5-Layer-MLP-InfiMNiST.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2016-04-20T21:52:04.548207",
     "start_time": "2016-04-20T21:52:04.544771"
    }
   },
   "source": [
    "## 5 Layer MLP w/InfiMNIST\n",
    "\n",
    "500 epochs of RMSProp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2016-05-04T22:35:05.559870",
     "start_time": "2016-05-04T22:35:03.489984"
    },
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2016-06-06T22:10:55.919425",
     "start_time": "2016-06-06T22:10:55.798303"
    },
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%run augmentmnist.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2016-05-04T22:35:05.569237",
     "start_time": "2016-05-04T22:35:05.566733"
    },
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "num_epochs = 500\n",
    "batch_size = 125\n",
    "\n",
    "learning_rate = 0.001\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2016-05-04T22:35:19.719645",
     "start_time": "2016-05-04T22:35:19.713990"
    },
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Network Parameters\n",
    "n_hidden_1 = 2500\n",
    "n_hidden_2 = 1500\n",
    "n_hidden_3 = 1000\n",
    "n_hidden_4 = 500\n",
    "n_input = 784 # MNIST data input (img shape: 28*28)\n",
    "n_classes = 10 # MNIST total classes (0-9 digits)\n",
    "\n",
    "std_0 = 1.0/np.sqrt(n_input)\n",
    "std_h1 = 1.0/np.sqrt(n_hidden_1)\n",
    "std_h2 = 1.0/np.sqrt(n_hidden_2)\n",
    "std_h3 = 1.0/np.sqrt(n_hidden_3)\n",
    "std_h4 = 1.0/np.sqrt(n_hidden_4)\n",
    "\n",
    "logfile = \"5-layer-mlp-infimnist\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2016-05-04T22:35:20.469254",
     "start_time": "2016-05-04T22:35:20.464080"
    },
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# tf Graph input\n",
    "x = tf.placeholder(\"float\", [None, n_input])\n",
    "y = tf.placeholder(\"float\", [None, n_classes])\n",
    "\n",
    "xt = tf.placeholder(\"float\", [None, n_input])\n",
    "yt = tf.placeholder(\"float\", [None, n_classes])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2016-05-04T22:35:21.195832",
     "start_time": "2016-05-04T22:35:21.191412"
    },
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def multilayer_perceptron(_X, _weights, _biases):\n",
    "    layer_1 = tf.nn.relu(tf.add(tf.matmul(_X, _weights['h1']), _biases['b1'])) \n",
    "    layer_2 = tf.nn.relu(tf.add(tf.matmul(layer_1, _weights['h2']), _biases['b2'])) \n",
    "    layer_3 = tf.nn.relu(tf.add(tf.matmul(layer_2, _weights['h3']), _biases['b3'])) \n",
    "    layer_4 = tf.nn.relu(tf.add(tf.matmul(layer_3, _weights['h4']), _biases['b4'])) \n",
    "    return tf.matmul(layer_4, weights['out']) + biases['out']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2016-05-04T22:35:21.913615",
     "start_time": "2016-05-04T22:35:21.855560"
    },
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Store layers weight & bias\n",
    "weights = {\n",
    "    'h1': tf.Variable(tf.random_normal([n_input, n_hidden_1], stddev=std_0)),\n",
    "    'h2': tf.Variable(tf.random_normal([n_hidden_1, n_hidden_2], stddev=std_h1)),\n",
    "    'h3': tf.Variable(tf.random_normal([n_hidden_2, n_hidden_3], stddev=std_h2)),\n",
    "    'h4': tf.Variable(tf.random_normal([n_hidden_3, n_hidden_4], stddev=std_h3)),\n",
    "    'out': tf.Variable(tf.random_normal([n_hidden_4, n_classes], stddev=std_h4))\n",
    "}\n",
    "biases = {\n",
    "    'b1': tf.Variable(tf.random_normal([n_hidden_1], stddev=0.1)),\n",
    "    'b2': tf.Variable(tf.random_normal([n_hidden_2], stddev=0.01)),\n",
    "    'b3': tf.Variable(tf.random_normal([n_hidden_3], stddev=0.01)),\n",
    "    'b4': tf.Variable(tf.random_normal([n_hidden_4], stddev=0.01)),\n",
    "    'out': tf.Variable(tf.random_normal([n_classes], stddev=0.001))\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2016-05-04T22:35:23.055468",
     "start_time": "2016-05-04T22:35:23.040509"
    },
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "mlp = multilayer_perceptron(x, weights, biases )\n",
    "mlp_test = multilayer_perceptron(xt, weights, biases )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2016-05-04T22:35:23.865189",
     "start_time": "2016-05-04T22:35:23.855758"
    },
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "cost =  tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(mlp, y)) \n",
    "gs = tf.get_variable(\"global_step\",[],trainable=False,initializer=tf.constant_initializer(0))\n",
    "lr = tf.constant(learning_rate) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2016-05-04T22:35:24.801765",
     "start_time": "2016-05-04T22:35:24.607145"
    },
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "train_op = tf.contrib.layers.optimize_loss(cost, global_step=gs, learning_rate=lr,optimizer=\"RMSProp\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2016-05-04T22:35:25.322532",
     "start_time": "2016-05-04T22:35:25.303706"
    },
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tf.Tensor 'ScalarSummary_2:0' shape=() dtype=string>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#with tf.name_scope(\"training accuracy\"):\n",
    "pred = tf.equal(tf.argmax(mlp, 1), tf.argmax(y, 1)) # Count correct predictions\n",
    "train_acc_op = tf.reduce_mean(tf.cast(pred, \"float\"))  # Cast boolean to float to average\n",
    "tf.scalar_summary(\"training accuracy\", train_acc_op)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2016-05-04T22:35:26.266848",
     "start_time": "2016-05-04T22:35:26.255677"
    },
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tf.Tensor 'ScalarSummary_3:0' shape=() dtype=string>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_pred = tf.equal(tf.argmax(mlp_test, 1), tf.argmax(yt, 1)) # Count correct predictions\n",
    "test_acc_op = tf.reduce_mean(tf.cast(test_pred, \"float\"))  # Cast boolean to float to average\n",
    "tf.scalar_summary(\"test 0 accuracy\", test_acc_op)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2016-05-04T22:35:28.052227",
     "start_time": "2016-05-04T22:35:27.823881"
    },
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[34m2-layer-mlp-alignmnist\u001b[m\u001b[m \u001b[34m2-layer-mlp-infimnist\u001b[m\u001b[m  \u001b[34m5-layer-mlp-alignmnist\u001b[m\u001b[m\r\n"
     ]
    }
   ],
   "source": [
    "!rm -rf ./logs/{logfile}\n",
    "!ls logs"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2016-05-04T22:35:29.913834",
     "start_time": "2016-05-04T22:35:29.911621"
    },
    "collapsed": true
   },
   "source": [
    "### Original MNIST Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "start_time": "2016-05-05T05:35:31.093Z"
    },
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Extracting ./train-images-idx3-ubyte.gz\n",
      "Extracting ./train-labels-idx1-ubyte.gz\n",
      "Extracting ./t10k-images-idx3-ubyte.gz\n",
      "Extracting ./t10k-labels-idx1-ubyte.gz\n"
     ]
    }
   ],
   "source": [
    "from tensorflow.examples.tutorials.mnist import input_data\n",
    "mnist = input_data.read_data_sets(\".\", one_hot=True)\n",
    "trX_0, trY_0 = mnist.train.images, mnist.train.labels\n",
    "teX_0, teY_0 = mnist.test.images, mnist.test.labels"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### InfiMNIST "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "infiminst = InfiMNIST()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "start_time": "2016-05-05T05:35:34.204Z"
    },
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "creating infimnist pat files 10000 - 69999\n",
      "updating 80000 - 139999\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "//anaconda/lib/python2.7/gzip.py:275: VisibleDeprecationWarning: converting an array with ndim > 0 to an index will result in an error in the future\n",
      "  chunk = self.extrabuf[offset: offset + size]\n",
      "<string>:59: VisibleDeprecationWarning: converting an array with ndim > 0 to an index will result in an error in the future\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(0, 0, 0.95796669, 0.95670003)\n",
      "creating infimnist pat files 710000 - 769999\n",
      "updating 780000 - 839999\n",
      "(10, 10, 0.98334998, 0.98049998)\n",
      "creating infimnist pat files 1410000 - 1469999\n",
      "updating 1480000 - 1539999\n",
      "(20, 20, 0.98963332, 0.98580003)\n",
      "creating infimnist pat files 2110000 - 2169999\n",
      "updating 2180000 - 2239999\n",
      "(30, 30, 0.98406667, 0.97549999)\n",
      "creating infimnist pat files 2810000 - 2869999\n",
      "updating 2880000 - 2939999\n",
      "(40, 40, 0.99360001, 0.98830003)\n",
      "creating infimnist pat files 3510000 - 3569999\n",
      "updating 3580000 - 3639999\n",
      "(50, 50, 0.98823333, 0.9813)\n",
      "creating infimnist pat files 4210000 - 4269999\n",
      "updating 4280000 - 4339999\n",
      "(60, 60, 0.98376667, 0.97820002)\n",
      "creating infimnist pat files 4910000 - 4969999\n",
      "updating 4980000 - 5039999\n",
      "(70, 70, 0.99488336, 0.9874)\n",
      "creating infimnist pat files 5610000 - 5669999\n",
      "updating 5680000 - 5739999\n",
      "(80, 80, 0.98431665, 0.97430003)\n",
      "creating infimnist pat files 6310000 - 6369999\n",
      "updating 6380000 - 6439999\n",
      "(90, 90, 0.99268335, 0.98750001)\n",
      "creating infimnist pat files 7010000 - 7069999\n",
      "updating 7080000 - 7139999\n",
      "(100, 100, 0.99216664, 0.9849)\n",
      "creating infimnist pat files 7710000 - 7769999\n",
      "updating 7780000 - 7839999\n",
      "(110, 110, 0.99258333, 0.98229998)\n",
      "creating infimnist pat files 8410000 - 8469999\n",
      "updating 8480000 - 8539999\n",
      "(120, 120, 0.98586667, 0.9781)\n",
      "creating infimnist pat files 9110000 - 9169999\n",
      "updating 9180000 - 9239999\n",
      "(130, 130, 0.97488332, 0.97000003)\n",
      "creating infimnist pat files 9810000 - 9869999\n",
      "updating 9880000 - 9939999\n",
      "(140, 140, 0.98906666, 0.98189998)\n",
      "creating infimnist pat files 10510000 - 10569999\n",
      "updating 10580000 - 10639999\n",
      "(150, 150, 0.9781, 0.97079998)\n",
      "creating infimnist pat files 11210000 - 11269999\n",
      "updating 11280000 - 11339999\n",
      "(160, 160, 0.97303331, 0.96439999)\n",
      "creating infimnist pat files 11910000 - 11969999\n",
      "updating 11980000 - 12039999\n",
      "(170, 170, 0.97001666, 0.96289998)\n",
      "creating infimnist pat files 12610000 - 12669999\n",
      "updating 12680000 - 12739999\n",
      "(180, 180, 0.90393335, 0.91070002)\n",
      "creating infimnist pat files 13310000 - 13369999\n",
      "updating 13380000 - 13439999\n",
      "(190, 190, 0.90428334, 0.91009998)\n",
      "creating infimnist pat files 14010000 - 14069999\n",
      "updating 14080000 - 14139999\n",
      "(200, 200, 0.83268332, 0.83660001)\n",
      "creating infimnist pat files 14710000 - 14769999\n",
      "updating 14780000 - 14839999\n",
      "(210, 210, 0.91461664, 0.91659999)\n",
      "creating infimnist pat files 15410000 - 15469999\n",
      "updating 15480000 - 15539999\n",
      "(220, 220, 0.92313331, 0.92150003)\n",
      "creating infimnist pat files 16110000 - 16169999\n",
      "updating 16180000 - 16239999\n",
      "(230, 230, 0.90691668, 0.91460001)\n",
      "creating infimnist pat files 16810000 - 16869999\n",
      "updating 16880000 - 16939999\n",
      "(240, 240, 0.86905003, 0.8836)\n",
      "creating infimnist pat files 17510000 - 17569999\n",
      "updating 17580000 - 17639999\n",
      "(250, 250, 0.89393336, 0.90799999)\n",
      "creating infimnist pat files 18210000 - 18269999\n",
      "updating 18280000 - 18339999\n",
      "(260, 260, 0.97143334, 0.9756)\n",
      "creating infimnist pat files 18910000 - 18969999\n",
      "updating 18980000 - 19039999\n",
      "(270, 270, 0.79661667, 0.8204)\n",
      "creating infimnist pat files 19610000 - 19669999\n",
      "updating 19680000 - 19739999\n",
      "(280, 280, 0.82286668, 0.84719998)\n",
      "creating infimnist pat files 20310000 - 20369999\n",
      "updating 20380000 - 20439999\n",
      "(290, 290, 0.88524997, 0.90469998)\n",
      "creating infimnist pat files 21010000 - 21069999\n",
      "updating 21080000 - 21139999\n",
      "(300, 300, 0.87124997, 0.88959998)\n",
      "creating infimnist pat files 21710000 - 21769999\n",
      "updating 21780000 - 21839999\n",
      "(310, 310, 0.87669998, 0.89029998)\n",
      "creating infimnist pat files 22410000 - 22469999\n",
      "updating 22480000 - 22539999\n",
      "(320, 320, 0.89416665, 0.91579998)\n",
      "creating infimnist pat files 23110000 - 23169999\n",
      "updating 23180000 - 23239999\n",
      "(330, 330, 0.89143336, 0.91109997)\n",
      "creating infimnist pat files 23810000 - 23869999\n",
      "updating 23880000 - 23939999\n",
      "(340, 340, 0.85863334, 0.8775)\n",
      "creating infimnist pat files 24510000 - 24569999\n",
      "updating 24580000 - 24639999\n",
      "(350, 350, 0.93846667, 0.94679999)\n",
      "creating infimnist pat files 25210000 - 25269999\n",
      "updating 25280000 - 25339999\n",
      "(360, 360, 0.90708333, 0.92079997)\n",
      "creating infimnist pat files 25910000 - 25969999\n",
      "updating 25980000 - 26039999\n",
      "(370, 370, 0.79788333, 0.8003)\n",
      "creating infimnist pat files 26610000 - 26669999\n",
      "updating 26680000 - 26739999\n",
      "(380, 380, 0.88611668, 0.88739997)\n",
      "creating infimnist pat files 27310000 - 27369999\n",
      "updating 27380000 - 27439999\n",
      "(390, 390, 0.81708336, 0.83039999)\n",
      "creating infimnist pat files 28010000 - 28069999\n",
      "updating 28080000 - 28139999\n",
      "(400, 400, 0.73680001, 0.75580001)\n",
      "creating infimnist pat files 28710000 - 28769999\n",
      "updating 28780000 - 28839999\n",
      "(410, 410, 0.78163332, 0.79290003)\n",
      "creating infimnist pat files 29410000 - 29469999\n",
      "updating 29480000 - 29539999\n",
      "(420, 420, 0.83845001, 0.84200001)\n",
      "creating infimnist pat files 30110000 - 30169999\n",
      "updating 30180000 - 30239999\n",
      "(430, 430, 0.84750003, 0.86440003)\n",
      "creating infimnist pat files 30810000 - 30869999\n",
      "updating 30880000 - 30939999\n",
      "(440, 440, 0.81484997, 0.81099999)\n",
      "creating infimnist pat files 31510000 - 31569999\n",
      "updating 31580000 - 31639999\n",
      "(450, 450, 0.85275, 0.8689)\n",
      "creating infimnist pat files 32210000 - 32269999\n",
      "updating 32280000 - 32339999\n",
      "(460, 460, 0.61098331, 0.62599999)\n",
      "creating infimnist pat files 32910000 - 32969999\n",
      "updating 32980000 - 33039999\n",
      "(470, 470, 0.57740003, 0.58579999)\n",
      "creating infimnist pat files 33610000 - 33669999\n",
      "updating 33680000 - 33739999\n",
      "(480, 480, 0.68045002, 0.68449998)\n",
      "creating infimnist pat files 34310000 - 34369999\n",
      "updating 34380000 - 34439999\n",
      "(490, 490, 0.82856667, 0.84579998)\n"
     ]
    }
   ],
   "source": [
    "test_accuracies = []\n",
    "train_accuracies = []\n",
    "with tf.Session() as sess:\n",
    "    # create a log writer. run 'tensorboard --logdir=./logs/{logfile}'\n",
    "    writer = tf.train.SummaryWriter(\"./logs/{0}\".format(logfile), sess.graph) # for 0.8\n",
    "    merged = tf.merge_all_summaries()\n",
    " \n",
    "    tf.initialize_all_variables().run()\n",
    "\n",
    "    i = -1\n",
    "    for epoch in range(num_epochs):\n",
    "        \n",
    "        i+=1\n",
    "        log=((i%10)==0)\n",
    "            \n",
    "        trX, trY = infiminst.next_epoch(log)\n",
    "        \n",
    "        for start, end in zip(range(0, len(trX), batch_size), range(batch_size, len(trX), batch_size)):\n",
    "            sess.run(train_op, feed_dict={x: trX[start:end], y: trY[start:end]})\n",
    "                 \n",
    "        summary, trn_acc, tst_acc = sess.run([merged, train_acc_op, test_acc_op], feed_dict={x: trX, y: trY, xt: teX_0, yt: teY_0})\n",
    "        writer.add_summary(summary, epoch)  \n",
    "        \n",
    "        if (log):\n",
    "            print(i,epoch, trn_acc, tst_acc)         \n",
    "        \n",
    "        train_accuracies.append(trn_acc)\n",
    "        test_accuracies.append(tst_acc)\n",
    "\n",
    "        writer.flush()\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import matplotlib\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x11d019190>]"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAEACAYAAAC9Gb03AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nfSAQCCw4QegDgUBgwfn/ATxKyA8rRQtuAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11d0191d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(train_accuracies)\n",
    "plt.plot(test_accuracies)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### the best error is"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.98909998"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.max(test_accuracies)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "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
}