalization')\n",
        "\n",
        "    # print(cm)\n",
        "\n",
        "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
        "    plt.title(title, fontsize=16)\n",
        "    plt.colorbar()\n",
        "    tick_marks = np.arange(len(classes))\n",
        "    plt.xticks(tick_marks, classes, rotation=45)\n",
        "    plt.yticks(tick_marks, classes)\n",
        "\n",
        "    fmt = '.2f' if normalize else 'd'\n",
        "    thresh = cm.max() / 2.\n",
        "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
        "        plt.text(j, i, format(cm[i, j], fmt), horizontalalignment=\"center\", color=\"white\" if cm[i, j] > thresh else \"black\")\n",
        "\n",
        "    plt.tight_layout()\n",
        "    plt.ylabel('True labels', fontsize=14)\n",
        "    plt.xlabel('Predicted labels', fontsize=14)\n",
        "\n",
        "    # Exporting plot image in PNG format.\n",
        "    plt.savefig(f'/content/drive/My Drive/1_LiveProjects/Project5_AgeGenderEmotion_Detection/1.1_age_input_output/output/cnn_logs/{export_as}.png', bbox_inches='tight');"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "J1XROLD_U91Z",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 620
        },
        "outputId": "33361272-b755-4ac1-c8b5-611a8077f9d3"
      },
      "source": [
        "# Plotting the confusion matrix using the function defined above.\n",
        "\n",
        "cm_plot_labels = ['1-2', '3-9', '10-20', '21-27', '28-45', '46-65', '66-116']\n",
        "\n",
        "plt.figure(figsize=(16,8))\n",
        "plot_confusion_matrix(conf_mat, cm_plot_labels, normalize=True,\n",
        "                      title=\"Confusion Matrix based on predictions from CNN model\",\n",
        "                      export_as=\"final_cnn_conf_mat_norm\"\n",
        "                     )\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Normalized confusion matrix\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x576 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3cpS3994e9GM"
      },
      "source": [
        "Below is the **final summary of age prediction CNN model**.\n",
        "\n",
        "\n",
        "As with any data science workflow, the deep learning approach presented above does have its own limitations as well. For instance, the original datasets used in this project only had about 33,000 images when combined together. Even though the training dataset was augmented to increase it's size from 23,440 images to 234,400 images, there is always a possibility that an **even larger training dataset with more variation in the images would have resulted in even better results**.\n",
        "\n",
        "Another approach to this project could be to use **Transfer Learning (using the architecture and layer weights from a pre-trained neural network)** instead of creating and training a neural network from scratch.\n",
        "\n",
        "Ref: https://towardsdatascience.com/age-detection-using-facial-images-traditional-machine-learning-vs-deep-learning-2437b2feeab2 by Prerak Agarwal"
      ]
    }
  ]
}