datasets/mnist/MNIST_-_Exploratory_Data_Analysis.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": null
      },
      "outputs": [],
      "source": [
        "# Copyright 2022 Google LLC\n",
        "#\n",
        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
        "# you may not use this file except in compliance with the License.\n",
        "# You may obtain a copy of the License at\n",
        "#\n",
        "#      http://www.apache.org/licenses/LICENSE-2.0\n",
        "#\n",
        "# Unless required by applicable law or agreed to in writing, software\n",
        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
        "# See the License for the specific language governing permissions and\n",
        "# limitations under the License."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": null
      },
      "source": [
        "[![View on GitHub][github-badge]][github-eda] [![Open In Colab][colab-badge]][colab-eda] [![Open in Binder][binder-badge]][binder-eda]\n",
        "\n",
        "[github-badge]: https://img.shields.io/badge/View-on%20GitHub-blue?logo=GitHub\n",
        "[colab-badge]: https://colab.research.google.com/assets/colab-badge.svg\n",
        "[binder-badge]: https://static.mybinder.org/badge_logo.svg\n",
        "\n",
        "[github-eda]: https://github.com/mbrukman/reimplementing-ml-papers/blob/main/datasets/mnist/MNIST_-_Exploratory_Data_Analysis.ipynb\n",
        "[colab-eda]: https://colab.research.google.com/github/mbrukman/reimplementing-ml-papers/blob/main/datasets/mnist/MNIST_-_Exploratory_Data_Analysis.ipynb\n",
        "[binder-eda]: https://mybinder.org/v2/gh/mbrukman/reimplementing-ml-papers/main?filepath=datasets/mnist/MNIST_-_Exploratory_Data_Analysis.ipynb"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": null
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "\n",
        "from matplotlib import pyplot as plt\n",
        "from tensorflow import keras"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": null
      },
      "outputs": [],
      "source": [
        "%pip install -q -U 'einops==0.4'\n",
        "import einops"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": null
      },
      "outputs": [],
      "source": [
        "%%capture\n",
        "# Load the MNIST dataset.\n",
        "(x_train_raw, y_train_raw), (x_test_raw, y_test_raw) = keras.datasets.mnist.load_data()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": null
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Raw data:\n",
            "Train x: (60000, 28, 28)\n",
            "      y: (60000,)\n",
            "Test  x: (10000, 28, 28)\n",
            "      y: (10000,)\n"
          ]
        }
      ],
      "source": [
        "# Examine the dataset shape.\n",
        "print('Raw data:')\n",
        "print(f'Train x: {x_train_raw.shape}')\n",
        "print(f'      y: {y_train_raw.shape}')\n",
        "print(f'Test  x: {x_test_raw.shape}')\n",
        "print(f'      y: {y_test_raw.shape}')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": null
      },
      "source": [
        "## Examine and update the input data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "height": 485
        },
        "id": null
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x576 with 25 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "# Let's see what some of the training images look like.\n",
        "plt.figure(figsize=(8, 8))\n",
        "for i in range(25):\n",
        "    plt.subplot(5, 5, i + 1)\n",
        "    plt.imshow(x_train_raw[i, :, :], cmap='Greys')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": null
      },
      "outputs": [],
      "source": [
        "# Scale train and test inputs by converting them from range of [0, 255] to [0.0, 1.0]\n",
        "x_train = x_train_raw.astype('float32') / 255.0\n",
        "x_test = x_test_raw.astype('float32') / 255.0"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": null
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "X train (raw): min=0, max=255\n",
            "     (scaled): min=0.0, max=1.0\n",
            "X test  (raw): min=0, max=255\n",
            "     (scaled): min=0.0, max=1.0\n"
          ]
        }
      ],
      "source": [
        "# Example statistics of X before and after scaling.\n",
        "print(f\"X train (raw): min={x_train_raw.min()}, max={x_train_raw.max()}\")\n",
        "print(f\"     (scaled): min={x_train.min()}, max={x_train.max()}\")\n",
        "print(f\"X test  (raw): min={x_test_raw.min()}, max={x_test_raw.max()}\")\n",
        "print(f\"     (scaled): min={x_test.min()}, max={x_test.max()}\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": null
      },
      "outputs": [],
      "source": [
        "# Add a channel dimension so that our inputs have the dimension (28, 28, 1)\n",
        "# rather than (28, 28). This is done by converting our input of\n",
        "# (batch, width, height) -> (batch, width, height, channels) with channels=1.\n",
        "#\n",
        "# In this specific case, this reshapes (60000, 28, 28) -> (60000, 28, 28, 1).\n",
        "#\n",
        "# Although this is similar to expanding dimensions via NumPy as follows:\n",
        "#\n",
        "#     x_train = np.expand_dims(x_train_raw, -1)\n",
        "#     x_test = np.expand_dims(x_test_raw, -1)\n",
        "#\n",
        "# the approach with `einops` makes it much more readable and understandable.\n",
        "x_train = einops.rearrange(x_train, 'b w h -> b w h ()')\n",
        "x_test = einops.rearrange(x_test, 'b w h -> b w h ()')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": null
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Dimension comparison:\n",
            "Train x (raw):      (60000, 28, 28)\n",
            "      x (expanded): (60000, 28, 28, 1)\n",
            "Test  x (raw):      (10000, 28, 28)\n",
            "      x (expanded): (10000, 28, 28, 1)\n"
          ]
        }
      ],
      "source": [
        "print('Dimension comparison:')\n",
        "print(f'Train x (raw):      {x_train_raw.shape}')\n",
        "print(f'      x (expanded): {x_train.shape}')\n",
        "print(f'Test  x (raw):      {x_test_raw.shape}')\n",
        "print(f'      x (expanded): {x_test.shape}')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": null
      },
      "source": [
        "## Examine and format the labels"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": null
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "y[0:4] = [5 0 4 1]\n"
          ]
        }
      ],
      "source": [
        "# Examine the format of the raw Y values.\n",
        "print(f'y[0:4] = {y_train_raw[0:4]}')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": null
      },
      "source": [
        "Since the $y$ values are the expected digits, but our network will output one-hot encoding of all 10 digits with probabilities of each value, we will need to convert it to a categorical value below."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": null
      },
      "outputs": [],
      "source": [
        "# Convert labels to one-hot encoding to match the output of LeNet-5 network.\n",
        "num_classes = 10\n",
        "y_train = keras.utils.to_categorical(y_train_raw, num_classes)\n",
        "y_test = keras.utils.to_categorical(y_test_raw, num_classes)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": null
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Processed data shape:\n",
            "Train y (raw):     (60000,)\n",
            "      y (one-hot): (60000, 10)\n",
            "Test  y (raw):     (10000,)\n",
            "      y (one-hot): (10000, 10)\n"
          ]
        }
      ],
      "source": [
        "print('Processed data shape:')\n",
        "print(f'Train y (raw):     {y_train_raw.shape}')\n",
        "print(f'      y (one-hot): {y_train.shape}')\n",
        "print(f'Test  y (raw):     {y_test_raw.shape}')\n",
        "print(f'      y (one-hot): {y_test.shape}')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": null
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Y (train) sample values before preprocessing:\n",
            "[5 0 4 1]\n",
            "\n",
            "Y (train) sample values after preprocessing:\n",
            "[[0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n",
            " [1. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            " [0. 0. 0. 0. 1. 0. 0. 0. 0. 0.]\n",
            " [0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]]\n"
          ]
        }
      ],
      "source": [
        "# Examine the format of the Y values before and after preprocessing.\n",
        "print('Y (train) sample values before preprocessing:')\n",
        "print(y_train_raw[0:4])\n",
        "\n",
        "print('\\nY (train) sample values after preprocessing:')\n",
        "print(y_train[0:4])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": null
      },
      "source": [
        "## Using the library to simplify input and label transformations\n",
        "\n",
        "This directory provides [`mnist_keras.py`][mnist-github] library which includes the\n",
        "above transformations for scaling the input and converting the labels to a\n",
        "categorical (one-hot) representation from the default numeric values, so you\n",
        "can use it to simplify the data setup for training and testing MNIST models.\n",
        "\n",
        "You can see examples of how the library can be used in the [LeNet notebooks][lenet].\n",
        "\n",
        "[mnist-github]: https://raw.githubusercontent.com/mbrukman/reimplementing-ml-papers/main/datasets/mnist/mnist_keras.py\n",
        "[lenet]: https://github.com/mbrukman/reimplementing-ml-papers/tree/main/lenet"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "name": "MNIST: Exploratory Data Analysis"
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}