Bidirectional LSTM.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAvwAAAEOCAIAAAB6iFRXAAAAA3NCSVQICAjb4U/gAAAAGXRFWHRT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\nCCGEcMJ/AT64UhaoJLx5AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Notebook created by: Nazmi Asri\n",
    "\n",
    "# Bidirectional LSTM: extension of LSTM that can improve model performance \n",
    "#     on sequence classification problems\n",
    "\n",
    "# In problems where all timesteps of the input sequence are available, \n",
    "# Bidirectional LSTMs train two instead of one LSTMs on the input sequence. \n",
    "# The first on the input sequence as-is and the second on a reversed copy of the input sequence. \n",
    "# This can provide additional context to the \n",
    "#     network and result in faster and even fuller learning on the problem.\n",
    "\n",
    "# 6 Parts:\n",
    "#     1) Bidirectional LSTMs\n",
    "#     2) Sequence Classification Problem\n",
    "#     3) LSTM for Sequence Classification\n",
    "#     4) Bidirectional LSTM for Sequence Classification\n",
    "#     5) Compare LSTM to Bidirectional LSTM\n",
    "#     6) Comparing Bidirectional LSTM merge modes\n",
    "\n",
    "# Source: http://machinelearningmastery.com/develop-bidirectional-lstm-sequence-classification-python-keras/\n",
    "\n",
    "from IPython.display import Image\n",
    "Image(\"RNN-bidirectional.png\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "# Import dependencies\n",
    "from random import random, seed\n",
    "from numpy import array\n",
    "from numpy import cumsum\n",
    "from pandas import DataFrame\n",
    "from matplotlib import pyplot\n",
    "from keras.models import Sequential\n",
    "from keras.layers import LSTM\n",
    "from keras.layers import Dense\n",
    "from keras.layers import TimeDistributed\n",
    "from keras.layers import Bidirectional"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TensorFlow version: v1.1.0\n",
      "Keras version: v2.0.4\n"
     ]
    }
   ],
   "source": [
    "# Check TensorFlow and Keras version, should be TF 1.1.0+ and Keras 2.0.4+\n",
    "import tensorflow\n",
    "import keras\n",
    "\n",
    "print(\"TensorFlow version: v\" + tensorflow.__version__)\n",
    "print(\"Keras version: v\" + keras.__version__)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# The idea of Bidirectional Recurrent Neural Network\n",
    "\n",
    "# Duplicating the first layer of recurrent neural network and now they have two \n",
    "# layers side-by-side, and providing the input sequence as-is as input to the first layer,\n",
    "# and providing the reverse input to the second layer\n",
    "\n",
    "# Bidirectional RNN Train the data using all available input from the past information and future information\n",
    "# RNN only capable of learning from the past information\n",
    "# The idea came from how can we understand on what we heard, on something that hasn't been said yet\n",
    "# It much more make sense with the light of future context\n",
    "\n",
    "# Much more significantly effective than unidirectional RNN "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Bidirectional LSTM in Keras\n",
    "\n",
    "# Bidirectional LSTM are supported in Keras using Bidirectional wrapper\n",
    "# it takes input as arguments, \n",
    "# and also the merge mode can be specify:\n",
    "#     1. sum: the output is added together\n",
    "#     2. mul: the output is multiplied together\n",
    "#     3. concat: the output is concatenated \n",
    "#     4. ave: avg output is taken (default)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0.8357651   0.43276707  0.76228008  0.00210605  0.44538719  0.72154003\n",
      "  0.22876222  0.9452707   0.90142746  0.03058998]\n",
      "[0 0 0 0 0 1 1 1 1 1]\n"
     ]
    }
   ],
   "source": [
    "# Sequence Classification Problem (without RNN)\n",
    "\n",
    "# Create a sequence classification instance\n",
    "def get_sequence(n_timesteps):\n",
    "\t# create a sequence of random numbers in [0,1]\n",
    "\tX = array([random() for _ in range(n_timesteps)])\n",
    "\t# calculate cut-off value to change class values\n",
    "\tlimit = n_timesteps/4.0\n",
    "\t# determine the class outcome for each item in cumulative sequence\n",
    "\ty = array([0 if x < limit else 1 for x in cumsum(X)])\n",
    "\treturn X, y\n",
    "\n",
    "# Print the results\n",
    "X, y = get_sequence(10)\n",
    "print(X)\n",
    "print(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1\n",
      "1s - loss: 0.6933 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6919 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6905 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6891 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6877 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6864 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6850 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6837 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6825 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6812 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6800 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6787 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6775 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6762 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6750 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6738 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6726 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6713 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6701 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6688 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6675 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6663 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6649 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6636 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6622 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6608 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6594 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6579 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6564 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6548 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6532 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6516 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6498 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6481 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6462 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6443 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6423 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6403 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6382 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6360 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6337 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6313 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6289 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6264 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6238 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6211 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6184 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6155 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6126 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6096 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6065 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6034 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6002 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5969 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5936 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5902 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5868 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5834 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5799 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5763 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5728 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5692 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5656 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5620 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5583 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5547 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5511 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5474 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5437 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5401 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5364 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5327 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5289 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5252 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5214 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5176 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5138 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5100 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5061 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5022 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4982 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4943 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4903 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4863 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4823 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4782 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4742 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4701 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4660 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4620 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4579 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4538 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4497 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4456 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4415 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4374 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4333 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4292 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4252 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4211 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4170 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4129 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4088 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4048 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4007 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3966 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3926 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3885 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3845 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3805 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3765 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3725 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3685 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3646 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3606 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3567 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3528 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3489 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3450 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3411 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3372 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3334 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3296 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3258 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3220 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3183 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3146 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3109 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3072 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3035 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2999 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2963 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2927 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2891 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2855 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2820 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2785 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2750 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2715 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2681 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2647 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2613 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2579 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2546 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2512 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2479 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2446 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2413 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2380 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2347 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2315 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2283 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2251 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2219 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2188 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2157 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2126 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2096 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2066 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2037 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2008 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1979 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1951 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1924 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1898 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1872 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1846 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1821 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1797 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1772 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1748 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1724 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1701 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1678 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1656 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1633 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1612 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1590 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1569 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1548 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1527 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1507 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1487 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1467 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1447 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1428 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1410 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1392 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1374 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1357 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1340 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1324 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1308 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1292 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1277 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1263 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1248 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1235 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1221 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1208 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1195 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1182 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1170 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1157 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1145 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1134 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1123 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1112 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1101 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1090 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1080 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1070 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1060 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1051 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1041 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1033 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1016 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1008 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1000 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0992 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0985 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0977 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0970 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0963 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0956 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0949 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0942 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0935 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0929 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0922 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0916 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0909 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0903 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0897 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0891 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0886 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0880 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0874 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0869 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0863 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0858 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0852 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0847 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0842 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0837 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0831 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0826 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0821 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0816 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0812 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0807 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0802 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0797 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0793 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0788 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0783 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0779 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0774 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0770 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0766 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0761 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0757 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0753 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0748 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0744 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0740 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0736 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0732 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0728 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0724 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0720 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0716 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0712 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0709 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0705 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0701 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0697 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0694 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0690 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0687 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0683 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0680 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0676 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0673 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0669 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0666 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0663 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0659 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0656 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0653 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0650 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0647 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0643 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0640 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0637 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0634 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0631 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0628 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0625 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0622 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0619 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0617 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0614 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0611 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0608 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0605 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0602 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0600 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0597 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0594 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0592 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0589 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0586 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0584 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0581 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0579 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0576 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0574 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0571 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0569 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0566 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0564 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0561 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0559 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0557 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0554 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0552 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0549 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0547 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0545 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0543 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0540 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0538 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0536 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0534 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0532 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0529 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0527 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0525 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0523 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0521 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0519 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0517 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0515 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0513 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0511 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0509 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0507 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0505 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0503 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0501 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0499 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0497 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0495 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0493 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0491 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0489 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0487 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0485 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0483 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0482 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0480 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0478 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0476 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0474 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0473 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0471 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0469 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0467 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0466 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0464 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0462 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0460 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0459 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0457 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0455 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0454 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0452 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0450 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0449 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0447 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0446 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0444 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0442 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0441 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0439 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0438 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0436 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0435 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0433 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0431 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0430 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0428 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0427 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0425 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0424 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0422 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0421 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0419 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0418 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0417 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0415 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0414 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0412 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0411 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0409 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0408 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0407 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0405 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0404 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0402 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0401 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0400 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0398 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0397 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0396 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0394 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0393 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0392 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0390 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0389 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0388 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0386 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0385 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0384 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0383 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0381 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0380 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0379 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0378 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0376 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0375 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0374 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0373 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0371 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0370 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0369 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0368 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0367 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0365 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0364 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0363 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0362 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0361 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0359 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0358 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0357 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0356 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0355 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0354 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0353 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0351 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0350 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0349 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0348 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0347 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0346 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0345 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0344 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0343 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0342 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0341 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0339 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0338 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0337 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0336 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0335 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0334 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0333 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0332 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0331 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0330 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0329 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0328 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0327 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0326 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0325 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0324 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0323 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0322 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0321 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0320 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0319 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0318 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0317 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0316 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0315 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0314 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0313 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0312 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0311 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0310 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0310 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0309 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0308 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0307 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0306 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0305 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0304 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0303 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0302 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0301 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0300 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0300 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0299 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0298 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0297 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0296 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0295 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0294 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0293 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0293 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0292 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0291 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0290 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0289 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0288 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0287 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0287 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0286 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0285 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0284 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0283 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0282 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0282 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0281 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0280 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0279 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0278 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0278 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0277 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0276 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0275 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0274 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0274 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0273 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0272 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0271 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0271 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0270 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0269 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0268 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0267 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0267 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0266 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0265 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0264 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0264 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0263 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0262 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0261 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0261 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0260 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0259 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0259 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0258 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0257 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0256 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0256 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0255 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0254 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0253 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0253 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0252 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0251 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0251 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0250 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0249 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0249 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0248 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0247 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0247 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0246 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0245 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0244 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0244 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0243 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0242 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0242 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0241 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0240 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0240 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0239 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0238 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0238 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0237 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0237 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0236 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0235 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0235 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0234 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0233 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0233 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0232 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0231 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0231 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0230 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0230 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0229 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0228 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0228 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0227 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0226 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0226 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0225 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0225 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0224 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0223 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0223 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0222 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0222 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0221 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0220 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0220 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0219 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0219 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0218 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0217 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0217 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0216 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0216 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0215 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0215 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0214 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0213 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0213 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0212 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0212 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0211 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0211 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0210 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0210 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0209 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0208 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0208 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0207 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0207 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0206 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0206 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0205 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0205 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0204 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0204 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0203 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0203 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0202 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0201 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0201 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0200 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0200 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0199 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0199 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0198 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0198 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0197 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0197 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0196 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0196 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0195 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0195 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0194 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0194 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0193 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0193 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0192 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0192 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0191 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0191 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0190 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0190 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0189 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0189 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0188 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0188 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0187 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0187 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0186 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0186 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0185 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0185 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0185 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0184 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0184 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0183 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0183 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0182 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0182 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0181 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0181 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0180 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0180 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0179 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0179 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0178 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0178 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0178 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0177 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0177 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0176 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0176 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0175 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0175 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0174 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0174 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0174 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0173 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0173 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0172 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0172 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0171 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0171 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0171 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0170 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0170 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0169 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0169 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0168 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0168 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0168 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0167 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0167 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0166 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0166 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0165 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0165 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0165 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0164 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0164 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0163 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0163 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0163 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0162 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0162 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0161 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0161 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0161 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0160 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0160 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0159 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0159 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0159 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0158 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0158 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0157 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0157 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0157 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0156 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0156 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0155 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0155 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0155 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0154 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0154 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0154 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0153 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0153 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0152 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0152 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0152 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0151 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0151 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0151 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0150 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0150 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0149 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0149 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0149 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0148 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0148 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0148 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0147 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0147 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0147 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0146 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0146 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0145 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0145 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0145 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0144 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0144 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0144 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0143 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0143 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0143 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0142 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0142 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0142 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0141 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0141 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0141 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0140 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0140 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0140 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0139 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0139 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0139 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0138 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0138 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0138 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0137 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0137 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0137 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0136 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0136 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0136 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0135 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0135 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0135 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0134 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0134 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0134 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0133 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0133 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0133 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0132 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0132 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0132 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0131 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0131 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0131 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0130 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0130 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0130 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0129 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0129 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0129 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0129 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0128 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0128 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0128 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0127 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0127 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0127 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0126 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0126 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0126 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0126 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0125 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0125 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0125 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0124 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0124 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0124 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0123 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0123 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0123 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0123 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0122 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0122 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0122 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0121 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0121 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0121 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0121 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0120 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0120 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0120 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0119 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0119 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0119 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0119 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0118 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0118 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0118 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0118 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0117 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0117 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0117 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0116 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0116 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0116 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0116 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0115 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0115 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0115 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0115 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0114 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0114 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0114 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0113 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0113 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0113 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0113 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0112 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0112 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0112 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0112 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0111 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0111 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0111 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0111 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0110 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0110 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0110 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0110 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0109 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0109 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0109 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0109 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0108 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0108 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0108 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0108 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0107 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0107 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0107 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0107 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0106 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0106 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0106 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0106 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0105 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0105 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0105 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0105 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0104 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0104 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0104 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0104 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0103 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0103 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0103 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0103 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0103 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0102 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0102 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0102 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0102 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0101 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0101 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0101 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0101 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0100 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0100 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0100 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0100 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0100 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0099 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0099 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0099 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0099 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0098 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0098 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0098 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0098 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0098 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0097 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0097 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0097 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0097 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0096 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0096 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0096 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0096 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0096 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0095 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0095 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0095 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0095 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0095 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0094 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0094 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0094 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0094 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0093 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0093 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0093 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0093 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0093 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0092 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0092 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0092 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0092 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0092 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0091 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0091 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0091 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0091 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0091 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0090 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0090 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0090 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0090 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0090 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0089 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0089 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0089 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0089 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0089 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0088 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0088 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0088 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0088 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0088 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0087 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0087 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0087 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0087 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0087 - acc: 1.0000\n",
      "Expected: [0] Predicted [0]\n",
      "Expected: [0] Predicted [0]\n",
      "Expected: [0] Predicted [0]\n",
      "Expected: [0] Predicted [0]\n",
      "Expected: [0] Predicted [0]\n",
      "Expected: [1] Predicted [1]\n",
      "Expected: [1] Predicted [1]\n",
      "Expected: [1] Predicted [1]\n",
      "Expected: [1] Predicted [1]\n",
      "Expected: [1] Predicted [1]\n"
     ]
    }
   ],
   "source": [
    "# LSTM for Sequence Classification Problem\n",
    "\n",
    "# Create a sequence classification instance\n",
    "def get_sequence(n_timesteps):\n",
    "\t# Define seed for consistent random value\n",
    "\tseed(1)\n",
    "\t# create a sequence of random numbers in [0,1]\n",
    "\tX = array([random() for _ in range(n_timesteps)])\n",
    "\t# calculate cut-off value to change class values\n",
    "\tlimit = n_timesteps/4.0\n",
    "\t# determine the class outcome for each item in cumulative sequence\n",
    "\ty = array([0 if x < limit else 1 for x in cumsum(X)])\n",
    "\t# reshape input and output data to be suitable for LSTMs\n",
    "\tX = X.reshape(1, n_timesteps, 1)\n",
    "\ty = y.reshape(1, n_timesteps, 1)\n",
    "\treturn X, y\n",
    "\n",
    "# For now we have 10 timesteps\n",
    "n_timesteps = 10\n",
    "\n",
    "# Reshape input and output data to be suitable for LSTMs [samples, timesteps, features]\n",
    "X,y = get_sequence(n_timesteps)\n",
    "X = X.reshape(1, n_timesteps, 1)\n",
    "y = y.reshape(1, n_timesteps, 1)\n",
    "\n",
    "# Define LSTM\n",
    "# TimeDistributed wrapper is used so one value per timestamps can be predicted,\n",
    "# given full sequence is provided as the input. \n",
    "# This required the LSTM hidden layer to return a sequence of values (one per timestamp)\n",
    "# rather than single value for the whole input sequence\n",
    "# Because this is a binary classification problem, we use binary cross entropy \n",
    "model = Sequential()\n",
    "model.add(LSTM(20, input_shape=(n_timesteps, 1), return_sequences=True))\n",
    "model.add(TimeDistributed(Dense(1, activation='sigmoid')))\n",
    "model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['acc'])\n",
    "\n",
    "# Train LSTM\n",
    "for epoch in range(1000):\n",
    "\t# generate new random sequence\n",
    "    # we don't want the model just memorize single sequence\n",
    "\tX,y = get_sequence(n_timesteps)\n",
    "\t# fit model for one epoch on this sequence\n",
    "\tmodel.fit(X, y, epochs=1, batch_size=1, verbose=2)\n",
    "    \n",
    "# Evaluate LSTM\n",
    "yhat = model.predict_classes(X, verbose=0)\n",
    "for i in range(n_timesteps):\n",
    "\tprint('Expected:', y[0, i], 'Predicted', yhat[0, i])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1\n",
      "2s - loss: 0.6886 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6864 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6842 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6820 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6799 - acc: 0.5000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6777 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6755 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6732 - acc: 0.6000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6710 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6688 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6665 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6642 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6618 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6594 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6570 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6545 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6519 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6493 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6466 - acc: 0.7000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6438 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6409 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6380 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6349 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6317 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6283 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6249 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6212 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6174 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6135 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6093 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6050 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.6004 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5956 - acc: 0.8000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5906 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5853 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5797 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5738 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5677 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5612 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5544 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5472 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5397 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5318 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5236 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5150 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.5060 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4966 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4868 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4767 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4662 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4554 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4443 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4329 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4213 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.4094 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3974 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3852 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3729 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3605 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3481 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3358 - acc: 0.9000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3236 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.3116 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2997 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2881 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2768 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2658 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2551 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2449 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2350 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2255 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2165 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.2079 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1997 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1919 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1845 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1775 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1709 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1646 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1587 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1531 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1478 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1428 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1380 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1335 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1292 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1251 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1212 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1175 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1140 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1107 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1075 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1044 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.1015 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0988 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0961 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0936 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0912 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0889 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0867 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0846 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0826 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0806 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0788 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0770 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0754 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0738 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0723 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0708 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0694 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0681 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0668 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0655 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0643 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0631 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0620 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0609 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0599 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0589 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0579 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0570 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0560 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0552 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0543 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0535 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0527 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0519 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0512 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0505 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0498 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0491 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0484 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0478 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0471 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0465 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0459 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0453 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0447 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0442 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0437 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0432 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0427 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0422 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0417 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0413 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0408 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0404 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0400 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0395 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0391 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0387 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0383 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0379 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0375 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0371 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0367 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0363 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0359 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0356 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0352 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0348 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0345 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0341 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0338 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0334 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0331 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0328 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0324 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0321 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0318 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0315 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0312 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0309 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0307 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0304 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0301 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0298 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0295 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0293 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0290 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0287 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0285 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0282 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0280 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0277 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0275 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0273 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0270 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0268 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0266 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0264 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0261 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0259 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0257 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0255 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0253 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0251 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0249 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0247 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0245 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0243 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0242 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0240 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0238 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0236 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0234 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0233 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0231 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0229 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0228 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0226 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0225 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0223 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0222 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0220 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0218 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0217 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0216 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0214 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0213 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0211 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0210 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0208 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0207 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0206 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0204 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0203 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0202 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0200 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0199 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0198 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0196 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0195 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0194 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0193 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0192 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0190 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0189 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0188 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0187 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0186 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0184 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0183 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0182 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0181 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0180 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0179 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0178 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0177 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0176 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0175 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0174 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0173 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0172 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0171 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0170 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0169 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0168 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0167 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0166 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0165 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0164 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0163 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0162 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0161 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0161 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0160 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0159 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0158 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0157 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0156 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0155 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0154 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0154 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0153 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0152 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0151 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0150 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0150 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0149 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0148 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0147 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0147 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0146 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0145 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0144 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0143 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0143 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0142 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0141 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0141 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0140 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0139 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0138 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0138 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0137 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0136 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0136 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0135 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0134 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0134 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0133 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0132 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0132 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0131 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0130 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0130 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0129 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0129 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0128 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0127 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0127 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0126 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0126 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0125 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0124 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0124 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0123 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0123 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0122 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0122 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0121 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0120 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0120 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0119 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0119 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0118 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0118 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0117 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0117 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0116 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0116 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0115 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0115 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0114 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0114 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0113 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0113 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0112 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0112 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0111 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0111 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0110 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0110 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0109 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0109 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0108 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0108 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0107 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0107 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0106 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0106 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0105 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0105 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0105 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0104 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0104 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0103 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0103 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0102 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0102 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0102 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0101 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0101 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0100 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0100 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0100 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0099 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0099 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0098 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0098 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0098 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0097 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0097 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0096 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0096 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0096 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0095 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0095 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0095 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0094 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0094 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0093 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0093 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0093 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0092 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0092 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0092 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0091 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0091 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0091 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0090 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0090 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0090 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0089 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0089 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0089 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0088 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0088 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0088 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0087 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0087 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0087 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0086 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0086 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0086 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0085 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0085 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0085 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0084 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0084 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0084 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0083 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0083 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0083 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0083 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0082 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0082 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0082 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0081 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0081 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0081 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0080 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0080 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0080 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0080 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0079 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0079 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0079 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0078 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0078 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0078 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0078 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0077 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0077 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0077 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0077 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0076 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0076 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0076 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0076 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0075 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0075 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0075 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0074 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0074 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0074 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0074 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0073 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0073 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0073 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0073 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0072 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0072 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0072 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0072 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0072 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0071 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0071 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0071 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0071 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0070 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0070 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0070 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0070 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0069 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0069 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0069 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0069 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0068 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0068 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0068 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0068 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0068 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0067 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0067 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0067 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0067 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0066 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0066 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0066 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0066 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0066 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0065 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0065 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0065 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0065 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0065 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0064 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0064 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0064 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0064 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0064 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0063 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0063 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0063 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0063 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0063 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0062 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0062 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0062 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0062 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0062 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0061 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0061 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0061 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0061 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0061 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0060 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0060 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0060 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0060 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0060 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0060 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0059 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0059 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0059 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0059 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0059 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0059 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0058 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0058 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0058 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0058 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0058 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0057 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0057 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0057 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0057 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0057 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0057 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0056 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0056 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0056 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0056 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0056 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0056 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0055 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0055 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0055 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0055 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0055 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0055 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0055 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0054 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0054 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0054 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0054 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0054 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0054 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0053 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0053 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0053 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0053 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0053 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0053 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0053 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0052 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0052 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0052 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0052 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0052 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0052 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0052 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0051 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0051 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0051 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0051 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0051 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0051 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0050 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0050 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0050 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0050 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0050 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0050 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0050 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0050 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0049 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0049 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0049 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0049 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0049 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0049 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0049 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0048 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0048 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0048 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0048 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0048 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0048 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0048 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0048 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0047 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0047 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0047 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0047 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0047 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0047 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0047 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0047 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0046 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0046 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0046 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0046 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0046 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0046 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0046 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0046 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0045 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0045 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0045 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0045 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0045 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0045 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0045 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0045 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0044 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0044 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0044 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0044 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0044 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0044 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0044 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0044 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0044 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0043 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0043 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0043 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0043 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0043 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0043 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0043 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0043 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0043 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0042 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0042 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0042 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0042 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0042 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0042 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0042 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0042 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0042 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0042 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0041 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0041 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0041 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0041 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0041 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0041 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0041 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0041 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0041 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0041 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0040 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0040 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0040 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0040 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0040 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0040 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0040 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0040 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0040 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0040 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0039 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0039 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0039 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0039 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0039 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0039 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0039 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0039 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0039 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0039 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0039 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0038 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0038 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0038 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0038 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0038 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0038 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0038 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0038 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0038 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0038 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0038 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0037 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0037 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0037 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0037 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0037 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0037 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0037 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0037 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0037 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0037 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0037 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0036 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0036 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0036 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0036 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0036 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0036 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0036 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0036 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0036 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0036 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0036 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0036 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0035 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0035 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0035 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0035 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0035 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0035 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0035 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0035 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0035 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0035 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0035 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0035 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0035 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0034 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0034 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0034 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0034 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0034 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0034 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0034 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0034 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0034 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0034 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0034 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0034 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0034 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0034 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0033 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0033 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0033 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0033 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0033 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0033 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0033 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0033 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0033 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0033 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0033 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0033 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0033 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0033 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0032 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0032 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0032 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0032 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0032 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0032 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0032 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0032 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0032 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0032 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0032 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0032 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0032 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0032 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0031 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0031 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0031 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0031 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0031 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0031 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0031 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0031 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0031 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0031 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0031 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0031 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0031 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0031 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0031 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0030 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0030 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0030 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0030 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0030 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0030 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0030 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0030 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0030 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0030 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0030 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0030 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0030 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0030 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0030 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0030 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0029 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0029 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0029 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0029 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0029 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0029 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0029 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0029 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0029 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0029 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0029 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0029 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0029 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0029 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0029 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0029 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0029 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0020 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0020 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0020 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0020 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0020 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0020 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0020 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0020 - acc: 1.0000\n",
      "Epoch 1/1\n",
      "0s - loss: 0.0020 - acc: 1.0000\n",
      "Expected: [0] Predicted [0]\n",
      "Expected: [0] Predicted [0]\n",
      "Expected: [0] Predicted [0]\n",
      "Expected: [0] Predicted [0]\n",
      "Expected: [0] Predicted [0]\n",
      "Expected: [1] Predicted [1]\n",
      "Expected: [1] Predicted [1]\n",
      "Expected: [1] Predicted [1]\n",
      "Expected: [1] Predicted [1]\n",
      "Expected: [1] Predicted [1]\n"
     ]
    }
   ],
   "source": [
    "# Bidirectional LSTM for Sequence Classification Problem\n",
    "\n",
    "# Reinitialize X and y\n",
    "X,y = get_sequence(n_timesteps)\n",
    "X = X.reshape(1, n_timesteps, 1)\n",
    "y = y.reshape(1, n_timesteps, 1)\n",
    "\n",
    "# Define Bidirectional LSTM\n",
    "# By adding Bidirectional wrapper, this will create 2 copies of the LSTM layer, one fit for the \n",
    "# input sequence as-is, and another one on a reversed copy of input sequence\n",
    "\n",
    "# This means that instead of TimeDistributed received 10 timesteps of 20 input (because of bidirectional)\n",
    "# it will now received 10 timesteps of 40 inputs (20 units + 20 units) outputs \n",
    "model_bLSTM = Sequential()\n",
    "model_bLSTM.add(Bidirectional(LSTM(n_timesteps*2, return_sequences=True), input_shape=(n_timesteps, 1)))\n",
    "model_bLSTM.add(TimeDistributed(Dense(1, activation='sigmoid')))\n",
    "model_bLSTM.compile(loss='binary_crossentropy', optimizer='adam', metrics=['acc'])\n",
    "\n",
    "# Train Bidirectional LSTM\n",
    "for epoch in range(1000):\n",
    "\t# generate new random sequence\n",
    "    # we don't want the model just memorize single sequence\n",
    "\tX,y = get_sequence(n_timesteps)\n",
    "\t# fit model for one epoch on this sequence\n",
    "\tmodel_bLSTM.fit(X, y, epochs=1, batch_size=1, verbose=2)\n",
    "\n",
    "# Evaluate LSTM\n",
    "yhat = model_bLSTM.predict_classes(X, verbose=0)\n",
    "for i in range(n_timesteps):\n",
    "\tprint('Expected:', y[0, i], 'Predicted', yhat[0, i])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "/* Put everything inside the global mpl namespace */\n",
       "window.mpl = {};\n",
       "\n",
       "\n",
       "mpl.get_websocket_type = function() {\n",
       "    if (typeof(WebSocket) !== 'undefined') {\n",
       "        return WebSocket;\n",
       "    } else if (typeof(MozWebSocket) !== 'undefined') {\n",
       "        return MozWebSocket;\n",
       "    } else {\n",
       "        alert('Your browser does not have WebSocket support.' +\n",
       "              'Please try Chrome, Safari or Firefox ≥ 6. ' +\n",
       "              'Firefox 4 and 5 are also supported but you ' +\n",
       "              'have to enable WebSockets in about:config.');\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure = function(figure_id, websocket, ondownload, parent_element) {\n",
       "    this.id = figure_id;\n",
       "\n",
       "    this.ws = websocket;\n",
       "\n",
       "    this.supports_binary = (this.ws.binaryType != undefined);\n",
       "\n",
       "    if (!this.supports_binary) {\n",
       "        var warnings = document.getElementById(\"mpl-warnings\");\n",
       "        if (warnings) {\n",
       "            warnings.style.display = 'block';\n",
       "            warnings.textContent = (\n",
       "                \"This browser does not support binary websocket messages. \" +\n",
       "                    \"Performance may be slow.\");\n",
       "        }\n",
       "    }\n",
       "\n",
       "    this.imageObj = new Image();\n",
       "\n",
       "    this.context = undefined;\n",
       "    this.message = undefined;\n",
       "    this.canvas = undefined;\n",
       "    this.rubberband_canvas = undefined;\n",
       "    this.rubberband_context = undefined;\n",
       "    this.format_dropdown = undefined;\n",
       "\n",
       "    this.image_mode = 'full';\n",
       "\n",
       "    this.root = $('<div/>');\n",
       "    this._root_extra_style(this.root)\n",
       "    this.root.attr('style', 'display: inline-block');\n",
       "\n",
       "    $(parent_element).append(this.root);\n",
       "\n",
       "    this._init_header(this);\n",
       "    this._init_canvas(this);\n",
       "    this._init_toolbar(this);\n",
       "\n",
       "    var fig = this;\n",
       "\n",
       "    this.waiting = false;\n",
       "\n",
       "    this.ws.onopen =  function () {\n",
       "            fig.send_message(\"supports_binary\", {value: fig.supports_binary});\n",
       "            fig.send_message(\"send_image_mode\", {});\n",
       "            if (mpl.ratio != 1) {\n",
       "                fig.send_message(\"set_dpi_ratio\", {'dpi_ratio': mpl.ratio});\n",
       "            }\n",
       "            fig.send_message(\"refresh\", {});\n",
       "        }\n",
       "\n",
       "    this.imageObj.onload = function() {\n",
       "            if (fig.image_mode == 'full') {\n",
       "                // Full images could contain transparency (where diff images\n",
       "                // almost always do), so we need to clear the canvas so that\n",
       "                // there is no ghosting.\n",
       "                fig.context.clearRect(0, 0, fig.canvas.width, fig.canvas.height);\n",
       "            }\n",
       "            fig.context.drawImage(fig.imageObj, 0, 0);\n",
       "        };\n",
       "\n",
       "    this.imageObj.onunload = function() {\n",
       "        this.ws.close();\n",
       "    }\n",
       "\n",
       "    this.ws.onmessage = this._make_on_message_function(this);\n",
       "\n",
       "    this.ondownload = ondownload;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_header = function() {\n",
       "    var titlebar = $(\n",
       "        '<div class=\"ui-dialog-titlebar ui-widget-header ui-corner-all ' +\n",
       "        'ui-helper-clearfix\"/>');\n",
       "    var titletext = $(\n",
       "        '<div class=\"ui-dialog-title\" style=\"width: 100%; ' +\n",
       "        'text-align: center; padding: 3px;\"/>');\n",
       "    titlebar.append(titletext)\n",
       "    this.root.append(titlebar);\n",
       "    this.header = titletext[0];\n",
       "}\n",
       "\n",
       "\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_canvas = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var canvas_div = $('<div/>');\n",
       "\n",
       "    canvas_div.attr('style', 'position: relative; clear: both; outline: 0');\n",
       "\n",
       "    function canvas_keyboard_event(event) {\n",
       "        return fig.key_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    canvas_div.keydown('key_press', canvas_keyboard_event);\n",
       "    canvas_div.keyup('key_release', canvas_keyboard_event);\n",
       "    this.canvas_div = canvas_div\n",
       "    this._canvas_extra_style(canvas_div)\n",
       "    this.root.append(canvas_div);\n",
       "\n",
       "    var canvas = $('<canvas/>');\n",
       "    canvas.addClass('mpl-canvas');\n",
       "    canvas.attr('style', \"left: 0; top: 0; z-index: 0; outline: 0\")\n",
       "\n",
       "    this.canvas = canvas[0];\n",
       "    this.context = canvas[0].getContext(\"2d\");\n",
       "\n",
       "    var backingStore = this.context.backingStorePixelRatio ||\n",
       "\tthis.context.webkitBackingStorePixelRatio ||\n",
       "\tthis.context.mozBackingStorePixelRatio ||\n",
       "\tthis.context.msBackingStorePixelRatio ||\n",
       "\tthis.context.oBackingStorePixelRatio ||\n",
       "\tthis.context.backingStorePixelRatio || 1;\n",
       "\n",
       "    mpl.ratio = (window.devicePixelRatio || 1) / backingStore;\n",
       "\n",
       "    var rubberband = $('<canvas/>');\n",
       "    rubberband.attr('style', \"position: absolute; left: 0; top: 0; z-index: 1;\")\n",
       "\n",
       "    var pass_mouse_events = true;\n",
       "\n",
       "    canvas_div.resizable({\n",
       "        start: function(event, ui) {\n",
       "            pass_mouse_events = false;\n",
       "        },\n",
       "        resize: function(event, ui) {\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "        stop: function(event, ui) {\n",
       "            pass_mouse_events = true;\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "    });\n",
       "\n",
       "    function mouse_event_fn(event) {\n",
       "        if (pass_mouse_events)\n",
       "            return fig.mouse_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    rubberband.mousedown('button_press', mouse_event_fn);\n",
       "    rubberband.mouseup('button_release', mouse_event_fn);\n",
       "    // Throttle sequential mouse events to 1 every 20ms.\n",
       "    rubberband.mousemove('motion_notify', mouse_event_fn);\n",
       "\n",
       "    rubberband.mouseenter('figure_enter', mouse_event_fn);\n",
       "    rubberband.mouseleave('figure_leave', mouse_event_fn);\n",
       "\n",
       "    canvas_div.on(\"wheel\", function (event) {\n",
       "        event = event.originalEvent;\n",
       "        event['data'] = 'scroll'\n",
       "        if (event.deltaY < 0) {\n",
       "            event.step = 1;\n",
       "        } else {\n",
       "            event.step = -1;\n",
       "        }\n",
       "        mouse_event_fn(event);\n",
       "    });\n",
       "\n",
       "    canvas_div.append(canvas);\n",
       "    canvas_div.append(rubberband);\n",
       "\n",
       "    this.rubberband = rubberband;\n",
       "    this.rubberband_canvas = rubberband[0];\n",
       "    this.rubberband_context = rubberband[0].getContext(\"2d\");\n",
       "    this.rubberband_context.strokeStyle = \"#000000\";\n",
       "\n",
       "    this._resize_canvas = function(width, height) {\n",
       "        // Keep the size of the canvas, canvas container, and rubber band\n",
       "        // canvas in synch.\n",
       "        canvas_div.css('width', width)\n",
       "        canvas_div.css('height', height)\n",
       "\n",
       "        canvas.attr('width', width * mpl.ratio);\n",
       "        canvas.attr('height', height * mpl.ratio);\n",
       "        canvas.attr('style', 'width: ' + width + 'px; height: ' + height + 'px;');\n",
       "\n",
       "        rubberband.attr('width', width);\n",
       "        rubberband.attr('height', height);\n",
       "    }\n",
       "\n",
       "    // Set the figure to an initial 600x600px, this will subsequently be updated\n",
       "    // upon first draw.\n",
       "    this._resize_canvas(600, 600);\n",
       "\n",
       "    // Disable right mouse context menu.\n",
       "    $(this.rubberband_canvas).bind(\"contextmenu\",function(e){\n",
       "        return false;\n",
       "    });\n",
       "\n",
       "    function set_focus () {\n",
       "        canvas.focus();\n",
       "        canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    window.setTimeout(set_focus, 100);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>')\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items) {\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) {\n",
       "            // put a spacer in here.\n",
       "            continue;\n",
       "        }\n",
       "        var button = $('<button/>');\n",
       "        button.addClass('ui-button ui-widget ui-state-default ui-corner-all ' +\n",
       "                        'ui-button-icon-only');\n",
       "        button.attr('role', 'button');\n",
       "        button.attr('aria-disabled', 'false');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "\n",
       "        var icon_img = $('<span/>');\n",
       "        icon_img.addClass('ui-button-icon-primary ui-icon');\n",
       "        icon_img.addClass(image);\n",
       "        icon_img.addClass('ui-corner-all');\n",
       "\n",
       "        var tooltip_span = $('<span/>');\n",
       "        tooltip_span.addClass('ui-button-text');\n",
       "        tooltip_span.html(tooltip);\n",
       "\n",
       "        button.append(icon_img);\n",
       "        button.append(tooltip_span);\n",
       "\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    var fmt_picker_span = $('<span/>');\n",
       "\n",
       "    var fmt_picker = $('<select/>');\n",
       "    fmt_picker.addClass('mpl-toolbar-option ui-widget ui-widget-content');\n",
       "    fmt_picker_span.append(fmt_picker);\n",
       "    nav_element.append(fmt_picker_span);\n",
       "    this.format_dropdown = fmt_picker[0];\n",
       "\n",
       "    for (var ind in mpl.extensions) {\n",
       "        var fmt = mpl.extensions[ind];\n",
       "        var option = $(\n",
       "            '<option/>', {selected: fmt === mpl.default_extension}).html(fmt);\n",
       "        fmt_picker.append(option)\n",
       "    }\n",
       "\n",
       "    // Add hover states to the ui-buttons\n",
       "    $( \".ui-button\" ).hover(\n",
       "        function() { $(this).addClass(\"ui-state-hover\");},\n",
       "        function() { $(this).removeClass(\"ui-state-hover\");}\n",
       "    );\n",
       "\n",
       "    var status_bar = $('<span class=\"mpl-message\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.request_resize = function(x_pixels, y_pixels) {\n",
       "    // Request matplotlib to resize the figure. Matplotlib will then trigger a resize in the client,\n",
       "    // which will in turn request a refresh of the image.\n",
       "    this.send_message('resize', {'width': x_pixels, 'height': y_pixels});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_message = function(type, properties) {\n",
       "    properties['type'] = type;\n",
       "    properties['figure_id'] = this.id;\n",
       "    this.ws.send(JSON.stringify(properties));\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_draw_message = function() {\n",
       "    if (!this.waiting) {\n",
       "        this.waiting = true;\n",
       "        this.ws.send(JSON.stringify({type: \"draw\", figure_id: this.id}));\n",
       "    }\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    var format_dropdown = fig.format_dropdown;\n",
       "    var format = format_dropdown.options[format_dropdown.selectedIndex].value;\n",
       "    fig.ondownload(fig, format);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_resize = function(fig, msg) {\n",
       "    var size = msg['size'];\n",
       "    if (size[0] != fig.canvas.width || size[1] != fig.canvas.height) {\n",
       "        fig._resize_canvas(size[0], size[1]);\n",
       "        fig.send_message(\"refresh\", {});\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_rubberband = function(fig, msg) {\n",
       "    var x0 = msg['x0'] / mpl.ratio;\n",
       "    var y0 = (fig.canvas.height - msg['y0']) / mpl.ratio;\n",
       "    var x1 = msg['x1'] / mpl.ratio;\n",
       "    var y1 = (fig.canvas.height - msg['y1']) / mpl.ratio;\n",
       "    x0 = Math.floor(x0) + 0.5;\n",
       "    y0 = Math.floor(y0) + 0.5;\n",
       "    x1 = Math.floor(x1) + 0.5;\n",
       "    y1 = Math.floor(y1) + 0.5;\n",
       "    var min_x = Math.min(x0, x1);\n",
       "    var min_y = Math.min(y0, y1);\n",
       "    var width = Math.abs(x1 - x0);\n",
       "    var height = Math.abs(y1 - y0);\n",
       "\n",
       "    fig.rubberband_context.clearRect(\n",
       "        0, 0, fig.canvas.width, fig.canvas.height);\n",
       "\n",
       "    fig.rubberband_context.strokeRect(min_x, min_y, width, height);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_figure_label = function(fig, msg) {\n",
       "    // Updates the figure title.\n",
       "    fig.header.textContent = msg['label'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_cursor = function(fig, msg) {\n",
       "    var cursor = msg['cursor'];\n",
       "    switch(cursor)\n",
       "    {\n",
       "    case 0:\n",
       "        cursor = 'pointer';\n",
       "        break;\n",
       "    case 1:\n",
       "        cursor = 'default';\n",
       "        break;\n",
       "    case 2:\n",
       "        cursor = 'crosshair';\n",
       "        break;\n",
       "    case 3:\n",
       "        cursor = 'move';\n",
       "        break;\n",
       "    }\n",
       "    fig.rubberband_canvas.style.cursor = cursor;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_message = function(fig, msg) {\n",
       "    fig.message.textContent = msg['message'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_draw = function(fig, msg) {\n",
       "    // Request the server to send over a new figure.\n",
       "    fig.send_draw_message();\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_image_mode = function(fig, msg) {\n",
       "    fig.image_mode = msg['mode'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Called whenever the canvas gets updated.\n",
       "    this.send_message(\"ack\", {});\n",
       "}\n",
       "\n",
       "// A function to construct a web socket function for onmessage handling.\n",
       "// Called in the figure constructor.\n",
       "mpl.figure.prototype._make_on_message_function = function(fig) {\n",
       "    return function socket_on_message(evt) {\n",
       "        if (evt.data instanceof Blob) {\n",
       "            /* FIXME: We get \"Resource interpreted as Image but\n",
       "             * transferred with MIME type text/plain:\" errors on\n",
       "             * Chrome.  But how to set the MIME type?  It doesn't seem\n",
       "             * to be part of the websocket stream */\n",
       "            evt.data.type = \"image/png\";\n",
       "\n",
       "            /* Free the memory for the previous frames */\n",
       "            if (fig.imageObj.src) {\n",
       "                (window.URL || window.webkitURL).revokeObjectURL(\n",
       "                    fig.imageObj.src);\n",
       "            }\n",
       "\n",
       "            fig.imageObj.src = (window.URL || window.webkitURL).createObjectURL(\n",
       "                evt.data);\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "        else if (typeof evt.data === 'string' && evt.data.slice(0, 21) == \"data:image/png;base64\") {\n",
       "            fig.imageObj.src = evt.data;\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        var msg = JSON.parse(evt.data);\n",
       "        var msg_type = msg['type'];\n",
       "\n",
       "        // Call the  \"handle_{type}\" callback, which takes\n",
       "        // the figure and JSON message as its only arguments.\n",
       "        try {\n",
       "            var callback = fig[\"handle_\" + msg_type];\n",
       "        } catch (e) {\n",
       "            console.log(\"No handler for the '\" + msg_type + \"' message type: \", msg);\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        if (callback) {\n",
       "            try {\n",
       "                // console.log(\"Handling '\" + msg_type + \"' message: \", msg);\n",
       "                callback(fig, msg);\n",
       "            } catch (e) {\n",
       "                console.log(\"Exception inside the 'handler_\" + msg_type + \"' callback:\", e, e.stack, msg);\n",
       "            }\n",
       "        }\n",
       "    };\n",
       "}\n",
       "\n",
       "// from http://stackoverflow.com/questions/1114465/getting-mouse-location-in-canvas\n",
       "mpl.findpos = function(e) {\n",
       "    //this section is from http://www.quirksmode.org/js/events_properties.html\n",
       "    var targ;\n",
       "    if (!e)\n",
       "        e = window.event;\n",
       "    if (e.target)\n",
       "        targ = e.target;\n",
       "    else if (e.srcElement)\n",
       "        targ = e.srcElement;\n",
       "    if (targ.nodeType == 3) // defeat Safari bug\n",
       "        targ = targ.parentNode;\n",
       "\n",
       "    // jQuery normalizes the pageX and pageY\n",
       "    // pageX,Y are the mouse positions relative to the document\n",
       "    // offset() returns the position of the element relative to the document\n",
       "    var x = e.pageX - $(targ).offset().left;\n",
       "    var y = e.pageY - $(targ).offset().top;\n",
       "\n",
       "    return {\"x\": x, \"y\": y};\n",
       "};\n",
       "\n",
       "/*\n",
       " * return a copy of an object with only non-object keys\n",
       " * we need this to avoid circular references\n",
       " * http://stackoverflow.com/a/24161582/3208463\n",
       " */\n",
       "function simpleKeys (original) {\n",
       "  return Object.keys(original).reduce(function (obj, key) {\n",
       "    if (typeof original[key] !== 'object')\n",
       "        obj[key] = original[key]\n",
       "    return obj;\n",
       "  }, {});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.mouse_event = function(event, name) {\n",
       "    var canvas_pos = mpl.findpos(event)\n",
       "\n",
       "    if (name === 'button_press')\n",
       "    {\n",
       "        this.canvas.focus();\n",
       "        this.canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    var x = canvas_pos.x * mpl.ratio;\n",
       "    var y = canvas_pos.y * mpl.ratio;\n",
       "\n",
       "    this.send_message(name, {x: x, y: y, button: event.button,\n",
       "                             step: event.step,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "\n",
       "    /* This prevents the web browser from automatically changing to\n",
       "     * the text insertion cursor when the button is pressed.  We want\n",
       "     * to control all of the cursor setting manually through the\n",
       "     * 'cursor' event from matplotlib */\n",
       "    event.preventDefault();\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    // Handle any extra behaviour associated with a key event\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.key_event = function(event, name) {\n",
       "\n",
       "    // Prevent repeat events\n",
       "    if (name == 'key_press')\n",
       "    {\n",
       "        if (event.which === this._key)\n",
       "            return;\n",
       "        else\n",
       "            this._key = event.which;\n",
       "    }\n",
       "    if (name == 'key_release')\n",
       "        this._key = null;\n",
       "\n",
       "    var value = '';\n",
       "    if (event.ctrlKey && event.which != 17)\n",
       "        value += \"ctrl+\";\n",
       "    if (event.altKey && event.which != 18)\n",
       "        value += \"alt+\";\n",
       "    if (event.shiftKey && event.which != 16)\n",
       "        value += \"shift+\";\n",
       "\n",
       "    value += 'k';\n",
       "    value += event.which.toString();\n",
       "\n",
       "    this._key_event_extra(event, name);\n",
       "\n",
       "    this.send_message(name, {key: value,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onclick = function(name) {\n",
       "    if (name == 'download') {\n",
       "        this.handle_save(this, null);\n",
       "    } else {\n",
       "        this.send_message(\"toolbar_button\", {name: name});\n",
       "    }\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onmouseover = function(tooltip) {\n",
       "    this.message.textContent = tooltip;\n",
       "};\n",
       "mpl.toolbar_items = [[\"Home\", \"Reset original view\", \"fa fa-home icon-home\", \"home\"], [\"Back\", \"Back to  previous view\", \"fa fa-arrow-left icon-arrow-left\", \"back\"], [\"Forward\", \"Forward to next view\", \"fa fa-arrow-right icon-arrow-right\", \"forward\"], [\"\", \"\", \"\", \"\"], [\"Pan\", \"Pan axes with left mouse, zoom with right\", \"fa fa-arrows icon-move\", \"pan\"], [\"Zoom\", \"Zoom to rectangle\", \"fa fa-square-o icon-check-empty\", \"zoom\"], [\"\", \"\", \"\", \"\"], [\"Download\", \"Download plot\", \"fa fa-floppy-o icon-save\", \"download\"]];\n",
       "\n",
       "mpl.extensions = [\"eps\", \"jpeg\", \"pdf\", \"png\", \"ps\", \"raw\", \"svg\", \"tif\"];\n",
       "\n",
       "mpl.default_extension = \"png\";var comm_websocket_adapter = function(comm) {\n",
       "    // Create a \"websocket\"-like object which calls the given IPython comm\n",
       "    // object with the appropriate methods. Currently this is a non binary\n",
       "    // socket, so there is still some room for performance tuning.\n",
       "    var ws = {};\n",
       "\n",
       "    ws.close = function() {\n",
       "        comm.close()\n",
       "    };\n",
       "    ws.send = function(m) {\n",
       "        //console.log('sending', m);\n",
       "        comm.send(m);\n",
       "    };\n",
       "    // Register the callback with on_msg.\n",
       "    comm.on_msg(function(msg) {\n",
       "        //console.log('receiving', msg['content']['data'], msg);\n",
       "        // Pass the mpl event to the overriden (by mpl) onmessage function.\n",
       "        ws.onmessage(msg['content']['data'])\n",
       "    });\n",
       "    return ws;\n",
       "}\n",
       "\n",
       "mpl.mpl_figure_comm = function(comm, msg) {\n",
       "    // This is the function which gets called when the mpl process\n",
       "    // starts-up an IPython Comm through the \"matplotlib\" channel.\n",
       "\n",
       "    var id = msg.content.data.id;\n",
       "    // Get hold of the div created by the display call when the Comm\n",
       "    // socket was opened in Python.\n",
       "    var element = $(\"#\" + id);\n",
       "    var ws_proxy = comm_websocket_adapter(comm)\n",
       "\n",
       "    function ondownload(figure, format) {\n",
       "        window.open(figure.imageObj.src);\n",
       "    }\n",
       "\n",
       "    var fig = new mpl.figure(id, ws_proxy,\n",
       "                           ondownload,\n",
       "                           element.get(0));\n",
       "\n",
       "    // Call onopen now - mpl needs it, as it is assuming we've passed it a real\n",
       "    // web socket which is closed, not our websocket->open comm proxy.\n",
       "    ws_proxy.onopen();\n",
       "\n",
       "    fig.parent_element = element.get(0);\n",
       "    fig.cell_info = mpl.find_output_cell(\"<div id='\" + id + \"'></div>\");\n",
       "    if (!fig.cell_info) {\n",
       "        console.error(\"Failed to find cell for figure\", id, fig);\n",
       "        return;\n",
       "    }\n",
       "\n",
       "    var output_index = fig.cell_info[2]\n",
       "    var cell = fig.cell_info[0];\n",
       "\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.handle_close = function(fig, msg) {\n",
       "    var width = fig.canvas.width/mpl.ratio\n",
       "    fig.root.unbind('remove')\n",
       "\n",
       "    // Update the output cell to use the data from the current canvas.\n",
       "    fig.push_to_output();\n",
       "    var dataURL = fig.canvas.toDataURL();\n",
       "    // Re-enable the keyboard manager in IPython - without this line, in FF,\n",
       "    // the notebook keyboard shortcuts fail.\n",
       "    IPython.keyboard_manager.enable()\n",
       "    $(fig.parent_element).html('<img src=\"' + dataURL + '\" width=\"' + width + '\">');\n",
       "    fig.close_ws(fig, msg);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.close_ws = function(fig, msg){\n",
       "    fig.send_message('closing', msg);\n",
       "    // fig.ws.close()\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.push_to_output = function(remove_interactive) {\n",
       "    // Turn the data on the canvas into data in the output cell.\n",
       "    var width = this.canvas.width/mpl.ratio\n",
       "    var dataURL = this.canvas.toDataURL();\n",
       "    this.cell_info[1]['text/html'] = '<img src=\"' + dataURL + '\" width=\"' + width + '\">';\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Tell IPython that the notebook contents must change.\n",
       "    IPython.notebook.set_dirty(true);\n",
       "    this.send_message(\"ack\", {});\n",
       "    var fig = this;\n",
       "    // Wait a second, then push the new image to the DOM so\n",
       "    // that it is saved nicely (might be nice to debounce this).\n",
       "    setTimeout(function () { fig.push_to_output() }, 1000);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>')\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items){\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) { continue; };\n",
       "\n",
       "        var button = $('<button class=\"btn btn-default\" href=\"#\" title=\"' + name + '\"><i class=\"fa ' + image + ' fa-lg\"></i></button>');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    // Add the status bar.\n",
       "    var status_bar = $('<span class=\"mpl-message\" style=\"text-align:right; float: right;\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "\n",
       "    // Add the close button to the window.\n",
       "    var buttongrp = $('<div class=\"btn-group inline pull-right\"></div>');\n",
       "    var button = $('<button class=\"btn btn-mini btn-primary\" href=\"#\" title=\"Stop Interaction\"><i class=\"fa fa-power-off icon-remove icon-large\"></i></button>');\n",
       "    button.click(function (evt) { fig.handle_close(fig, {}); } );\n",
       "    button.mouseover('Stop Interaction', toolbar_mouse_event);\n",
       "    buttongrp.append(button);\n",
       "    var titlebar = this.root.find($('.ui-dialog-titlebar'));\n",
       "    titlebar.prepend(buttongrp);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(el){\n",
       "    var fig = this\n",
       "    el.on(\"remove\", function(){\n",
       "\tfig.close_ws(fig, {});\n",
       "    });\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(el){\n",
       "    // this is important to make the div 'focusable\n",
       "    el.attr('tabindex', 0)\n",
       "    // reach out to IPython and tell the keyboard manager to turn it's self\n",
       "    // off when our div gets focus\n",
       "\n",
       "    // location in version 3\n",
       "    if (IPython.notebook.keyboard_manager) {\n",
       "        IPython.notebook.keyboard_manager.register_events(el);\n",
       "    }\n",
       "    else {\n",
       "        // location in version 2\n",
       "        IPython.keyboard_manager.register_events(el);\n",
       "    }\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    var manager = IPython.notebook.keyboard_manager;\n",
       "    if (!manager)\n",
       "        manager = IPython.keyboard_manager;\n",
       "\n",
       "    // Check for shift+enter\n",
       "    if (event.shiftKey && event.which == 13) {\n",
       "        this.canvas_div.blur();\n",
       "        // select the cell after this one\n",
       "        var index = IPython.notebook.find_cell_index(this.cell_info[0]);\n",
       "        IPython.notebook.select(index + 1);\n",
       "    }\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    fig.ondownload(fig, null);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.find_output_cell = function(html_output) {\n",
       "    // Return the cell and output element which can be found *uniquely* in the notebook.\n",
       "    // Note - this is a bit hacky, but it is done because the \"notebook_saving.Notebook\"\n",
       "    // IPython event is triggered only after the cells have been serialised, which for\n",
       "    // our purposes (turning an active figure into a static one), is too late.\n",
       "    var cells = IPython.notebook.get_cells();\n",
       "    var ncells = cells.length;\n",
       "    for (var i=0; i<ncells; i++) {\n",
       "        var cell = cells[i];\n",
       "        if (cell.cell_type === 'code'){\n",
       "            for (var j=0; j<cell.output_area.outputs.length; j++) {\n",
       "                var data = cell.output_area.outputs[j];\n",
       "                if (data.data) {\n",
       "                    // IPython >= 3 moved mimebundle to data attribute of output\n",
       "                    data = data.data;\n",
       "                }\n",
       "                if (data['text/html'] == html_output) {\n",
       "                    return [cell, data, j];\n",
       "                }\n",
       "            }\n",
       "        }\n",
       "    }\n",
       "}\n",
       "\n",
       "// Register the function which deals with the matplotlib target/channel.\n",
       "// The kernel may be null if the page has been refreshed.\n",
       "if (IPython.notebook.kernel != null) {\n",
       "    IPython.notebook.kernel.comm_manager.register_target('matplotlib', mpl.mpl_figure_comm);\n",
       "}\n"
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"data:image/png;base64,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\" width=\"640\">"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib notebook\n",
    "# Compare LSTM to Bidirectional LSTM\n",
    "\n",
    "def get_lstm_model(n_timesteps, backwards):\n",
    "\tmodel = Sequential()\n",
    "\tmodel.add(LSTM(20, input_shape=(n_timesteps, 1), return_sequences=True, go_backwards=backwards))\n",
    "\tmodel.add(TimeDistributed(Dense(1, activation='sigmoid')))\n",
    "\tmodel.compile(loss='binary_crossentropy', optimizer='adam')\n",
    "\treturn model\n",
    "\n",
    "def get_bi_lstm_model(n_timesteps, mode):\n",
    "\tmodel = Sequential()\n",
    "\tmodel.add(Bidirectional(LSTM(20, return_sequences=True), input_shape=(n_timesteps, 1), merge_mode=mode))\n",
    "\tmodel.add(TimeDistributed(Dense(1, activation='sigmoid')))\n",
    "\tmodel.compile(loss='binary_crossentropy', optimizer='adam')\n",
    "\treturn model\n",
    "\n",
    "def train_model(model, n_timesteps):\n",
    "\tloss = list()\n",
    "\tfor _ in range(250):\n",
    "\t\t# generate new random sequence\n",
    "\t\tX,y = get_sequence(n_timesteps)\n",
    "\t\t# fit model for one epoch on this sequence\n",
    "\t\thist = model.fit(X, y, epochs=1, batch_size=1, verbose=0)\n",
    "\t\tloss.append(hist.history['loss'][0])\n",
    "\treturn loss\n",
    "\n",
    "n_timesteps = 10\n",
    "results = DataFrame()\n",
    "# lstm forwards\n",
    "model = get_lstm_model(n_timesteps, False)\n",
    "results['lstm_forw'] = train_model(model, n_timesteps)\n",
    "# lstm backwards\n",
    "model = get_lstm_model(n_timesteps, True)\n",
    "results['lstm_back'] = train_model(model, n_timesteps)\n",
    "# bidirectional concat\n",
    "model = get_bi_lstm_model(n_timesteps, 'concat')\n",
    "results['bilstm_con'] = train_model(model, n_timesteps)\n",
    "# line plot of results\n",
    "results.plot()\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "/* Put everything inside the global mpl namespace */\n",
       "window.mpl = {};\n",
       "\n",
       "\n",
       "mpl.get_websocket_type = function() {\n",
       "    if (typeof(WebSocket) !== 'undefined') {\n",
       "        return WebSocket;\n",
       "    } else if (typeof(MozWebSocket) !== 'undefined') {\n",
       "        return MozWebSocket;\n",
       "    } else {\n",
       "        alert('Your browser does not have WebSocket support.' +\n",
       "              'Please try Chrome, Safari or Firefox ≥ 6. ' +\n",
       "              'Firefox 4 and 5 are also supported but you ' +\n",
       "              'have to enable WebSockets in about:config.');\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure = function(figure_id, websocket, ondownload, parent_element) {\n",
       "    this.id = figure_id;\n",
       "\n",
       "    this.ws = websocket;\n",
       "\n",
       "    this.supports_binary = (this.ws.binaryType != undefined);\n",
       "\n",
       "    if (!this.supports_binary) {\n",
       "        var warnings = document.getElementById(\"mpl-warnings\");\n",
       "        if (warnings) {\n",
       "            warnings.style.display = 'block';\n",
       "            warnings.textContent = (\n",
       "                \"This browser does not support binary websocket messages. \" +\n",
       "                    \"Performance may be slow.\");\n",
       "        }\n",
       "    }\n",
       "\n",
       "    this.imageObj = new Image();\n",
       "\n",
       "    this.context = undefined;\n",
       "    this.message = undefined;\n",
       "    this.canvas = undefined;\n",
       "    this.rubberband_canvas = undefined;\n",
       "    this.rubberband_context = undefined;\n",
       "    this.format_dropdown = undefined;\n",
       "\n",
       "    this.image_mode = 'full';\n",
       "\n",
       "    this.root = $('<div/>');\n",
       "    this._root_extra_style(this.root)\n",
       "    this.root.attr('style', 'display: inline-block');\n",
       "\n",
       "    $(parent_element).append(this.root);\n",
       "\n",
       "    this._init_header(this);\n",
       "    this._init_canvas(this);\n",
       "    this._init_toolbar(this);\n",
       "\n",
       "    var fig = this;\n",
       "\n",
       "    this.waiting = false;\n",
       "\n",
       "    this.ws.onopen =  function () {\n",
       "            fig.send_message(\"supports_binary\", {value: fig.supports_binary});\n",
       "            fig.send_message(\"send_image_mode\", {});\n",
       "            if (mpl.ratio != 1) {\n",
       "                fig.send_message(\"set_dpi_ratio\", {'dpi_ratio': mpl.ratio});\n",
       "            }\n",
       "            fig.send_message(\"refresh\", {});\n",
       "        }\n",
       "\n",
       "    this.imageObj.onload = function() {\n",
       "            if (fig.image_mode == 'full') {\n",
       "                // Full images could contain transparency (where diff images\n",
       "                // almost always do), so we need to clear the canvas so that\n",
       "                // there is no ghosting.\n",
       "                fig.context.clearRect(0, 0, fig.canvas.width, fig.canvas.height);\n",
       "            }\n",
       "            fig.context.drawImage(fig.imageObj, 0, 0);\n",
       "        };\n",
       "\n",
       "    this.imageObj.onunload = function() {\n",
       "        this.ws.close();\n",
       "    }\n",
       "\n",
       "    this.ws.onmessage = this._make_on_message_function(this);\n",
       "\n",
       "    this.ondownload = ondownload;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_header = function() {\n",
       "    var titlebar = $(\n",
       "        '<div class=\"ui-dialog-titlebar ui-widget-header ui-corner-all ' +\n",
       "        'ui-helper-clearfix\"/>');\n",
       "    var titletext = $(\n",
       "        '<div class=\"ui-dialog-title\" style=\"width: 100%; ' +\n",
       "        'text-align: center; padding: 3px;\"/>');\n",
       "    titlebar.append(titletext)\n",
       "    this.root.append(titlebar);\n",
       "    this.header = titletext[0];\n",
       "}\n",
       "\n",
       "\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(canvas_div) {\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_canvas = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var canvas_div = $('<div/>');\n",
       "\n",
       "    canvas_div.attr('style', 'position: relative; clear: both; outline: 0');\n",
       "\n",
       "    function canvas_keyboard_event(event) {\n",
       "        return fig.key_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    canvas_div.keydown('key_press', canvas_keyboard_event);\n",
       "    canvas_div.keyup('key_release', canvas_keyboard_event);\n",
       "    this.canvas_div = canvas_div\n",
       "    this._canvas_extra_style(canvas_div)\n",
       "    this.root.append(canvas_div);\n",
       "\n",
       "    var canvas = $('<canvas/>');\n",
       "    canvas.addClass('mpl-canvas');\n",
       "    canvas.attr('style', \"left: 0; top: 0; z-index: 0; outline: 0\")\n",
       "\n",
       "    this.canvas = canvas[0];\n",
       "    this.context = canvas[0].getContext(\"2d\");\n",
       "\n",
       "    var backingStore = this.context.backingStorePixelRatio ||\n",
       "\tthis.context.webkitBackingStorePixelRatio ||\n",
       "\tthis.context.mozBackingStorePixelRatio ||\n",
       "\tthis.context.msBackingStorePixelRatio ||\n",
       "\tthis.context.oBackingStorePixelRatio ||\n",
       "\tthis.context.backingStorePixelRatio || 1;\n",
       "\n",
       "    mpl.ratio = (window.devicePixelRatio || 1) / backingStore;\n",
       "\n",
       "    var rubberband = $('<canvas/>');\n",
       "    rubberband.attr('style', \"position: absolute; left: 0; top: 0; z-index: 1;\")\n",
       "\n",
       "    var pass_mouse_events = true;\n",
       "\n",
       "    canvas_div.resizable({\n",
       "        start: function(event, ui) {\n",
       "            pass_mouse_events = false;\n",
       "        },\n",
       "        resize: function(event, ui) {\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "        stop: function(event, ui) {\n",
       "            pass_mouse_events = true;\n",
       "            fig.request_resize(ui.size.width, ui.size.height);\n",
       "        },\n",
       "    });\n",
       "\n",
       "    function mouse_event_fn(event) {\n",
       "        if (pass_mouse_events)\n",
       "            return fig.mouse_event(event, event['data']);\n",
       "    }\n",
       "\n",
       "    rubberband.mousedown('button_press', mouse_event_fn);\n",
       "    rubberband.mouseup('button_release', mouse_event_fn);\n",
       "    // Throttle sequential mouse events to 1 every 20ms.\n",
       "    rubberband.mousemove('motion_notify', mouse_event_fn);\n",
       "\n",
       "    rubberband.mouseenter('figure_enter', mouse_event_fn);\n",
       "    rubberband.mouseleave('figure_leave', mouse_event_fn);\n",
       "\n",
       "    canvas_div.on(\"wheel\", function (event) {\n",
       "        event = event.originalEvent;\n",
       "        event['data'] = 'scroll'\n",
       "        if (event.deltaY < 0) {\n",
       "            event.step = 1;\n",
       "        } else {\n",
       "            event.step = -1;\n",
       "        }\n",
       "        mouse_event_fn(event);\n",
       "    });\n",
       "\n",
       "    canvas_div.append(canvas);\n",
       "    canvas_div.append(rubberband);\n",
       "\n",
       "    this.rubberband = rubberband;\n",
       "    this.rubberband_canvas = rubberband[0];\n",
       "    this.rubberband_context = rubberband[0].getContext(\"2d\");\n",
       "    this.rubberband_context.strokeStyle = \"#000000\";\n",
       "\n",
       "    this._resize_canvas = function(width, height) {\n",
       "        // Keep the size of the canvas, canvas container, and rubber band\n",
       "        // canvas in synch.\n",
       "        canvas_div.css('width', width)\n",
       "        canvas_div.css('height', height)\n",
       "\n",
       "        canvas.attr('width', width * mpl.ratio);\n",
       "        canvas.attr('height', height * mpl.ratio);\n",
       "        canvas.attr('style', 'width: ' + width + 'px; height: ' + height + 'px;');\n",
       "\n",
       "        rubberband.attr('width', width);\n",
       "        rubberband.attr('height', height);\n",
       "    }\n",
       "\n",
       "    // Set the figure to an initial 600x600px, this will subsequently be updated\n",
       "    // upon first draw.\n",
       "    this._resize_canvas(600, 600);\n",
       "\n",
       "    // Disable right mouse context menu.\n",
       "    $(this.rubberband_canvas).bind(\"contextmenu\",function(e){\n",
       "        return false;\n",
       "    });\n",
       "\n",
       "    function set_focus () {\n",
       "        canvas.focus();\n",
       "        canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    window.setTimeout(set_focus, 100);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>')\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items) {\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) {\n",
       "            // put a spacer in here.\n",
       "            continue;\n",
       "        }\n",
       "        var button = $('<button/>');\n",
       "        button.addClass('ui-button ui-widget ui-state-default ui-corner-all ' +\n",
       "                        'ui-button-icon-only');\n",
       "        button.attr('role', 'button');\n",
       "        button.attr('aria-disabled', 'false');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "\n",
       "        var icon_img = $('<span/>');\n",
       "        icon_img.addClass('ui-button-icon-primary ui-icon');\n",
       "        icon_img.addClass(image);\n",
       "        icon_img.addClass('ui-corner-all');\n",
       "\n",
       "        var tooltip_span = $('<span/>');\n",
       "        tooltip_span.addClass('ui-button-text');\n",
       "        tooltip_span.html(tooltip);\n",
       "\n",
       "        button.append(icon_img);\n",
       "        button.append(tooltip_span);\n",
       "\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    var fmt_picker_span = $('<span/>');\n",
       "\n",
       "    var fmt_picker = $('<select/>');\n",
       "    fmt_picker.addClass('mpl-toolbar-option ui-widget ui-widget-content');\n",
       "    fmt_picker_span.append(fmt_picker);\n",
       "    nav_element.append(fmt_picker_span);\n",
       "    this.format_dropdown = fmt_picker[0];\n",
       "\n",
       "    for (var ind in mpl.extensions) {\n",
       "        var fmt = mpl.extensions[ind];\n",
       "        var option = $(\n",
       "            '<option/>', {selected: fmt === mpl.default_extension}).html(fmt);\n",
       "        fmt_picker.append(option)\n",
       "    }\n",
       "\n",
       "    // Add hover states to the ui-buttons\n",
       "    $( \".ui-button\" ).hover(\n",
       "        function() { $(this).addClass(\"ui-state-hover\");},\n",
       "        function() { $(this).removeClass(\"ui-state-hover\");}\n",
       "    );\n",
       "\n",
       "    var status_bar = $('<span class=\"mpl-message\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.request_resize = function(x_pixels, y_pixels) {\n",
       "    // Request matplotlib to resize the figure. Matplotlib will then trigger a resize in the client,\n",
       "    // which will in turn request a refresh of the image.\n",
       "    this.send_message('resize', {'width': x_pixels, 'height': y_pixels});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_message = function(type, properties) {\n",
       "    properties['type'] = type;\n",
       "    properties['figure_id'] = this.id;\n",
       "    this.ws.send(JSON.stringify(properties));\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.send_draw_message = function() {\n",
       "    if (!this.waiting) {\n",
       "        this.waiting = true;\n",
       "        this.ws.send(JSON.stringify({type: \"draw\", figure_id: this.id}));\n",
       "    }\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    var format_dropdown = fig.format_dropdown;\n",
       "    var format = format_dropdown.options[format_dropdown.selectedIndex].value;\n",
       "    fig.ondownload(fig, format);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.figure.prototype.handle_resize = function(fig, msg) {\n",
       "    var size = msg['size'];\n",
       "    if (size[0] != fig.canvas.width || size[1] != fig.canvas.height) {\n",
       "        fig._resize_canvas(size[0], size[1]);\n",
       "        fig.send_message(\"refresh\", {});\n",
       "    };\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_rubberband = function(fig, msg) {\n",
       "    var x0 = msg['x0'] / mpl.ratio;\n",
       "    var y0 = (fig.canvas.height - msg['y0']) / mpl.ratio;\n",
       "    var x1 = msg['x1'] / mpl.ratio;\n",
       "    var y1 = (fig.canvas.height - msg['y1']) / mpl.ratio;\n",
       "    x0 = Math.floor(x0) + 0.5;\n",
       "    y0 = Math.floor(y0) + 0.5;\n",
       "    x1 = Math.floor(x1) + 0.5;\n",
       "    y1 = Math.floor(y1) + 0.5;\n",
       "    var min_x = Math.min(x0, x1);\n",
       "    var min_y = Math.min(y0, y1);\n",
       "    var width = Math.abs(x1 - x0);\n",
       "    var height = Math.abs(y1 - y0);\n",
       "\n",
       "    fig.rubberband_context.clearRect(\n",
       "        0, 0, fig.canvas.width, fig.canvas.height);\n",
       "\n",
       "    fig.rubberband_context.strokeRect(min_x, min_y, width, height);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_figure_label = function(fig, msg) {\n",
       "    // Updates the figure title.\n",
       "    fig.header.textContent = msg['label'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_cursor = function(fig, msg) {\n",
       "    var cursor = msg['cursor'];\n",
       "    switch(cursor)\n",
       "    {\n",
       "    case 0:\n",
       "        cursor = 'pointer';\n",
       "        break;\n",
       "    case 1:\n",
       "        cursor = 'default';\n",
       "        break;\n",
       "    case 2:\n",
       "        cursor = 'crosshair';\n",
       "        break;\n",
       "    case 3:\n",
       "        cursor = 'move';\n",
       "        break;\n",
       "    }\n",
       "    fig.rubberband_canvas.style.cursor = cursor;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_message = function(fig, msg) {\n",
       "    fig.message.textContent = msg['message'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_draw = function(fig, msg) {\n",
       "    // Request the server to send over a new figure.\n",
       "    fig.send_draw_message();\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_image_mode = function(fig, msg) {\n",
       "    fig.image_mode = msg['mode'];\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Called whenever the canvas gets updated.\n",
       "    this.send_message(\"ack\", {});\n",
       "}\n",
       "\n",
       "// A function to construct a web socket function for onmessage handling.\n",
       "// Called in the figure constructor.\n",
       "mpl.figure.prototype._make_on_message_function = function(fig) {\n",
       "    return function socket_on_message(evt) {\n",
       "        if (evt.data instanceof Blob) {\n",
       "            /* FIXME: We get \"Resource interpreted as Image but\n",
       "             * transferred with MIME type text/plain:\" errors on\n",
       "             * Chrome.  But how to set the MIME type?  It doesn't seem\n",
       "             * to be part of the websocket stream */\n",
       "            evt.data.type = \"image/png\";\n",
       "\n",
       "            /* Free the memory for the previous frames */\n",
       "            if (fig.imageObj.src) {\n",
       "                (window.URL || window.webkitURL).revokeObjectURL(\n",
       "                    fig.imageObj.src);\n",
       "            }\n",
       "\n",
       "            fig.imageObj.src = (window.URL || window.webkitURL).createObjectURL(\n",
       "                evt.data);\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "        else if (typeof evt.data === 'string' && evt.data.slice(0, 21) == \"data:image/png;base64\") {\n",
       "            fig.imageObj.src = evt.data;\n",
       "            fig.updated_canvas_event();\n",
       "            fig.waiting = false;\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        var msg = JSON.parse(evt.data);\n",
       "        var msg_type = msg['type'];\n",
       "\n",
       "        // Call the  \"handle_{type}\" callback, which takes\n",
       "        // the figure and JSON message as its only arguments.\n",
       "        try {\n",
       "            var callback = fig[\"handle_\" + msg_type];\n",
       "        } catch (e) {\n",
       "            console.log(\"No handler for the '\" + msg_type + \"' message type: \", msg);\n",
       "            return;\n",
       "        }\n",
       "\n",
       "        if (callback) {\n",
       "            try {\n",
       "                // console.log(\"Handling '\" + msg_type + \"' message: \", msg);\n",
       "                callback(fig, msg);\n",
       "            } catch (e) {\n",
       "                console.log(\"Exception inside the 'handler_\" + msg_type + \"' callback:\", e, e.stack, msg);\n",
       "            }\n",
       "        }\n",
       "    };\n",
       "}\n",
       "\n",
       "// from http://stackoverflow.com/questions/1114465/getting-mouse-location-in-canvas\n",
       "mpl.findpos = function(e) {\n",
       "    //this section is from http://www.quirksmode.org/js/events_properties.html\n",
       "    var targ;\n",
       "    if (!e)\n",
       "        e = window.event;\n",
       "    if (e.target)\n",
       "        targ = e.target;\n",
       "    else if (e.srcElement)\n",
       "        targ = e.srcElement;\n",
       "    if (targ.nodeType == 3) // defeat Safari bug\n",
       "        targ = targ.parentNode;\n",
       "\n",
       "    // jQuery normalizes the pageX and pageY\n",
       "    // pageX,Y are the mouse positions relative to the document\n",
       "    // offset() returns the position of the element relative to the document\n",
       "    var x = e.pageX - $(targ).offset().left;\n",
       "    var y = e.pageY - $(targ).offset().top;\n",
       "\n",
       "    return {\"x\": x, \"y\": y};\n",
       "};\n",
       "\n",
       "/*\n",
       " * return a copy of an object with only non-object keys\n",
       " * we need this to avoid circular references\n",
       " * http://stackoverflow.com/a/24161582/3208463\n",
       " */\n",
       "function simpleKeys (original) {\n",
       "  return Object.keys(original).reduce(function (obj, key) {\n",
       "    if (typeof original[key] !== 'object')\n",
       "        obj[key] = original[key]\n",
       "    return obj;\n",
       "  }, {});\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.mouse_event = function(event, name) {\n",
       "    var canvas_pos = mpl.findpos(event)\n",
       "\n",
       "    if (name === 'button_press')\n",
       "    {\n",
       "        this.canvas.focus();\n",
       "        this.canvas_div.focus();\n",
       "    }\n",
       "\n",
       "    var x = canvas_pos.x * mpl.ratio;\n",
       "    var y = canvas_pos.y * mpl.ratio;\n",
       "\n",
       "    this.send_message(name, {x: x, y: y, button: event.button,\n",
       "                             step: event.step,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "\n",
       "    /* This prevents the web browser from automatically changing to\n",
       "     * the text insertion cursor when the button is pressed.  We want\n",
       "     * to control all of the cursor setting manually through the\n",
       "     * 'cursor' event from matplotlib */\n",
       "    event.preventDefault();\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    // Handle any extra behaviour associated with a key event\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.key_event = function(event, name) {\n",
       "\n",
       "    // Prevent repeat events\n",
       "    if (name == 'key_press')\n",
       "    {\n",
       "        if (event.which === this._key)\n",
       "            return;\n",
       "        else\n",
       "            this._key = event.which;\n",
       "    }\n",
       "    if (name == 'key_release')\n",
       "        this._key = null;\n",
       "\n",
       "    var value = '';\n",
       "    if (event.ctrlKey && event.which != 17)\n",
       "        value += \"ctrl+\";\n",
       "    if (event.altKey && event.which != 18)\n",
       "        value += \"alt+\";\n",
       "    if (event.shiftKey && event.which != 16)\n",
       "        value += \"shift+\";\n",
       "\n",
       "    value += 'k';\n",
       "    value += event.which.toString();\n",
       "\n",
       "    this._key_event_extra(event, name);\n",
       "\n",
       "    this.send_message(name, {key: value,\n",
       "                             guiEvent: simpleKeys(event)});\n",
       "    return false;\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onclick = function(name) {\n",
       "    if (name == 'download') {\n",
       "        this.handle_save(this, null);\n",
       "    } else {\n",
       "        this.send_message(\"toolbar_button\", {name: name});\n",
       "    }\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.toolbar_button_onmouseover = function(tooltip) {\n",
       "    this.message.textContent = tooltip;\n",
       "};\n",
       "mpl.toolbar_items = [[\"Home\", \"Reset original view\", \"fa fa-home icon-home\", \"home\"], [\"Back\", \"Back to  previous view\", \"fa fa-arrow-left icon-arrow-left\", \"back\"], [\"Forward\", \"Forward to next view\", \"fa fa-arrow-right icon-arrow-right\", \"forward\"], [\"\", \"\", \"\", \"\"], [\"Pan\", \"Pan axes with left mouse, zoom with right\", \"fa fa-arrows icon-move\", \"pan\"], [\"Zoom\", \"Zoom to rectangle\", \"fa fa-square-o icon-check-empty\", \"zoom\"], [\"\", \"\", \"\", \"\"], [\"Download\", \"Download plot\", \"fa fa-floppy-o icon-save\", \"download\"]];\n",
       "\n",
       "mpl.extensions = [\"eps\", \"jpeg\", \"pdf\", \"png\", \"ps\", \"raw\", \"svg\", \"tif\"];\n",
       "\n",
       "mpl.default_extension = \"png\";var comm_websocket_adapter = function(comm) {\n",
       "    // Create a \"websocket\"-like object which calls the given IPython comm\n",
       "    // object with the appropriate methods. Currently this is a non binary\n",
       "    // socket, so there is still some room for performance tuning.\n",
       "    var ws = {};\n",
       "\n",
       "    ws.close = function() {\n",
       "        comm.close()\n",
       "    };\n",
       "    ws.send = function(m) {\n",
       "        //console.log('sending', m);\n",
       "        comm.send(m);\n",
       "    };\n",
       "    // Register the callback with on_msg.\n",
       "    comm.on_msg(function(msg) {\n",
       "        //console.log('receiving', msg['content']['data'], msg);\n",
       "        // Pass the mpl event to the overriden (by mpl) onmessage function.\n",
       "        ws.onmessage(msg['content']['data'])\n",
       "    });\n",
       "    return ws;\n",
       "}\n",
       "\n",
       "mpl.mpl_figure_comm = function(comm, msg) {\n",
       "    // This is the function which gets called when the mpl process\n",
       "    // starts-up an IPython Comm through the \"matplotlib\" channel.\n",
       "\n",
       "    var id = msg.content.data.id;\n",
       "    // Get hold of the div created by the display call when the Comm\n",
       "    // socket was opened in Python.\n",
       "    var element = $(\"#\" + id);\n",
       "    var ws_proxy = comm_websocket_adapter(comm)\n",
       "\n",
       "    function ondownload(figure, format) {\n",
       "        window.open(figure.imageObj.src);\n",
       "    }\n",
       "\n",
       "    var fig = new mpl.figure(id, ws_proxy,\n",
       "                           ondownload,\n",
       "                           element.get(0));\n",
       "\n",
       "    // Call onopen now - mpl needs it, as it is assuming we've passed it a real\n",
       "    // web socket which is closed, not our websocket->open comm proxy.\n",
       "    ws_proxy.onopen();\n",
       "\n",
       "    fig.parent_element = element.get(0);\n",
       "    fig.cell_info = mpl.find_output_cell(\"<div id='\" + id + \"'></div>\");\n",
       "    if (!fig.cell_info) {\n",
       "        console.error(\"Failed to find cell for figure\", id, fig);\n",
       "        return;\n",
       "    }\n",
       "\n",
       "    var output_index = fig.cell_info[2]\n",
       "    var cell = fig.cell_info[0];\n",
       "\n",
       "};\n",
       "\n",
       "mpl.figure.prototype.handle_close = function(fig, msg) {\n",
       "    var width = fig.canvas.width/mpl.ratio\n",
       "    fig.root.unbind('remove')\n",
       "\n",
       "    // Update the output cell to use the data from the current canvas.\n",
       "    fig.push_to_output();\n",
       "    var dataURL = fig.canvas.toDataURL();\n",
       "    // Re-enable the keyboard manager in IPython - without this line, in FF,\n",
       "    // the notebook keyboard shortcuts fail.\n",
       "    IPython.keyboard_manager.enable()\n",
       "    $(fig.parent_element).html('<img src=\"' + dataURL + '\" width=\"' + width + '\">');\n",
       "    fig.close_ws(fig, msg);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.close_ws = function(fig, msg){\n",
       "    fig.send_message('closing', msg);\n",
       "    // fig.ws.close()\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.push_to_output = function(remove_interactive) {\n",
       "    // Turn the data on the canvas into data in the output cell.\n",
       "    var width = this.canvas.width/mpl.ratio\n",
       "    var dataURL = this.canvas.toDataURL();\n",
       "    this.cell_info[1]['text/html'] = '<img src=\"' + dataURL + '\" width=\"' + width + '\">';\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.updated_canvas_event = function() {\n",
       "    // Tell IPython that the notebook contents must change.\n",
       "    IPython.notebook.set_dirty(true);\n",
       "    this.send_message(\"ack\", {});\n",
       "    var fig = this;\n",
       "    // Wait a second, then push the new image to the DOM so\n",
       "    // that it is saved nicely (might be nice to debounce this).\n",
       "    setTimeout(function () { fig.push_to_output() }, 1000);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._init_toolbar = function() {\n",
       "    var fig = this;\n",
       "\n",
       "    var nav_element = $('<div/>')\n",
       "    nav_element.attr('style', 'width: 100%');\n",
       "    this.root.append(nav_element);\n",
       "\n",
       "    // Define a callback function for later on.\n",
       "    function toolbar_event(event) {\n",
       "        return fig.toolbar_button_onclick(event['data']);\n",
       "    }\n",
       "    function toolbar_mouse_event(event) {\n",
       "        return fig.toolbar_button_onmouseover(event['data']);\n",
       "    }\n",
       "\n",
       "    for(var toolbar_ind in mpl.toolbar_items){\n",
       "        var name = mpl.toolbar_items[toolbar_ind][0];\n",
       "        var tooltip = mpl.toolbar_items[toolbar_ind][1];\n",
       "        var image = mpl.toolbar_items[toolbar_ind][2];\n",
       "        var method_name = mpl.toolbar_items[toolbar_ind][3];\n",
       "\n",
       "        if (!name) { continue; };\n",
       "\n",
       "        var button = $('<button class=\"btn btn-default\" href=\"#\" title=\"' + name + '\"><i class=\"fa ' + image + ' fa-lg\"></i></button>');\n",
       "        button.click(method_name, toolbar_event);\n",
       "        button.mouseover(tooltip, toolbar_mouse_event);\n",
       "        nav_element.append(button);\n",
       "    }\n",
       "\n",
       "    // Add the status bar.\n",
       "    var status_bar = $('<span class=\"mpl-message\" style=\"text-align:right; float: right;\"/>');\n",
       "    nav_element.append(status_bar);\n",
       "    this.message = status_bar[0];\n",
       "\n",
       "    // Add the close button to the window.\n",
       "    var buttongrp = $('<div class=\"btn-group inline pull-right\"></div>');\n",
       "    var button = $('<button class=\"btn btn-mini btn-primary\" href=\"#\" title=\"Stop Interaction\"><i class=\"fa fa-power-off icon-remove icon-large\"></i></button>');\n",
       "    button.click(function (evt) { fig.handle_close(fig, {}); } );\n",
       "    button.mouseover('Stop Interaction', toolbar_mouse_event);\n",
       "    buttongrp.append(button);\n",
       "    var titlebar = this.root.find($('.ui-dialog-titlebar'));\n",
       "    titlebar.prepend(buttongrp);\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._root_extra_style = function(el){\n",
       "    var fig = this\n",
       "    el.on(\"remove\", function(){\n",
       "\tfig.close_ws(fig, {});\n",
       "    });\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._canvas_extra_style = function(el){\n",
       "    // this is important to make the div 'focusable\n",
       "    el.attr('tabindex', 0)\n",
       "    // reach out to IPython and tell the keyboard manager to turn it's self\n",
       "    // off when our div gets focus\n",
       "\n",
       "    // location in version 3\n",
       "    if (IPython.notebook.keyboard_manager) {\n",
       "        IPython.notebook.keyboard_manager.register_events(el);\n",
       "    }\n",
       "    else {\n",
       "        // location in version 2\n",
       "        IPython.keyboard_manager.register_events(el);\n",
       "    }\n",
       "\n",
       "}\n",
       "\n",
       "mpl.figure.prototype._key_event_extra = function(event, name) {\n",
       "    var manager = IPython.notebook.keyboard_manager;\n",
       "    if (!manager)\n",
       "        manager = IPython.keyboard_manager;\n",
       "\n",
       "    // Check for shift+enter\n",
       "    if (event.shiftKey && event.which == 13) {\n",
       "        this.canvas_div.blur();\n",
       "        // select the cell after this one\n",
       "        var index = IPython.notebook.find_cell_index(this.cell_info[0]);\n",
       "        IPython.notebook.select(index + 1);\n",
       "    }\n",
       "}\n",
       "\n",
       "mpl.figure.prototype.handle_save = function(fig, msg) {\n",
       "    fig.ondownload(fig, null);\n",
       "}\n",
       "\n",
       "\n",
       "mpl.find_output_cell = function(html_output) {\n",
       "    // Return the cell and output element which can be found *uniquely* in the notebook.\n",
       "    // Note - this is a bit hacky, but it is done because the \"notebook_saving.Notebook\"\n",
       "    // IPython event is triggered only after the cells have been serialised, which for\n",
       "    // our purposes (turning an active figure into a static one), is too late.\n",
       "    var cells = IPython.notebook.get_cells();\n",
       "    var ncells = cells.length;\n",
       "    for (var i=0; i<ncells; i++) {\n",
       "        var cell = cells[i];\n",
       "        if (cell.cell_type === 'code'){\n",
       "            for (var j=0; j<cell.output_area.outputs.length; j++) {\n",
       "                var data = cell.output_area.outputs[j];\n",
       "                if (data.data) {\n",
       "                    // IPython >= 3 moved mimebundle to data attribute of output\n",
       "                    data = data.data;\n",
       "                }\n",
       "                if (data['text/html'] == html_output) {\n",
       "                    return [cell, data, j];\n",
       "                }\n",
       "            }\n",
       "        }\n",
       "    }\n",
       "}\n",
       "\n",
       "// Register the function which deals with the matplotlib target/channel.\n",
       "// The kernel may be null if the page has been refreshed.\n",
       "if (IPython.notebook.kernel != null) {\n",
       "    IPython.notebook.kernel.comm_manager.register_target('matplotlib', mpl.mpl_figure_comm);\n",
       "}\n"
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<img src=\"data:image/png;base64,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\" width=\"640\">"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib notebook\n",
    "# Comparing Bidirectional LSTM merge modes\n",
    "\n",
    "n_timesteps = 10\n",
    "results = DataFrame()\n",
    "# sum merge\n",
    "model = get_bi_lstm_model(n_timesteps, 'sum')\n",
    "results['bilstm_sum'] = train_model(model, n_timesteps)\n",
    "# mul merge\n",
    "model = get_bi_lstm_model(n_timesteps, 'mul')\n",
    "results['bilstm_mul'] = train_model(model, n_timesteps)\n",
    "# avg merge\n",
    "model = get_bi_lstm_model(n_timesteps, 'ave')\n",
    "results['bilstm_ave'] = train_model(model, n_timesteps)\n",
    "# concat merge\n",
    "model = get_bi_lstm_model(n_timesteps, 'concat')\n",
    "results['bilstm_con'] = train_model(model, n_timesteps)\n",
    "# line plot of results\n",
    "results.plot()\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [conda root]",
   "language": "python",
   "name": "conda-root-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}