KerasMLP-MNIST/keras-mlp-mnist.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import random\n",
    "import numpy as np\n",
    "import tensorflow as tf \n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's configure all random numbers generators to support determinism and obtain reproducible results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# settings for reproducibility\n",
    "seed = 42\n",
    "random.seed(seed)\n",
    "tf.random.set_seed(seed)\n",
    "np.random.seed(seed)\n",
    "\n",
    "os.environ['TF_DETERMINISTIC_OPS'] = '1'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load Data\n",
    "\n",
    "Keras comes with the MNIST data loader. Keras has a function `mnist.load_data()` which downloads the data from its servers if it is not present already. The data loaded using this function is divided into training and test sets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.datasets import mnist\n",
    "(train_images, train_labels), (test_images, test_labels) = mnist.load_data()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Checkout the data\n",
    "\n",
    "The data consists of handwritten numbers ranging from 0 to 9, along with their ground truth. It has 60,000 train samples and 10,000 test samples. Each sample is a 28x28 grayscale image."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.utils import to_categorical"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training data shape :  (60000, 28, 28) (60000,)\n",
      "Testing data shape :  (10000, 28, 28) (10000,)\n",
      "Total number of outputs :  10\n",
      "Output classes :  [0 1 2 3 4 5 6 7 8 9]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Ground Truth : 7')"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "print('Training data shape : ', train_images.shape, train_labels.shape)\n",
    "\n",
    "print('Testing data shape : ', test_images.shape, test_labels.shape)\n",
    "\n",
    "# Find the unique numbers from the train labels\n",
    "classes = np.unique(train_labels)\n",
    "classes_num = len(classes)\n",
    "print('Total number of outputs : ', classes_num)\n",
    "print('Output classes : ', classes)\n",
    "\n",
    "plt.figure(figsize=[10,5])\n",
    "\n",
    "# Display the first image in training data\n",
    "plt.subplot(121)\n",
    "plt.imshow(train_images[0,:,:], cmap='gray')\n",
    "plt.title(\"Ground Truth : {}\".format(train_labels[0]))\n",
    "\n",
    "# Display the first image in testing data\n",
    "plt.subplot(122)\n",
    "plt.imshow(test_images[0,:,:], cmap='gray')\n",
    "plt.title(\"Ground Truth : {}\".format(test_labels[0]))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Process the data\n",
    "\n",
    "* The images are grayscale and the pixel values range from 0 to 255.\n",
    "* Convert each image matrix ( 28x28 ) to an array ( 28*28 = 784 dimenstional ) which will be fed to the network as a single feature.\n",
    "* We convert the data to float and **scale** the values between 0 to 1.\n",
    "* We also convert the labels from integer to **categorical ( one-hot ) encoding** since that is the format required by Keras to perform multiclass classification. One-hot encoding is a type of boolean representation of integer data. It converts the integer to an array of all zeros except a 1 at the index of the number. For example, using a one-hot encoding of 10 classes, the integer 5 will be encoded as 0000010000"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Change from matrix to array of dimension 28x28 to array of dimention 784\n",
    "dim_data = np.prod(train_images.shape[1:])\n",
    "train_data = train_images.reshape(train_images.shape[0], dim_data)\n",
    "test_data = test_images.reshape(test_images.shape[0], dim_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Change to float datatype\n",
    "train_data = train_data.astype('float32')\n",
    "test_data = test_data.astype('float32')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Scale the data to lie between 0 to 1\n",
    "train_data /= 255\n",
    "test_data /= 255"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Change the labels from integer to categorical data\n",
    "train_labels_one_hot = to_categorical(train_labels)\n",
    "test_labels_one_hot = to_categorical(test_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original label 0 :  5\n",
      "After conversion to categorical ( one-hot ) :  [0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n"
     ]
    }
   ],
   "source": [
    "# Display the change for category label using one-hot encoding\n",
    "print('Original label 0 : ', train_labels[0])\n",
    "print('After conversion to categorical ( one-hot ) : ', train_labels_one_hot[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Create the network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "model.add(Dense(512, activation='relu', input_shape=(dim_data,)))\n",
    "model.add(Dense(512, activation='relu'))\n",
    "model.add(Dense(classes_num, activation='softmax'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Configure the Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train the Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/20\n",
      "60000/60000 [==============================] - 1s 24us/sample - loss: 0.2700 - accuracy: 0.9166 - val_loss: 0.1934 - val_accuracy: 0.9364\n",
      "Epoch 2/20\n",
      "60000/60000 [==============================] - 1s 13us/sample - loss: 0.0932 - accuracy: 0.9710 - val_loss: 0.0896 - val_accuracy: 0.9732\n",
      "Epoch 3/20\n",
      "60000/60000 [==============================] - 1s 14us/sample - loss: 0.0583 - accuracy: 0.9819 - val_loss: 0.0871 - val_accuracy: 0.9731\n",
      "Epoch 4/20\n",
      "60000/60000 [==============================] - 1s 14us/sample - loss: 0.0404 - accuracy: 0.9875 - val_loss: 0.0729 - val_accuracy: 0.9800\n",
      "Epoch 5/20\n",
      "60000/60000 [==============================] - 1s 14us/sample - loss: 0.0288 - accuracy: 0.9907 - val_loss: 0.0805 - val_accuracy: 0.9801\n",
      "Epoch 6/20\n",
      "60000/60000 [==============================] - 1s 13us/sample - loss: 0.0220 - accuracy: 0.9926 - val_loss: 0.0915 - val_accuracy: 0.9764\n",
      "Epoch 7/20\n",
      "60000/60000 [==============================] - 1s 14us/sample - loss: 0.0162 - accuracy: 0.9948 - val_loss: 0.0914 - val_accuracy: 0.9794\n",
      "Epoch 8/20\n",
      "60000/60000 [==============================] - 1s 14us/sample - loss: 0.0142 - accuracy: 0.9954 - val_loss: 0.1080 - val_accuracy: 0.9760\n",
      "Epoch 9/20\n",
      "60000/60000 [==============================] - 1s 14us/sample - loss: 0.0115 - accuracy: 0.9963 - val_loss: 0.1410 - val_accuracy: 0.9709\n",
      "Epoch 10/20\n",
      "60000/60000 [==============================] - 1s 13us/sample - loss: 0.0102 - accuracy: 0.9967 - val_loss: 0.0998 - val_accuracy: 0.9812\n",
      "Epoch 11/20\n",
      "60000/60000 [==============================] - 1s 13us/sample - loss: 0.0087 - accuracy: 0.9971 - val_loss: 0.0872 - val_accuracy: 0.9837\n",
      "Epoch 12/20\n",
      "60000/60000 [==============================] - 1s 13us/sample - loss: 0.0078 - accuracy: 0.9974 - val_loss: 0.1102 - val_accuracy: 0.9809\n",
      "Epoch 13/20\n",
      "60000/60000 [==============================] - 1s 13us/sample - loss: 0.0066 - accuracy: 0.9980 - val_loss: 0.1006 - val_accuracy: 0.9838\n",
      "Epoch 14/20\n",
      "60000/60000 [==============================] - 1s 13us/sample - loss: 0.0062 - accuracy: 0.9983 - val_loss: 0.1112 - val_accuracy: 0.9830\n",
      "Epoch 15/20\n",
      "60000/60000 [==============================] - 1s 13us/sample - loss: 0.0050 - accuracy: 0.9985 - val_loss: 0.1305 - val_accuracy: 0.9796\n",
      "Epoch 16/20\n",
      "60000/60000 [==============================] - 1s 13us/sample - loss: 0.0055 - accuracy: 0.9983 - val_loss: 0.1254 - val_accuracy: 0.9821\n",
      "Epoch 17/20\n",
      "60000/60000 [==============================] - 1s 13us/sample - loss: 0.0052 - accuracy: 0.9982 - val_loss: 0.1283 - val_accuracy: 0.9821\n",
      "Epoch 18/20\n",
      "60000/60000 [==============================] - 1s 12us/sample - loss: 0.0038 - accuracy: 0.9988 - val_loss: 0.1313 - val_accuracy: 0.9826\n",
      "Epoch 19/20\n",
      "60000/60000 [==============================] - 1s 13us/sample - loss: 0.0039 - accuracy: 0.9988 - val_loss: 0.1507 - val_accuracy: 0.9800\n",
      "Epoch 20/20\n",
      "60000/60000 [==============================] - 1s 13us/sample - loss: 0.0038 - accuracy: 0.9988 - val_loss: 0.1589 - val_accuracy: 0.9801\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(train_data, train_labels_one_hot, batch_size=256, epochs=20, verbose=1, \n",
    "                   validation_data=(test_data, test_labels_one_hot))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Plot the loss and accuracy curves"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Loss Curves')"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=[8,6])\n",
    "plt.plot(history.history['loss'],'r',linewidth=3.0)\n",
    "plt.plot(history.history['val_loss'],'b',linewidth=3.0)\n",
    "plt.legend(['Training loss', 'Validation Loss'],fontsize=18)\n",
    "plt.xlabel('Epochs ',fontsize=16)\n",
    "plt.ylabel('Loss',fontsize=16)\n",
    "plt.title('Loss Curves',fontsize=16)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Accuracy Curves')"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=[8,6])\n",
    "plt.plot(history.history['accuracy'],'r',linewidth=3.0)\n",
    "plt.plot(history.history['val_accuracy'],'b',linewidth=3.0)\n",
    "plt.legend(['Training Accuracy', 'Validation Accuracy'],fontsize=18)\n",
    "plt.xlabel('Epochs ',fontsize=16)\n",
    "plt.ylabel('Accuracy',fontsize=16)\n",
    "plt.title('Accuracy Curves',fontsize=16)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Evaluate the trained network on test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10000/10000 [==============================] - 0s 47us/sample - loss: 0.1589 - accuracy: 0.9801\n"
     ]
    }
   ],
   "source": [
    "[test_loss, test_acc] = model.evaluate(test_data, test_labels_one_hot)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Evaluation result on Test Data : Loss = 0.15887983631565528, accuracy = 0.9800999760627747\n"
     ]
    }
   ],
   "source": [
    "print(\"Evaluation result on Test Data : Loss = {}, accuracy = {}\".format(test_loss, test_acc))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Create a new network with Dropout Regularization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.layers import Dropout\n",
    "model_reg = Sequential()\n",
    "model_reg.add(Dense(512, activation='relu', input_shape=(dim_data,)))\n",
    "model_reg.add(Dropout(0.5))\n",
    "model_reg.add(Dense(512, activation='relu'))\n",
    "model_reg.add(Dropout(0.5))\n",
    "model_reg.add(Dense(classes_num, activation='softmax'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_reg.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/20\n",
      "60000/60000 [==============================] - 1s 21us/sample - loss: 0.3770 - accuracy: 0.8833 - val_loss: 0.1518 - val_accuracy: 0.9533\n",
      "Epoch 2/20\n",
      "60000/60000 [==============================] - 1s 13us/sample - loss: 0.1702 - accuracy: 0.9491 - val_loss: 0.1094 - val_accuracy: 0.9650\n",
      "Epoch 3/20\n",
      "60000/60000 [==============================] - 1s 14us/sample - loss: 0.1298 - accuracy: 0.9617 - val_loss: 0.1003 - val_accuracy: 0.9717\n",
      "Epoch 4/20\n",
      "60000/60000 [==============================] - 1s 14us/sample - loss: 0.1114 - accuracy: 0.9665 - val_loss: 0.0783 - val_accuracy: 0.9765\n",
      "Epoch 5/20\n",
      "60000/60000 [==============================] - 1s 14us/sample - loss: 0.0981 - accuracy: 0.9709 - val_loss: 0.0821 - val_accuracy: 0.9778\n",
      "Epoch 6/20\n",
      "60000/60000 [==============================] - 1s 14us/sample - loss: 0.0868 - accuracy: 0.9744 - val_loss: 0.0819 - val_accuracy: 0.9773\n",
      "Epoch 7/20\n",
      "60000/60000 [==============================] - 1s 15us/sample - loss: 0.0799 - accuracy: 0.9758 - val_loss: 0.0778 - val_accuracy: 0.9791\n",
      "Epoch 8/20\n",
      "60000/60000 [==============================] - 1s 14us/sample - loss: 0.0751 - accuracy: 0.9778 - val_loss: 0.0728 - val_accuracy: 0.9810\n",
      "Epoch 9/20\n",
      "60000/60000 [==============================] - 1s 13us/sample - loss: 0.0729 - accuracy: 0.9785 - val_loss: 0.0782 - val_accuracy: 0.9787\n",
      "Epoch 10/20\n",
      "60000/60000 [==============================] - 1s 14us/sample - loss: 0.0686 - accuracy: 0.9797 - val_loss: 0.0765 - val_accuracy: 0.9833\n",
      "Epoch 11/20\n",
      "60000/60000 [==============================] - 1s 14us/sample - loss: 0.0649 - accuracy: 0.9816 - val_loss: 0.0764 - val_accuracy: 0.9815\n",
      "Epoch 12/20\n",
      "60000/60000 [==============================] - 1s 14us/sample - loss: 0.0626 - accuracy: 0.9818 - val_loss: 0.0720 - val_accuracy: 0.9835\n",
      "Epoch 13/20\n",
      "60000/60000 [==============================] - 1s 13us/sample - loss: 0.0608 - accuracy: 0.9824 - val_loss: 0.0755 - val_accuracy: 0.9831\n",
      "Epoch 14/20\n",
      "60000/60000 [==============================] - 1s 13us/sample - loss: 0.0599 - accuracy: 0.9827 - val_loss: 0.0729 - val_accuracy: 0.9845\n",
      "Epoch 15/20\n",
      "60000/60000 [==============================] - 1s 13us/sample - loss: 0.0557 - accuracy: 0.9839 - val_loss: 0.0787 - val_accuracy: 0.9834\n",
      "Epoch 16/20\n",
      "60000/60000 [==============================] - 1s 14us/sample - loss: 0.0525 - accuracy: 0.9845 - val_loss: 0.0763 - val_accuracy: 0.9837\n",
      "Epoch 17/20\n",
      "60000/60000 [==============================] - 1s 14us/sample - loss: 0.0528 - accuracy: 0.9845 - val_loss: 0.0768 - val_accuracy: 0.9850\n",
      "Epoch 18/20\n",
      "60000/60000 [==============================] - 1s 14us/sample - loss: 0.0498 - accuracy: 0.9856 - val_loss: 0.0803 - val_accuracy: 0.9838\n",
      "Epoch 19/20\n",
      "60000/60000 [==============================] - 1s 13us/sample - loss: 0.0500 - accuracy: 0.9855 - val_loss: 0.0786 - val_accuracy: 0.9845\n",
      "Epoch 20/20\n",
      "60000/60000 [==============================] - 1s 13us/sample - loss: 0.0490 - accuracy: 0.9858 - val_loss: 0.0814 - val_accuracy: 0.9833\n"
     ]
    }
   ],
   "source": [
    "history_reg = model_reg.fit(train_data, train_labels_one_hot, batch_size=256, epochs=20, verbose=1, \n",
    "                            validation_data=(test_data, test_labels_one_hot))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Loss Curves')"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=[8,6])\n",
    "plt.plot(history_reg.history['loss'],'r',linewidth=3.0)\n",
    "plt.plot(history_reg.history['val_loss'],'b',linewidth=3.0)\n",
    "plt.legend(['Training loss', 'Validation Loss'],fontsize=18)\n",
    "plt.xlabel('Epochs ',fontsize=16)\n",
    "plt.ylabel('Loss',fontsize=16)\n",
    "plt.title('Loss Curves',fontsize=16)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Accuracy Curves')"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=[8,6])\n",
    "plt.plot(history_reg.history['accuracy'],'r',linewidth=3.0)\n",
    "plt.plot(history_reg.history['val_accuracy'],'b',linewidth=3.0)\n",
    "plt.legend(['Training Accuracy', 'Validation Accuracy'],fontsize=18)\n",
    "plt.xlabel('Epochs ',fontsize=16)\n",
    "plt.ylabel('Accuracy',fontsize=16)\n",
    "plt.title('Accuracy Curves',fontsize=16)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10000/10000 [==============================] - 0s 46us/sample - loss: 0.0814 - accuracy: 0.9833\n"
     ]
    }
   ],
   "source": [
    "[test_loss, test_acc] = model_reg.evaluate(test_data, test_labels_one_hot)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Evaluation result on Test Data : Loss = 0.08135654206477057, accuracy = 0.983299970626831\n"
     ]
    }
   ],
   "source": [
    "print(\"Evaluation result on Test Data : Loss = {}, accuracy = {}\".format(test_loss, test_acc))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Predict the first image from test data\n",
    "\n",
    "We have seen that the first image is the number 7. Let us predict using the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1.29905996e-18, 1.10607663e-14, 1.40711756e-11, 1.03725896e-10,\n",
       "        7.29663463e-19, 1.92765065e-14, 4.65404149e-28, 1.00000000e+00,\n",
       "        6.88956165e-16, 3.75895357e-11]], dtype=float32)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Predict the probabilities for each class \n",
    "model_reg.predict(test_data[[0],:])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model prediction: 2\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Ground Truth : 2')"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Predict the most likely class\n",
    "print(\"Model prediction: {}\".format(model_reg.predict_classes(test_data[[1],:])[0]))\n",
    "\n",
    "# Display the predicted image\n",
    "plt.imshow(test_images[1], cmap='gray')\n",
    "plt.title(\"Ground Truth : {}\".format(test_labels[1]))"
   ]
  }
 ],
 "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": "3.7.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}