assignment_1_notebook.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "file_extension": ".py",
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.6"
    },
    "mimetype": "text/x-python",
    "name": "python",
    "npconvert_exporter": "python",
    "pygments_lexer": "ipython3",
    "version": 3,
    "colab": {
      "name": "220079842_assignment1_solution.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "v81NrNYlAzIm"
      },
      "source": [
        "# SIT744 Assignment 1: Image Classification with Deep Feedforward Neural Network\n",
        "\n",
        "<div class=\"alert-info\">\n",
        "    <p>Due: <strong>8:00 pm (AEST) 19 April 2021</strong>  (Monday)</p>\n",
        "\n",
        "\n",
        "This is an <strong>individual</strong> assignment. It contributes <strong>30%</strong> to your final mark. Read the assignment instruction carefully.\n",
        "\n",
        "<h2> What to submit </h2>\n",
        "\n",
        "<p>\n",
        "This assignment is to be completed individually and submitted to CloudDeakin. <strong>By the due date, you are required to submit the following files to the corresponding Assignment (Dropbox) in CloudDeakin</strong>:\n",
        "\n",
        "<ol>\n",
        "<li>\t<strong>[YourID]_assignment1_solution.ipynp</strong>:  This is your Python notebook solution source file. </li>\n",
        "<li>\t<strong>[YourID]_assingment1_output.html</strong>: This is the output of your Python notebook solution <emph>exported</emph> in HTML format.</li>\n",
        "<li>\tExtra files needed to complete your assignment, if any (e.g., images used in your answers).</li>\n",
        "</ol>\n",
        "</p>\n",
        "\n",
        "<p>\n",
        "For example, if your student ID is: 123456, you will then need to submit the following files:\n",
        "<ul>\n",
        "<li> 123456_assignment1_solution.ipynp </li>\n",
        "<li> 123456_assignment1_output.html</li>\n",
        "</ul>\n",
        "</p>\n",
        "\n",
        "<h2> Marking criteria </h2>\n",
        "\n",
        "<p>\n",
        "Your submission will be marked using the following criteria.\n",
        "\n",
        "<ul>\n",
        "<li> Showing good effort through completed tasks.</li>\n",
        "<li> Applying deep learning theory to design suitable deep learning solutions for the tasks.</li>\n",
        "<li> Critically evaluating and reflecting on the pros and cons of various design decisions.</li>\n",
        "<li> Demonstrating creativity and resourcefulness in providing unique individual solutions. (<em>Warning: Highly similar solutions will be investigated for collusion.</em>)</li>\n",
        "<li> Showing attention to details through a good quality assignment report.</li>\n",
        "</ul>\n",
        "</p>\n",
        "\n",
        "<p>\n",
        "Indicative weights of various tasks are provided below, but the assignment will be marked by the overall quality per the above criteria. \n",
        "</p>\n",
        "</div>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5QKV13ZpxFel"
      },
      "source": [
        "GIT_USERNAME = \"akritikts\"\n",
        "GIT_TOKEN = \"9c057450d445230b40b31c08ed177bd1b3946a71\"\n",
        "GIT_REPOSITORY = \"DeepLearning\"\n",
        "\n",
        "#!mkdir \"{PROJECT_PATH}\"\n",
        "!git clone https://{GIT_TOKEN}@github.com/{GIT_USERNAME}/{GIT_REPOSITORY}.git "
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9T3o8w5jvP_Y",
        "outputId": "67fcc2d7-f00c-498c-c021-5ce405ceff5a"
      },
      "source": [
        "\n",
        "\n",
        "\n",
        "\n",
        "!mkdir ./temp\n",
        "!git clone \"https://{GIT_TOKEN}@github.com/{GIT_USERNAME}/{GIT_REPOSITORY}.git\" ./temp\n",
        "!rsync -aP --exclude=data//* ./temp\n",
        "\n",
        "%cd ./temp\n",
        "!git add .\n",
        "!git commit -m '\"{loading data, exploring data and building the input pipeline}\"'\n",
        "!git config --global user.email \"{akritikts@gmail.com}\"\n",
        "!git config --global user.name \"{akritikts}\"\n",
        "!git push -u origin master --force\n",
        "%cd /content\n",
        "!rm -rf ./temp"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Cloning into './temp'...\n",
            "warning: You appear to have cloned an empty repository.\n",
            "sending incremental file list\n",
            "drwxr-xr-x          4,096 2021/03/21 12:40:52 temp\n",
            "drwxr-xr-x          4,096 2021/03/21 12:40:52 temp/.git\n",
            "-rw-r--r--             23 2021/03/21 12:40:52 temp/.git/HEAD\n",
            "-rw-r--r--            311 2021/03/21 12:40:52 temp/.git/config\n",
            "-rw-r--r--             73 2021/03/21 12:40:52 temp/.git/description\n",
            "drwxr-xr-x          4,096 2021/03/21 12:40:52 temp/.git/branches\n",
            "drwxr-xr-x          4,096 2021/03/21 12:40:52 temp/.git/hooks\n",
            "-rwxr-xr-x            478 2021/03/21 12:40:52 temp/.git/hooks/applypatch-msg.sample\n",
            "-rwxr-xr-x            896 2021/03/21 12:40:52 temp/.git/hooks/commit-msg.sample\n",
            "-rwxr-xr-x          3,327 2021/03/21 12:40:52 temp/.git/hooks/fsmonitor-watchman.sample\n",
            "-rwxr-xr-x            189 2021/03/21 12:40:52 temp/.git/hooks/post-update.sample\n",
            "-rwxr-xr-x            424 2021/03/21 12:40:52 temp/.git/hooks/pre-applypatch.sample\n",
            "-rwxr-xr-x          1,642 2021/03/21 12:40:52 temp/.git/hooks/pre-commit.sample\n",
            "-rwxr-xr-x          1,348 2021/03/21 12:40:52 temp/.git/hooks/pre-push.sample\n",
            "-rwxr-xr-x          4,898 2021/03/21 12:40:52 temp/.git/hooks/pre-rebase.sample\n",
            "-rwxr-xr-x            544 2021/03/21 12:40:52 temp/.git/hooks/pre-receive.sample\n",
            "-rwxr-xr-x          1,492 2021/03/21 12:40:52 temp/.git/hooks/prepare-commit-msg.sample\n",
            "-rwxr-xr-x          3,610 2021/03/21 12:40:52 temp/.git/hooks/update.sample\n",
            "drwxr-xr-x          4,096 2021/03/21 12:40:52 temp/.git/info\n",
            "-rw-r--r--            240 2021/03/21 12:40:52 temp/.git/info/exclude\n",
            "drwxr-xr-x          4,096 2021/03/21 12:40:52 temp/.git/objects\n",
            "drwxr-xr-x          4,096 2021/03/21 12:40:52 temp/.git/objects/info\n",
            "drwxr-xr-x          4,096 2021/03/21 12:40:52 temp/.git/objects/pack\n",
            "drwxr-xr-x          4,096 2021/03/21 12:40:52 temp/.git/refs\n",
            "drwxr-xr-x          4,096 2021/03/21 12:40:52 temp/.git/refs/heads\n",
            "drwxr-xr-x          4,096 2021/03/21 12:40:52 temp/.git/refs/tags\n",
            "/content/temp\n",
            "On branch master\n",
            "\n",
            "Initial commit\n",
            "\n",
            "nothing to commit\n",
            "error: src refspec master does not match any.\n",
            "error: failed to push some refs to 'https://9c057450d445230b40b31c08ed177bd1b3946a71@github.com/akritikts/DeepLearning.git'\n",
            "/content\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JOr2bP8kAzKE"
      },
      "source": [
        "## Assignment objective\n",
        "\n",
        "\n",
        "\n",
        "This assignment is for you to demonstrate the knowledge in deep learning that you have acquired from the lectures and practical lab materials. Most tasks in this assignment are straightforward applications of the practical materials in weeks 1-5. Going through these materials before attempting this assignment is highly recommended.\n",
        "\n",
        "In this assignment, you are going to work with the Fashion-MNIST dataset for image recognition. The dataset contains 10 classes of 28x28 grayscale images. You will see some examples in the visualization task below. \n",
        "\n",
        "This assignment consists of five tasks.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cDqvFgbRAzKG"
      },
      "source": [
        "## Task 1 Load the data\n",
        "\n",
        "*(weight ~5%)*\n",
        "\n",
        "Load the Fashion MNIST dataset (https://github.com/zalandoresearch/fashion-mnist). You may get the data via Keras (keras.datasets) or Tensorflow Datasets (tfds). "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Krg_ro9M-t_k",
        "outputId": "b9670d6c-352a-4ba4-b4f6-e421b6cb87e2"
      },
      "source": [
        "import tensorflow as tf\n",
        "import tensorflow_datasets\n",
        "import math\n",
        "import numpy as np\n",
        "\n",
        "# Loading the data using tensorflow\n",
        "fashion_mnist = tf.keras.datasets.fashion_mnist\n",
        "\n",
        "(train_images, train_labels), (test_images, test_labels) = fashion_mnist.load_data()\n",
        "\n",
        "# class names\n",
        "class_names = ['T-shirt/top', 'Trouser', 'Pullover', 'Dress', 'Coat',\n",
        "               'Sandal', 'Shirt', 'Sneaker', 'Bag', 'Ankle boot']"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-labels-idx1-ubyte.gz\n",
            "32768/29515 [=================================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-images-idx3-ubyte.gz\n",
            "26427392/26421880 [==============================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-labels-idx1-ubyte.gz\n",
            "8192/5148 [===============================================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-images-idx3-ubyte.gz\n",
            "4423680/4422102 [==============================] - 0s 0us/step\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nFbty6i5AzKm"
      },
      "source": [
        "## Task 2 Understand the data\n",
        "\n",
        "*(weight ~15%)*\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 826
        },
        "id": "Vg6WnS54Cl_G",
        "outputId": "fadaba34-12ad-4885-f5c4-3453240b854e"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# 25 images from the train set in the form of 5x5 matrix\n",
        "print(\"25 images from the train set in the form of 5x5 matrix\")\n",
        "plt.figure(figsize=(5,5))\n",
        "for i in range(25):\n",
        "    plt.subplot(5,5,i+1)\n",
        "    plt.xticks([])\n",
        "    plt.yticks([])\n",
        "    plt.grid(False)\n",
        "    plt.imshow(train_images[i])\n",
        "plt.show()\n",
        "\n",
        "# training data\n",
        "print(\"Data type of training data: \", train_images.dtype)\n",
        "print(\"Shape of training data: \", train_images.shape)\n",
        "print(\"Data type of training labels: \", train_labels.dtype)\n",
        "print(\"Shape of training labels: \", train_labels.shape)\n",
        "print(\"Length of training data: \", len(train_labels))\n",
        "\n",
        "# test data\n",
        "print(\"Data type of test data: \", test_images.dtype)\n",
        "print(\"Shape of test data: \", test_images.shape)\n",
        "print(\"Data type of test labels: \", test_labels.dtype)\n",
        "print(\"Shape of test labels: \", test_labels.shape)\n",
        "print(\"Length of test data: \", len(test_images))\n",
        "\n",
        "# unique labels in the data\n",
        "print(\"Labels  in the data: \", np.unique(test_labels))\n",
        "\n",
        "# size of each image\n",
        "print(\"Size of each image\", train_images[0].shape)\n",
        "\n",
        "# memory required for training data\n",
        "print(\"Memory required for training data: \", train_images.size * train_images.itemsize, \"bytes\")\n",
        "\n",
        "\n",
        "# plotting the numeric range of input\n",
        "print(\"\\nPlotting the numeric range of input: [0-255]\")\n",
        "plt.figure()\n",
        "plt.imshow(train_images[100])\n",
        "plt.colorbar()\n",
        "plt.grid(False)\n",
        "plt.show()\n",
        "\n",
        "\n",
        "# The numeric range of input data is in the range [0,255], these values need to be normalized to the range [0,1]. \n",
        "# Normalisation ensures that each input parameter (pixel, in this case) has a similar data distribution. \n",
        "# This makes convergence faster while training the network.\n",
        "\n",
        "# normalisation\n",
        "train_images = train_images / 255.0\n",
        "test_images = test_images / 255.0"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "25 images from the train set in the form of 5x5 matrix\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 360x360 with 25 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Data type of training data:  uint8\n",
            "Shape of training data:  (60000, 28, 28)\n",
            "Data type of training labels:  uint8\n",
            "Shape of training labels:  (60000,)\n",
            "Length of training data:  60000\n",
            "Data type of test data:  uint8\n",
            "Shape of test data:  (10000, 28, 28)\n",
            "Data type of test labels:  uint8\n",
            "Shape of test labels:  (10000,)\n",
            "Length of test data:  10000\n",
            "Labels  in the data:  [0 1 2 3 4 5 6 7 8 9]\n",
            "Size of each image (28, 28)\n",
            "Memory required for training data:  47040000 bytes\n",
            "\n",
            "Plotting the numeric range of input: [0-255]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PLYAtx-ZAzKo"
      },
      "source": [
        "Display 25 images from the train set in the form of 5x5 matrix.\n",
        "\n",
        "Answer the following questions:\n",
        "\n",
        "1. What are the data types and shapes of the features and the label? \n",
        "2. What are the unique labels in this dataset?\n",
        "3. How many training images and how many test images?\n",
        "4. What is the size of each image? How much memory is required for holding the whole training data.\n",
        "5. Find out the numeric range of the input. Do we need to rescale the input? Why?\n",
        " "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-tkl-PwsrXip"
      },
      "source": [
        "## Task 3 Construct an input pipeline\n",
        "\n",
        "*(weight ~15%)*\n",
        "\n",
        "Creat train/validate/test data splits and construct tf.data pipelines. Make sure that the training data is batched. \n",
        "\n",
        "- How do you determine the batch size?\n",
        "- Do we need to shuffle the training data? If yes, how do you determine the buffer size?\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hVmNO1PzOU6E"
      },
      "source": [
        "# train-data shuffling ensures that the model gets to see all permutauions of input in every batch.\n",
        "\n",
        "# shuffle train data\n",
        "from sklearn.utils import shuffle\n",
        "train_images, train_labels = shuffle(train_images, train_labels)\n",
        "\n",
        "# combine training features and labels\n",
        "train_dataset = []\n",
        "for i in range(len(train_images)):\n",
        "   train_dataset.append([train_images[i], train_labels[i]])\n",
        "\n",
        "# combine test features and labels\n",
        "test_dataset = []\n",
        "for i in range(len(test_images)):\n",
        "   train_dataset.append([test_images[i], test_labels[i]])\n",
        "\n",
        "# batch size should be in powers of 2 and the larger the better, provided that the batch fits into the (GPU) memory;\n",
        "# small batches can offer a regularizing effect, perhaps due to the noise they add to the learning process.\n",
        "# larger batches provide a more accurate estimate of the gradient.\n",
        "\n",
        "# defining the batch size\n",
        "BATCH_SIZE = 32\n"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lnE7ZFS_AzLg"
      },
      "source": [
        "## Task 4 Construct a deep forward neural network\n",
        "\n",
        "*(weight ~35%)*"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "X6gil-HshhHI"
      },
      "source": [
        ""
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "h2qU873qfGVY"
      },
      "source": [
        ""
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ChlhkwhkAzLi"
      },
      "source": [
        "### Task 4.1 Setting up a model for training"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yuxhQM6jAzLl"
      },
      "source": [
        "Construct a deep feedforward neural network. You need to decide and report the following configurations:\n",
        "\n",
        "- Output layer: \n",
        "    - How many output nodes?\n",
        "    - Which activation function?\n",
        "- Hidden layers:\n",
        "    - How many hidden layers?\n",
        "    - How many nodes in each layer?\n",
        "    - Which activation function for each layer?\n",
        "- Input layer\n",
        "    - What is the input size?\n",
        "- The loss function\n",
        "- The metrics for model evaluation (which may be different from the loss function)\n",
        "- The optimiser\n",
        "\n",
        "Justify your model design decisions.\n",
        "\n",
        "Plot the model structure `using keras.utils.plot_model` or similar tools.\n",
        "\n",
        "What is the number of trainable parameters in the model. Explain how the total number can be estimated from the model configurations."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MZ2BVK5tAzMM"
      },
      "source": [
        "### Task 4.2 Fitting the model\n",
        "\n",
        "Before you fit the model. Think about what initialisation method have you chosen? If you did not specify the initialisation method, find out what is the default one. Choose a layer and visualise its initial weights. (Hint: You may use UMAP to visualise a collection of high-dimension vectors.)\n",
        "\n",
        "Decide and report the following training setting:\n",
        "\n",
        "1. The training batch size\n",
        "2. The number of training epochs (at least 1,000 epochs recommended)\n",
        "3. The learning rate. If you used momentum or a learning rate schedule, please report the configuration as well.\n",
        "\n",
        "Now fit the model. Show how the training loss changes. How did you decide when to stop training?\n",
        "\n",
        "After fitting the model, visualise the model weights again. How did the weights change? Why?\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9C2VXCM-4Nw9"
      },
      "source": [
        "### Task 4.3 Check the training using TensorBoard\n",
        "\n",
        "Use TensorBoard to visualise the training process. Show screenshots of your TensorBoard output.\n",
        "\n",
        "**Optional task:** Record the gradients during training and use TensorBoard to visualise the gradients."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4EgXZCBNAzMs"
      },
      "source": [
        "## Task 5 Overfitting and regularisation\n",
        "\n",
        "*(weight ~30%)*\n",
        "\n",
        "Go back to the previous task. Plot the training and validation loss and accuracy. Answer the following questions:\n",
        "\n",
        "1. Do you see overfitting or underfitting? Why?\n",
        "2. If you see overfitting, at which epoch did it happen?\n",
        "\n",
        "Now retrain the model with only 200 training examples. (Make sure that you  reinitialise the weights before retraining.) Do you see overfitting? How did the validation loss and accuracy change?\n",
        "\n",
        "Neural networks are overparametrised models, meaning there can be more parameters than the training examples. Some form of regularisation is almost always necessary to obtain a useful model. Below are some options:\n",
        "\n",
        "1. Add dropout\n",
        "2. Add Batch Normalisation\n",
        "3. Add layer-specific weight regularizations\n",
        "4. Change the learning rate\n",
        "\n",
        "In addition, you may also try changing the weight initialisation method.\n",
        "\n",
        "Apply different regularisation techniques to the model training. You may also try other techniques for improving training such as learning rate scheduling (see https://www.tensorflow.org/guide/keras/train_and_evaluate#using_learning_rate_schedules).\n",
        "\n",
        "Run **five or more** experiments of different training configurations and record the validation accuracy achieved in the Markdown table below. You may modify the table heading to match your experiment design.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "d1LRIBrEgu92"
      },
      "source": [
        "\n",
        "|Dropout (rate)   | Batch Normalisation (Y/N)  | Optimiser  | Learning Rate  | Number of Epochs   |  Validation Accuracy |\n",
        "|---|---|---|---|---|---|\n",
        "|   |   |   |   |   |   |\n",
        "|   |   |   |   |   |   |\n",
        "|   |   |   |   |   |   |\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dLjJA98Qgxg5"
      },
      "source": [
        "\n",
        "Answer the following questions:\n",
        "\n",
        "1. Which configuration achieved the best validation accuracy? Report the test accuracy of your final model.\n",
        "2. Which setting had the most impact and which one had the least impact?"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UgzZh0jN5ULb"
      },
      "source": [
        "## Task 6 Generate HTML output\n",
        "\n",
        "Use *nbconvert* to convert your completed notebook into an HTML file and name it **[YourID]_assingment1_output.html**."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cWbr7jjEAzNb"
      },
      "source": [
        "---\n",
        "**END OF ASSIGNMENT ONE**\n",
        "\n"
      ]
    }
  ]
}