notebook/MNIST Classifier.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "MNIST Classification task is the \"hello world\" of Computer Vision. The task is to identify (or classify) the greyscale images of hand-written digits. The dataset can be easily loaded using Keras."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from keras.datasets import mnist\n",
    "\n",
    "(X_train, y_train), (X_test, y_test) = mnist.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of samples in training data = 60000.\n",
      "Dimensions of each sample are 28X28\n",
      "Number of samples in test data = 10000\n"
     ]
    }
   ],
   "source": [
    "n_train, height, width = X_train.shape\n",
    "n_test, _, _ = X_test.shape\n",
    "print(\"Number of samples in training data = {}.\\nDimensions of each sample are {}X{}\".format(n_train, height, width))\n",
    "print(\"Number of samples in test data = {}\".format(n_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(60000, 28, 28)\n",
      "(28, 28)\n"
     ]
    }
   ],
   "source": [
    "print(X_train.shape)\n",
    "print(X_train[0].shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP4AAAD8CAYAAABXXhlaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nED8QaCCC+IFAAxHEDwQaiP8PYrfpjRk/ZYUAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f1dded0b588>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.imshow(X_train[0], cmap=plt.get_cmap('gray'))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[5 0 4 ..., 5 6 8]\n"
     ]
    }
   ],
   "source": [
    "print(y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We need to process the data before feeding to Keras. \n",
    "\n",
    "* Reshape data. Keras layers for 2-d convolution expects another dimension corresponding to pixel. For greyscale image, we have only one pixel dimension. For RGB, there would be 3 pixel dimensions.\n",
    "* Normalize image data so that all pixel values are in range [0, 1] instead of [0, 255]. \n",
    "* Transform output values to be one-hot vectors instead of single values."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from keras.utils.np_utils import to_categorical\n",
    "from keras import backend as K\n",
    "\n",
    "# Reshaping data\n",
    "\n",
    "# this if condition is needed because the input shape is specified differently for theano and tensorflow backend.\n",
    "if K.image_dim_ordering() == 'th':\n",
    "    X_train = X_train.reshape(n_train, 1, height, width).astype('float32')\n",
    "    X_test = X_test.reshape(n_test, 1, height, width).astype('float32')\n",
    "    input_shape = (1, height, width)\n",
    "else:\n",
    "    X_train = X_train.reshape(n_train, height, width, 1).astype('float32')\n",
    "    X_test = X_test.reshape(n_test, height, width, 1).astype('float32')\n",
    "    input_shape = (height, width, 1)\n",
    "    \n",
    "\n",
    "# Normalizing data\n",
    "X_train /= 255\n",
    "X_test /= 255\n",
    "\n",
    "# Transforming output variables\n",
    "n_classes = 10\n",
    "y_train = to_categorical(y_train, n_classes)\n",
    "y_test = to_categorical(y_test, n_classes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(60000, 28, 28, 1)\n",
      "(28, 28, 1)\n"
     ]
    }
   ],
   "source": [
    "print(X_train.shape)\n",
    "print(X_train[0].shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 0.  0.  0. ...,  0.  0.  0.]\n",
      " [ 1.  0.  0. ...,  0.  0.  0.]\n",
      " [ 0.  0.  0. ...,  0.  0.  0.]\n",
      " ..., \n",
      " [ 0.  0.  0. ...,  0.  0.  0.]\n",
      " [ 0.  0.  0. ...,  0.  0.  0.]\n",
      " [ 0.  0.  0. ...,  0.  1.  0.]]\n"
     ]
    }
   ],
   "source": [
    "print(y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0.  0.  0.  0.  0.  1.  0.  0.  0.  0.]\n",
      "[ 1.  0.  0.  0.  0.  0.  0.  0.  0.  0.]\n"
     ]
    }
   ],
   "source": [
    "print(y_train[0])\n",
    "print(y_train[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will create a sequential Keras model where different layers would be added in sequence."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from keras.models import Sequential\n",
    "model = Sequential()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Architecture:\n",
    " * Convolution layers\n",
    "     * Convolution layer would have 32 feature maps.\n",
    "     * Each feature map will be of size 5x5.\n",
    "     * This will be followed by RELU activation.\n",
    " * Pooling layer\n",
    "     * Pool Size of 2x2\n",
    " * Regularization layer (Dropout)\n",
    "     * Exclude 25% of neurons in the layer to reduce overfitting.\n",
    " * Flatten layer to convert the matrix data to vectors.\n",
    " * Fully connected (dense) layer\n",
    "     * 128 neurons and RELU activation.\n",
    " * Regularization layer (Dropout)\n",
    "     * Exclude 25% of neurons in the layer to reduce overfitting.\n",
    " * Fully connected (dense) layer\n",
    "     * 10 neurons (1 for each class)\n",
    "     * Softmax classifier convolution filters)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from keras.layers import Activation\n",
    "from keras.layers.convolutional import Convolution2D, MaxPooling2D\n",
    "from keras.layers import Dropout, Flatten, Dense\n",
    "\n",
    "# Convolution Layer\n",
    "\n",
    "model.add(Convolution2D(32, 3, 3, border_mode='valid', input_shape=input_shape))\n",
    "model.add(Activation('relu'))\n",
    "model.add(Convolution2D(32, 5, 5, activation='relu'))\n",
    "\n",
    "# Pooling\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "\n",
    "# Regularization Layer\n",
    "model.add(Dropout(0.25))\n",
    "\n",
    "# Flatten Layer\n",
    "model.add(Flatten())\n",
    "\n",
    "# Fully Connected\n",
    "model.add(Dense(128))\n",
    "model.add(Activation('relu'))\n",
    "\n",
    "# Regularization Layer\n",
    "model.add(Dropout(0.5))\n",
    "\n",
    "# Fully Connected with softmax\n",
    "model.add(Dense(n_classes))\n",
    "model.add(Activation('softmax'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Specify the loss function, the optimizer to use and the metric to track."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "model.compile(\n",
    "    loss='categorical_crossentropy',\n",
    "    optimizer='adam',\n",
    "    metrics=['accuracy']\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we can start the training. For demo purpose, we will train it only for 2 epochs. Feel free to train for more epochs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/2\n",
      "60000/60000 [==============================] - 322s - loss: 0.2527 - acc: 0.9235 - val_loss: 0.0543 - val_acc: 0.9831\n",
      "Epoch 2/2\n",
      "60000/60000 [==============================] - 331s - loss: 0.0902 - acc: 0.9733 - val_loss: 0.0378 - val_acc: 0.9874\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7f1daf5df5c0>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(X_train, y_train, batch_size=128, nb_epoch=2, validation_data=(X_test, y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10000/10000 [==============================] - 17s    \n",
      "Loss: 0.0378334802061\n",
      "Accuracy: 0.9874\n"
     ]
    }
   ],
   "source": [
    "loss, accuracy = model.evaluate(X_test, y_test)\n",
    "print('Loss:', loss)\n",
    "print('Accuracy:', accuracy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.4.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}