1_notmnist.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "5hIbr52I7Z7U"
   },
   "source": [
    "Deep Learning\n",
    "=============\n",
    "\n",
    "Assignment 1\n",
    "------------\n",
    "\n",
    "The objective of this assignment is to learn about simple data curation practices, and familiarize you with some of the data we'll be reusing later.\n",
    "\n",
    "This notebook uses the [notMNIST](http://yaroslavvb.blogspot.com/2011/09/notmnist-dataset.html) dataset to be used with python experiments. This dataset is designed to look like the classic [MNIST](http://yann.lecun.com/exdb/mnist/) dataset, while looking a little more like real data: it's a harder task, and the data is a lot less 'clean' than MNIST."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "cellView": "both",
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "collapsed": true,
    "id": "apJbCsBHl-2A"
   },
   "outputs": [],
   "source": [
    "# These are all the modules we'll be using later. Make sure you can import them\n",
    "# before proceeding further.\n",
    "from __future__ import print_function\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import os\n",
    "import sys\n",
    "import tarfile\n",
    "from IPython.display import display, Image\n",
    "from scipy import ndimage\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from six.moves.urllib.request import urlretrieve\n",
    "from six.moves import cPickle as pickle\n",
    "\n",
    "# Config the matlotlib backend as plotting inline in IPython\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "jNWGtZaXn-5j"
   },
   "source": [
    "First, we'll download the dataset to our local machine. The data consists of characters rendered in a variety of fonts on a 28x28 image. The labels are limited to 'A' through 'J' (10 classes). The training set has about 500k and the testset 19000 labelled examples. Given these sizes, it should be possible to train models quickly on any machine."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "cellView": "both",
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "output_extras": [
      {
       "item_id": 1
      }
     ]
    },
    "colab_type": "code",
    "collapsed": false,
    "executionInfo": {
     "elapsed": 186058,
     "status": "ok",
     "timestamp": 1444485672507,
     "user": {
      "color": "#1FA15D",
      "displayName": "Vincent Vanhoucke",
      "isAnonymous": false,
      "isMe": true,
      "permissionId": "05076109866853157986",
      "photoUrl": "//lh6.googleusercontent.com/-cCJa7dTDcgQ/AAAAAAAAAAI/AAAAAAAACgw/r2EZ_8oYer4/s50-c-k-no/photo.jpg",
      "sessionId": "2a0a5e044bb03b66",
      "userId": "102167687554210253930"
     },
     "user_tz": 420
    },
    "id": "EYRJ4ICW6-da",
    "outputId": "0d0f85df-155f-4a89-8e7e-ee32df36ec8d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found and verified notMNIST_large.tar.gz\n",
      "Found and verified notMNIST_small.tar.gz\n"
     ]
    }
   ],
   "source": [
    "url = 'http://commondatastorage.googleapis.com/books1000/'\n",
    "last_percent_reported = None\n",
    "\n",
    "def download_progress_hook(count, blockSize, totalSize):\n",
    "  \"\"\"A hook to report the progress of a download. This is mostly intended for users with\n",
    "  slow internet connections. Reports every 1% change in download progress.\n",
    "  \"\"\"\n",
    "  global last_percent_reported\n",
    "  percent = int(count * blockSize * 100 / totalSize)\n",
    "\n",
    "  if last_percent_reported != percent:\n",
    "    if percent % 5 == 0:\n",
    "      sys.stdout.write(\"%s%%\" % percent)\n",
    "      sys.stdout.flush()\n",
    "    else:\n",
    "      sys.stdout.write(\".\")\n",
    "      sys.stdout.flush()\n",
    "      \n",
    "    last_percent_reported = percent\n",
    "        \n",
    "def maybe_download(filename, expected_bytes, force=False):\n",
    "  \"\"\"Download a file if not present, and make sure it's the right size.\"\"\"\n",
    "  if force or not os.path.exists(filename):\n",
    "    print('Attempting to download:', filename) \n",
    "    filename, _ = urlretrieve(url + filename, filename, reporthook=download_progress_hook)\n",
    "    print('\\nDownload Complete!')\n",
    "  statinfo = os.stat(filename)\n",
    "  if statinfo.st_size == expected_bytes:\n",
    "    print('Found and verified', filename)\n",
    "  else:\n",
    "    raise Exception(\n",
    "      'Failed to verify ' + filename + '. Can you get to it with a browser?')\n",
    "  return filename\n",
    "\n",
    "train_filename = maybe_download('notMNIST_large.tar.gz', 247336696)\n",
    "test_filename = maybe_download('notMNIST_small.tar.gz', 8458043)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "cC3p0oEyF8QT"
   },
   "source": [
    "Extract the dataset from the compressed .tar.gz file.\n",
    "This should give you a set of directories, labelled A through J."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "cellView": "both",
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "output_extras": [
      {
       "item_id": 1
      }
     ]
    },
    "colab_type": "code",
    "collapsed": false,
    "executionInfo": {
     "elapsed": 186055,
     "status": "ok",
     "timestamp": 1444485672525,
     "user": {
      "color": "#1FA15D",
      "displayName": "Vincent Vanhoucke",
      "isAnonymous": false,
      "isMe": true,
      "permissionId": "05076109866853157986",
      "photoUrl": "//lh6.googleusercontent.com/-cCJa7dTDcgQ/AAAAAAAAAAI/AAAAAAAACgw/r2EZ_8oYer4/s50-c-k-no/photo.jpg",
      "sessionId": "2a0a5e044bb03b66",
      "userId": "102167687554210253930"
     },
     "user_tz": 420
    },
    "id": "H8CBE-WZ8nmj",
    "outputId": "ef6c790c-2513-4b09-962e-27c79390c762"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "notMNIST_large already present - Skipping extraction of notMNIST_large.tar.gz.\n",
      "['notMNIST_large/A', 'notMNIST_large/B', 'notMNIST_large/C', 'notMNIST_large/D', 'notMNIST_large/E', 'notMNIST_large/F', 'notMNIST_large/G', 'notMNIST_large/H', 'notMNIST_large/I', 'notMNIST_large/J']\n",
      "notMNIST_small already present - Skipping extraction of notMNIST_small.tar.gz.\n",
      "['notMNIST_small/A', 'notMNIST_small/B', 'notMNIST_small/C', 'notMNIST_small/D', 'notMNIST_small/E', 'notMNIST_small/F', 'notMNIST_small/G', 'notMNIST_small/H', 'notMNIST_small/I', 'notMNIST_small/J']\n"
     ]
    }
   ],
   "source": [
    "num_classes = 10\n",
    "np.random.seed(133)\n",
    "\n",
    "def maybe_extract(filename, force=False):\n",
    "  root = os.path.splitext(os.path.splitext(filename)[0])[0]  # remove .tar.gz\n",
    "  if os.path.isdir(root) and not force:\n",
    "    # You may override by setting force=True.\n",
    "    print('%s already present - Skipping extraction of %s.' % (root, filename))\n",
    "  else:\n",
    "    print('Extracting data for %s. This may take a while. Please wait.' % root)\n",
    "    tar = tarfile.open(filename)\n",
    "    sys.stdout.flush()\n",
    "    tar.extractall()\n",
    "    tar.close()\n",
    "  data_folders = [\n",
    "    os.path.join(root, d) for d in sorted(os.listdir(root))\n",
    "    if os.path.isdir(os.path.join(root, d))]\n",
    "  if len(data_folders) != num_classes:\n",
    "    raise Exception(\n",
    "      'Expected %d folders, one per class. Found %d instead.' % (\n",
    "        num_classes, len(data_folders)))\n",
    "  print(data_folders)\n",
    "  return data_folders\n",
    "  \n",
    "train_folders = maybe_extract(train_filename)\n",
    "test_folders = maybe_extract(test_filename)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "4riXK3IoHgx6"
   },
   "source": [
    "---\n",
    "Problem 1\n",
    "---------\n",
    "\n",
    "Let's take a peek at some of the data to make sure it looks sensible. Each exemplar should be an image of a character A through J rendered in a different font. Display a sample of the images that we just downloaded. Hint: you can use the package IPython.display.\n",
    "\n",
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAABwAAAAcCAAAAABXZoBIAAAA5UlEQVR4nM2SMU4CARRE3//7TQwn\nINl4AWNha2NhYjyCBZ2VpYcx8Qwegw4OAAcQaiuRsLB/LBaQ1Q2tTvV/JjOZTAbAuHieLrTHYvpy\niQGA8/ihlLQajcejlaTU8gkHKLhZal0rNcOMmVL1WtUdBY5zfVqFA4YZBnhUJ1c4AaixR2j3GILG\n+Te0TXME/5o03HedAhAHt8jvRn6ScRZsopM0+m/bo0v5+erkfa9LKXt/MHTb017aThtxJO2mnfZv\nunWsFREAw3CcbIbVhpqiB5PhXCkd7lZKzYeTQZTnddFlW/aL8gtV/maYrzbgSgAAAABJRU5ErkJg\ngg==\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAABwAAAAcCAAAAABXZoBIAAABtUlEQVR4nG2QzYsNYBTGf+d933tv\n3PkIU2R0iVmIxEbjo6yHslR2SpRsxNiNrKdMSlIW1Gwk/4DFbLAQq2mGjYgyhhllxPXR/XjveSyG\ncbt3nuX5dc5znscEyOY//aJU7q+9ez4780FEOQCSmhrrLUB568EL7+VfH49WCOEfbGiEZIDRc0Ny\n/bi5ibACj5GAEFLkdM5N1/wBYtsmAIQi42p5Q4s7CZ2QwI7PcjV1B5ZPt8nD2yoiMRK7oXm5AGS+\nqxtGhgdQTr9H3do8zUKIhcA9NVyzw4Sub0kTynpzZi0B/g9LXlrTW9lzai9T1554iK0VaBTPH948\n1AfKceDQwivRXl8srtty5OLDquSu+uR2onV4GpVLL+T1rOo5s7aGzCymAMWT08rNlu73EDq7TUa8\nWlfOejbYFQVS5ERNXtej/m6IJSbkaurKKhBjtyRpobPbZeXlB1aDxj7AmV4FBuMsLudBp6fFFBiT\n56bu/o1ylFKMMaZkGOtvq9GSbvV15Sztv/5RLp85TrAERC5vWPpZ6t04uG1oVz/44tTk0xxdJlCj\noFojZ1f+9mVhbu71yyUstOAPPMURzsJBxO8AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAABwAAAAcCAAAAABXZoBIAAABlklEQVR4nHWSO2tUQRiGn++bs6tG\ndGUNLhqiaJFCDYgixi5eIJ1ia7NBEAt/gWBjK1or2hptTAorEUtBrIIBXYmilVdciK5mL+fyWpwF\nV/fsWw0887zzzTBGP07G3lMnpia2lNvr7U7rw2qj8ZdxemlNA0n0KALAPD18cxag+XbtZxirTW53\n40WumV2NlejHnZPjAIRtR+q3Xh4HMDYuKu5pYQIDj4IDEDlglB+rmyUXsBAsP8Y8ypeB2+qpeybf\n+m8ClxWnqhMNIZzpjmLdoDTMCCyoq9djWAFkJlWieUIR44F6erOpWNzTVKxrI8SLSpUeZfgWAPfU\nU6PYw/cDy2mx6LuBdyNafSvQKm7FE4xxVAy/A/tGQJ4o1tfa/29gBuBPsXRHndIAtRCUV01+VqrO\nLCEKbu4higyoVPOuK+pmal0aMGvn7n6cwcHwpbOxu60sPvvWLFWqUwcOTVf5crBpAmPzfWVJT1Ly\nez2WpEzt6/0JzTm/MvCZ4/cP53f2JwPztDw3d2xXZUPr16fVV8+XW+AZwB/hcbLRISRy4gAAAABJ\nRU5ErkJggg==\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAABwAAAAcCAAAAABXZoBIAAABZUlEQVR4nH3QsUuVURzG8e/vvOc2\nSEZvXm1ykDCCChqKLi0iSIuLOAua44WmIopWIWi3JfoLSmgTBKkhER3EwCFyUEFRMMKLROj7nvM0\n3GskvKdn/Zzfc37n2NvGb6eIMwBCq7W/ur55Yi4AfFYRohRDCCGEKEnlyuN6+3T3lV6eSE1fz/M8\nvzp4f/rdobQ7jgPA05QmyWjH6HnakmbIwKxGU3qENzMzl/nMuPVNeo47m5z6OwmuxvUdFXc7zecT\niwvfn+FfVCIU9v4LwwPVqCzOkz+sRsQy3EziQUlvEo8CPQmECD6JgP6HiXe2Y0l0EBNoXM74mZzs\n8/xIoKMB2wkMjHC8UI1eoyNsfPVV22Rl/RW8ljesvZ5Z5zqLZf3DbeY+mpMEINRJKBlbHWJxEs5q\n1Wmsdd9oTNwhvHn5y2Sz906L/mts7nuBXbzU1yU7XJhdzqKwpQf//pKOj7bW1j7t4ST4Awfoj/cW\n+IzxAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAABwAAAAcCAAAAABXZoBIAAABp0lEQVR4nGWSwWpUQRBFT/V7M2+Y\nTYgxGBVBJCCCqGv/QP2BgFkMfoI/koVrF2E24kI/QCG6EjEQ0G1ARgjRaJCQhJDp7roumjeTGWvT\nFJe63DrVQKi4/sZdM5WUBkAFayPNaVG/n2DA1aGULyqek3buUEHz7M9E8xRjTC5puEAFrO/sHZzN\nep58WYcAWCdWi8tLm7c8yH69Ojw4Gh+NdjHEpLaVlbTVtlV5ajACdWmqykFSnojCNTERYAbyIk5t\nYs5crBqwkAUYy4/7naZpml51+fMwOECAwckcvp+3LZTZ5oUKPk8ppZTONVrFivPND4rTOc9j7d9r\ns6wdX6DqLun7/Tbhw3fbX3/8nYKPuxsrrWbB6fa6W3c9ePj2vHe8v3dKCQq1m8ZjIiAO3wNU7pM9\nZZiqFl+esGvxiVDYZWYQlchWJq2ubIpYXgOmbrewTek/ttB0AePGQOcxe2iWriwt9q9tFLgrB/Pf\nT9Jmr4irp3JJHmNMKaUY45leFuvAg5RSzjnn9jhJry1YOfWjOcest10zakDoUzRZ6DT9hUs1uH18\nOg7OP00lZ3bzW95OAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAABwAAAAcCAAAAABXZoBIAAABeklEQVR4nGWSUW4UMRBEq+2a2QSi\nSAhFQhyNU3AbLsNB4BsUECjRTlbZWc+4u/jweIjAX7aeq9yqsn11YBX2dRV9l2be/E54N3SqPP9K\nnb7iUqjbq36WzXO7aRZGsxQoXwwAYMBIGABYUuJyOjjiwZopPG8elmqmL+ZCbVBw7zCvTktJQGow\nQbYrkzHW5MLaYMB3GKuzPF1X6SkBgLj5AzCWW5qZwVKf1jpMZqAiQtZCkBQdwkIsU13lP5vteFh3\nJec7KkL1zYfN9vMDe5IRam/idoNbPADMzFimWNLxUwJg9W0+dWUangtrORRh62lSrb2euFTmcRAw\ntIGGLSoAaThk1stYZJcGHX1aZZ+d6zOLdO7Kv3A8L4QkQAIAAfuXkASGu0OuzbYHL7gHyzHNjmNu\nCWHpkONUCMXmCEgvbQO0zCLYoFZZ7ErSuM7T+2vdHSkAAvdSTufK6fvh8SMej/nFx24pXCa7/5Fr\nBfkvA/Ca3240GOJ/JN7/AdRQ4DvCBcmVAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAABwAAAAcCAAAAABXZoBIAAABR0lEQVR4nMWSvUoDQRSFz53JGpWQ\niEknytoIaSLBCHmMlL6HPoGFvoGdb2JhYeEPKCqmUCFBFHYNIWA2EtidORa7G92wvbeay8e5czgc\ngVR323W3Vi0pEwwHve7l7RDpNI6eSJITv+9PSJIPh1tOAt8DWhP5+zsVXWkd+JEhvy7WE0hrSfZc\nAIDbI2lMsqFAAQBoRUA0ACgwERZiBmMB0MSLpBCZsYZCsflQJYfzYPhyfffYaK6EqVvy1604GoB2\nJE/JUC9QTKgEADn3J4wBAJundDbi4AxUESMvC9fO3OQ1fj099+fPAgCm99PpZ1RUudDb6wMFmFwI\nrUHD+YQynmdJ5c5/wJya4E9NCEGmJpzVRH18C61Z6jTLutzsLBtLiYJUun3cJUmOvb4fkCT7J+3F\nmAmk1mrXN2urJcXJaPD2fHPlpS35ASchoO++VQffAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAABwAAAAcCAAAAABXZoBIAAACFklEQVR4nF2QXUiTcRjFz//9712s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\nALfo50Jvt684AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAABwAAAAcCAAAAABXZoBIAAABZElEQVR4nGNgQAWia178h4Jbs1Cl\nGAWd4XL/N2qwoEj+187lhTEvH7mBIsnIp+fDDGNv2sSIIvnf3RMm9+vuyZuoNrLP/gqz8E2WAiOK\nJHfsCajUv69nZFDlGCSOfvoHlT2TxosixSgT9xpm6I8FPGgaPU7/gEnujGFgYGBggLuWUdbYBMb+\nt207439kyf++vv+hRn07eOEtipks4qv+wgx94SnMiGIsd6o2E9SINyfPomlUvvELpnGvMydMFEIp\nh4izQjW+vrAX1ReMQY9/wzTOc0SYx8DAwMBoYCMD88XnfUfgkkwMDAwM//38/kP5XxZd+Y1iKJvu\nFpiZfx/qcyICjoWBgYE3Rwfq///3Nz/+juIYLpMnsMj4v1GFFdmDDP+10vihJv2/fvwOqjc4Ej7D\nNbZposQUC4O9Mw+U/evhieuoGhkmvIFpfJurhBbFIQdhZn47K4smx3DiFUzj+VxeNDkA2tDohPBS\np1sAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAABwAAAAcCAAAAABXZoBIAAABJElEQVR4nOWTvS9EQRTFf3dmluza\nlQgbW2yJQjQoJNSCWkUl6w9R6hQKW22ikahFrSCi1ihlol7ESoh9b2YU+97uxNc/4FR3zpl77r3z\ngfXhF3hrgN023zHRAqwP9R806lmmyIDrxxow3onSGaGE1GWxA0xZy13IiG7SaR5E1qZRkkJutb1e\nGo3rmvNoMb/Wd+2JOupGxa0BJt7qv4yj+AP/RTTElxkJGjCEAeGStHeCIt4F581ZTcZz77FCpQwo\ncaG4LHrELFXl6DUTu8nHjQheTW42FpOLJvepix+YKIbm9m3otFYA68NUMcewprJ68pLavWkQjXU+\nyqztXPq360YVlAajRLae86FnN2bap8dX76KDB7h9fHL9H+AeDhcUorP2PwEwVXn0SFtFBwAAAABJ\nRU5ErkJggg==\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import os, random\n",
    "from IPython import display\n",
    "\n",
    "for dir in train_folders:\n",
    "    relativepath = dir + \"/\" + random.choice(os.listdir(dir))\n",
    "    image = display.Image(filename=relativepath)\n",
    "    display.display(image)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "PBdkjESPK8tw"
   },
   "source": [
    "Now let's load the data in a more manageable format. Since, depending on your computer setup you might not be able to fit it all in memory, we'll load each class into a separate dataset, store them on disk and curate them independently. Later we'll merge them into a single dataset of manageable size.\n",
    "\n",
    "We'll convert the entire dataset into a 3D array (image index, x, y) of floating point values, normalized to have approximately zero mean and standard deviation ~0.5 to make training easier down the road. \n",
    "\n",
    "A few images might not be readable, we'll just skip them."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "cellView": "both",
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "output_extras": [
      {
       "item_id": 30
      }
     ]
    },
    "colab_type": "code",
    "collapsed": false,
    "executionInfo": {
     "elapsed": 399874,
     "status": "ok",
     "timestamp": 1444485886378,
     "user": {
      "color": "#1FA15D",
      "displayName": "Vincent Vanhoucke",
      "isAnonymous": false,
      "isMe": true,
      "permissionId": "05076109866853157986",
      "photoUrl": "//lh6.googleusercontent.com/-cCJa7dTDcgQ/AAAAAAAAAAI/AAAAAAAACgw/r2EZ_8oYer4/s50-c-k-no/photo.jpg",
      "sessionId": "2a0a5e044bb03b66",
      "userId": "102167687554210253930"
     },
     "user_tz": 420
    },
    "id": "h7q0XhG3MJdf",
    "outputId": "92c391bb-86ff-431d-9ada-315568a19e59"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "notMNIST_large/A.pickle already present - Skipping pickling.\n",
      "notMNIST_large/B.pickle already present - Skipping pickling.\n",
      "notMNIST_large/C.pickle already present - Skipping pickling.\n",
      "notMNIST_large/D.pickle already present - Skipping pickling.\n",
      "notMNIST_large/E.pickle already present - Skipping pickling.\n",
      "notMNIST_large/F.pickle already present - Skipping pickling.\n",
      "notMNIST_large/G.pickle already present - Skipping pickling.\n",
      "notMNIST_large/H.pickle already present - Skipping pickling.\n",
      "notMNIST_large/I.pickle already present - Skipping pickling.\n",
      "notMNIST_large/J.pickle already present - Skipping pickling.\n",
      "notMNIST_small/A.pickle already present - Skipping pickling.\n",
      "notMNIST_small/B.pickle already present - Skipping pickling.\n",
      "notMNIST_small/C.pickle already present - Skipping pickling.\n",
      "notMNIST_small/D.pickle already present - Skipping pickling.\n",
      "notMNIST_small/E.pickle already present - Skipping pickling.\n",
      "notMNIST_small/F.pickle already present - Skipping pickling.\n",
      "notMNIST_small/G.pickle already present - Skipping pickling.\n",
      "notMNIST_small/H.pickle already present - Skipping pickling.\n",
      "notMNIST_small/I.pickle already present - Skipping pickling.\n",
      "notMNIST_small/J.pickle already present - Skipping pickling.\n"
     ]
    }
   ],
   "source": [
    "image_size = 28  # Pixel width and height.\n",
    "pixel_depth = 255.0  # Number of levels per pixel.\n",
    "\n",
    "def load_letter(folder, min_num_images):\n",
    "  \"\"\"Load the data for a single letter label.\"\"\"\n",
    "  image_files = os.listdir(folder)\n",
    "  dataset = np.ndarray(shape=(len(image_files), image_size, image_size),\n",
    "                         dtype=np.float32)\n",
    "  print(folder)\n",
    "  num_images = 0\n",
    "  for image in image_files:\n",
    "    image_file = os.path.join(folder, image)\n",
    "    try:\n",
    "      image_data = (ndimage.imread(image_file).astype(float) - \n",
    "                    pixel_depth / 2) / pixel_depth\n",
    "      if image_data.shape != (image_size, image_size):\n",
    "        raise Exception('Unexpected image shape: %s' % str(image_data.shape))\n",
    "      dataset[num_images, :, :] = image_data\n",
    "      num_images = num_images + 1\n",
    "    except IOError as e:\n",
    "      print('Could not read:', image_file, ':', e, '- it\\'s ok, skipping.')\n",
    "    \n",
    "  dataset = dataset[0:num_images, :, :]\n",
    "  if num_images < min_num_images:\n",
    "    raise Exception('Many fewer images than expected: %d < %d' %\n",
    "                    (num_images, min_num_images))\n",
    "    \n",
    "  print('Full dataset tensor:', dataset.shape)\n",
    "  print('Mean:', np.mean(dataset))\n",
    "  print('Standard deviation:', np.std(dataset))\n",
    "  return dataset\n",
    "        \n",
    "def maybe_pickle(data_folders, min_num_images_per_class, force=False):\n",
    "  dataset_names = []\n",
    "  for folder in data_folders:\n",
    "    set_filename = folder + '.pickle'\n",
    "    dataset_names.append(set_filename)\n",
    "    if os.path.exists(set_filename) and not force:\n",
    "      # You may override by setting force=True.\n",
    "      print('%s already present - Skipping pickling.' % set_filename)\n",
    "    else:\n",
    "      print('Pickling %s.' % set_filename)\n",
    "      dataset = load_letter(folder, min_num_images_per_class)\n",
    "      try:\n",
    "        with open(set_filename, 'wb') as f:\n",
    "          pickle.dump(dataset, f, pickle.HIGHEST_PROTOCOL)\n",
    "      except Exception as e:\n",
    "        print('Unable to save data to', set_filename, ':', e)\n",
    "  \n",
    "  return dataset_names\n",
    "\n",
    "train_datasets = maybe_pickle(train_folders, 45000)\n",
    "test_datasets = maybe_pickle(test_folders, 1800)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "vUdbskYE2d87"
   },
   "source": [
    "---\n",
    "Problem 2\n",
    "---------\n",
    "\n",
    "Let's verify that the data still looks good. Displaying a sample of the labels and images from the ndarray. Hint: you can use matplotlib.pyplot.\n",
    "\n",
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[-0.0372549   0.44509804  0.5         0.49607843  0.5         0.5         0.5\n",
      "   0.5         0.5         0.5         0.5         0.5         0.5         0.5\n",
      "   0.5         0.5         0.5         0.5         0.5         0.5         0.5\n",
      "   0.5         0.5         0.5         0.49607843  0.5         0.44117647\n",
      "  -0.06862745]\n",
      " [ 0.46862745  0.5         0.48039216  0.48039216  0.48039216  0.49607843\n",
      "   0.5         0.5         0.5         0.48823529  0.48039216  0.48039216\n",
      "   0.48039216  0.48039216  0.48039216  0.48039216  0.48039216  0.48039216\n",
      "   0.48039216  0.48039216  0.48039216  0.48039216  0.48039216  0.48431373\n",
      "   0.5         0.5         0.5         0.42941177]\n",
      " [ 0.30784315  0.5         0.49607843  0.5         0.5         0.5         0.5\n",
      "   0.5         0.5         0.5         0.5         0.5         0.5         0.5\n",
      "   0.5         0.5         0.5         0.5         0.5         0.5         0.5\n",
      "   0.5         0.5         0.5         0.5         0.5         0.49607843\n",
      "   0.49607843]\n",
      " [-0.39803922 -0.10784314 -0.02156863  0.00588235 -0.02941176  0.37843138\n",
      "   0.5         0.48431373  0.5         0.1627451  -0.03333334  0.00980392\n",
      "   0.00196078  0.00196078  0.00196078  0.00196078  0.00196078  0.00196078\n",
      "   0.00196078  0.00196078  0.00196078  0.00588235 -0.01372549  0.05294118\n",
      "   0.5         0.5         0.5         0.49607843]\n",
      " [-0.5        -0.5        -0.5        -0.48431373 -0.5         0.25294119\n",
      "   0.5         0.46862745  0.5        -0.17450981 -0.5        -0.48431373\n",
      "  -0.5        -0.5        -0.5        -0.5        -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5        -0.49607843 -0.5        -0.39411765  0.5         0.5\n",
      "   0.5         0.5       ]\n",
      " [-0.49607843 -0.48431373 -0.48039216 -0.46470588 -0.5         0.26078433\n",
      "   0.5         0.46862745  0.5        -0.15882353 -0.5        -0.46470588\n",
      "  -0.48039216 -0.48039216 -0.48039216 -0.48039216 -0.47254902 -0.46470588\n",
      "  -0.46470588 -0.47647059 -0.48039216 -0.47647059 -0.5        -0.37843138\n",
      "   0.5         0.5         0.5         0.5       ]\n",
      " [-0.5        -0.5        -0.5        -0.48431373 -0.5         0.25294119\n",
      "   0.5         0.46862745  0.5        -0.17450981 -0.5        -0.48431373\n",
      "  -0.5        -0.5        -0.5        -0.5        -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5        -0.49607843 -0.5        -0.39411765  0.49607843\n",
      "   0.5         0.5         0.49607843]\n",
      " [-0.5        -0.5        -0.5        -0.48431373 -0.5         0.25294119\n",
      "   0.5         0.46862745  0.5        -0.17450981 -0.5        -0.48431373\n",
      "  -0.5        -0.5        -0.49607843 -0.5        -0.33137256 -0.0254902\n",
      "  -0.09215686 -0.44509804 -0.5        -0.49215686 -0.5        -0.40588236\n",
      "   0.47254902  0.49607843  0.48431373  0.5       ]\n",
      " [-0.5        -0.5        -0.5        -0.48431373 -0.5         0.25294119\n",
      "   0.5         0.46862745  0.5        -0.17058824 -0.5        -0.48039216\n",
      "  -0.49607843 -0.48823529 -0.5        -0.42156863  0.35882354  0.5         0.5\n",
      "   0.10784314 -0.5        -0.48823529 -0.5        -0.47647059  0.28431374\n",
      "   0.5         0.5         0.35882354]\n",
      " [-0.5        -0.5        -0.5        -0.48431373 -0.5         0.25294119\n",
      "   0.5         0.46862745  0.5        -0.17450981 -0.5        -0.48431373\n",
      "  -0.5        -0.48823529 -0.5        -0.28431374  0.5         0.46862745\n",
      "   0.5         0.33137256 -0.5        -0.49215686 -0.49607843 -0.5\n",
      "  -0.37058824 -0.04509804 -0.04509804 -0.35490197]\n",
      " [-0.5        -0.5        -0.5        -0.48431373 -0.5         0.25294119\n",
      "   0.5         0.47254902  0.5        -0.13529412 -0.5        -0.42941177\n",
      "  -0.44509804 -0.43333334 -0.49607843 -0.21764706  0.5         0.48039216\n",
      "   0.5         0.32745099 -0.5        -0.49215686 -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5        -0.5       ]\n",
      " [-0.5        -0.5        -0.5        -0.48431373 -0.5         0.25294119\n",
      "   0.5         0.48823529  0.5         0.46078432  0.43725491  0.44117647\n",
      "   0.44117647  0.44117647  0.43725491  0.45686275  0.5         0.48823529\n",
      "   0.5         0.32745099 -0.5        -0.49215686 -0.5        -0.5\n",
      "  -0.49607843 -0.48039216 -0.48431373 -0.49215686]\n",
      " [-0.5        -0.5        -0.5        -0.48431373 -0.5         0.25294119\n",
      "   0.5         0.48823529  0.5         0.5         0.5         0.5         0.5\n",
      "   0.5         0.5         0.5         0.5         0.49215686  0.5\n",
      "   0.32745099 -0.5        -0.49215686 -0.5        -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5       ]\n",
      " [-0.5        -0.5        -0.5        -0.48431373 -0.5         0.25294119\n",
      "   0.5         0.48823529  0.5         0.46078432  0.43725491  0.44117647\n",
      "   0.44117647  0.44117647  0.43725491  0.45686275  0.5         0.48823529\n",
      "   0.5         0.32745099 -0.5        -0.49215686 -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5        -0.5       ]\n",
      " [-0.5        -0.5        -0.5        -0.48431373 -0.5         0.25294119\n",
      "   0.5         0.47254902  0.5        -0.13529412 -0.5        -0.42941177\n",
      "  -0.44509804 -0.43333334 -0.49607843 -0.21372549  0.5         0.47647059\n",
      "   0.5         0.32745099 -0.5        -0.49215686 -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5        -0.5       ]\n",
      " [-0.5        -0.5        -0.5        -0.48431373 -0.5         0.25294119\n",
      "   0.5         0.46862745  0.5        -0.17450981 -0.5        -0.48431373\n",
      "  -0.5        -0.48823529 -0.5        -0.28039217  0.5         0.46862745\n",
      "   0.5         0.32745099 -0.5        -0.49215686 -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5        -0.5       ]\n",
      " [-0.5        -0.5        -0.5        -0.48431373 -0.5         0.25294119\n",
      "   0.5         0.46862745  0.5        -0.17058824 -0.5        -0.48039216\n",
      "  -0.49607843 -0.48823529 -0.5        -0.41764706  0.36274511  0.5         0.5\n",
      "   0.13529412 -0.5        -0.49215686 -0.5        -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5       ]\n",
      " [-0.5        -0.5        -0.5        -0.48431373 -0.5         0.25294119\n",
      "   0.5         0.46862745  0.5        -0.17450981 -0.5        -0.48431373\n",
      "  -0.5        -0.5        -0.49607843 -0.5        -0.32745099 -0.02941176\n",
      "  -0.07254902 -0.42156863 -0.5        -0.49607843 -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5        -0.5       ]\n",
      " [-0.5        -0.5        -0.5        -0.48431373 -0.5         0.25294119\n",
      "   0.5         0.46862745  0.5        -0.17450981 -0.5        -0.48431373\n",
      "  -0.5        -0.5        -0.5        -0.5        -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5        -0.5        -0.5        -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5       ]\n",
      " [-0.5        -0.5        -0.5        -0.48431373 -0.5         0.25294119\n",
      "   0.5         0.46862745  0.5        -0.17450981 -0.5        -0.48431373\n",
      "  -0.5        -0.5        -0.5        -0.5        -0.49215686 -0.48039216\n",
      "  -0.48431373 -0.49607843 -0.5        -0.5        -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5        -0.5       ]\n",
      " [-0.5        -0.5        -0.5        -0.48431373 -0.5         0.25294119\n",
      "   0.5         0.46862745  0.5        -0.17450981 -0.5        -0.48431373\n",
      "  -0.5        -0.5        -0.5        -0.5        -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5        -0.5        -0.5        -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5       ]\n",
      " [-0.5        -0.5        -0.5        -0.48431373 -0.5         0.25294119\n",
      "   0.5         0.46862745  0.5        -0.17450981 -0.5        -0.48431373\n",
      "  -0.5        -0.5        -0.5        -0.5        -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5        -0.5        -0.5        -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5       ]\n",
      " [-0.49215686 -0.48431373 -0.48039216 -0.46470588 -0.5         0.26078433\n",
      "   0.5         0.46862745  0.5        -0.15882353 -0.5        -0.46470588\n",
      "  -0.48039216 -0.48039216 -0.48039216 -0.48039216 -0.48823529 -0.5        -0.5\n",
      "  -0.5        -0.5        -0.5        -0.5        -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5       ]\n",
      " [-0.5        -0.5        -0.5        -0.48431373 -0.5         0.25294119\n",
      "   0.5         0.46862745  0.5        -0.17450981 -0.5        -0.48431373\n",
      "  -0.5        -0.5        -0.5        -0.5        -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5        -0.5        -0.5        -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5       ]\n",
      " [-0.39411765 -0.0882353  -0.00196078  0.00980392 -0.02941176  0.37843138\n",
      "   0.5         0.48431373  0.5         0.1627451  -0.03333334  0.00980392\n",
      "   0.00196078  0.00196078  0.00196078 -0.00588235 -0.20588236 -0.5        -0.5\n",
      "  -0.5        -0.5        -0.5        -0.5        -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5       ]\n",
      " [ 0.31960785  0.5         0.49607843  0.5         0.5         0.5         0.5\n",
      "   0.5         0.5         0.5         0.5         0.5         0.5         0.5\n",
      "   0.5         0.49607843  0.5        -0.1509804  -0.5        -0.48431373\n",
      "  -0.5        -0.5        -0.5        -0.5        -0.5        -0.5        -0.5\n",
      "  -0.5       ]\n",
      " [ 0.46470588  0.5         0.48039216  0.48039216  0.48039216  0.49607843\n",
      "   0.5         0.5         0.5         0.48823529  0.48039216  0.48039216\n",
      "   0.48039216  0.48039216  0.48039216  0.48039216  0.5        -0.06078431\n",
      "  -0.5        -0.48431373 -0.5        -0.5        -0.5        -0.5        -0.5\n",
      "  -0.5        -0.5        -0.5       ]\n",
      " [-0.04509804  0.44509804  0.5         0.49607843  0.5         0.5         0.5\n",
      "   0.5         0.5         0.5         0.5         0.5         0.5         0.5\n",
      "   0.49607843  0.44901961  0.14313726 -0.44117647 -0.5        -0.49607843\n",
      "  -0.5        -0.5        -0.5        -0.5        -0.5        -0.5        -0.5\n",
      "  -0.5       ]]\n",
      "(28, 28)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x10a3e2f90>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP4AAAD8CAYAAABXXhlaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nK4iZ+DNmvIKYiT9jxiuImfgzZryC+H+nQQYhQAjDBgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x103e1af10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# load one pickle and check what we can find\n",
    "obj2 = pickle.load(open('notMNIST_large/F.pickle', 'r'))\n",
    "sampleBimage = obj2[1111]\n",
    "print(sampleBimage)\n",
    "print(sampleBimage.shape)\n",
    "plt.imshow(sampleBimage)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "cYznx5jUwzoO"
   },
   "source": [
    "---\n",
    "Problem 3\n",
    "---------\n",
    "Another check: we expect the data to be balanced across classes. Verify that.\n",
    "\n",
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "A: 1872\n",
      "B: 1873\n",
      "C: 1873\n",
      "D: 1873\n",
      "E: 1873\n",
      "F: 1872\n",
      "G: 1872\n",
      "H: 1872\n",
      "I: 1872\n",
      "J: 1872\n"
     ]
    }
   ],
   "source": [
    "for l in ['A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J']:\n",
    "    letters = pickle.load( open( 'notMNIST_small/{0}.pickle'.format(l), \"rb\" ) )\n",
    "    print('{0}: {1}'.format(l, len(letters)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "LA7M7K22ynCt"
   },
   "source": [
    "Merge and prune the training data as needed. Depending on your computer setup, you might not be able to fit it all in memory, and you can tune `train_size` as needed. The labels will be stored into a separate array of integers 0 through 9.\n",
    "\n",
    "Also create a validation dataset for hyperparameter tuning."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "cellView": "both",
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "output_extras": [
      {
       "item_id": 1
      }
     ]
    },
    "colab_type": "code",
    "collapsed": false,
    "executionInfo": {
     "elapsed": 411281,
     "status": "ok",
     "timestamp": 1444485897869,
     "user": {
      "color": "#1FA15D",
      "displayName": "Vincent Vanhoucke",
      "isAnonymous": false,
      "isMe": true,
      "permissionId": "05076109866853157986",
      "photoUrl": "//lh6.googleusercontent.com/-cCJa7dTDcgQ/AAAAAAAAAAI/AAAAAAAACgw/r2EZ_8oYer4/s50-c-k-no/photo.jpg",
      "sessionId": "2a0a5e044bb03b66",
      "userId": "102167687554210253930"
     },
     "user_tz": 420
    },
    "id": "s3mWgZLpyuzq",
    "outputId": "8af66da6-902d-4719-bedc-7c9fb7ae7948"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training: (200000, 28, 28) (200000,)\n",
      "Validation: (10000, 28, 28) (10000,)\n",
      "Testing: (10000, 28, 28) (10000,)\n"
     ]
    }
   ],
   "source": [
    "def make_arrays(nb_rows, img_size):\n",
    "  if nb_rows:\n",
    "    dataset = np.ndarray((nb_rows, img_size, img_size), dtype=np.float32)\n",
    "    labels = np.ndarray(nb_rows, dtype=np.int32)\n",
    "  else:\n",
    "    dataset, labels = None, None\n",
    "  return dataset, labels\n",
    "\n",
    "def merge_datasets(pickle_files, train_size, valid_size=0):\n",
    "  num_classes = len(pickle_files)\n",
    "  valid_dataset, valid_labels = make_arrays(valid_size, image_size)\n",
    "  train_dataset, train_labels = make_arrays(train_size, image_size)\n",
    "  vsize_per_class = valid_size // num_classes\n",
    "  tsize_per_class = train_size // num_classes\n",
    "    \n",
    "  start_v, start_t = 0, 0\n",
    "  end_v, end_t = vsize_per_class, tsize_per_class\n",
    "  end_l = vsize_per_class+tsize_per_class\n",
    "  for label, pickle_file in enumerate(pickle_files):       \n",
    "    try:\n",
    "      with open(pickle_file, 'rb') as f:\n",
    "        letter_set = pickle.load(f)\n",
    "        # let's shuffle the letters to have random validation and training set\n",
    "        np.random.shuffle(letter_set)\n",
    "        if valid_dataset is not None:\n",
    "          valid_letter = letter_set[:vsize_per_class, :, :]\n",
    "          valid_dataset[start_v:end_v, :, :] = valid_letter\n",
    "          valid_labels[start_v:end_v] = label\n",
    "          start_v += vsize_per_class\n",
    "          end_v += vsize_per_class\n",
    "                    \n",
    "        train_letter = letter_set[vsize_per_class:end_l, :, :]\n",
    "        train_dataset[start_t:end_t, :, :] = train_letter\n",
    "        train_labels[start_t:end_t] = label\n",
    "        start_t += tsize_per_class\n",
    "        end_t += tsize_per_class\n",
    "    except Exception as e:\n",
    "      print('Unable to process data from', pickle_file, ':', e)\n",
    "      raise\n",
    "    \n",
    "  return valid_dataset, valid_labels, train_dataset, train_labels\n",
    "            \n",
    "            \n",
    "train_size = 200000\n",
    "valid_size = 10000\n",
    "test_size = 10000\n",
    "\n",
    "valid_dataset, valid_labels, train_dataset, train_labels = merge_datasets(\n",
    "  train_datasets, train_size, valid_size)\n",
    "_, _, test_dataset, test_labels = merge_datasets(test_datasets, test_size)\n",
    "\n",
    "print('Training:', train_dataset.shape, train_labels.shape)\n",
    "print('Validation:', valid_dataset.shape, valid_labels.shape)\n",
    "print('Testing:', test_dataset.shape, test_labels.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "GPTCnjIcyuKN"
   },
   "source": [
    "Next, we'll randomize the data. It's important to have the labels well shuffled for the training and test distributions to match."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "cellView": "both",
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "collapsed": true,
    "id": "6WZ2l2tN2zOL"
   },
   "outputs": [],
   "source": [
    "def randomize(dataset, labels):\n",
    "  permutation = np.random.permutation(labels.shape[0])\n",
    "  shuffled_dataset = dataset[permutation,:,:]\n",
    "  shuffled_labels = labels[permutation]\n",
    "  return shuffled_dataset, shuffled_labels\n",
    "train_dataset, train_labels = randomize(train_dataset, train_labels)\n",
    "test_dataset, test_labels = randomize(test_dataset, test_labels)\n",
    "valid_dataset, valid_labels = randomize(valid_dataset, valid_labels)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "puDUTe6t6USl"
   },
   "source": [
    "---\n",
    "Problem 4\n",
    "---------\n",
    "Convince yourself that the data is still good after shuffling!\n",
    "\n",
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP4AAAD8CAYAAABXXhlaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nI/GPOOI1xJH4RxzxGuL/Ag27xkPJ6rFCAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10a566110>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "x = 12\n",
    "plt.imshow(train_dataset[x,:,:])\n",
    "train_labels[12]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "tIQJaJuwg5Hw"
   },
   "source": [
    "Finally, let's save the data for later reuse:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "cellView": "both",
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "collapsed": true,
    "id": "QiR_rETzem6C"
   },
   "outputs": [],
   "source": [
    "pickle_file = 'notMNIST.pickle'\n",
    "\n",
    "try:\n",
    "  f = open(pickle_file, 'wb')\n",
    "  save = {\n",
    "    'train_dataset': train_dataset,\n",
    "    'train_labels': train_labels,\n",
    "    'valid_dataset': valid_dataset,\n",
    "    'valid_labels': valid_labels,\n",
    "    'test_dataset': test_dataset,\n",
    "    'test_labels': test_labels,\n",
    "    }\n",
    "  pickle.dump(save, f, pickle.HIGHEST_PROTOCOL)\n",
    "  f.close()\n",
    "except Exception as e:\n",
    "  print('Unable to save data to', pickle_file, ':', e)\n",
    "  raise"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "cellView": "both",
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "output_extras": [
      {
       "item_id": 1
      }
     ]
    },
    "colab_type": "code",
    "collapsed": false,
    "executionInfo": {
     "elapsed": 413065,
     "status": "ok",
     "timestamp": 1444485899688,
     "user": {
      "color": "#1FA15D",
      "displayName": "Vincent Vanhoucke",
      "isAnonymous": false,
      "isMe": true,
      "permissionId": "05076109866853157986",
      "photoUrl": "//lh6.googleusercontent.com/-cCJa7dTDcgQ/AAAAAAAAAAI/AAAAAAAACgw/r2EZ_8oYer4/s50-c-k-no/photo.jpg",
      "sessionId": "2a0a5e044bb03b66",
      "userId": "102167687554210253930"
     },
     "user_tz": 420
    },
    "id": "hQbLjrW_iT39",
    "outputId": "b440efc6-5ee1-4cbc-d02d-93db44ebd956"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Compressed pickle size: 690800441\n"
     ]
    }
   ],
   "source": [
    "statinfo = os.stat(pickle_file)\n",
    "print('Compressed pickle size:', statinfo.st_size)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "gE_cRAQB33lk"
   },
   "source": [
    "---\n",
    "Problem 5\n",
    "---------\n",
    "\n",
    "By construction, this dataset might contain a lot of overlapping samples, including training data that's also contained in the validation and test set! Overlap between training and test can skew the results if you expect to use your model in an environment where there is never an overlap, but are actually ok if you expect to see training samples recur when you use it.\n",
    "Measure how much overlap there is between training, validation and test samples.\n",
    "\n",
    "Optional questions:\n",
    "- What about near duplicates between datasets? (images that are almost identical)\n",
    "- Create a sanitized validation and test set, and compare your accuracy on those in subsequent assignments.\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "L8oww1s4JMQx"
   },
   "source": [
    "---\n",
    "Problem 6\n",
    "---------\n",
    "\n",
    "Let's get an idea of what an off-the-shelf classifier can give you on this data. It's always good to check that there is something to learn, and that it's a problem that is not so trivial that a canned solution solves it.\n",
    "\n",
    "Train a simple model on this data using 50, 100, 1000 and 5000 training samples. Hint: you can use the LogisticRegression model from sklearn.linear_model.\n",
    "\n",
    "Optional question: train an off-the-shelf model on all the data!\n",
    "\n",
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Converting matricies to vectors\n",
      "Done\n"
     ]
    }
   ],
   "source": [
    "print(\"Converting matricies to vectors\")\n",
    "\n",
    "x_train = train_dataset.reshape(train_dataset.shape[0], train_dataset.shape[1] * train_dataset.shape[2])\n",
    "y_train = train_labels\n",
    "\n",
    "x_test = test_dataset.reshape(test_dataset.shape[0], test_dataset.shape[1] * test_dataset.shape[2])\n",
    "y_test = test_labels\n",
    "\n",
    "x_valid = valid_dataset.reshape(valid_dataset.shape[0], valid_dataset.shape[1] * valid_dataset.shape[2])\n",
    "y_valid = valid_labels\n",
    "\n",
    "print(\"Done\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training size 50. Classifier rate: 39.86.\n",
      "Training size 100. Classifier rate: 63.02.\n",
      "Training size 1000. Classifier rate: 86.37.\n",
      "Training size 5000. Classifier rate: 88.33.\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.metrics import mean_squared_error, make_scorer\n",
    "from sklearn.grid_search import GridSearchCV\n",
    "\n",
    "dataset_sizes = [50, 100, 1000, 5000]\n",
    "\n",
    "def scoring_function(y_true, y_pred):\n",
    "    return np.mean(y_true.ravel() == y_pred.ravel()) * 100\n",
    "\n",
    "for size in dataset_sizes:\n",
    "    # Optimal parameters were found by applying GridSearchCV\n",
    "    regressor = LogisticRegression(C=0.008, penalty='l2', tol=0.00005, solver='newton-cg', multi_class='multinomial')\n",
    "    regressor.fit(x_train[:size], y_train[:size])\n",
    "    \n",
    "    pred = regressor.predict(x_test)\n",
    "    clf_rate = scoring_function(y_test, pred)\n",
    "    print(\"Training size {}. Classifier rate: {}.\".format(size, clf_rate))\n"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "default_view": {},
   "name": "1_notmnist.ipynb",
   "provenance": [],
   "version": "0.3.2",
   "views": {}
  },
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}