Real_Time_Age_Prediction_using_RESNET_50.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Real Time Age Prediction using RESNET 50.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CX9zDq8VZd7V",
        "colab_type": "text"
      },
      "source": [
        "##*Real Time Age Prediction Using Pretrained ResNet50*\n",
        "By Nakshatra Singh"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "COr4jUmnZmXM",
        "colab_type": "text"
      },
      "source": [
        "This notebook will show you how to use pretrained resnet50 for *Real Time Age Prediction*."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0tL_3nWWZxBz",
        "colab_type": "text"
      },
      "source": [
        "##Using Google GPU for Training\n",
        "\n",
        "Google colab offers free GPUs and TPUs! Since we'll be training a large model it's best to take advantage of this (in this case we'll use GPU), otherwise training can take long time.\n",
        "\n",
        "A GPU can be added by going to the menu and selecting:\n",
        "\n",
        "`Edit -> Notebook Settings -> Hardware Accelerator -> (GPU)`\n",
        "\n",
        "Then run the following cell to confirm that a GPU is detected. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RIum5s4ZZykq",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "d1f94bd8-692e-4645-d2d7-533511d131d2"
      },
      "source": [
        "import tensorflow as tf\n",
        "# Get the device GPU name \n",
        "device_name = tf.test.gpu_device_name()\n",
        "\n",
        "# The device name should look like the following:\n",
        "if device_name == '/device:GPU:0':\n",
        "  print('Found GPU at : {}'.format(device_name)) \n",
        "else:\n",
        "  raise SystemError('GPU not found!')"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found GPU at : /device:GPU:0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "i9mLlwYZ_x8x",
        "colab_type": "text"
      },
      "source": [
        "###**1. ResNet50**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EyMVcDl8Ulec",
        "colab_type": "text"
      },
      "source": [
        "ResNet is one of the most powerful deep neural networks which has achieved fantabulous performance results in the ILSVRC 2015 classification challenge. ResNet has achieved excellent generalization performance on other recognition tasks and won the first place on ImageNet detection, ImageNet localization, COCO detection and COCO segmentation in ILSVRC and COCO 2015 competitions. There are many variants of ResNet architecture i.e. same concept but with a different number of layers. We have ResNet-18, ResNet-34, ResNet-50, ResNet-101, ResNet-110, ResNet-152, ResNet-164, ResNet-1202 etc. The name ResNet followed by a two or more digit number simply implies the ResNet architecture with a certain number of neural network layers.\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oPkTrjrLZ2gK",
        "colab_type": "text"
      },
      "source": [
        "###**2. Retrieve and Inspect Dataset**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8eVMUzCrZ6G4",
        "colab_type": "text"
      },
      "source": [
        "Let's download the dataset which is uploaded on my [{Github Repository}](/~https://github.com/nakshatrasinghh/Image-Dataset-for-Age-Prediction)."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nwyZKpoAPnAV",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 121
        },
        "outputId": "82df57d4-07b5-4084-d12c-afc13ff3b3d6"
      },
      "source": [
        "!git clone /~https://github.com/nakshatrasinghh/Image-Dataset-for-Age-Prediction.git"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Cloning into 'Image-Dataset-for-Age-Prediction'...\n",
            "remote: Enumerating objects: 23797, done.\u001b[K\n",
            "remote: Total 23797 (delta 0), reused 0 (delta 0), pack-reused 23797\u001b[K\n",
            "Receiving objects: 100% (23797/23797), 1.27 GiB | 41.69 MiB/s, done.\n",
            "Resolving deltas: 100% (4/4), done.\n",
            "Checking out files: 100% (24109/24109), done.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hoMhhbcmaghQ",
        "colab_type": "text"
      },
      "source": [
        "###**3. Ktrain**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9TOYjDS3_6oT",
        "colab_type": "text"
      },
      "source": [
        "ktrain is a lightweight wrapper for the deep learning library TensorFlow Keras (and other libraries) to help build, train, and deploy neural networks and other machine learning models. Inspired by ML framework extensions like fastai and ludwig, it is designed to make deep learning and AI more accessible and easier to apply for both newcomers and experienced practitioners. With only a few lines of code, ktrain allows you to easily and quickly.\n",
        "\n",
        "Next, let's install the *`Ktrain`* library from this [Github Repository](/~https://github.com/amaiya/ktrain)."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sUER5OZ7s8L5",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "95769c5e-1db6-4341-8442-c254143b4a14"
      },
      "source": [
        "!pip install ktrain "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting ktrain\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/b9/b1/f61f03d2bbe4f8615b9643090994914b418569d6f87b9805328b43840e50/ktrain-0.21.3.tar.gz (25.3MB)\n",
            "\u001b[K     |████████████████████████████████| 25.3MB 129kB/s \n",
            "\u001b[?25hRequirement already satisfied: scikit-learn>=0.21.3 in /usr/local/lib/python3.6/dist-packages (from ktrain) (0.22.2.post1)\n",
            "Requirement already satisfied: matplotlib>=3.0.0 in /usr/local/lib/python3.6/dist-packages (from ktrain) (3.2.2)\n",
            "Requirement already satisfied: pandas>=1.0.1 in /usr/local/lib/python3.6/dist-packages (from ktrain) (1.0.5)\n",
            "Requirement already satisfied: fastprogress>=0.1.21 in /usr/local/lib/python3.6/dist-packages (from ktrain) (1.0.0)\n",
            "Collecting keras_bert>=0.81.0\n",
            "  Downloading https://files.pythonhosted.org/packages/e2/7f/95fabd29f4502924fa3f09ff6538c5a7d290dfef2c2fe076d3d1a16e08f0/keras-bert-0.86.0.tar.gz\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.6/dist-packages (from ktrain) (2.23.0)\n",
            "Requirement already satisfied: joblib in /usr/local/lib/python3.6/dist-packages (from ktrain) (0.16.0)\n",
            "Collecting langdetect\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/56/a3/8407c1e62d5980188b4acc45ef3d94b933d14a2ebc9ef3505f22cf772570/langdetect-1.0.8.tar.gz (981kB)\n",
            "\u001b[K     |████████████████████████████████| 983kB 48.5MB/s \n",
            "\u001b[?25hRequirement already satisfied: jieba in /usr/local/lib/python3.6/dist-packages (from ktrain) (0.42.1)\n",
            "Collecting cchardet\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/1e/c5/7e1a0d7b4afd83d6f8de794fce82820ec4c5136c6d52e14000822681a842/cchardet-2.1.6-cp36-cp36m-manylinux2010_x86_64.whl (241kB)\n",
            "\u001b[K     |████████████████████████████████| 245kB 45.9MB/s \n",
            "\u001b[?25hRequirement already satisfied: networkx>=2.3 in /usr/local/lib/python3.6/dist-packages (from ktrain) (2.5)\n",
            "Requirement already satisfied: bokeh in /usr/local/lib/python3.6/dist-packages (from ktrain) (2.1.1)\n",
            "Collecting seqeval\n",
            "  Downloading https://files.pythonhosted.org/packages/34/91/068aca8d60ce56dd9ba4506850e876aba5e66a6f2f29aa223224b50df0de/seqeval-0.0.12.tar.gz\n",
            "Requirement already satisfied: packaging in /usr/local/lib/python3.6/dist-packages (from ktrain) (20.4)\n",
            "Collecting transformers>=3.1.0\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/ae/05/c8c55b600308dc04e95100dc8ad8a244dd800fe75dfafcf1d6348c6f6209/transformers-3.1.0-py3-none-any.whl (884kB)\n",
            "\u001b[K     |████████████████████████████████| 890kB 49.9MB/s \n",
            "\u001b[?25hRequirement already satisfied: ipython in /usr/local/lib/python3.6/dist-packages (from ktrain) (5.5.0)\n",
            "Collecting syntok\n",
            "  Downloading https://files.pythonhosted.org/packages/8c/76/a49e73a04b3e3a14ce232e8e28a1587f8108baa665644fe8c40e307e792e/syntok-1.3.1.tar.gz\n",
            "Collecting whoosh\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/ba/19/24d0f1f454a2c1eb689ca28d2f178db81e5024f42d82729a4ff6771155cf/Whoosh-2.7.4-py2.py3-none-any.whl (468kB)\n",
            "\u001b[K     |████████████████████████████████| 471kB 41.6MB/s \n",
            "\u001b[?25hRequirement already satisfied: scipy>=0.17.0 in /usr/local/lib/python3.6/dist-packages (from scikit-learn>=0.21.3->ktrain) (1.4.1)\n",
            "Requirement already satisfied: numpy>=1.11.0 in /usr/local/lib/python3.6/dist-packages (from scikit-learn>=0.21.3->ktrain) (1.18.5)\n",
            "Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib>=3.0.0->ktrain) (2.4.7)\n",
            "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib>=3.0.0->ktrain) (1.2.0)\n",
            "Requirement already satisfied: python-dateutil>=2.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib>=3.0.0->ktrain) (2.8.1)\n",
            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.6/dist-packages (from matplotlib>=3.0.0->ktrain) (0.10.0)\n",
            "Requirement already satisfied: pytz>=2017.2 in /usr/local/lib/python3.6/dist-packages (from pandas>=1.0.1->ktrain) (2018.9)\n",
            "Requirement already satisfied: Keras>=2.4.3 in /usr/local/lib/python3.6/dist-packages (from keras_bert>=0.81.0->ktrain) (2.4.3)\n",
            "Collecting keras-transformer>=0.38.0\n",
            "  Downloading https://files.pythonhosted.org/packages/89/6c/d6f0c164f4cc16fbc0d0fea85f5526e87a7d2df7b077809e422a7e626150/keras-transformer-0.38.0.tar.gz\n",
            "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.6/dist-packages (from requests->ktrain) (2.10)\n",
            "Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.6/dist-packages (from requests->ktrain) (1.24.3)\n",
            "Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.6/dist-packages (from requests->ktrain) (3.0.4)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.6/dist-packages (from requests->ktrain) (2020.6.20)\n",
            "Requirement already satisfied: six in /usr/local/lib/python3.6/dist-packages (from langdetect->ktrain) (1.15.0)\n",
            "Requirement already satisfied: decorator>=4.3.0 in /usr/local/lib/python3.6/dist-packages (from networkx>=2.3->ktrain) (4.4.2)\n",
            "Requirement already satisfied: typing-extensions>=3.7.4 in /usr/local/lib/python3.6/dist-packages (from bokeh->ktrain) (3.7.4.3)\n",
            "Requirement already satisfied: pillow>=4.0 in /usr/local/lib/python3.6/dist-packages (from bokeh->ktrain) (7.0.0)\n",
            "Requirement already satisfied: Jinja2>=2.7 in /usr/local/lib/python3.6/dist-packages (from bokeh->ktrain) (2.11.2)\n",
            "Requirement already satisfied: PyYAML>=3.10 in /usr/local/lib/python3.6/dist-packages (from bokeh->ktrain) (3.13)\n",
            "Requirement already satisfied: tornado>=5.1 in /usr/local/lib/python3.6/dist-packages (from bokeh->ktrain) (5.1.1)\n",
            "Requirement already satisfied: filelock in /usr/local/lib/python3.6/dist-packages (from transformers>=3.1.0->ktrain) (3.0.12)\n",
            "Collecting sentencepiece!=0.1.92\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/d4/a4/d0a884c4300004a78cca907a6ff9a5e9fe4f090f5d95ab341c53d28cbc58/sentencepiece-0.1.91-cp36-cp36m-manylinux1_x86_64.whl (1.1MB)\n",
            "\u001b[K     |████████████████████████████████| 1.1MB 40.1MB/s \n",
            "\u001b[?25hRequirement already satisfied: regex!=2019.12.17 in /usr/local/lib/python3.6/dist-packages (from transformers>=3.1.0->ktrain) (2019.12.20)\n",
            "Requirement already satisfied: tqdm>=4.27 in /usr/local/lib/python3.6/dist-packages (from transformers>=3.1.0->ktrain) (4.41.1)\n",
            "Requirement already satisfied: dataclasses; python_version < \"3.7\" in /usr/local/lib/python3.6/dist-packages (from transformers>=3.1.0->ktrain) (0.7)\n",
            "Collecting sacremoses\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/7d/34/09d19aff26edcc8eb2a01bed8e98f13a1537005d31e95233fd48216eed10/sacremoses-0.0.43.tar.gz (883kB)\n",
            "\u001b[K     |████████████████████████████████| 890kB 50.3MB/s \n",
            "\u001b[?25hCollecting tokenizers==0.8.1.rc2\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/80/83/8b9fccb9e48eeb575ee19179e2bdde0ee9a1904f97de5f02d19016b8804f/tokenizers-0.8.1rc2-cp36-cp36m-manylinux1_x86_64.whl (3.0MB)\n",
            "\u001b[K     |████████████████████████████████| 3.0MB 51.0MB/s \n",
            "\u001b[?25hRequirement already satisfied: traitlets>=4.2 in /usr/local/lib/python3.6/dist-packages (from ipython->ktrain) (4.3.3)\n",
            "Requirement already satisfied: pickleshare in /usr/local/lib/python3.6/dist-packages (from ipython->ktrain) (0.7.5)\n",
            "Requirement already satisfied: pexpect; sys_platform != \"win32\" in /usr/local/lib/python3.6/dist-packages (from ipython->ktrain) (4.8.0)\n",
            "Requirement already satisfied: prompt-toolkit<2.0.0,>=1.0.4 in /usr/local/lib/python3.6/dist-packages (from ipython->ktrain) (1.0.18)\n",
            "Requirement already satisfied: setuptools>=18.5 in /usr/local/lib/python3.6/dist-packages (from ipython->ktrain) (50.3.0)\n",
            "Requirement already satisfied: simplegeneric>0.8 in /usr/local/lib/python3.6/dist-packages (from ipython->ktrain) (0.8.1)\n",
            "Requirement already satisfied: pygments in /usr/local/lib/python3.6/dist-packages (from ipython->ktrain) (2.6.1)\n",
            "Requirement already satisfied: h5py in /usr/local/lib/python3.6/dist-packages (from Keras>=2.4.3->keras_bert>=0.81.0->ktrain) (2.10.0)\n",
            "Collecting keras-pos-embd>=0.11.0\n",
            "  Downloading https://files.pythonhosted.org/packages/09/70/b63ed8fc660da2bb6ae29b9895401c628da5740c048c190b5d7107cadd02/keras-pos-embd-0.11.0.tar.gz\n",
            "Collecting keras-multi-head>=0.27.0\n",
            "  Downloading https://files.pythonhosted.org/packages/e6/32/45adf2549450aca7867deccfa04af80a0ab1ca139af44b16bc669e0e09cd/keras-multi-head-0.27.0.tar.gz\n",
            "Collecting keras-layer-normalization>=0.14.0\n",
            "  Downloading https://files.pythonhosted.org/packages/a4/0e/d1078df0494bac9ce1a67954e5380b6e7569668f0f3b50a9531c62c1fc4a/keras-layer-normalization-0.14.0.tar.gz\n",
            "Collecting keras-position-wise-feed-forward>=0.6.0\n",
            "  Downloading https://files.pythonhosted.org/packages/e3/59/f0faa1037c033059e7e9e7758e6c23b4d1c0772cd48de14c4b6fd4033ad5/keras-position-wise-feed-forward-0.6.0.tar.gz\n",
            "Collecting keras-embed-sim>=0.8.0\n",
            "  Downloading https://files.pythonhosted.org/packages/57/ef/61a1e39082c9e1834a2d09261d4a0b69f7c818b359216d4e1912b20b1c86/keras-embed-sim-0.8.0.tar.gz\n",
            "Requirement already satisfied: MarkupSafe>=0.23 in /usr/local/lib/python3.6/dist-packages (from Jinja2>=2.7->bokeh->ktrain) (1.1.1)\n",
            "Requirement already satisfied: click in /usr/local/lib/python3.6/dist-packages (from sacremoses->transformers>=3.1.0->ktrain) (7.1.2)\n",
            "Requirement already satisfied: ipython-genutils in /usr/local/lib/python3.6/dist-packages (from traitlets>=4.2->ipython->ktrain) (0.2.0)\n",
            "Requirement already satisfied: ptyprocess>=0.5 in /usr/local/lib/python3.6/dist-packages (from pexpect; sys_platform != \"win32\"->ipython->ktrain) (0.6.0)\n",
            "Requirement already satisfied: wcwidth in /usr/local/lib/python3.6/dist-packages (from prompt-toolkit<2.0.0,>=1.0.4->ipython->ktrain) (0.2.5)\n",
            "Collecting keras-self-attention==0.46.0\n",
            "  Downloading https://files.pythonhosted.org/packages/15/6b/c804924a056955fa1f3ff767945187103cfc851ba9bd0fc5a6c6bc18e2eb/keras-self-attention-0.46.0.tar.gz\n",
            "Building wheels for collected packages: ktrain, keras-bert, langdetect, seqeval, syntok, keras-transformer, sacremoses, keras-pos-embd, keras-multi-head, keras-layer-normalization, keras-position-wise-feed-forward, keras-embed-sim, keras-self-attention\n",
            "  Building wheel for ktrain (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for ktrain: filename=ktrain-0.21.3-cp36-none-any.whl size=25270203 sha256=89ee19522361fbe417b44c8a704f4fe6bf140b5a1d63febf86688c05cf9df70f\n",
            "  Stored in directory: /root/.cache/pip/wheels/d6/ef/6d/9428365b27ac1e84a77214c5718e7f8b2180007af593791247\n",
            "  Building wheel for keras-bert (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for keras-bert: filename=keras_bert-0.86.0-cp36-none-any.whl size=34145 sha256=e5258aa3a691a68925fb89f13eaff30f7ffab4302c06f8a60da1041b05ff25ef\n",
            "  Stored in directory: /root/.cache/pip/wheels/66/f0/b1/748128b58562fc9e31b907bb5e2ab6a35eb37695e83911236b\n",
            "  Building wheel for langdetect (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for langdetect: filename=langdetect-1.0.8-cp36-none-any.whl size=993195 sha256=e59ce1cbf0410374ecf333777a2d6142c2c7add40772ce93b1a7965a2582497e\n",
            "  Stored in directory: /root/.cache/pip/wheels/8d/b3/aa/6d99de9f3841d7d3d40a60ea06e6d669e8e5012e6c8b947a57\n",
            "  Building wheel for seqeval (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for seqeval: filename=seqeval-0.0.12-cp36-none-any.whl size=7423 sha256=ac72f27b138ff0ac57dbb6b845f3d35f665771c4bbbae63928371984e35a1556\n",
            "  Stored in directory: /root/.cache/pip/wheels/4f/32/0a/df3b340a82583566975377d65e724895b3fad101a3fb729f68\n",
            "  Building wheel for syntok (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for syntok: filename=syntok-1.3.1-cp36-none-any.whl size=20919 sha256=8358ba147ee52641c055148db5c41161f2d11782ca3f37c6204e20e1cd0f830a\n",
            "  Stored in directory: /root/.cache/pip/wheels/51/c6/a4/be1920586c49469846bcd2888200bdecfe109ec421dab9be2d\n",
            "  Building wheel for keras-transformer (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for keras-transformer: filename=keras_transformer-0.38.0-cp36-none-any.whl size=12942 sha256=5a9ecb35b26a32c82a7c6acf09049ff647b481357afdc10725c9eb3aa13d0027\n",
            "  Stored in directory: /root/.cache/pip/wheels/e5/fb/3a/37b2b9326c799aa010ae46a04ddb04f320d8c77c0b7e837f4e\n",
            "  Building wheel for sacremoses (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for sacremoses: filename=sacremoses-0.0.43-cp36-none-any.whl size=893257 sha256=10260aa5da216b0acfbbbfb34be4c6e44c87a889d9fc061745ebdbfdcf9b94fe\n",
            "  Stored in directory: /root/.cache/pip/wheels/29/3c/fd/7ce5c3f0666dab31a50123635e6fb5e19ceb42ce38d4e58f45\n",
            "  Building wheel for keras-pos-embd (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for keras-pos-embd: filename=keras_pos_embd-0.11.0-cp36-none-any.whl size=7554 sha256=3cd76f03203cb9f80a031c06f1a10e15aa26240411bf7ff9c292f1cb51de9a7c\n",
            "  Stored in directory: /root/.cache/pip/wheels/5b/a1/a0/ce6b1d49ba1a9a76f592e70cf297b05c96bc9f418146761032\n",
            "  Building wheel for keras-multi-head (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for keras-multi-head: filename=keras_multi_head-0.27.0-cp36-none-any.whl size=15612 sha256=6ab887c551baae11748436fdf36448abbbddc3af939153a323a6745e1184b101\n",
            "  Stored in directory: /root/.cache/pip/wheels/b5/b4/49/0a0c27dcb93c13af02fea254ff51d1a43a924dd4e5b7a7164d\n",
            "  Building wheel for keras-layer-normalization (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for keras-layer-normalization: filename=keras_layer_normalization-0.14.0-cp36-none-any.whl size=5268 sha256=2467ee7b552f7886d16676a77aeccdb93a08cdb7466812bf8a07bfbefbc218ec\n",
            "  Stored in directory: /root/.cache/pip/wheels/54/80/22/a638a7d406fd155e507aa33d703e3fa2612b9eb7bb4f4fe667\n",
            "  Building wheel for keras-position-wise-feed-forward (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for keras-position-wise-feed-forward: filename=keras_position_wise_feed_forward-0.6.0-cp36-none-any.whl size=5626 sha256=16e272ca3f737128f725946ee305307db79020755ea7f81745a7a46f9b48c2a9\n",
            "  Stored in directory: /root/.cache/pip/wheels/39/e2/e2/3514fef126a00574b13bc0b9e23891800158df3a3c19c96e3b\n",
            "  Building wheel for keras-embed-sim (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for keras-embed-sim: filename=keras_embed_sim-0.8.0-cp36-none-any.whl size=4559 sha256=a1cc99a583bbbd0398715c80d1dbdc68f4353f813c19aba1a5c7a13bc742f57a\n",
            "  Stored in directory: /root/.cache/pip/wheels/49/45/8b/c111f6cc8bec253e984677de73a6f4f5d2f1649f42aac191c8\n",
            "  Building wheel for keras-self-attention (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for keras-self-attention: filename=keras_self_attention-0.46.0-cp36-none-any.whl size=17278 sha256=0355b13b7001a2ea54618ccc17a8158c5e3239af5065c99351cb7f1c88e80f34\n",
            "  Stored in directory: /root/.cache/pip/wheels/d2/2e/80/fec4c05eb23c8e13b790e26d207d6e0ffe8013fad8c6bdd4d2\n",
            "Successfully built ktrain keras-bert langdetect seqeval syntok keras-transformer sacremoses keras-pos-embd keras-multi-head keras-layer-normalization keras-position-wise-feed-forward keras-embed-sim keras-self-attention\n",
            "Installing collected packages: keras-pos-embd, keras-self-attention, keras-multi-head, keras-layer-normalization, keras-position-wise-feed-forward, keras-embed-sim, keras-transformer, keras-bert, langdetect, cchardet, seqeval, sentencepiece, sacremoses, tokenizers, transformers, syntok, whoosh, ktrain\n",
            "Successfully installed cchardet-2.1.6 keras-bert-0.86.0 keras-embed-sim-0.8.0 keras-layer-normalization-0.14.0 keras-multi-head-0.27.0 keras-pos-embd-0.11.0 keras-position-wise-feed-forward-0.6.0 keras-self-attention-0.46.0 keras-transformer-0.38.0 ktrain-0.21.3 langdetect-1.0.8 sacremoses-0.0.43 sentencepiece-0.1.91 seqeval-0.0.12 syntok-1.3.1 tokenizers-0.8.1rc2 transformers-3.1.0 whoosh-2.7.4\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hEvmUKK-bAEw",
        "colab_type": "text"
      },
      "source": [
        "Now, let's do some basic imports for this notebook. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "byHwEkyitILh",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "%matplotlib inline\n",
        "import os\n",
        "import ktrain\n",
        "from ktrain import vision as vis"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LwFmtivnAEuw",
        "colab_type": "text"
      },
      "source": [
        "We'll be using `Regular Expression` to parse through the filename."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tm0prqarAAA_",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "8e295837-3067-4248-bf8a-b6fbd18c90a3"
      },
      "source": [
        "import re\n",
        "pattern = r'([^/]+)_\\d+_\\d+_\\d+.jpg$' # Excludes everything after the bracket, d represents digits\n",
        "\n",
        "p = re.compile(pattern)\n",
        "r = p.search('100_0_0_20170112213500903.jpg') # Example filename from the dataset.\n",
        "print(r.group(1))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "100\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "91CPPrewccOu",
        "colab_type": "text"
      },
      "source": [
        "Let's define our path for the working dataset directory."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1NiPJURIA5Cw",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "DATADIR = '/content/Image-Dataset-for-Age-Prediction/Images'"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vFSD-OEXc7h8",
        "colab_type": "text"
      },
      "source": [
        "`images_from_fname` is a function in vision package which renders all the images from a filename. We'll pass our complete working directory here and all the heavy-lifting will be done by the function. This function always returns 3 variables (including 2 tuples).\n",
        "\n",
        "Note: Default train-test split is 10% unless specified."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AC1jqcBfkXLt",
        "colab_type": "text"
      },
      "source": [
        "###**4. Model Building and Training with Ktrain**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "R8XQjy8pBA-Z",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 314
        },
        "outputId": "73756564-019d-40e4-8922-d52a99a874ae"
      },
      "source": [
        "(train_data), (test_data), preproc = vis.images_from_fname(DATADIR,\n",
        "                      pattern=pattern,\n",
        "                      is_regression=True,\n",
        "                      random_state=42)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ktrain/vision/data.py:718: UserWarning: Could not extract target for /content/Image-Dataset-for-Age-Prediction/Images/39_1_20170116174525125.jpg -  skipping this file\n",
            "  warnings.warn('Could not extract target for %s -  skipping this file'% (fname))\n",
            "/usr/local/lib/python3.6/dist-packages/ktrain/vision/data.py:718: UserWarning: Could not extract target for /content/Image-Dataset-for-Age-Prediction/Images/61_3_20170109150557335.jpg -  skipping this file\n",
            "  warnings.warn('Could not extract target for %s -  skipping this file'% (fname))\n",
            "/usr/local/lib/python3.6/dist-packages/ktrain/vision/data.py:718: UserWarning: Could not extract target for /content/Image-Dataset-for-Age-Prediction/Images/53__0_20170116184028385.jpg -  skipping this file\n",
            "  warnings.warn('Could not extract target for %s -  skipping this file'% (fname))\n",
            "/usr/local/lib/python3.6/dist-packages/ktrain/vision/data.py:718: UserWarning: Could not extract target for /content/Image-Dataset-for-Age-Prediction/Images/61_1_20170109142408075.jpg -  skipping this file\n",
            "  warnings.warn('Could not extract target for %s -  skipping this file'% (fname))\n",
            "/usr/local/lib/python3.6/dist-packages/ktrain/vision/data.py:718: UserWarning: Could not extract target for /content/Image-Dataset-for-Age-Prediction/Images/24_0_1_20170116220224657 .jpg -  skipping this file\n",
            "  warnings.warn('Could not extract target for %s -  skipping this file'% (fname))\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Found 21690 validated image filenames.\n",
            "Found 2411 validated image filenames.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ktrain/utils.py:664: UserWarning: is_regression=True was supplied but targets are strings - casting to floats\n",
            "  warnings.warn('is_regression=True was supplied but targets are strings - casting to floats')\n",
            "/usr/local/lib/python3.6/dist-packages/ktrain/utils.py:565: UserWarning: Task is being treated as REGRESSION because either class_names argument was not supplied or is_regression=True. If this is incorrect, change accordingly.\n",
            "  'If this is incorrect, change accordingly.')\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "teDbyPYpeBuO",
        "colab_type": "text"
      },
      "source": [
        "The filenames which were incorrectly encoded are just ignored and not taken in training."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "R_CGPN-heUYx",
        "colab_type": "text"
      },
      "source": [
        "Let's see what all models are ready for us to use."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yl8xdYr_BX-7",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 155
        },
        "outputId": "29d0a291-334f-47d7-a151-bf5c0d3ebfb5"
      },
      "source": [
        "vis.print_image_regression_models() "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "pretrained_resnet50: 50-layer Residual Network (pretrained on ImageNet)\n",
            "resnet50: 50-layer Resididual Network (randomly initialized)\n",
            "pretrained_mobilenet: MobileNet Neural Network (pretrained on ImageNet)\n",
            "mobilenet: MobileNet Neural Network (randomly initialized)\n",
            "pretrained_inception: Inception Version 3  (pretrained on ImageNet)\n",
            "inception: Inception Version 3 (randomly initialized)\n",
            "wrn22: 22-layer Wide Residual Network (randomly initialized)\n",
            "default_cnn: a default LeNet-like Convolutional Neural Network\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ymT7dXlmeZXj",
        "colab_type": "text"
      },
      "source": [
        "We'll be using `pretrained_resnet50`."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EOSPz3VEB1Ml",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 158
        },
        "outputId": "cb870dc0-cd6f-4c67-cb8a-f725d20851e3"
      },
      "source": [
        "model = vis.image_regression_model('pretrained_resnet50',\n",
        "                                   train_data = train_data,\n",
        "                                   val_data = test_data)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "The normalization scheme has been changed for use with a pretrained_resnet50 model. If you decide to use a different model, please reload your dataset with a ktrain.vision.data.images_from* function.\n",
            "\n",
            "Is Multi-Label? False\n",
            "Is Regression? True\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/resnet/resnet50_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "94773248/94765736 [==============================] - 1s 0us/step\n",
            "pretrained_resnet50 model created.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "06PjZyKeiZCX",
        "colab_type": "text"
      },
      "source": [
        "We'll define our learner which is used in training the model and getting predictions."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2cwzoYmiCBbL",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "learner = ktrain.get_learner(model = model,\n",
        "                             train_data = train_data,\n",
        "                             val_data = test_data,\n",
        "                             batch_size=64)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Y1V94N-lix4w",
        "colab_type": "text"
      },
      "source": [
        "Now, we'll fit the learner with onecycle policy having parameters - learning rate=0.00001, epochs=2. You can change these hyperparameters which can change performance, I would recommend not to lower the learning rate, it will reduce your accuracy and if you want you can run it for more epochs."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2fvUkIdICTQV",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 155
        },
        "outputId": "93acbbd0-5767-438a-c8c1-0f690f4d6e99"
      },
      "source": [
        "learner.fit_onecycle(1e-5, 2)   # Training on all 50 layers. "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "\n",
            "begin training using onecycle policy with max lr of 1e-05...\n",
            "Epoch 1/2\n",
            "339/339 [==============================] - 363s 1s/step - loss: 923.9599 - mae: 24.1285 - val_loss: 440.4344 - val_mae: 15.8978\n",
            "Epoch 2/2\n",
            "339/339 [==============================] - 350s 1s/step - loss: 362.7765 - mae: 14.5246 - val_loss: 347.7791 - val_mae: 14.1301\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.keras.callbacks.History at 0x7f68520a3b38>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mXKTRPGSjIg-",
        "colab_type": "text"
      },
      "source": [
        "Let's freeze the first 15 layers of the model and train it again using the last 35 layers."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sNWQXhcfCfRF",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 155
        },
        "outputId": "14fd9c81-dba3-46b1-cbe3-9ce0d75173a2"
      },
      "source": [
        "learner.freeze(15)               # Freezing first 15 layers of ResNet50.\n",
        "\n",
        "learner.fit_onecycle(1e-5, 2)    # Training last 35 layers again."
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "\n",
            "begin training using onecycle policy with max lr of 1e-05...\n",
            "Epoch 1/2\n",
            "339/339 [==============================] - 420s 1s/step - loss: 172.3675 - mae: 10.3238 - val_loss: 112.2607 - val_mae: 7.7747\n",
            "Epoch 2/2\n",
            "339/339 [==============================] - 424s 1s/step - loss: 80.4011 - mae: 6.8333 - val_loss: 89.6877 - val_mae: 7.0222\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.keras.callbacks.History at 0x7f6841cf19b0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CNVPsMoRjSRw",
        "colab_type": "text"
      },
      "source": [
        "Let's get our predictor using `get_predictor`."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RgfoIdo2DnQK",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "predictor = ktrain.get_predictor(learner.model, preproc)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "H6gwZhxvjoOA",
        "colab_type": "text"
      },
      "source": [
        "Let's take a random sample from the test set, just to see how are model works."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lDrfLbV9EUPB",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "69d648a5-bb7c-457c-88be-23118bfcd7a4"
      },
      "source": [
        "test_data.filenames[10] "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'88_1_0_20170120230330380.jpg'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 54
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "W0BJxsWhjt6s",
        "colab_type": "text"
      },
      "source": [
        "I'll write a helper function which will help us in visualizing the result a little better."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8irJEvCxEo92",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def show_prediction(fname):\n",
        "  fname = DATADIR + '/' + fname\n",
        "  pred = round(predictor.predict_filename(fname)[0])\n",
        "  actual = int(p.search(fname).group(1))\n",
        "  vis.show_image(fname)\n",
        "  print(\"Predicted Age: %s || Actual Age: %s\" %   (pred, actual))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2kQRjkREj01B",
        "colab_type": "text"
      },
      "source": [
        "Let's run the function on the sample image present in our test set."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aLoAmcLLFYkN",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        },
        "outputId": "d9b01fa2-c7ad-477b-e248-8c04c89bfff8"
      },
      "source": [
        "show_prediction('88_1_0_20170120230330380.jpg')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Predicted Age: 87.0 || Actual Age: 88\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "v-d4MtKGj7pc",
        "colab_type": "text"
      },
      "source": [
        "WooHoo! The model works brilliantly and is predicting the age quite accurately.  "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xi52XG-TGE0F",
        "colab_type": "text"
      },
      "source": [
        "###**5. Real-Time Age Prediction**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1mhii_6Ekv-H",
        "colab_type": "text"
      },
      "source": [
        "This a google code snippet which I found in the code snippet section on the right, you can get this code by just typing `Camera Capture`."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LK6aDYtHGDlE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from IPython.display import display, Javascript\n",
        "from google.colab.output import eval_js\n",
        "from base64 import b64decode\n",
        "\n",
        "def take_photo(filename='photo.jpg', quality=0.8):\n",
        "  js = Javascript('''\n",
        "    async function takePhoto(quality) {\n",
        "      const div = document.createElement('div');\n",
        "      const capture = document.createElement('button');\n",
        "      capture.textContent = 'Capture';\n",
        "      div.appendChild(capture);\n",
        "\n",
        "      const video = document.createElement('video');\n",
        "      video.style.display = 'block';\n",
        "      const stream = await navigator.mediaDevices.getUserMedia({video: true});\n",
        "\n",
        "      document.body.appendChild(div);\n",
        "      div.appendChild(video);\n",
        "      video.srcObject = stream;\n",
        "      await video.play();\n",
        "\n",
        "      // Resize the output to fit the video element.\n",
        "      google.colab.output.setIframeHeight(document.documentElement.scrollHeight, true);\n",
        "\n",
        "      // Wait for Capture to be clicked.\n",
        "      await new Promise((resolve) => capture.onclick = resolve);\n",
        "\n",
        "      const canvas = document.createElement('canvas');\n",
        "      canvas.width = video.videoWidth;\n",
        "      canvas.height = video.videoHeight;\n",
        "      canvas.getContext('2d').drawImage(video, 0, 0);\n",
        "      stream.getVideoTracks()[0].stop();\n",
        "      div.remove();\n",
        "      return canvas.toDataURL('image/jpeg', quality);\n",
        "    }\n",
        "    ''')\n",
        "  display(js)\n",
        "  data = eval_js('takePhoto({})'.format(quality))\n",
        "  binary = b64decode(data.split(',')[1])\n",
        "  with open(filename, 'wb') as f:\n",
        "    f.write(binary)\n",
        "  return filename"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SA9RPW6Uvfxn",
        "colab_type": "text"
      },
      "source": [
        "Run this cell to take a picture using your camera and it saves the image as `photo.jpg`. Click on the *capture* button to take the photo and continue running the other cells."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "G1sdCCEtGDjd",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from IPython.display import Image\n",
        "try:\n",
        "  filename = take_photo()\n",
        "  print('Saved to {}'.format(filename))\n",
        "  \n",
        "  # Show the image which was just taken.\n",
        "  display(Image(filename))\n",
        "except Exception as err:\n",
        "  # Errors will be thrown if the user does not have a webcam or if they do not\n",
        "  # grant the page permission to access it.\n",
        "  print(str(err))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "H4K5i0TnlKHX",
        "colab_type": "text"
      },
      "source": [
        "Now, let's write a helper function which will input the image you captured and show real-time prediction using the `predictor` we defined earlier."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rrQB8ALgGUoE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def real_prediction(fname):\n",
        "  pred = round(predictor.predict_filename(fname)[0])\n",
        "  vis.show_image(fname)\n",
        "  print(\"Predicted Age: %s\" % (pred)) "
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_Xm5zExKlmuO",
        "colab_type": "text"
      },
      "source": [
        "Calling the helper function on the captured image."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eKcPHTcDJ5vT",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "real_prediction('photo.jpg')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yXpRutXVgSYn",
        "colab_type": "text"
      },
      "source": [
        "There you go! I hope the results were accurate, if they weren't then it could be because of *hues of light and the position in your image*, but I did notice the model is slightlynot accurate for younger ages, we can surely increase the epochs to enhance model performance."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-B7m_8g4luRJ",
        "colab_type": "text"
      },
      "source": [
        "###**6. Model Saving**\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QRSn9XE4lycz",
        "colab_type": "text"
      },
      "source": [
        "Let's save the model in a folder called `Age_Prediction_Model`."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7FsG08hOKLwW",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "predictor.save('/content/Age_Prediction_Model')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IbrjDtiSl7Ut",
        "colab_type": "text"
      },
      "source": [
        "Since it contains a preproc model and the final tensorflow model (.h5), we'll zip the files in another folder so we can download them directly from colab. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "q3wquxtNKRva",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 69
        },
        "outputId": "e2f3bfef-eb43-4570-9768-c774e2362f86"
      },
      "source": [
        "!zip -r /content/Age_Prediction_Model_zip.zip /content/Age_Prediction_Model"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "  adding: content/Age_Prediction_Model/ (stored 0%)\n",
            "  adding: content/Age_Prediction_Model/tf_model.h5 (deflated 9%)\n",
            "  adding: content/Age_Prediction_Model/tf_model.preproc (deflated 41%)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jQtzVZf-0f6c",
        "colab_type": "text"
      },
      "source": [
        "###**Summary**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FcM0bDPL0lLZ",
        "colab_type": "text"
      },
      "source": [
        "This notebook illustrated how to use a pretrained resnet model for predicting age in real-time. You can definitely change the hyperparameters to achieve better results. The model suffers in predicting ages of younger people, this could be due to many reasons: \n",
        "\n",
        "- Less data for young ages\n",
        "- Hues of light falling on the face\n",
        "- Positioning of the object\n",
        "\n",
        "Special thanks to [amaiya](/~https://github.com/amaiya) for giving this huge contribution, this library has helped me build complex models with just few lines of code. Please check out the [repo](/~https://github.com/amaiya/ktrain) for more details."
      ]
    }
  ]
}