cnnfashionMNIST.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import tensorflow_datasets as tfds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Warning: Setting shuffle_files=True because split=TRAIN and shuffle_files=None. This behavior will be deprecated on 2019-08-06, at which point shuffle_files=False will be the default for all splits.\n"
     ]
    }
   ],
   "source": [
    "# tfds.load returns a tf.data.Dataset object and a tf.core.DatasetInfo object. The latter contains info about how the set is split \n",
    "# into test and train, number of examples and names of features\n",
    "dataset, metadata = tfds.load('fashion_mnist', as_supervised=True, with_info=True) \n",
    "train_dataset, test_dataset = dataset['train'], dataset['test']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAEKCAYAAADdIIPUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAFNpJREFUeJzt3W2MlWV6B/D/Bc4AM7wzMA4IHfEVAhbsRNvaKpv1bY0umrpmMd2wiVn8oJuusbGGfli+NJpatca0NKwQ0ey6bt21UENWwHR1l1bCSEZQQBhgFJxhXuT9bYCZqx/mYXeUea5rPG/PYa//LyEznP+559wcuDhz5nru+xZVBRHFMyTrCRBRNlj8REGx+ImCYvETBcXiJwqKxU8UFIufKCgWP1FQLH6ioC4p5YPV1NRofX19KR8yhK6urtSspqbGHHvu3DkzP3z4sJkPGWK/fpw4cSI1GzNmjDl29OjRZk4XamlpQVdXlwzmvnkVv4jcCeAFAEMBvKSqT1v3r6+vR2NjYz4PmbMsL2MWGdTfRc6WL1+emj300EPm2M7OTjNftWqVmVdVVZn5xo0bU7O77rrLHHvHHXeYeT68fw/F/jsrloaGhkHfN+dv+0VkKIB/A/AtADMBLBCRmbl+PSIqrXze898AoFlV96jqGQA/BzC/MNMiomLLp/inANjX7/f7k9u+REQWiUijiDR632ISUenkU/wDvSm64I2Uqi5T1QZVbZg4cWIeD0dEhZRP8e8HMLXf7y8D0JrfdIioVPIp/k0ArhKRy0WkEsB3AawuzLSIqNhybvWp6jkReRTA2+hr9a1Q1Y8LNrMCK+fWTVtbm5k/88wzZt7U1JSavfTSS+bYlpYWMz958qSZz54928zHjRuXmj377LPm2HfffdfMH3zwQTOfNWtWaub9e+jt7TVzb3w5/3s7L68+v6quAbCmQHMhohLi5b1EQbH4iYJi8RMFxeInCorFTxQUi58oqJKu58+S17f11qVbXn31VTNfs8buhu7du9fML7300pzz1lb7ostbb73VzL0199u3bzdzaz+Bnp4ec+z777+fV3755ZenZo8//rg5duZMe4HqH8NJV3zlJwqKxU8UFIufKCgWP1FQLH6ioFj8REGFafXl08oDgKVLl6Zmb7/9tjnWajkBwKlTp8x8xIgRZn706NHUbMOGDebYLVu2mPmMGTPMfPjw4WZeWVmZmlVUVJhjvRZnd3e3mVttxkceecQc++KLL5q5tVwYuDh2B+YrP1FQLH6ioFj8REGx+ImCYvETBcXiJwqKxU8UVJg+f77eeOON1GzUqFHm2GHDhpm51acH/H74+PHjUzPvCO4PP/zQzL0+v7cs1zqlyet1nz592sw7OjrMfMqUC06PG/Rjr1ixwsyfe+45My+HPr6Hr/xEQbH4iYJi8RMFxeInCorFTxQUi58oKBY/UVB59flFpAXAMQA9AM6pakMhJpWFTz/91MyrqqpSM2/ttjUWAK699loznzp1qplbj3/77bebY5944gkzX758uZl71zjU1dWlZtXV1eZYr88/evRoM7f6/J2dnebYAwcOmHkxt4IvlUJc5PMNVe0qwNchohIq//+eiKgo8i1+BbBWRD4QkUWFmBARlUa+3/bfpKqtIjIJwDoR2aGq7/W/Q/KfwiIAmDZtWp4PR0SFktcrv6q2Jh87ALwJ4IYB7rNMVRtUtcFa5EFEpZVz8YtItYiMOv85gNsBfFSoiRFRceXzbX8tgDeTpYuXAPiZqv66ILMioqLLufhVdQ+APy3gXDLV3t5u5taae29ffm9t944dO8y8q8vupB47diw18/a+b2lpMfNPPvnEzOfMmWPmu3btSs28vQa8583r1VvXTwwdOtQc+/HHH5v5559/bubetRnlgK0+oqBY/ERBsfiJgmLxEwXF4icKisVPFBS37k54rR1rm2ivbWQtawX8rbm9rz927NjUzFt66h017eVHjhwxc2tprDVvADh8+LCZe1ui7969OzVrbW01x3pbknvtV7b6iKhssfiJgmLxEwXF4icKisVPFBSLnygoFj9RUOzzJ/bs2WPmI0aMSM3OnDljjm1qajLzkSNHmrnXc7a27u7u7jbHen36cePGmbl3xLf13Fxyif3P7/jx42Zu/Z0AwNatW3Me611bsWnTJjOfO3eumZcDvvITBcXiJwqKxU8UFIufKCgWP1FQLH6ioFj8REGxz5/Yt2+fmVtr7r0tqL/44gsz98Z7x2CfPXs2NfP2CvCOkvauYRg+fHjO40+ePGmOtf5cAFBZWZnzY3tf27sOoLGx0cwXLSr/oyv5yk8UFIufKCgWP1FQLH6ioFj8REGx+ImCYvETBeX2+UVkBYC7AXSo6qzktvEAXgdQD6AFwAOqeqh40yy+bdu2mbl1DLa3f/xTTz1l5o899piZ19bWmrnXs7acOnXKzL298w8ePGjmVi/fu77B2xvfe94ffvjh1GzJkiXm2IkTJ5q5dfT4xWIwr/wvA7jzK7c9CeAdVb0KwDvJ74noIuIWv6q+B+Cr/73PB7Ay+XwlgHsLPC8iKrJc3/PXqmobACQfJxVuSkRUCkX/gZ+ILBKRRhFp7OzsLPbDEdEg5Vr87SJSBwDJx9RTLFV1mao2qGqD90MUIiqdXIt/NYCFyecLAawqzHSIqFTc4heR1wD8H4BrRGS/iDwE4GkAt4nILgC3Jb8noouI2+dX1QUp0TcLPJdMXXbZZWZu9XW9s969r11dXW3ms2fPNvPe3t7UzNvz/+jRo2Y+aZL9s9wZM2aYuXUdwZgxY8yxXn7ixAkzb25uTs28MwPq6urM3Lru42LBK/yIgmLxEwXF4icKisVPFBSLnygoFj9RUGG27vZaWla7DLC3cp48ebI5dsqUKWZeVVVl5t4x21butbS8ZbVe7h2jbbXEvDakZ+zYsWZ+3XXXpWZLly41x4qIme/cudPMvb8zbzlyKfCVnygoFj9RUCx+oqBY/ERBsfiJgmLxEwXF4icKKkyff8uWLWZ+4MABM29vb0/NbrnlFnPs/v3788q96wAOHUrfNX3o0KHmWO/4cG9p6969e83c2vrb69N724p71yBYcx83bpw5dvfu3Wbubae+ceNGM7/55pvNvBT4yk8UFIufKCgWP1FQLH6ioFj8REGx+ImCYvETBRWmz+8d9+w5ffp0anbvvfY5pV4v3Ftz7/XDrV5+RUWFOdZTU1OTV27Nzeu1e3sFeNdHWM/7vHnzzLFr1641c+/6B2vbcIB9fiLKEIufKCgWP1FQLH6ioFj8REGx+ImCYvETBeX2+UVkBYC7AXSo6qzktiUAfgCgM7nbYlVdU6xJFsK+ffvM3Dvu+cyZM6nZ9ddfb45dt26dmXu9+CNHjpi51Q/31vN7R0176/2t9fre1x8yxH7t8ebm/dmsazvmz59vjl21apWZe+c8dHZ2mnk5GMwr/8sA7hzg9udVdU7yq6wLn4gu5Ba/qr4H4GAJ5kJEJZTPe/5HRWSLiKwQEfs6TSIqO7kW/1IAVwCYA6ANwLNpdxSRRSLSKCKNF8P7IKIocip+VW1X1R5V7QXwEwA3GPddpqoNqtowceLEXOdJRAWWU/GLSP8lTfcB+Kgw0yGiUhlMq+81APMA1IjIfgA/BjBPROYAUAAtAB4u4hyJqAjc4lfVBQPcvLwIcymqzz77zMy9XrvXz7Zs377dzMeMGWPmo0ePNvPKysrUzNsrwOtXe2vuJ0yYYObW43t/blU18/Hjx5v5tm3bUrN77rnHHHvwoN3gmj59upl7eziUA17hRxQUi58oKBY/UVAsfqKgWPxEQbH4iYIKs3V3R0eHmXstL0tra6uZe8uJvXact0W1ta348OHDzbFtbW1m7j1v3tHmVovUa5+ePHnSzGfPnm3mmzZtSs2847+HDRtm5mfPnjVzr01ZDvjKTxQUi58oKBY/UVAsfqKgWPxEQbH4iYJi8RMFFabP7y2LnTx5spnv3r07NfOWf3q9dO+4556eHjO3etIiYo71rgPwnhevX25tO+4dPe5dP+HtDGVd/+Bde1FdXW3m9fX1Zl5VVWXm5YCv/ERBsfiJgmLxEwXF4icKisVPFBSLnygoFj9RUGH6/O3t7WbuHUVtrd/2et1eH/+KK64wc2/bcWvrbq/P7/WzvWOwvesEvOcmn6/tbd09d+7c1My7RsD7c+/cudPMvee9HPCVnygoFj9RUCx+oqBY/ERBsfiJgmLxEwXF4icKyu3zi8hUAK8AuBRAL4BlqvqCiIwH8DqAegAtAB5Q1UPFm2p+vL7roUP21K019Zs3bzbHeuv9vb0GWlpazPzcuXOpmXcmQFdXl5l7R1E3NzebubU3v9fH9/b19474tq7t2LVrlznWm5t3HcDFYDCv/OcAPK6qMwD8OYBHRGQmgCcBvKOqVwF4J/k9EV0k3OJX1TZV3Zx8fgzAdgBTAMwHsDK520oA9xZrkkRUeF/rPb+I1AOYC2AjgFpVbQP6/oMAMKnQkyOi4hl08YvISAC/BPAjVT36NcYtEpFGEWns7OzMZY5EVASDKn4RqUBf4f9UVX+V3NwuInVJXgdgwBMdVXWZqjaoaoO3mIKISsctfun7MflyANtV9bl+0WoAC5PPFwJYVfjpEVGxDGZJ700Avgdgq4g0JbctBvA0gF+IyEMAPgPwneJMsTC8rZS9VuDIkSNTM68V57XTamtrzdxrO+Uz1jsG2zNixAgzt7bPnjBhgjn2xIkTZu4dk239nb711lvmWG8psrcUuru728zLgVv8qvo7AGnP4jcLOx0iKhVe4UcUFIufKCgWP1FQLH6ioFj8REGx+ImCCrN1t9eX9Vj9cq8n7PWzvWW3+Wx/3dvbm9fXVlUzt5YTA8CxY8dSM29Z7JkzZ8zcu0bh6quvTs3Wr19vjvV4y4nzuTajVPjKTxQUi58oKBY/UVAsfqKgWPxEQbH4iYJi8RMFFabPbx1jDQBHjhwxc6vfPWSI/X+o10uvqKgwc2/NvDXe+9pev9q7PsLa5wCwtyX3rn/weNcBWNdPeLtKedc3eNuKe8eHlwO+8hMFxeInCorFTxQUi58oKBY/UVAsfqKgWPxEQYXp80+aZB8l6PXSrX62tTc94J8Z4LGOBwfsNfX59MIBf/95b029NTfvWHTvsb1rM+rr61Mz79oMbz2+d43CtGnTzLwc8JWfKCgWP1FQLH6ioFj8REGx+ImCYvETBcXiJwrK7fOLyFQArwC4FEAvgGWq+oKILAHwAwCdyV0Xq+qaYk00X7W1tWbure+2zoI/fvy4Odbb2947h97rSVt783v79nvr1q199wF/731rPwFvjwWPd21GPucdeLx9Di6GffsHc5HPOQCPq+pmERkF4AMRWZdkz6vqvxRvekRULG7xq2obgLbk82Mish3AlGJPjIiK62u95xeRegBzAWxMbnpURLaIyAoRGZcyZpGINIpIY2dn50B3IaIMDLr4RWQkgF8C+JGqHgWwFMAVAOag7zuDZwcap6rLVLVBVRu899VEVDqDKn4RqUBf4f9UVX8FAKrarqo9qtoL4CcAbijeNImo0NziFxEBsBzAdlV9rt/tdf3udh+Ajwo/PSIqlsH8tP8mAN8DsFVEmpLbFgNYICJzACiAFgAPF2WGBbJjxw4zb2trM3OrtVNXV5eaAf722dOnTzdz7+3SqFGjUjNvObDVwgSAa665xsz37dtn5q2tramZtxzYW9LrsVqkV155pTm2qanJzJubm8187NixZl4OBvPT/t8BkAGisu3pE5GPV/gRBcXiJwqKxU8UFIufKCgWP1FQLH6ioMJs3X3jjTea+YYNG8z86NGjqdmaNex6/rFZvHixma9fv97M77///kJOpyj4yk8UFIufKCgWP1FQLH6ioFj8REGx+ImCYvETBSXe1s0FfTCRTgCf9rupBkBXySbw9ZTr3Mp1XgDnlqtCzu1PVHVQ++WVtPgveHCRRlVtyGwChnKdW7nOC+DccpXV3PhtP1FQLH6ioLIu/mUZP76lXOdWrvMCOLdcZTK3TN/zE1F2sn7lJ6KMZFL8InKniHwiIs0i8mQWc0gjIi0islVEmkSkMeO5rBCRDhH5qN9t40VknYjsSj4OeExaRnNbIiKfJ89dk4jcldHcporI/4jIdhH5WET+Lrk90+fOmFcmz1vJv+0XkaEAdgK4DcB+AJsALFDVbSWdSAoRaQHQoKqZ94RF5GYAxwG8oqqzktv+GcBBVX06+Y9znKr+Q5nMbQmA41mf3JwcKFPX/2RpAPcC+D4yfO6MeT2ADJ63LF75bwDQrKp7VPUMgJ8DmJ/BPMqeqr4H4OBXbp4PYGXy+Ur0/eMpuZS5lQVVbVPVzcnnxwCcP1k60+fOmFcmsij+KQD6H/OyH+V15LcCWCsiH4jIoqwnM4Da5Nj088enT8p4Pl/lntxcSl85WbpsnrtcTrwutCyKf6DTf8qp5XCTql4P4FsAHkm+vaXBGdTJzaUywMnSZSHXE68LLYvi3w9gar/fXwYg/UC3ElPV1uRjB4A3UX6nD7efPyQ1+diR8Xx+r5xObh7oZGmUwXNXTideZ1H8mwBcJSKXi0glgO8CWJ3BPC4gItXJD2IgItUAbkf5nT68GsDC5POFAFZlOJcvKZeTm9NOlkbGz125nXidyUU+SSvjXwEMBbBCVf+p5JMYgIhMR9+rPdC3s/HPspybiLwGYB76Vn21A/gxgP8C8AsA0wB8BuA7qlryH7ylzG0e+r51/f3JzeffY5d4bn8F4LcAtgLoTW5ejL7315k9d8a8FiCD541X+BEFxSv8iIJi8RMFxeInCorFTxQUi58oKBZ/UCLyj8nKsi3JSrIbkxWNNQPc99tpqy9FZJ6I/GXxZ0yFFuaUXvoDEfkLAHcDuF5Vu5OCr0y7v6quxgAXYonIJejr7R8H8L/FmS0VC4s/pjoAXaraDQDnly/3XYCGH4rIPQAq0HcRzA4R+T76ljk/KiIvo28139zk400AekTkbwH8UFV/W+o/DOWG3/bHtBbAVBHZKSL/LiK39Mu6koVNSwH8fcr4qwHcqqp/A+A/ADyvqnNY+BcXFn9AqnocwJ8BWASgE8Dryas7AJxfBPMBgPqUL/GfqtpTzDlS8fHb/qCS4v0NgN+IyFb8YcFLd/KxB+n/Pk4Ud3ZUCnzlD0hErhGRq/rdNAdfPkbt6zgGYFT+s6JSY/HHNBLAShHZJiJbAMwEsCTHr/XfAO5L2oV/XagJUvFxVR9RUHzlJwqKxU8UFIufKCgWP1FQLH6ioFj8REGx+ImCYvETBfX/9HSwveDtP7UAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Human readable labels for the classes\n",
    "class_names = ['T-shirt/top', 'Trouser', 'Pullover', 'Dress', 'Coat',\n",
    "               'Sandal',      'Shirt',   'Sneaker',  'Bag',   'Ankle boot']\n",
    "# Print example\n",
    "for (image, label) in train_dataset.take(1):\n",
    "    image = image.numpy().reshape((28,28))\n",
    "    plt.imshow(image, cmap= plt.cm.binary)\n",
    "    plt.xlabel(class_names[label])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "60000 10000\n"
     ]
    }
   ],
   "source": [
    "# Size of training and test sets\n",
    "train_size = metadata.splits['train'].num_examples\n",
    "test_size = metadata.splits['test'].num_examples\n",
    "# Normalise data\n",
    "def minmaxScaling(images, labels):\n",
    "    images = tf.cast(images, tf.float32)\n",
    "    images = images/255\n",
    "    return images, labels\n",
    "\n",
    "train_dataset = train_dataset.map(minmaxScaling)\n",
    "test_dataset = test_dataset.map(minmaxScaling)\n",
    "print(train_size, test_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Convolutional Model setup\n",
    "model = tf.keras.Sequential([\n",
    "    tf.keras.layers.Conv2D(filters= 16, kernel_size = 2, padding = 'same', activation=('relu')),\n",
    "    tf.keras.layers.MaxPooling2D(pool_size=2, padding='same'),\n",
    "    #tf.keras.layers.Conv2D(filters= 32, kernel_size = 2, padding = 'same', activation=('relu')),\n",
    "    #tf.keras.layers.MaxPooling2D(pool_size=2, padding='same'),\n",
    "    tf.keras.layers.Flatten(),\n",
    "    tf.keras.layers.Dense(units=10, input_shape=([1]), activation=('softmax'))\n",
    "])\n",
    "# Compile model\n",
    "model.compile(loss='sparse_categorical_crossentropy', \n",
    "              optimizer = 'adam',\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Shuffling and splitting of training set \n",
    "BATCH_SIZE = 32\n",
    "train_dataset = train_dataset.repeat().shuffle(train_size).batch(BATCH_SIZE)\n",
    "test_dataset_batched = test_dataset.batch(BATCH_SIZE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "WARNING:tensorflow:From /home/erick/anaconda3/envs/env1/lib/python3.7/site-packages/tensorflow/python/ops/math_grad.py:1250: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.where in 2.0, which has the same broadcast rule as np.where\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /home/erick/anaconda3/envs/env1/lib/python3.7/site-packages/tensorflow/python/ops/math_grad.py:1250: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.where in 2.0, which has the same broadcast rule as np.where\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1875/1875 [==============================] - 55s 29ms/step - loss: 0.4918 - accuracy: 0.7753\n",
      "Epoch 2/5\n",
      "1875/1875 [==============================] - 44s 24ms/step - loss: 0.3569 - accuracy: 0.8711\n",
      "Epoch 3/5\n",
      "1875/1875 [==============================] - 48s 26ms/step - loss: 0.3176 - accuracy: 0.8855\n",
      "Epoch 4/5\n",
      "1875/1875 [==============================] - 49s 26ms/step - loss: 0.2972 - accuracy: 0.8937\n",
      "Epoch 5/5\n",
      "1875/1875 [==============================] - 53s 28ms/step - loss: 0.2818 - accuracy: 0.8971\n"
     ]
    }
   ],
   "source": [
    "# Training model \n",
    "trained = model.fit(train_dataset, epochs=5, verbose=True, steps_per_epoch=np.ceil(train_size/BATCH_SIZE))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 6s 20ms/step - loss: 0.3116 - accuracy: 0.8875\n"
     ]
    }
   ],
   "source": [
    "# Evaluation on test set\n",
    "test_loss, test_acc = model.evaluate(test_dataset_batched, steps=np.ceil(test_size/BATCH_SIZE))\n",
    "predictions = model.predict(test_dataset_batched)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 9 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Print example and its predicted class\n",
    "plt.figure(figsize=(10,10))\n",
    "\n",
    "i = 0\n",
    "for (image, label) in test_dataset.take(9):\n",
    "    image = image.numpy().reshape((28,28))\n",
    "    plt.subplot(3,3,i+1)\n",
    "    plt.imshow(image, cmap= plt.cm.binary)\n",
    "    plt.xlabel((class_names[label], class_names[predictions[i].argmax()]))\n",
    "    i += 1\n",
    "plt.show()"
   ]
  }
 ],
 "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}