improved_variational_autoencoder.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Variational Autoencoder (VAE) for real-valued input data\n",
    "\n",
    "This notebook contains the class definition of a VAE with two hidden layers in the encoder and decoder. The layer sizes and the dimension of the model can be chosen freely. The model is defined for real-valued input data and uses a sigmoid output layer.\n",
    "\n",
    "For details on the individual aspects, look at the other notebook. The functionality of this model is exactly the same as before. However, I tried to improve the implementation:\n",
    "\n",
    "- To structure the class I looked at existing TensorFlow models and a [post](https://blog.metaflow.fr/tensorflow-a-proposal-of-good-practices-for-files-folders-and-models-architecture-f23171501ae3) on good practices for model architectures\n",
    "- I used tf.layers.dense to avoid initializing weights and biases manually\n",
    "- I used the Xavier initializer provided by tf.contrib\n",
    "- The model is saved and can be restored\n",
    "- The implementation includes tensorboard to aid debugging and visalization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Extracting MNIST_data/train-images-idx3-ubyte.gz\n",
      "Extracting MNIST_data/train-labels-idx1-ubyte.gz\n",
      "Extracting MNIST_data/t10k-images-idx3-ubyte.gz\n",
      "Extracting MNIST_data/t10k-labels-idx1-ubyte.gz\n"
     ]
    }
   ],
   "source": [
    "% matplotlib inline\n",
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from datetime import datetime\n",
    "\n",
    "# Import the MNIST Data\n",
    "from tensorflow.examples.tutorials.mnist import input_data\n",
    "mnist = input_data.read_data_sets('MNIST_data', one_hot=True)\n",
    "n_samples = mnist.train.num_examples"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Class definition"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "class variational_autoencoder:\n",
    "    \n",
    "    def __init__(self, config):\n",
    "        self.config = config\n",
    "        self.initializer = tf.contrib.layers.xavier_initializer()\n",
    "        self.activation_fct = self.config['activation_fct']\n",
    "        self.random_seed = self.config['random_seed']\n",
    "        self.learning_rate = self.config['learning_rate']      \n",
    "        \n",
    "        self.sess = tf.InteractiveSession()\n",
    "        self.build_graph()\n",
    "        \n",
    "    def _create_placeholders(self):\n",
    "        \"\"\"\n",
    "        Creates placeholder for the input data\n",
    "        \"\"\"\n",
    "        with tf.name_scope(\"data\"):\n",
    "            self.data = tf.placeholder(tf.float32, shape=[None, self.config['n_input']])\n",
    "            \n",
    "                \n",
    "    def _create_encoder(self):    \n",
    "        \"\"\"\n",
    "        Computes the output of the encoder network for a given input vector\n",
    "        \n",
    "        The encoder network is used to approximate the true posterior p(z|x)\n",
    "        using the variational distribution q_phi(z|x). Usually, q_phi(z|x) is\n",
    "        taken to be a Gaussian distribution with a diagonal covariance matrix\n",
    "        whose mean and variance vectors are parametrized by a neural network\n",
    "        with input x.\n",
    "        So the encoder network takes x as an input and produces a vector of \n",
    "        means and a vector of variances (or more precise a vector of log squared variances) \n",
    "        of a Gaussian distribution. From this we can sample values of z, i.e. z ~ q_phi(z|x)\n",
    "        \"\"\"\n",
    "        with tf.name_scope(\"encoder\"):\n",
    "            \n",
    "            layer1 = tf.layers.dense(self.data, self.config['h1_encoder'], activation=self.activation_fct, kernel_initializer=self.initializer)\n",
    "            layer2 = tf.layers.dense(layer1, self.config['h2_encoder'], activation=self.activation_fct, kernel_initializer=self.initializer)\n",
    "            means = tf.layers.dense(layer2, self.config['n_latent'], activation=None, kernel_initializer=self.initializer)\n",
    "            log_sigmas_sq = tf.layers.dense(layer2, self.config['n_latent'], activation=None, kernel_initializer=self.initializer)\n",
    "\n",
    "            # Vector of means\n",
    "            self.encoder_mean = means\n",
    "            # Vector of log squared variances\n",
    "            self.encoder_loq_sigma_sq = log_sigmas_sq\n",
    "\n",
    "    def _create_decoder(self):      \n",
    "        \"\"\"\n",
    "        Computes the output of the decoder network (i.e. the reconstruction) for some latent vector z.\n",
    "        The latent vector is sampled from the encoder using the reparametrization trick.\n",
    "        The decoder network takes a latent variable z as an input and reproduces the input x. \n",
    "        \"\"\"\n",
    "        with tf.name_scope(\"decoder\"):\n",
    "            # Reparametrization trick\n",
    "            eps = tf.random_normal(tf.shape(self.encoder_loq_sigma_sq), dtype=tf.float32)\n",
    "            self.z = self.encoder_mean + tf.sqrt(tf.exp(self.encoder_loq_sigma_sq)) * eps\n",
    "            # Layer activations\n",
    "            layer1 = tf.layers.dense(self.z, self.config['h1_decoder'], activation=self.activation_fct, kernel_initializer=self.initializer)\n",
    "            layer2 = tf.layers.dense(layer1, self.config['h2_decoder'], activation=self.activation_fct, kernel_initializer=self.initializer)\n",
    "            # Reconstruction                       \n",
    "            self.logits = tf.layers.dense(layer2, self.config['n_input'])   \n",
    "            self.reconstr = tf.sigmoid(self.logits)\n",
    "            \n",
    "    def _create_loss(self):\n",
    "        \"\"\"\n",
    "        Computes the loss of the network. The loss function has two terms\n",
    "        1) The reconstruction loss: -log(p(x|z)) which rewards a good reconstruction\n",
    "\n",
    "        2) The latent loss: Kullback-Leibler divergence between q_phi(z|x) and the prior\n",
    "        p(z) (which is given by a standard normal distribution).\n",
    "        This loss acts as a regularizer because it penalizes q_phi(z|x) for\n",
    "        deviating from a standard normal distribution.\n",
    "        \"\"\"\n",
    "        with tf.name_scope(\"loss\"):\n",
    "            with tf.name_scope(\"latent_loss\"):             \n",
    "                self.latent_loss = -0.5*tf.reduce_sum(1 + self.encoder_loq_sigma_sq\n",
    "                                             - tf.square(self.encoder_mean) \n",
    "                                             - tf.exp(self.encoder_loq_sigma_sq))\n",
    "\n",
    "            with tf.name_scope(\"reconstruction_loss\"):\n",
    "                self.reconstr_loss = tf.reduce_sum(\n",
    "                    tf.nn.sigmoid_cross_entropy_with_logits(labels=self.data, logits=self.logits))\n",
    "\n",
    "            \n",
    "            self.loss = self.latent_loss + self.reconstr_loss\n",
    "        \n",
    "    def _create_optimizer(self):\n",
    "        \"\"\"\n",
    "        Minimizes the loss using an Adam optimizer\n",
    "        \"\"\"\n",
    "        with tf.name_scope(\"optimizer\"):\n",
    "            self.optimizer = tf.train.AdamOptimizer(self.learning_rate).minimize(self.loss)\n",
    "\n",
    "    def _create_summaries(self):\n",
    "        \"\"\"\n",
    "        Adds summaries for visualization in TensorBoard\n",
    "        \"\"\"\n",
    "        with tf.name_scope(\"summaries\"):\n",
    "            tf.summary.scalar(\"loss\", self.loss)    \n",
    "            tf.summary.scalar(\"latent_loss\", self.latent_loss)\n",
    "            tf.summary.scalar(\"reconstruction_loss\", self.reconstr_loss)\n",
    "            self.summary_op = tf.summary.merge_all()\n",
    "            \n",
    "    def build_graph(self):\n",
    "        \"\"\"\n",
    "        Creates the VAE network by creating placeholders and setting up the \n",
    "        encoder and decoder network, the loss, the optimizer, summaries for\n",
    "        tensorboard and initializing all variables.\n",
    "        \"\"\"\n",
    "        print(\"Creating placeholders...\")\n",
    "        self._create_placeholders()\n",
    "        print(\"Create encoder...\")\n",
    "        self._create_encoder()\n",
    "        print(\"Create decoder...\")\n",
    "        self._create_decoder()\n",
    "        print(\"Create loss...\")\n",
    "        self._create_loss()\n",
    "        print(\"Create optimizer...\")\n",
    "        self._create_optimizer()\n",
    "        print(\"Create summaries...\")\n",
    "        self._create_summaries()\n",
    "        print(\"Initializing all variables...\")\n",
    "        with tf.name_scope(\"init_variables\"):\n",
    "            self.sess.run(tf.global_variables_initializer())\n",
    "        \n",
    "        \n",
    "    def batch_fit(self, input_data): \n",
    "        \"\"\"\n",
    "        Trains the model on a given batch of input data.\n",
    "        \"\"\"\n",
    "        with tf.name_scope(\"batch_fit\"):\n",
    "            self.sess.run(self.optimizer, {self.data:input_data})\n",
    "            return self\n",
    "            \n",
    "    def get_loss(self, input_data):\n",
    "        \"\"\"\n",
    "        Returns the loss of the model\n",
    "        \"\"\"\n",
    "        with tf.name_scope(\"get_loss\"):\n",
    "            return self.sess.run(self.loss, {self.data:input_data})\n",
    "    \n",
    "    def reconstruct(self, input_data):\n",
    "        \"\"\"\n",
    "        Reconstructs a given batch of input data\n",
    "        \"\"\"\n",
    "        with tf.name_scope(\"reconstruct_data\"):\n",
    "            return self.sess.run(self.reconstr, {self.data:input_data})\n",
    "    \n",
    "    def generate(self, batch_size):\n",
    "        \"\"\"\n",
    "        Generates new data by sampling from the latent dimension and running the decoder\n",
    "        \"\"\"\n",
    "        with tf.name_scope('generate_data'):\n",
    "            z = np.random.normal(size=(batch_size, self.config['n_latent']))\n",
    "\n",
    "            new_data = self.sess.run(self.reconstr, {self.z: z})\n",
    "            return new_data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model initalization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Creating placeholders...\n",
      "Create encoder...\n",
      "Create decoder...\n",
      "Create loss...\n",
      "Create optimizer...\n",
      "Create summaries...\n",
      "Initializing all variables...\n"
     ]
    }
   ],
   "source": [
    "tf.reset_default_graph()\n",
    "\n",
    "n_epochs = 40\n",
    "batch_size = 128\n",
    "display_step = 5\n",
    "\n",
    "net_architecture = dict(\n",
    "    \n",
    "    n_input = 784,\n",
    "    \n",
    "    h1_encoder = 500,\n",
    "    h2_encoder = 500,\n",
    "    \n",
    "    h1_decoder = 500,\n",
    "    h2_decoder = 500,\n",
    "    \n",
    "    n_latent = 20,    \n",
    "    \n",
    "    activation_fct = tf.nn.elu, #tf.nn.softplus,\n",
    "    learning_rate = 0.001,\n",
    "    random_seed = 123\n",
    ")\n",
    "\n",
    "\n",
    "model = variational_autoencoder(net_architecture)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TensorBoard log directory:  /tmp/tensorboard/run-20170620125133\n",
      "Epoch:  0 | Loss:  14915.1 | Latent loss:  2850.59 | Reconstr. loss:  12064.6\n",
      "Epoch:  1 | Loss:  14210.7 | Latent loss:  3032.94 | Reconstr. loss:  11177.8\n",
      "Epoch:  2 | Loss:  13975.1 | Latent loss:  3055.83 | Reconstr. loss:  10919.3\n",
      "Epoch:  3 | Loss:  13503.8 | Latent loss:  3108.58 | Reconstr. loss:  10395.2\n",
      "Epoch:  4 | Loss:  13885.2 | Latent loss:  3189.24 | Reconstr. loss:  10696.0\n",
      "Epoch:  5 | Loss:  13676.9 | Latent loss:  3204.29 | Reconstr. loss:  10472.6\n",
      "Epoch:  6 | Loss:  13682.8 | Latent loss:  3172.06 | Reconstr. loss:  10510.8\n",
      "Epoch:  7 | Loss:  13881.0 | Latent loss:  3262.33 | Reconstr. loss:  10618.7\n",
      "Epoch:  8 | Loss:  13282.7 | Latent loss:  3144.47 | Reconstr. loss:  10138.2\n",
      "Epoch:  9 | Loss:  13386.8 | Latent loss:  3222.39 | Reconstr. loss:  10164.4\n",
      "Epoch:  10 | Loss:  13063.7 | Latent loss:  3072.59 | Reconstr. loss:  9991.07\n",
      "Epoch:  11 | Loss:  12977.5 | Latent loss:  3070.71 | Reconstr. loss:  9906.76\n",
      "Epoch:  12 | Loss:  13028.7 | Latent loss:  3111.48 | Reconstr. loss:  9917.18\n",
      "Epoch:  13 | Loss:  13731.4 | Latent loss:  3263.84 | Reconstr. loss:  10467.6\n",
      "Epoch:  14 | Loss:  13056.8 | Latent loss:  3161.38 | Reconstr. loss:  9895.45\n",
      "Epoch:  15 | Loss:  12907.9 | Latent loss:  3159.68 | Reconstr. loss:  9748.23\n",
      "Epoch:  16 | Loss:  13041.6 | Latent loss:  3283.13 | Reconstr. loss:  9758.5\n",
      "Epoch:  17 | Loss:  13220.1 | Latent loss:  3158.94 | Reconstr. loss:  10061.2\n",
      "Epoch:  18 | Loss:  12878.2 | Latent loss:  3152.86 | Reconstr. loss:  9725.3\n",
      "Epoch:  19 | Loss:  13380.0 | Latent loss:  3203.26 | Reconstr. loss:  10176.7\n",
      "Epoch:  20 | Loss:  13098.0 | Latent loss:  3232.03 | Reconstr. loss:  9865.99\n",
      "Epoch:  21 | Loss:  12987.3 | Latent loss:  3065.79 | Reconstr. loss:  9921.51\n",
      "Epoch:  22 | Loss:  13269.1 | Latent loss:  3145.52 | Reconstr. loss:  10123.5\n",
      "Epoch:  23 | Loss:  12609.7 | Latent loss:  3089.27 | Reconstr. loss:  9520.41\n",
      "Epoch:  24 | Loss:  12656.9 | Latent loss:  3044.83 | Reconstr. loss:  9612.05\n",
      "Epoch:  25 | Loss:  13057.5 | Latent loss:  3257.4 | Reconstr. loss:  9800.05\n",
      "Epoch:  26 | Loss:  12989.9 | Latent loss:  3140.3 | Reconstr. loss:  9849.62\n",
      "Epoch:  27 | Loss:  12864.7 | Latent loss:  3125.28 | Reconstr. loss:  9739.41\n",
      "Epoch:  28 | Loss:  12761.1 | Latent loss:  3110.38 | Reconstr. loss:  9650.72\n",
      "Epoch:  29 | Loss:  12411.3 | Latent loss:  3011.11 | Reconstr. loss:  9400.15\n",
      "Epoch:  30 | Loss:  12952.3 | Latent loss:  3182.33 | Reconstr. loss:  9769.98\n",
      "Epoch:  31 | Loss:  12796.0 | Latent loss:  3091.21 | Reconstr. loss:  9704.81\n",
      "Epoch:  32 | Loss:  13124.5 | Latent loss:  3171.2 | Reconstr. loss:  9953.26\n",
      "Epoch:  33 | Loss:  12901.6 | Latent loss:  3056.14 | Reconstr. loss:  9845.47\n",
      "Epoch:  34 | Loss:  12703.1 | Latent loss:  3102.63 | Reconstr. loss:  9600.46\n",
      "Epoch:  35 | Loss:  13147.3 | Latent loss:  3192.66 | Reconstr. loss:  9954.6\n",
      "Epoch:  36 | Loss:  11984.7 | Latent loss:  2982.45 | Reconstr. loss:  9002.24\n",
      "Epoch:  37 | Loss:  12743.6 | Latent loss:  3129.0 | Reconstr. loss:  9614.62\n",
      "Epoch:  38 | Loss:  12818.5 | Latent loss:  3096.38 | Reconstr. loss:  9722.12\n",
      "Epoch:  39 | Loss:  12289.8 | Latent loss:  2969.19 | Reconstr. loss:  9320.63\n"
     ]
    }
   ],
   "source": [
    "# We create a string with the current data and time to get a unique\n",
    "# log-directory every time the model is run\n",
    "now = datetime.utcnow().strftime(\"%Y%m%d%H%M%S\")\n",
    "root_logdir = \"/tmp/tensorboard\"\n",
    "result_dir = \"{}/run-{}\".format(root_logdir, now)\n",
    "\n",
    "print(\"TensorBoard log directory: \", result_dir)\n",
    "\n",
    "# Add a writer for tensorboard\n",
    "writer = tf.summary.FileWriter(result_dir, model.sess.graph)\n",
    "saver = tf.train.Saver()\n",
    "\n",
    "tf.set_random_seed(model.random_seed)\n",
    "\n",
    "for epoch in range(n_epochs):\n",
    "    n_batches = int(mnist.train.num_examples/batch_size)\n",
    "    \n",
    "    for i in range(n_batches):\n",
    "\n",
    "        batch_x, _ = mnist.train.next_batch(batch_size)\n",
    "        # Fit the model on the current batch\n",
    "        model.batch_fit(batch_x)\n",
    "        # Compute the loss and add summaries\n",
    "        loss, logits, latent_l, reconstr_l, summary = model.sess.run([model.loss, model.logits, model.latent_loss,\n",
    "                                                                      model.reconstr_loss, model.summary_op],\n",
    "                                                                     {model.data: batch_x})\n",
    "        #loss, summary = model.sess.run([model.loss, model.summary_op], {model.data: batch_x})\n",
    "        writer.add_summary(summary, i)\n",
    "      \n",
    "    # Save the current model\n",
    "    save_path = saver.save(model.sess, \"/tmp/my_model.ckpt\")\n",
    "    print(\"Epoch: \", epoch, \"| Loss: \", loss, \"| Latent loss: \", latent_l, \"| Reconstr. loss: \", reconstr_l)\n",
    "\n",
    "# Save the final model \n",
    "save_path = saver.save(model.sess, \"/tmp/my_model_final.ckpt\")    \n",
    "writer.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Reconstruction\n",
    "\n",
    "To test the reconstruction capability of the model we visualize 5 original MNIST digits and their reconstructions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZYAAANYCAYAAADnhct7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkUz9fw1nkVRJrbM3AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb3c22bb940>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "x_sample, _ = mnist.test.next_batch(batch_size)\n",
    "x_reconstruct = model.reconstruct(x_sample)\n",
    "\n",
    "fig = plt.figure(figsize=(8,12))\n",
    "for i in range(5):\n",
    "    ax1 = fig.add_subplot(5,2,2*i+1)\n",
    "    ax1.set_title(\"Test input\")\n",
    "    ax1.imshow(x_sample[i].reshape(28,28), cmap='gray')\n",
    "    \n",
    "    ax2 = fig.add_subplot(5,2,2*i+2)\n",
    "    ax2.set_title(\"Reconstruction\")\n",
    "    ax2.imshow(x_reconstruct[i].reshape(28,28), cmap='gray')\n",
    "    \n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Generating new data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZYAAANYCAYAAADnhct7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nVKlhERGRVKlhERGRVP1/obQivnGrIwIAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb3c205e748>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "new_data = model.generate(batch_size)\n",
    "\n",
    "fig = plt.figure(figsize=(8,12))\n",
    "\n",
    "for i in range(10):\n",
    "    ax = fig.add_subplot(5,2,i+1)\n",
    "    ax.imshow(new_data[i].reshape(28,28), cmap='gray')\n",
    "    ax.set_title(\"Generated digit\")\n",
    "\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Restoring and testing a saved model\n",
    "\n",
    "To restore a saved model run all code up to the training part. When initializing the model be careful to use the same layer sizes as in the model you want to restore."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZYAAANYCAYAAADnhct7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nBElEQVTce7NIaepSTgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb3ad9aa160>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "saver = tf.train.Saver()\n",
    "with tf.Session() as sess:\n",
    "\n",
    "    # Restore model weights from previously saved model\n",
    "    saver.restore(model.sess, '/tmp/my_model_final.ckpt')\n",
    "    #saver.restore(model.sess, '/home/anna-lena/dllibrary/code/saved_models/500_500_100_final.ckpt')\n",
    "    \n",
    "    new_data = model.generate(batch_size)\n",
    "\n",
    "    fig = plt.figure(figsize=(8,12))\n",
    "\n",
    "    for i in range(10):\n",
    "        ax = fig.add_subplot(5,2,i+1)\n",
    "        ax.imshow(new_data[i].reshape(28,28), cmap='gray')\n",
    "        ax.set_title(\"Generated digit\")\n",
    "\n",
    "    plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [default]",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.4"
  },
  "toc": {
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}