mnist_conv_gen.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Deep learning with Python Francois Chollet 2018\n",
    "\n",
    "Train a convolution network to classify digits using the MNIST dataset with augmenting the training set using image generation."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load the train and test sets\n",
    "\n",
    "We start by loading the train and test sets consisting of 60000 and 10000 28x28 pixel grayscale images. Each pixel has a value between 0 and 255 stored in numpy uint8."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The shape of the training set is: (60000, 28, 28)\n",
      "The shape of the test set is: (10000, 28, 28)\n",
      "The data type of the training set is: uint8\n",
      "The data type of the test set is: uint8\n"
     ]
    }
   ],
   "source": [
    "from keras import datasets\n",
    "\n",
    "(orig_train_images, orig_train_labels), (orig_test_images, orig_test_labels) = datasets.mnist.load_data()\n",
    "\n",
    "print(f'The shape of the training set is: {orig_train_images.shape}')\n",
    "print(f'The shape of the test set is: {orig_test_images.shape}')\n",
    "print(f'The data type of the training set is: {orig_train_images.dtype}')\n",
    "print(f'The data type of the test set is: {orig_test_images.dtype}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Example images\n",
    "\n",
    "Let's plot first images from the training and test sets with their associated labels to see what they look like."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAALIAAACcCAYAAADbLPEuAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAADsNJREFUeJztnXuwVdV9xz9fQVSUh/gAgyJCGRAfxU4Ea52ogxZJZHxmRhKUVGKaVlo7Y2isrWlM1ZIoJmWgI87EF1HUJmlRmlQ6AWSskWp9RihqTISLKBFBEPEB/vrHXud2r809h8s9+9xzWPw+M2fO/u219lq/vc/3rPXb++yzfzIzHGdvZ79mO+A4ZeBCdpLAhewkgQvZSQIXspMELmQnCbpFyJLukHRD2XVbBUk/lzS12X7sDkkHSFopaVCw75F0Uye3XSbpq13st0vbBn//V9KRu6u7WyFL+q2kc/bUiTxm9nUz+4ey63YHkr4i6Yladcxsopnd210+1cHXgOVm9lazHakQju9OSe/nXmcBmNlHwF3AN3fXTt0jsqSe9bbhdBt/CsxvthMd8EszOyT3WpYrewCYKumAWg3UFLKk+cAQ4NHwTflrSUMlmaRpktYAS0Ldf5H0lqT3JC2XdEKunfYpTNJZktokXStpg6T1kv6ki3UPk/SopC2SnpZ0U7XRU9KBkn4kaaOkzaH+wFDWT9IPQ/vrQjs9JB0P3AH8Ydj/zVXabp86wwjzX5K+H/p5XdLpYf3asB9Tc9t+QdJzYR/WSvp2oe0rJL0R/L4hP0NK2k/SdZJ+HcofljSgio9DgOHAiirlh0paJOl3kjaF5aML1YZL+u/wGS/M9yXpNElPhn1+oTKq1ouZtQGbgNNq1aspZDO7HFgDTArflO/lis8EjgcmBPvnwAjgSOBZ4P4aTQ8C+gGDgWnAXEmHdqHuXGBbqDM1vKoxNbRzDHAY8HVgeyi7F9gB/B5wCvDHwFfNbFWoVxkx+tdoP8844MXQzwPAg8Cpof0pwBxJh4S624ArgP7AF4A/k3QhgKTRwD8DXwaOyh2HCn8JXEj2WXyG7AOfW8Wnk4DXzWxHlfL9gLuBY8kGr+3AnEKdK4ArQ187gNnBz8HAvwM3AQOAbwA/kXRElb6KnCLpHUmvhC9rcZZfBfx+zRbMrOYL+C1wTs4eChgwrMY2/UOdfsG+B7gpLJ8VDlLPXP0NwGl7UhfoAXwCjMyV3QQ8UcWnK4EngZML6wcCHwEH5dZNBpaG5a9UazNXfxmZ8Cv1X82VnRSOxcDcuo3AmCpt/QD4flj+FrAgV9Yb+LjyeYQPeHyu/KhwTHp20O6XgacK69qPdQf1xwCbCvs4M2ePDr70IIth5xe2fwyYWjw+HfQzDDiO7It0ErAS+JtCnfuBb9X6DOqJkddWFsI0PDNMcVvIxA9weJVtNxZGhg+AQ/aw7hFAz7wfheUi88kO7oOS3pT0PUn7k41A+wPrw7S4GZhHNrN0lbdzy9sBzKy47hAASeMkLQ1T+ntkM0DluH0mv09m9gHZl6DCscC/5vxeBewk+3IW2QT0qeawpN6S5oUwZguwHOgvqUeuWv74vkF23A4Pfnyx4kfw5QyyL1ZNzOx1M/uNmX1qZi8B3wEuLVTrA3QY1lXojJCr3R6XX/8l4ALgHLLpb2hYr06031V+Rza95eO4Y6pVNrNPzOxGMxsNnA6cTzZVriUbkQ83s/7h1dfMKjF+o28PfAB4BDjGzPqRxeSV47ae3P5JOogsXKmwFpiY87u/mR1oZus66OdFYFgH03aFa4GRwDgz6wt8rtJtrk7++A4hG/3fCX7ML/hxsJnN3P3u74Kxq26OB16otVFnhPw22fBfiz5kYthINv3d0ol268LMdgI/Bb4dRpNRZMLsEElnSzopjDBbyD6EnWa2HlgMzJLUN5xADZd0Ztj0beBoSb0atCt9gHfN7ENJY8kGhQo/BiaFk8VewI3EH/IdwM2Sjg37eISkCzrqxLKTpleBsTX82A5sDidxf99BnSmSRkvqTTZy/jh8Dj8Kfk4Is/OByk7UiyeLuyBpYu6kexRwA7AwVz6YLO5+qlY7nRHyPwJ/F6aMb1Spcx/ZVLOOLMap2WmJTCebAd4iCx0WkH2hOmIQmTC2kE3Bj5N9AJB9AXqR+b4p1KtMi0uAl4G3JL1T/i7w58B3JG0li4kfrhSY2cvAX5CdLK4HtpKdI1T28Z/IRvPFYfunyE40qzEPuLxK2Q+Ag8hG2KeA/+igznyyuPot4ECyk03MbC3ZjHw92Uy5FphB5/Q1HnhR0jbgZ2SDU34g/BJwr2XXlKsiS+jGeknfBQaZWcv/ytYVwpWOzcAIM/tNF7Y/AHiO7ARxfdn+lU3w9wXgc2a2oWbdvVnIYSrqBbxEdnnrZ2Rnx//WVMdKRNIk4BdkIcUsshH3D2xv/uAawN5+01AfsqloG9mUPItcfJUIFwBvhtcI4DIX8a7s1SOy41Ro6Igs6TxJqyW9Jum6Rvbl7Ns0bEQOl7leAc4F2oCngclmtrIhHTr7NI28c20s8JqZvQ4g6UGyeK+qkCV5nLP38Y6ZdfaeiobRyNBiMPFPmm3EN7w4afBGsx2Axo7IHf08vcuIK+lrZDd8O06XaaSQ24h/mz+a7BJShJndCdwJHlo4XaeRocXTwAhJx4X7BC4j+znVcUqnYSOyme2QNJ3s1skewF3h3gHHKZ2W+kHEQ4u9kv8xs88224m9/SdqxwFcyE4iuJCdJHAhO0ngQnaSwIXsJIEL2UkCF7KTBC5kJwlcyE4SuJCdJHAhO0ngD+kuiR49ekR2v379Or3t9OnTI7t3796RPXLkyMi++uqrI/u2226L7MmTJ0f2hx9+GNkzZ/7/I9luvPHGTvvZyviI7CSBC9lJAheykwQeIweGDBkS2b16xU+RPf300yP7jDPOiOz+/eOsDJdccklpvrW1tUX27NmzI/uiiy6K7K1bt0b2Cy/EjxZ+/PHHS/OtVfAR2UkCF7KTBC5kJwn22T+fjhkzJrKXLFkS2XtyHbhsPv3008i+8sorI/v999+vuf369fEzvDdt2hTZq1evrsO7XfA/nzpOWbiQnSRwITtJsM9eR16zZk1kb9y4MbLLjJFXrIjTP2/eHOc+PPvssyP7448/juz581sxD3pr4SOykwQuZCcJXMhOEuyzMfK7774b2TNmzIjs888/P7Kfe+65yC7e71Dk+eefb18+99xzo7Jt27ZF9gknnBDZ11xzTc22nV3xEdlJglKELOkuSRsk/Sq3boCk/5T0ang/tIy+HKcjyhqR7wHOK6y7DviFmY0gS0HrefachlHavRaShgKLzOzEYK8GzjKz9ZKOApaZ2cgaTbTUg7779u0b2cV7fOfNmxfZ06ZNi+wpU6a0Ly9YsKBk71qK5O+1GFjJQB/ej2xgX84+TtOvWnh6MqcMGjkivx1CCsL7ho4qmdmdZvbZVpienL2XRo7IjwBTgZnhfWED+yqdLVu21Cx/7733apZfddVV7csPPfRQVFa839ipn7Iuvy0AfgmMlNQmaRqZgM+V9CpZYvWZtdpwnHooZUQ2s8lVisaX0b7j7A7/Zc9Jgn32P3v1cvDBB0f2o48+Gtlnnnlm+/LEiROjssWLFzfOse4n+evIjtNtuJCdJHAhO0ngMXJJDB8+PLKfffbZ9uXif/SWLl0a2c8880xkz507N7Jb6TPqAI+RHacsXMhOEnho0SDyj3q9++67o7I+ffrU3Pb666+P7Pvuuy+yi4/EajIeWjhOWbiQnSRwITtJ4DFyN3DiiSdG9u233x7Z48fXvreq+Leqm2++ObLXrVtXh3d14zGy45SFC9lJAheykwQeIzeBYiqzSZMmRXbxurOkyC6miSg+kqub8RjZccrChewkgQvZSQKPkVuQjz76KLJ79oz/I7xjx47InjBhQmQvW7asIX5VwWNkxykLF7KTBC5kJwma/hDDfYGTTz45si+99NLIPvXUUyO7GBMXWblyZWQvX768Du/SwEdkJwlcyE4SuJCdJPAYuSRGjoyzSkyfPr19+eKLL47KBg0atEdt79y5M7KL/9nzx9T6iOwkQt1ClnSMpKWSVkl6WdI1Yb2nJ3O6jTJG5B3AtWZ2PHAacLWk0Xh6MqcbqTtGDhmbKtmbtkpaBQwGLgDOCtXuBZYB36y3v2ZRjGsnT46fbZ6PiQGGDh3a5b6Kj9Aq/kfvkUce6XLbqVJqjBxy7Z0CrMDTkzndSGlXLSQdAvwE+Csz21L8V0ON7Tw9mVM3ZSXD2Z9MxPeb2U/Dak9P5nQbdY/IyobeHwKrzCz/wIa9Kj3ZwIEDI3v06NGRPWfOnMgeNWpUl/tasWJFZN96662RvXBhfKj8OvHuKSO0+CPgcuAlSc+HddeTCfjhkKpsDfDFEvpynA4p46rFE0C1gNjTkzndgv+y5yTBPnOvxYABAyK7+Dy1MWPGRPawYcPq6u/JJ59sX541a1ZU9thjj0X29u3b6+rL8RHZSQQXspMELmQnCZKKkceNG9e+PGPGjKhs7NixkT148OC6+vrggw8ie/bs2ZF9yy23tC9v27atrr6c3eMjspMELmQnCZIKLfIpwfLLnaH4F/tFixZFdvExVcVLasXspk734iOykwQuZCcJXMhOEvhjZZ168cfKOk5ZuJCdJHAhO0ngQnaSwIXsJIEL2UkCF7KTBC5kJwlcyE4SuJCdJHAhO0nQavcjvwO8ARwellsR9y3m2G7ur0Na6qahCpKeaYUbUTrCfWtNPLRwksCF7CRBqwr5zmY7UAP3rQVpyRjZcfaUVh2RHWePaDkhSzpP0mpJr0lqakozSXdJ2iDpV7l1Tc8f6LkNd6WlhCypBzAXmAiMBiaHnH3N4h7gvMK6Vsgf6LkNC7SUkIGxwGtm9rqZfQw8SJavrymY2XLg3cLqC8jyBhLeL+xWp8jSvZnZs2F5K5DPbdhU35pFqwl5MLA2Z7eFda1ES+UP9NyGGa0m5I5ykfhllSoUcxs2259m0mpCbgOOydlHA282yZdqdCp/YKOpJ7dhirSakJ8GRkg6TlIv4DKyfH2tRCV/IDQpf2Anchs2zbemYWYt9QI+D7wC/Br42yb7soAsYfwnZLPFNOAwsisCr4b3AU3w6wyykOtF4Pnw+nwr+Nasl/+y5yRBq4UWjtMlXMhOEriQnSRwITtJ4EJ2ksCF7CSBC9lJAheykwT/B2OSi/MU/T3QAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 144x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 144x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "def show_image(title, images, labels):\n",
    "    from matplotlib import pyplot\n",
    "    pyplot.rcParams[\"figure.figsize\"] = (2,2)\n",
    "    pyplot.title(f'{title} set image (label {labels[0]})')\n",
    "    pyplot.imshow(images[0], cmap=\"gray\")\n",
    "    pyplot.show()\n",
    "\n",
    "show_image('training', orig_train_images, orig_train_labels)\n",
    "show_image('test', orig_test_images, orig_test_labels)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Shape the training and test set images\n",
    "\n",
    "We need to shape the images to be suitable for training the network. Images are converted from 28x28 2D tensors with 0-255 numpy uint8s to 28*28 1D tensors with 0-1 numpy float32s. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The shape of the original training set: (60000, 28, 28)\n",
      "The shape of the actual training set: (60000, 28, 28, 1)\n",
      "The shape of the original test set: (10000, 28, 28)\n",
      "The shape of the actual test set: (10000, 28, 28, 1)\n",
      "The data type of the original training set: uint8\n",
      "The data type of the actual training set: float32\n",
      "The data type of the original test set: uint8\n",
      "The data type of the actual test set: float32\n"
     ]
    }
   ],
   "source": [
    "train_images = orig_train_images.reshape((orig_train_images.shape[0], 28, 28, 1)).astype('float32') / 255\n",
    "test_images = orig_test_images.reshape((orig_test_images.shape[0], 28, 28, 1)).astype('float32') / 255\n",
    "\n",
    "print(f'The shape of the original training set: {orig_train_images.shape}')\n",
    "print(f'The shape of the actual training set: {train_images.shape}')\n",
    "print(f'The shape of the original test set: {orig_test_images.shape}')\n",
    "print(f'The shape of the actual test set: {test_images.shape}')\n",
    "print(f'The data type of the original training set: {orig_train_images.dtype}')\n",
    "print(f'The data type of the actual training set: {train_images.dtype}')\n",
    "print(f'The data type of the original test set: {orig_test_images.dtype}')\n",
    "print(f'The data type of the actual test set: {test_images.dtype}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Augment the training set by generating additional images"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "\n",
    "gen_train_images = ImageDataGenerator(\n",
    "    rotation_range=20,\n",
    "    shear_range=0.2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Shape the training and test set labels\n",
    "\n",
    "We need to also shape the labels to be suitable for training the network. Scalar labels are replaced with 1D binary tensors with value 0 or 1 for each label. If the value is 1 then the image is associated with that label."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The shape of the original train labels: (60000,)\n",
      "The shape of the actual train labels: (60000, 10)\n",
      "The shape of the original test labels: (10000,)\n",
      "The shape of the actual test labels: (10000, 10)\n",
      "First original train label: 5\n",
      "First actual train label: [0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n",
      "First original test label: 7\n",
      "First actual test label: [0. 0. 0. 0. 0. 0. 0. 1. 0. 0.]\n"
     ]
    }
   ],
   "source": [
    "from keras.utils import to_categorical\n",
    "\n",
    "train_labels = to_categorical(orig_train_labels)\n",
    "test_labels = to_categorical(orig_test_labels)\n",
    "\n",
    "print(f'The shape of the original train labels: {orig_train_labels.shape}')\n",
    "print(f'The shape of the actual train labels: {train_labels.shape}')\n",
    "print(f'The shape of the original test labels: {orig_test_labels.shape}')\n",
    "print(f'The shape of the actual test labels: {test_labels.shape}')\n",
    "print(f'First original train label: {orig_train_labels[0]}')\n",
    "print(f'First actual train label: {train_labels[0]}')\n",
    "print(f'First original test label: {orig_test_labels[0]}')\n",
    "print(f'First actual test label: {test_labels[0]}')\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Define and train the network\n",
    "\n",
    "We use a sequential network. The input shape for the first layer must corresponds to the shape of the images (28*28 1D tensors). Once the network layers have been created we define the optimizer and loss function using the compile function. Finally, we train the network using the fit function and print out the loss and accuracy of the model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_16 (Conv2D)           (None, 26, 26, 32)        320       \n",
      "_________________________________________________________________\n",
      "max_pooling2d_11 (MaxPooling (None, 13, 13, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_17 (Conv2D)           (None, 11, 11, 64)        18496     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_12 (MaxPooling (None, 5, 5, 64)          0         \n",
      "_________________________________________________________________\n",
      "conv2d_18 (Conv2D)           (None, 3, 3, 64)          36928     \n",
      "_________________________________________________________________\n",
      "flatten_6 (Flatten)          (None, 576)               0         \n",
      "_________________________________________________________________\n",
      "dense_11 (Dense)             (None, 64)                36928     \n",
      "_________________________________________________________________\n",
      "dense_12 (Dense)             (None, 10)                650       \n",
      "=================================================================\n",
      "Total params: 93,322\n",
      "Trainable params: 93,322\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Epoch 1/7\n",
      "1875/1875 [==============================] - 20s 11ms/step - loss: 0.1667 - acc: 0.9473 - val_loss: 0.0458 - val_acc: 0.9849\n",
      "Epoch 2/7\n",
      "1875/1875 [==============================] - 17s 9ms/step - loss: 0.0596 - acc: 0.9815 - val_loss: 0.0527 - val_acc: 0.9824\n",
      "Epoch 3/7\n",
      "1875/1875 [==============================] - 17s 9ms/step - loss: 0.0493 - acc: 0.9851 - val_loss: 0.0325 - val_acc: 0.9896\n",
      "Epoch 4/7\n",
      "1875/1875 [==============================] - 17s 9ms/step - loss: 0.0420 - acc: 0.9870 - val_loss: 0.0287 - val_acc: 0.9911\n",
      "Epoch 5/7\n",
      "1875/1875 [==============================] - 17s 9ms/step - loss: 0.0366 - acc: 0.9891 - val_loss: 0.0415 - val_acc: 0.9877\n",
      "Epoch 6/7\n",
      "1875/1875 [==============================] - 17s 9ms/step - loss: 0.0372 - acc: 0.9891 - val_loss: 0.0534 - val_acc: 0.9894\n",
      "Epoch 7/7\n",
      "1875/1875 [==============================] - 18s 10ms/step - loss: 0.0348 - acc: 0.9899 - val_loss: 0.0324 - val_acc: 0.9921\n",
      "10000/10000 [==============================] - 1s 88us/step\n",
      "Loss: 0.03243839122550844\n",
      "Accuracy: 0.9921\n"
     ]
    }
   ],
   "source": [
    "import math\n",
    "from keras import models\n",
    "from keras import layers\n",
    "\n",
    "epochs = 7\n",
    "\n",
    "network = models.Sequential()\n",
    "network.add(layers.Conv2D(32, (3, 3), activation='relu', input_shape=(28, 28, 1)))\n",
    "network.add(layers.MaxPooling2D(2, 2))\n",
    "network.add(layers.Conv2D(64, (3, 3), activation='relu'))\n",
    "network.add(layers.MaxPooling2D(2, 2))\n",
    "network.add(layers.Conv2D(64, (3, 3), activation='relu'))\n",
    "network.add(layers.Flatten())\n",
    "network.add(layers.Dense(64, activation='relu'))\n",
    "network.add(layers.Dense(10, activation='softmax'))\n",
    "network.summary()\n",
    "\n",
    "network.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "history = network.fit_generator(gen_train_images.flow(train_images, train_labels, batch_size=32), \n",
    "               steps_per_epoch=len(train_images) / 32, epochs=epochs, validation_data=(test_images, test_labels))\n",
    "\n",
    "loss, accuracy = network.evaluate(test_images, test_labels)\n",
    "\n",
    "print(f'Loss: {loss}')\n",
    "print(f'Accuracy: {accuracy}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model overfitting\n",
    "\n",
    "There is very little evidence of overfitting in our model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from matplotlib import pyplot\n",
    "\n",
    "pyplot.rcParams[\"figure.figsize\"] = (5,5)\n",
    "\n",
    "train_accuracy = history.history['acc']\n",
    "validation_accuracy = history.history['val_acc']\n",
    "\n",
    "train_loss = history.history['loss']\n",
    "validation_loss = history.history['val_loss']\n",
    "\n",
    "epocs = range(1, len(train_loss) + 1)\n",
    "\n",
    "pyplot.plot(epocs, train_accuracy, 'bo', label='Training accuracy')\n",
    "pyplot.plot(epocs, validation_accuracy, 'b', label='Validation accuracy')\n",
    "pyplot.title('Training and validation accuracy')\n",
    "pyplot.xlabel('Epochs')\n",
    "pyplot.xlabel('Accuracy')\n",
    "pyplot.legend()\n",
    "pyplot.show()\n",
    "\n",
    "pyplot.plot(epocs, train_loss, 'bo', label='Training loss')\n",
    "pyplot.plot(epocs, validation_loss, 'b', label='Validation loss')\n",
    "pyplot.title('Training and validation loss')\n",
    "pyplot.xlabel('Epochs')\n",
    "pyplot.xlabel('Loss')\n",
    "pyplot.legend()\n",
    "pyplot.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Invalid prediction counts\n",
    "\n",
    "Lets show how may times each digit in the test set was incorrectly classified as one of the other digits."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy\n",
    "from matplotlib import pyplot\n",
    "\n",
    "predicted_labels = network.predict(test_images)\n",
    "\n",
    "## Find invalid predictions.\n",
    "\n",
    "invalid_prediction = []\n",
    "invalid_prediction_cnt = numpy.zeros((10,10), dtype=int)\n",
    "\n",
    "for i,v in enumerate(predicted_labels):\n",
    "    test_label = numpy.where(test_labels[i] == numpy.amax(test_labels[i]))[0].item(0)\n",
    "    predicted_test_label = numpy.where(predicted_labels[i] == numpy.amax(predicted_labels[i]))[0].item(0)\n",
    "    if test_label != predicted_test_label:\n",
    "        invalid_prediction_cnt[test_label][predicted_test_label] += 1        \n",
    "        invalid_prediction.append((i, test_label, predicted_test_label))\n",
    "        \n",
    "total_invalid_prediction_cnt = numpy.sum(invalid_prediction_cnt)        \n",
    "\n",
    "## Plot invalid prediction counts.\n",
    "\n",
    "x_test_labels = numpy.arange(10)\n",
    "y_predicted_labels = numpy.arange(10)\n",
    "\n",
    "fig, ax = pyplot.subplots(figsize=(8,8))\n",
    "\n",
    "ax.imshow(invalid_prediction_cnt, cmap=\"coolwarm\")\n",
    "ax.set_title(f'Invalid predictions ({total_invalid_prediction_cnt})')\n",
    "ax.set_xticks(x_test_labels)\n",
    "ax.set_yticks(y_predicted_labels)\n",
    "ax.set_xlabel('Test label')\n",
    "ax.set_ylabel('Predicted label')\n",
    "ax.invert_yaxis()\n",
    "\n",
    "for i in range(len(x_test_labels)):\n",
    "    for j in range(len(y_predicted_labels)):\n",
    "        # Show number of invalid predictions.\n",
    "        text = ax.text(j, i, invalid_prediction_cnt[i, j],\n",
    "                       ha=\"center\", va=\"center\", color=\"w\")\n",
    "\n",
    "fig.tight_layout()\n",
    "pyplot.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Invalid predictions\n",
    "\n",
    "Finally lets show all the images in the test set that were categorized incorrectly."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x2160 with 79 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pyplot.rcParams[\"figure.figsize\"] = (30,30)\n",
    "\n",
    "ncols = 10\n",
    "nrows = math.ceil(total_invalid_prediction_cnt / ncols)\n",
    "\n",
    "for i in range(total_invalid_prediction_cnt):\n",
    "    idx, test_label, predicted_test_label = invalid_prediction[i]\n",
    "    row = math.floor(i / ncols)\n",
    "    col = i % ncols\n",
    "    ax = pyplot.subplot2grid((nrows,ncols),(row,col))\n",
    "    ax.axis('off')\n",
    "    ax.set_title(f'{test_label}->{predicted_test_label}')\n",
    "    ax.imshow(orig_test_images[idx], cmap=\"gray\")\n",
    "pyplot.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}