sensor_fusion_nn.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(142, 16) (142, 4) (42, 16) (42, 4)\n"
     ]
    }
   ],
   "source": [
    "data_x = np.loadtxt('data/input_x.csv', dtype=float)\n",
    "data_y = np.loadtxt('data/input_y.csv', dtype=float)\n",
    "\n",
    "data_x_test = np.loadtxt('data/input_x_test.csv', dtype=float)\n",
    "data_y_test = np.loadtxt('data/input_y_test.csv', dtype=float)\n",
    "\n",
    "print(data_x.shape, data_y.shape, data_x_test.shape, data_y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 1. 1. 1. 0.] [0. 0. 0. 1.]\n"
     ]
    }
   ],
   "source": [
    "print(data_x[5], data_y[5])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /home/darshanakg/.conda/envs/tf_gpu/lib/python3.7/site-packages/tensorflow/python/ops/resource_variable_ops.py:435: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Colocations handled automatically by placer.\n",
      "WARNING:tensorflow:From /home/darshanakg/.conda/envs/tf_gpu/lib/python3.7/site-packages/tensorflow/python/keras/layers/core.py:143: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n"
     ]
    }
   ],
   "source": [
    "model = tf.keras.models.Sequential([\n",
    "    tf.keras.layers.Dense(16, activation=tf.nn.relu, input_shape=(16,)),\n",
    "    tf.keras.layers.Dropout(0.2),\n",
    "    tf.keras.layers.Dense(4, activation=tf.nn.softmax)\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense (Dense)                (None, 16)                272       \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 16)                0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 4)                 68        \n",
      "=================================================================\n",
      "Total params: 340\n",
      "Trainable params: 340\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/300\n",
      "142/142 [==============================] - 0s 668us/sample - loss: 0.0219 - acc: 0.9648\n",
      "Epoch 2/300\n",
      "142/142 [==============================] - 0s 67us/sample - loss: 0.0196 - acc: 0.9648\n",
      "Epoch 3/300\n",
      "142/142 [==============================] - 0s 77us/sample - loss: 0.0192 - acc: 0.9648\n",
      "Epoch 4/300\n",
      "142/142 [==============================] - 0s 74us/sample - loss: 0.0141 - acc: 0.9859\n",
      "Epoch 5/300\n",
      "142/142 [==============================] - 0s 79us/sample - loss: 0.0121 - acc: 0.9859\n",
      "Epoch 6/300\n",
      "142/142 [==============================] - 0s 90us/sample - loss: 0.0149 - acc: 0.9718\n",
      "Epoch 7/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0177 - acc: 0.9718\n",
      "Epoch 8/300\n",
      "142/142 [==============================] - 0s 82us/sample - loss: 0.0189 - acc: 0.9648\n",
      "Epoch 9/300\n",
      "142/142 [==============================] - 0s 91us/sample - loss: 0.0142 - acc: 0.9930\n",
      "Epoch 10/300\n",
      "142/142 [==============================] - 0s 84us/sample - loss: 0.0137 - acc: 0.9789\n",
      "Epoch 11/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0149 - acc: 0.9648\n",
      "Epoch 12/300\n",
      "142/142 [==============================] - 0s 83us/sample - loss: 0.0133 - acc: 0.9789\n",
      "Epoch 13/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0166 - acc: 0.9789\n",
      "Epoch 14/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0132 - acc: 0.9859\n",
      "Epoch 15/300\n",
      "142/142 [==============================] - 0s 77us/sample - loss: 0.0115 - acc: 0.9789\n",
      "Epoch 16/300\n",
      "142/142 [==============================] - 0s 81us/sample - loss: 0.0149 - acc: 0.9789\n",
      "Epoch 17/300\n",
      "142/142 [==============================] - 0s 93us/sample - loss: 0.0157 - acc: 0.9789\n",
      "Epoch 18/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0176 - acc: 0.9648\n",
      "Epoch 19/300\n",
      "142/142 [==============================] - 0s 79us/sample - loss: 0.0152 - acc: 0.9718\n",
      "Epoch 20/300\n",
      "142/142 [==============================] - 0s 94us/sample - loss: 0.0153 - acc: 0.9859\n",
      "Epoch 21/300\n",
      "142/142 [==============================] - 0s 118us/sample - loss: 0.0129 - acc: 0.9859\n",
      "Epoch 22/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0135 - acc: 0.9859\n",
      "Epoch 23/300\n",
      "142/142 [==============================] - 0s 80us/sample - loss: 0.0115 - acc: 0.9930\n",
      "Epoch 24/300\n",
      "142/142 [==============================] - 0s 77us/sample - loss: 0.0130 - acc: 0.9718\n",
      "Epoch 25/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0125 - acc: 1.0000\n",
      "Epoch 26/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0154 - acc: 0.9648\n",
      "Epoch 27/300\n",
      "142/142 [==============================] - 0s 72us/sample - loss: 0.0119 - acc: 0.9859\n",
      "Epoch 28/300\n",
      "142/142 [==============================] - 0s 67us/sample - loss: 0.0134 - acc: 0.9718\n",
      "Epoch 29/300\n",
      "142/142 [==============================] - 0s 68us/sample - loss: 0.0098 - acc: 0.9930\n",
      "Epoch 30/300\n",
      "142/142 [==============================] - 0s 81us/sample - loss: 0.0169 - acc: 0.9577\n",
      "Epoch 31/300\n",
      "142/142 [==============================] - 0s 87us/sample - loss: 0.0131 - acc: 0.9930\n",
      "Epoch 32/300\n",
      "142/142 [==============================] - 0s 93us/sample - loss: 0.0128 - acc: 0.9718\n",
      "Epoch 33/300\n",
      "142/142 [==============================] - 0s 82us/sample - loss: 0.0114 - acc: 1.0000\n",
      "Epoch 34/300\n",
      "142/142 [==============================] - 0s 84us/sample - loss: 0.0133 - acc: 0.9789\n",
      "Epoch 35/300\n",
      "142/142 [==============================] - 0s 69us/sample - loss: 0.0123 - acc: 0.9859\n",
      "Epoch 36/300\n",
      "142/142 [==============================] - 0s 68us/sample - loss: 0.0128 - acc: 0.9789\n",
      "Epoch 37/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0113 - acc: 0.9859\n",
      "Epoch 38/300\n",
      "142/142 [==============================] - 0s 89us/sample - loss: 0.0088 - acc: 0.9930\n",
      "Epoch 39/300\n",
      "142/142 [==============================] - 0s 98us/sample - loss: 0.0114 - acc: 0.9859\n",
      "Epoch 40/300\n",
      "142/142 [==============================] - 0s 79us/sample - loss: 0.0123 - acc: 0.9789\n",
      "Epoch 41/300\n",
      "142/142 [==============================] - 0s 95us/sample - loss: 0.0117 - acc: 0.9859\n",
      "Epoch 42/300\n",
      "142/142 [==============================] - 0s 84us/sample - loss: 0.0130 - acc: 0.9859\n",
      "Epoch 43/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0133 - acc: 0.9859\n",
      "Epoch 44/300\n",
      "142/142 [==============================] - 0s 92us/sample - loss: 0.0110 - acc: 0.9930\n",
      "Epoch 45/300\n",
      "142/142 [==============================] - 0s 105us/sample - loss: 0.0122 - acc: 0.9648\n",
      "Epoch 46/300\n",
      "142/142 [==============================] - 0s 88us/sample - loss: 0.0138 - acc: 0.9789\n",
      "Epoch 47/300\n",
      "142/142 [==============================] - 0s 84us/sample - loss: 0.0097 - acc: 0.9859\n",
      "Epoch 48/300\n",
      "142/142 [==============================] - 0s 81us/sample - loss: 0.0118 - acc: 0.9859\n",
      "Epoch 49/300\n",
      "142/142 [==============================] - 0s 86us/sample - loss: 0.0112 - acc: 0.9859\n",
      "Epoch 50/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0104 - acc: 0.9859\n",
      "Epoch 51/300\n",
      "142/142 [==============================] - 0s 71us/sample - loss: 0.0067 - acc: 0.9930\n",
      "Epoch 52/300\n",
      "142/142 [==============================] - ETA: 0s - loss: 0.0081 - acc: 0.968 - 0s 84us/sample - loss: 0.0142 - acc: 0.9789\n",
      "Epoch 53/300\n",
      "142/142 [==============================] - 0s 110us/sample - loss: 0.0119 - acc: 0.9859\n",
      "Epoch 54/300\n",
      "142/142 [==============================] - 0s 83us/sample - loss: 0.0121 - acc: 0.9859\n",
      "Epoch 55/300\n",
      "142/142 [==============================] - 0s 83us/sample - loss: 0.0121 - acc: 0.9859\n",
      "Epoch 56/300\n",
      "142/142 [==============================] - 0s 90us/sample - loss: 0.0096 - acc: 0.9930\n",
      "Epoch 57/300\n",
      "142/142 [==============================] - 0s 68us/sample - loss: 0.0065 - acc: 0.9930\n",
      "Epoch 58/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0068 - acc: 1.0000\n",
      "Epoch 59/300\n",
      "142/142 [==============================] - 0s 78us/sample - loss: 0.0106 - acc: 0.9930\n",
      "Epoch 60/300\n",
      "142/142 [==============================] - 0s 77us/sample - loss: 0.0111 - acc: 0.9859\n",
      "Epoch 61/300\n",
      "142/142 [==============================] - 0s 97us/sample - loss: 0.0076 - acc: 0.9930\n",
      "Epoch 62/300\n",
      "142/142 [==============================] - 0s 67us/sample - loss: 0.0082 - acc: 1.0000\n",
      "Epoch 63/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0060 - acc: 1.0000\n",
      "Epoch 64/300\n",
      "142/142 [==============================] - 0s 83us/sample - loss: 0.0076 - acc: 0.9930\n",
      "Epoch 65/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0095 - acc: 0.9930\n",
      "Epoch 66/300\n",
      "142/142 [==============================] - 0s 67us/sample - loss: 0.0070 - acc: 1.0000\n",
      "Epoch 67/300\n",
      "142/142 [==============================] - 0s 91us/sample - loss: 0.0086 - acc: 0.9930\n",
      "Epoch 68/300\n",
      "142/142 [==============================] - 0s 80us/sample - loss: 0.0066 - acc: 1.0000\n",
      "Epoch 69/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0087 - acc: 0.9930\n",
      "Epoch 70/300\n",
      "142/142 [==============================] - 0s 81us/sample - loss: 0.0118 - acc: 0.9789\n",
      "Epoch 71/300\n",
      "142/142 [==============================] - 0s 72us/sample - loss: 0.0080 - acc: 1.0000\n",
      "Epoch 72/300\n",
      "142/142 [==============================] - 0s 74us/sample - loss: 0.0114 - acc: 0.9859\n",
      "Epoch 73/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0053 - acc: 1.0000\n",
      "Epoch 74/300\n",
      "142/142 [==============================] - 0s 72us/sample - loss: 0.0105 - acc: 0.9859\n",
      "Epoch 75/300\n",
      "142/142 [==============================] - 0s 70us/sample - loss: 0.0090 - acc: 0.9859\n",
      "Epoch 76/300\n",
      "142/142 [==============================] - 0s 80us/sample - loss: 0.0099 - acc: 0.9859\n",
      "Epoch 77/300\n",
      "142/142 [==============================] - 0s 89us/sample - loss: 0.0093 - acc: 0.9859\n",
      "Epoch 78/300\n",
      "142/142 [==============================] - 0s 80us/sample - loss: 0.0112 - acc: 0.9859\n",
      "Epoch 79/300\n",
      "142/142 [==============================] - 0s 77us/sample - loss: 0.0074 - acc: 0.9859\n",
      "Epoch 80/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0075 - acc: 0.9930\n",
      "Epoch 81/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0084 - acc: 1.0000\n",
      "Epoch 82/300\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "142/142 [==============================] - 0s 85us/sample - loss: 0.0092 - acc: 0.9930\n",
      "Epoch 83/300\n",
      "142/142 [==============================] - 0s 92us/sample - loss: 0.0077 - acc: 0.9930\n",
      "Epoch 84/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0075 - acc: 0.9930\n",
      "Epoch 85/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0054 - acc: 1.0000\n",
      "Epoch 86/300\n",
      "142/142 [==============================] - 0s 86us/sample - loss: 0.0068 - acc: 0.9859\n",
      "Epoch 87/300\n",
      "142/142 [==============================] - 0s 70us/sample - loss: 0.0086 - acc: 0.9859\n",
      "Epoch 88/300\n",
      "142/142 [==============================] - 0s 71us/sample - loss: 0.0089 - acc: 0.9930\n",
      "Epoch 89/300\n",
      "142/142 [==============================] - 0s 84us/sample - loss: 0.0068 - acc: 1.0000\n",
      "Epoch 90/300\n",
      "142/142 [==============================] - 0s 68us/sample - loss: 0.0066 - acc: 0.9930\n",
      "Epoch 91/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0086 - acc: 0.9930\n",
      "Epoch 92/300\n",
      "142/142 [==============================] - 0s 70us/sample - loss: 0.0078 - acc: 0.9930\n",
      "Epoch 93/300\n",
      "142/142 [==============================] - 0s 79us/sample - loss: 0.0063 - acc: 1.0000\n",
      "Epoch 94/300\n",
      "142/142 [==============================] - 0s 74us/sample - loss: 0.0099 - acc: 0.9789\n",
      "Epoch 95/300\n",
      "142/142 [==============================] - 0s 74us/sample - loss: 0.0062 - acc: 0.9930\n",
      "Epoch 96/300\n",
      "142/142 [==============================] - 0s 78us/sample - loss: 0.0088 - acc: 0.9930\n",
      "Epoch 97/300\n",
      "142/142 [==============================] - 0s 71us/sample - loss: 0.0093 - acc: 0.9930\n",
      "Epoch 98/300\n",
      "142/142 [==============================] - 0s 101us/sample - loss: 0.0116 - acc: 0.9718\n",
      "Epoch 99/300\n",
      "142/142 [==============================] - 0s 67us/sample - loss: 0.0055 - acc: 1.0000\n",
      "Epoch 100/300\n",
      "142/142 [==============================] - 0s 72us/sample - loss: 0.0060 - acc: 0.9859\n",
      "Epoch 101/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0061 - acc: 0.9930\n",
      "Epoch 102/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0063 - acc: 1.0000\n",
      "Epoch 103/300\n",
      "142/142 [==============================] - 0s 79us/sample - loss: 0.0071 - acc: 1.0000\n",
      "Epoch 104/300\n",
      "142/142 [==============================] - 0s 68us/sample - loss: 0.0084 - acc: 1.0000\n",
      "Epoch 105/300\n",
      "142/142 [==============================] - 0s 78us/sample - loss: 0.0067 - acc: 0.9930\n",
      "Epoch 106/300\n",
      "142/142 [==============================] - 0s 67us/sample - loss: 0.0078 - acc: 1.0000\n",
      "Epoch 107/300\n",
      "142/142 [==============================] - 0s 71us/sample - loss: 0.0045 - acc: 1.0000\n",
      "Epoch 108/300\n",
      "142/142 [==============================] - 0s 78us/sample - loss: 0.0071 - acc: 1.0000\n",
      "Epoch 109/300\n",
      "142/142 [==============================] - 0s 72us/sample - loss: 0.0074 - acc: 0.9930\n",
      "Epoch 110/300\n",
      "142/142 [==============================] - 0s 63us/sample - loss: 0.0066 - acc: 1.0000\n",
      "Epoch 111/300\n",
      "142/142 [==============================] - 0s 110us/sample - loss: 0.0069 - acc: 0.9859\n",
      "Epoch 112/300\n",
      "142/142 [==============================] - 0s 67us/sample - loss: 0.0063 - acc: 1.0000\n",
      "Epoch 113/300\n",
      "142/142 [==============================] - 0s 69us/sample - loss: 0.0082 - acc: 0.9930\n",
      "Epoch 114/300\n",
      "142/142 [==============================] - 0s 83us/sample - loss: 0.0115 - acc: 0.9859\n",
      "Epoch 115/300\n",
      "142/142 [==============================] - 0s 65us/sample - loss: 0.0066 - acc: 0.9930\n",
      "Epoch 116/300\n",
      "142/142 [==============================] - 0s 69us/sample - loss: 0.0057 - acc: 0.9930\n",
      "Epoch 117/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0073 - acc: 1.0000\n",
      "Epoch 118/300\n",
      "142/142 [==============================] - 0s 69us/sample - loss: 0.0079 - acc: 0.9859\n",
      "Epoch 119/300\n",
      "142/142 [==============================] - 0s 65us/sample - loss: 0.0077 - acc: 0.9930\n",
      "Epoch 120/300\n",
      "142/142 [==============================] - 0s 74us/sample - loss: 0.0065 - acc: 0.9930\n",
      "Epoch 121/300\n",
      "142/142 [==============================] - 0s 82us/sample - loss: 0.0044 - acc: 1.0000\n",
      "Epoch 122/300\n",
      "142/142 [==============================] - 0s 74us/sample - loss: 0.0060 - acc: 1.0000\n",
      "Epoch 123/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0083 - acc: 0.9859\n",
      "Epoch 124/300\n",
      "142/142 [==============================] - 0s 72us/sample - loss: 0.0045 - acc: 1.0000\n",
      "Epoch 125/300\n",
      "142/142 [==============================] - 0s 71us/sample - loss: 0.0065 - acc: 0.9930\n",
      "Epoch 126/300\n",
      "142/142 [==============================] - 0s 74us/sample - loss: 0.0041 - acc: 1.0000\n",
      "Epoch 127/300\n",
      "142/142 [==============================] - 0s 72us/sample - loss: 0.0058 - acc: 0.9930\n",
      "Epoch 128/300\n",
      "142/142 [==============================] - 0s 77us/sample - loss: 0.0045 - acc: 0.9930\n",
      "Epoch 129/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0039 - acc: 1.0000\n",
      "Epoch 130/300\n",
      "142/142 [==============================] - 0s 72us/sample - loss: 0.0065 - acc: 1.0000\n",
      "Epoch 131/300\n",
      "142/142 [==============================] - 0s 72us/sample - loss: 0.0069 - acc: 0.9930\n",
      "Epoch 132/300\n",
      "142/142 [==============================] - 0s 72us/sample - loss: 0.0065 - acc: 0.9859\n",
      "Epoch 133/300\n",
      "142/142 [==============================] - 0s 69us/sample - loss: 0.0098 - acc: 0.9859\n",
      "Epoch 134/300\n",
      "142/142 [==============================] - 0s 65us/sample - loss: 0.0065 - acc: 0.9930\n",
      "Epoch 135/300\n",
      "142/142 [==============================] - 0s 70us/sample - loss: 0.0080 - acc: 0.9930\n",
      "Epoch 136/300\n",
      "142/142 [==============================] - 0s 82us/sample - loss: 0.0054 - acc: 1.0000\n",
      "Epoch 137/300\n",
      "142/142 [==============================] - 0s 72us/sample - loss: 0.0065 - acc: 0.9859\n",
      "Epoch 138/300\n",
      "142/142 [==============================] - 0s 66us/sample - loss: 0.0039 - acc: 1.0000\n",
      "Epoch 139/300\n",
      "142/142 [==============================] - 0s 72us/sample - loss: 0.0052 - acc: 0.9859\n",
      "Epoch 140/300\n",
      "142/142 [==============================] - 0s 74us/sample - loss: 0.0030 - acc: 1.0000\n",
      "Epoch 141/300\n",
      "142/142 [==============================] - 0s 72us/sample - loss: 0.0044 - acc: 0.9930\n",
      "Epoch 142/300\n",
      "142/142 [==============================] - 0s 67us/sample - loss: 0.0061 - acc: 0.9930\n",
      "Epoch 143/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0062 - acc: 0.9930\n",
      "Epoch 144/300\n",
      "142/142 [==============================] - 0s 68us/sample - loss: 0.0039 - acc: 1.0000\n",
      "Epoch 145/300\n",
      "142/142 [==============================] - 0s 68us/sample - loss: 0.0069 - acc: 0.9930\n",
      "Epoch 146/300\n",
      "142/142 [==============================] - 0s 69us/sample - loss: 0.0055 - acc: 1.0000\n",
      "Epoch 147/300\n",
      "142/142 [==============================] - 0s 78us/sample - loss: 0.0086 - acc: 0.9859\n",
      "Epoch 148/300\n",
      "142/142 [==============================] - 0s 80us/sample - loss: 0.0036 - acc: 1.0000\n",
      "Epoch 149/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0049 - acc: 0.9930\n",
      "Epoch 150/300\n",
      "142/142 [==============================] - 0s 68us/sample - loss: 0.0051 - acc: 0.9859\n",
      "Epoch 151/300\n",
      "142/142 [==============================] - 0s 71us/sample - loss: 0.0065 - acc: 0.9859\n",
      "Epoch 152/300\n",
      "142/142 [==============================] - 0s 83us/sample - loss: 0.0053 - acc: 0.9930\n",
      "Epoch 153/300\n",
      "142/142 [==============================] - 0s 69us/sample - loss: 0.0050 - acc: 1.0000\n",
      "Epoch 154/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0067 - acc: 0.9930\n",
      "Epoch 155/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 156/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0037 - acc: 1.0000\n",
      "Epoch 157/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0079 - acc: 0.9789\n",
      "Epoch 158/300\n",
      "142/142 [==============================] - 0s 79us/sample - loss: 0.0088 - acc: 0.9859\n",
      "Epoch 159/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0053 - acc: 0.9930\n",
      "Epoch 160/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0060 - acc: 1.0000\n",
      "Epoch 161/300\n",
      "142/142 [==============================] - 0s 66us/sample - loss: 0.0055 - acc: 0.9859\n",
      "Epoch 162/300\n",
      "142/142 [==============================] - 0s 68us/sample - loss: 0.0047 - acc: 1.0000\n",
      "Epoch 163/300\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "142/142 [==============================] - 0s 86us/sample - loss: 0.0054 - acc: 1.0000\n",
      "Epoch 164/300\n",
      "142/142 [==============================] - 0s 84us/sample - loss: 0.0068 - acc: 0.9930\n",
      "Epoch 165/300\n",
      "142/142 [==============================] - 0s 65us/sample - loss: 0.0085 - acc: 0.9789\n",
      "Epoch 166/300\n",
      "142/142 [==============================] - 0s 67us/sample - loss: 0.0062 - acc: 0.9859\n",
      "Epoch 167/300\n",
      "142/142 [==============================] - 0s 70us/sample - loss: 0.0075 - acc: 0.9789\n",
      "Epoch 168/300\n",
      "142/142 [==============================] - 0s 86us/sample - loss: 0.0047 - acc: 0.9930\n",
      "Epoch 169/300\n",
      "142/142 [==============================] - 0s 72us/sample - loss: 0.0039 - acc: 1.0000\n",
      "Epoch 170/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0063 - acc: 0.9930\n",
      "Epoch 171/300\n",
      "142/142 [==============================] - 0s 74us/sample - loss: 0.0049 - acc: 1.0000\n",
      "Epoch 172/300\n",
      "142/142 [==============================] - 0s 74us/sample - loss: 0.0035 - acc: 1.0000\n",
      "Epoch 173/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0058 - acc: 0.9930\n",
      "Epoch 174/300\n",
      "142/142 [==============================] - 0s 64us/sample - loss: 0.0046 - acc: 0.9930\n",
      "Epoch 175/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0036 - acc: 1.0000\n",
      "Epoch 176/300\n",
      "142/142 [==============================] - 0s 74us/sample - loss: 0.0057 - acc: 0.9930\n",
      "Epoch 177/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0036 - acc: 1.0000\n",
      "Epoch 178/300\n",
      "142/142 [==============================] - 0s 72us/sample - loss: 0.0066 - acc: 0.9859\n",
      "Epoch 179/300\n",
      "142/142 [==============================] - 0s 72us/sample - loss: 0.0033 - acc: 1.0000\n",
      "Epoch 180/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0029 - acc: 1.0000\n",
      "Epoch 181/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0038 - acc: 1.0000\n",
      "Epoch 182/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0034 - acc: 1.0000\n",
      "Epoch 183/300\n",
      "142/142 [==============================] - 0s 78us/sample - loss: 0.0036 - acc: 0.9930\n",
      "Epoch 184/300\n",
      "142/142 [==============================] - 0s 90us/sample - loss: 0.0072 - acc: 0.9859\n",
      "Epoch 185/300\n",
      "142/142 [==============================] - 0s 77us/sample - loss: 0.0064 - acc: 0.9859\n",
      "Epoch 186/300\n",
      "142/142 [==============================] - 0s 67us/sample - loss: 0.0038 - acc: 1.0000\n",
      "Epoch 187/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0048 - acc: 0.9930\n",
      "Epoch 188/300\n",
      "142/142 [==============================] - 0s 78us/sample - loss: 0.0041 - acc: 0.9930\n",
      "Epoch 189/300\n",
      "142/142 [==============================] - 0s 85us/sample - loss: 0.0064 - acc: 0.9930\n",
      "Epoch 190/300\n",
      "142/142 [==============================] - 0s 79us/sample - loss: 0.0043 - acc: 1.0000\n",
      "Epoch 191/300\n",
      "142/142 [==============================] - 0s 68us/sample - loss: 0.0044 - acc: 1.0000\n",
      "Epoch 192/300\n",
      "142/142 [==============================] - 0s 63us/sample - loss: 0.0043 - acc: 0.9930\n",
      "Epoch 193/300\n",
      "142/142 [==============================] - 0s 74us/sample - loss: 0.0061 - acc: 0.9930\n",
      "Epoch 194/300\n",
      "142/142 [==============================] - 0s 70us/sample - loss: 0.0039 - acc: 1.0000\n",
      "Epoch 195/300\n",
      "142/142 [==============================] - 0s 66us/sample - loss: 0.0037 - acc: 1.0000\n",
      "Epoch 196/300\n",
      "142/142 [==============================] - 0s 77us/sample - loss: 0.0040 - acc: 1.0000\n",
      "Epoch 197/300\n",
      "142/142 [==============================] - 0s 71us/sample - loss: 0.0045 - acc: 0.9930\n",
      "Epoch 198/300\n",
      "142/142 [==============================] - 0s 83us/sample - loss: 0.0044 - acc: 0.9930\n",
      "Epoch 199/300\n",
      "142/142 [==============================] - 0s 70us/sample - loss: 0.0046 - acc: 1.0000\n",
      "Epoch 200/300\n",
      "142/142 [==============================] - 0s 85us/sample - loss: 0.0046 - acc: 0.9930\n",
      "Epoch 201/300\n",
      "142/142 [==============================] - 0s 78us/sample - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 202/300\n",
      "142/142 [==============================] - 0s 87us/sample - loss: 0.0049 - acc: 0.9930\n",
      "Epoch 203/300\n",
      "142/142 [==============================] - 0s 72us/sample - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 204/300\n",
      "142/142 [==============================] - 0s 62us/sample - loss: 0.0055 - acc: 0.9930\n",
      "Epoch 205/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0034 - acc: 1.0000\n",
      "Epoch 206/300\n",
      "142/142 [==============================] - 0s 71us/sample - loss: 0.0038 - acc: 1.0000\n",
      "Epoch 207/300\n",
      "142/142 [==============================] - 0s 70us/sample - loss: 0.0045 - acc: 0.9930\n",
      "Epoch 208/300\n",
      "142/142 [==============================] - 0s 71us/sample - loss: 0.0031 - acc: 1.0000\n",
      "Epoch 209/300\n",
      "142/142 [==============================] - 0s 72us/sample - loss: 0.0047 - acc: 0.9930\n",
      "Epoch 210/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 211/300\n",
      "142/142 [==============================] - 0s 69us/sample - loss: 0.0065 - acc: 0.9859\n",
      "Epoch 212/300\n",
      "142/142 [==============================] - 0s 79us/sample - loss: 0.0051 - acc: 0.9930\n",
      "Epoch 213/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 214/300\n",
      "142/142 [==============================] - 0s 74us/sample - loss: 0.0045 - acc: 0.9930\n",
      "Epoch 215/300\n",
      "142/142 [==============================] - 0s 82us/sample - loss: 0.0020 - acc: 1.0000\n",
      "Epoch 216/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0045 - acc: 0.9930\n",
      "Epoch 217/300\n",
      "142/142 [==============================] - 0s 77us/sample - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 218/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 219/300\n",
      "142/142 [==============================] - 0s 89us/sample - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 220/300\n",
      "142/142 [==============================] - 0s 93us/sample - loss: 0.0040 - acc: 0.9930\n",
      "Epoch 221/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0050 - acc: 0.9930\n",
      "Epoch 222/300\n",
      "142/142 [==============================] - 0s 74us/sample - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 223/300\n",
      "142/142 [==============================] - 0s 71us/sample - loss: 0.0051 - acc: 0.9930\n",
      "Epoch 224/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0032 - acc: 1.0000\n",
      "Epoch 225/300\n",
      "142/142 [==============================] - 0s 70us/sample - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 226/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0052 - acc: 0.9930\n",
      "Epoch 227/300\n",
      "142/142 [==============================] - 0s 69us/sample - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 228/300\n",
      "142/142 [==============================] - 0s 67us/sample - loss: 0.0041 - acc: 1.0000\n",
      "Epoch 229/300\n",
      "142/142 [==============================] - 0s 74us/sample - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 230/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0044 - acc: 1.0000\n",
      "Epoch 231/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0035 - acc: 1.0000\n",
      "Epoch 232/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0053 - acc: 1.0000\n",
      "Epoch 233/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 234/300\n",
      "142/142 [==============================] - 0s 81us/sample - loss: 0.0038 - acc: 0.9930\n",
      "Epoch 235/300\n",
      "142/142 [==============================] - 0s 99us/sample - loss: 0.0040 - acc: 1.0000\n",
      "Epoch 236/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 237/300\n",
      "142/142 [==============================] - 0s 67us/sample - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 238/300\n",
      "142/142 [==============================] - 0s 80us/sample - loss: 0.0025 - acc: 0.9930\n",
      "Epoch 239/300\n",
      "142/142 [==============================] - 0s 87us/sample - loss: 0.0041 - acc: 1.0000\n",
      "Epoch 240/300\n",
      "142/142 [==============================] - 0s 69us/sample - loss: 0.0036 - acc: 1.0000\n",
      "Epoch 241/300\n",
      "142/142 [==============================] - 0s 65us/sample - loss: 0.0049 - acc: 0.9930\n",
      "Epoch 242/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0040 - acc: 1.0000\n",
      "Epoch 243/300\n",
      "142/142 [==============================] - 0s 77us/sample - loss: 0.0043 - acc: 0.9930\n",
      "Epoch 244/300\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0035 - acc: 1.0000\n",
      "Epoch 245/300\n",
      "142/142 [==============================] - 0s 65us/sample - loss: 0.0014 - acc: 1.0000\n",
      "Epoch 246/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0032 - acc: 0.9930\n",
      "Epoch 247/300\n",
      "142/142 [==============================] - 0s 77us/sample - loss: 0.0039 - acc: 0.9930\n",
      "Epoch 248/300\n",
      "142/142 [==============================] - 0s 80us/sample - loss: 0.0034 - acc: 1.0000\n",
      "Epoch 249/300\n",
      "142/142 [==============================] - 0s 65us/sample - loss: 0.0060 - acc: 0.9859\n",
      "Epoch 250/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0048 - acc: 1.0000\n",
      "Epoch 251/300\n",
      "142/142 [==============================] - 0s 79us/sample - loss: 0.0040 - acc: 1.0000\n",
      "Epoch 252/300\n",
      "142/142 [==============================] - 0s 80us/sample - loss: 0.0028 - acc: 1.0000\n",
      "Epoch 253/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0038 - acc: 1.0000\n",
      "Epoch 254/300\n",
      "142/142 [==============================] - 0s 68us/sample - loss: 0.0025 - acc: 0.9930\n",
      "Epoch 255/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0037 - acc: 1.0000\n",
      "Epoch 256/300\n",
      "142/142 [==============================] - 0s 78us/sample - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 257/300\n",
      "142/142 [==============================] - 0s 68us/sample - loss: 0.0017 - acc: 1.0000\n",
      "Epoch 258/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0045 - acc: 0.9930\n",
      "Epoch 259/300\n",
      "142/142 [==============================] - 0s 70us/sample - loss: 0.0047 - acc: 0.9930\n",
      "Epoch 260/300\n",
      "142/142 [==============================] - 0s 74us/sample - loss: 0.0044 - acc: 1.0000\n",
      "Epoch 261/300\n",
      "142/142 [==============================] - 0s 67us/sample - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 262/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0040 - acc: 1.0000\n",
      "Epoch 263/300\n",
      "142/142 [==============================] - 0s 72us/sample - loss: 0.0027 - acc: 1.0000\n",
      "Epoch 264/300\n",
      "142/142 [==============================] - 0s 69us/sample - loss: 0.0038 - acc: 1.0000\n",
      "Epoch 265/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0026 - acc: 1.0000\n",
      "Epoch 266/300\n",
      "142/142 [==============================] - 0s 78us/sample - loss: 0.0029 - acc: 1.0000\n",
      "Epoch 267/300\n",
      "142/142 [==============================] - 0s 84us/sample - loss: 0.0029 - acc: 0.9930\n",
      "Epoch 268/300\n",
      "142/142 [==============================] - 0s 67us/sample - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 269/300\n",
      "142/142 [==============================] - 0s 77us/sample - loss: 0.0036 - acc: 0.9930\n",
      "Epoch 270/300\n",
      "142/142 [==============================] - 0s 71us/sample - loss: 0.0050 - acc: 0.9859\n",
      "Epoch 271/300\n",
      "142/142 [==============================] - 0s 77us/sample - loss: 0.0054 - acc: 0.9930\n",
      "Epoch 272/300\n",
      "142/142 [==============================] - 0s 84us/sample - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 273/300\n",
      "142/142 [==============================] - 0s 69us/sample - loss: 0.0051 - acc: 0.9930\n",
      "Epoch 274/300\n",
      "142/142 [==============================] - 0s 82us/sample - loss: 0.0054 - acc: 0.9930\n",
      "Epoch 275/300\n",
      "142/142 [==============================] - 0s 77us/sample - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 276/300\n",
      "142/142 [==============================] - 0s 81us/sample - loss: 0.0019 - acc: 1.0000\n",
      "Epoch 277/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0015 - acc: 1.0000\n",
      "Epoch 278/300\n",
      "142/142 [==============================] - 0s 70us/sample - loss: 0.0025 - acc: 1.0000\n",
      "Epoch 279/300\n",
      "142/142 [==============================] - 0s 80us/sample - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 280/300\n",
      "142/142 [==============================] - 0s 78us/sample - loss: 0.0036 - acc: 0.9930\n",
      "Epoch 281/300\n",
      "142/142 [==============================] - 0s 71us/sample - loss: 0.0041 - acc: 0.9930\n",
      "Epoch 282/300\n",
      "142/142 [==============================] - 0s 70us/sample - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 283/300\n",
      "142/142 [==============================] - 0s 71us/sample - loss: 0.0015 - acc: 1.0000\n",
      "Epoch 284/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0031 - acc: 0.9930\n",
      "Epoch 285/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 286/300\n",
      "142/142 [==============================] - 0s 63us/sample - loss: 0.0015 - acc: 1.0000\n",
      "Epoch 287/300\n",
      "142/142 [==============================] - 0s 78us/sample - loss: 0.0024 - acc: 1.0000\n",
      "Epoch 288/300\n",
      "142/142 [==============================] - 0s 76us/sample - loss: 0.0028 - acc: 0.9930\n",
      "Epoch 289/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 290/300\n",
      "142/142 [==============================] - 0s 75us/sample - loss: 0.0017 - acc: 1.0000\n",
      "Epoch 291/300\n",
      "142/142 [==============================] - 0s 67us/sample - loss: 0.0040 - acc: 0.9930\n",
      "Epoch 292/300\n",
      "142/142 [==============================] - 0s 73us/sample - loss: 0.0033 - acc: 0.9930\n",
      "Epoch 293/300\n",
      "142/142 [==============================] - 0s 61us/sample - loss: 0.0023 - acc: 1.0000\n",
      "Epoch 294/300\n",
      "142/142 [==============================] - 0s 68us/sample - loss: 0.0037 - acc: 1.0000\n",
      "Epoch 295/300\n",
      "142/142 [==============================] - 0s 68us/sample - loss: 0.0038 - acc: 0.9859\n",
      "Epoch 296/300\n",
      "142/142 [==============================] - 0s 65us/sample - loss: 0.0022 - acc: 1.0000\n",
      "Epoch 297/300\n",
      "142/142 [==============================] - 0s 69us/sample - loss: 0.0021 - acc: 1.0000\n",
      "Epoch 298/300\n",
      "142/142 [==============================] - 0s 65us/sample - loss: 0.0030 - acc: 1.0000\n",
      "Epoch 299/300\n",
      "142/142 [==============================] - 0s 61us/sample - loss: 0.0019 - acc: 1.0000\n",
      "Epoch 300/300\n",
      "142/142 [==============================] - 0s 79us/sample - loss: 0.0026 - acc: 0.9930\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x7f1d244f4e48>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.compile(optimizer='adam',\n",
    "              loss='mean_squared_error',\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "model.fit(data_x, data_y, epochs=300)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "42/42 [==============================] - 0s 944us/sample - loss: 0.0014 - acc: 1.0000\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.00143392366312799, 1.0]"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.evaluate(data_x_test, data_y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = model.predict(data_x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "command_classes = ['count', 'color', 'focus', 'no_op']\n",
    "words_file = open(\"data/words.txt\", \"r\")\n",
    "words = words_file.read().split('\\n')[:-1]\n",
    "\n",
    "def get_phrase(x):\n",
    "    phrase = \"\"\n",
    "    for i in range(len(words)):\n",
    "        if x[i] == 1:\n",
    "            phrase += words[i] + \" \"\n",
    "    return phrase"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('no_op', 'there color is this the book cup bottle phone '),\n",
       " ('no_op', 'how many are what color this book bottle pen '),\n",
       " ('no_op', 'how are there color this the cup laptop pen '),\n",
       " ('no_op', 'are highlight the cup phone '),\n",
       " ('no_op', 'what this the cup pen '),\n",
       " ('no_op', 'is laptop bottle pen '),\n",
       " ('no_op', 'what cup laptop bottle pen '),\n",
       " ('no_op', 'what color the book pen '),\n",
       " ('focus', 'highlight cup '),\n",
       " ('color', 'color book '),\n",
       " ('no_op', 'many are there what highlight the laptop bottle '),\n",
       " ('focus', 'highlight the pen '),\n",
       " ('focus', 'color the phone '),\n",
       " ('no_op', 'there what color is highlight the book pen phone '),\n",
       " ('no_op', 'are is the book cup bottle pen '),\n",
       " ('focus', 'highlight phone '),\n",
       " ('no_op', 'there what is this highlight the book cup laptop bottle '),\n",
       " ('no_op', 'book cup bottle '),\n",
       " ('no_op', 'there is this the book pen phone '),\n",
       " ('no_op', 'this highlight the pen '),\n",
       " ('no_op', 'how many are is cup laptop bottle pen '),\n",
       " ('no_op', 'there color is this the book laptop bottle '),\n",
       " ('count', 'how many book '),\n",
       " ('count', 'how many bottle '),\n",
       " ('no_op',\n",
       "  'there what color is this highlight the book cup laptop bottle pen '),\n",
       " ('count', 'how many cup '),\n",
       " ('no_op', 'is this book cup laptop bottle pen phone '),\n",
       " ('color', 'what color is this laptop '),\n",
       " ('no_op', 'how what color is highlight book cup laptop pen phone '),\n",
       " ('no_op', 'is this the laptop pen phone '),\n",
       " ('no_op', 'this laptop pen '),\n",
       " ('focus', 'highlight book '),\n",
       " ('no_op', 'are is the book laptop phone '),\n",
       " ('no_op', 'there the book cup bottle phone '),\n",
       " ('no_op', 'color is highlight the book laptop bottle phone '),\n",
       " ('no_op', 'this the cup bottle phone '),\n",
       " ('no_op', 'are there what is this the book laptop phone '),\n",
       " ('no_op', 'color cup bottle pen '),\n",
       " ('no_op', 'are color is the bottle pen '),\n",
       " ('focus', 'color the laptop '),\n",
       " ('no_op', 'how there what is this book cup '),\n",
       " ('no_op', 'there color is this the book cup pen phone '),\n",
       " ('no_op', 'how there this highlight bottle phone '),\n",
       " ('count', 'how many are there bottle '),\n",
       " ('no_op', 'many are there color is this highlight the bottle '),\n",
       " ('no_op', 'this the book laptop bottle pen '),\n",
       " ('no_op', 'are the cup laptop pen phone '),\n",
       " ('count', 'how many are there book '),\n",
       " ('focus', 'highlight the phone '),\n",
       " ('color', 'color cup '),\n",
       " ('focus', 'highlight pen '),\n",
       " ('no_op', 'what color this highlight the book bottle '),\n",
       " ('no_op',\n",
       "  'how many are there color is highlight book cup laptop bottle pen phone '),\n",
       " ('no_op', 'what is the book cup laptop bottle pen '),\n",
       " ('no_op', 'are there what is book pen '),\n",
       " ('no_op', 'there book cup pen phone '),\n",
       " ('no_op', 'what color book laptop pen '),\n",
       " ('no_op', 'this highlight the book pen phone '),\n",
       " ('no_op', 'what color this highlight book laptop bottle pen phone '),\n",
       " ('focus', 'highlight the laptop '),\n",
       " ('no_op', 'are color this the cup laptop bottle pen phone '),\n",
       " ('no_op', 'how what color cup bottle '),\n",
       " ('focus', 'color the book '),\n",
       " ('no_op', 'are this pen '),\n",
       " ('no_op', 'there is the laptop phone '),\n",
       " ('count', 'how many are there cup '),\n",
       " ('no_op', 'color this book bottle pen '),\n",
       " ('no_op', 'laptop '),\n",
       " ('count', 'how many laptop '),\n",
       " ('no_op', 'are there what color highlight bottle pen '),\n",
       " ('no_op', 'many are there what this the book laptop bottle pen phone '),\n",
       " ('focus', 'highlight the book '),\n",
       " ('no_op', 'book cup laptop phone '),\n",
       " ('focus', 'highlight the bottle '),\n",
       " ('no_op', 'there what color is bottle pen phone '),\n",
       " ('no_op', 'color this the book bottle pen '),\n",
       " ('no_op', 'what color the cup laptop pen '),\n",
       " ('no_op', 'this highlight the cup bottle pen phone '),\n",
       " ('no_op', 'is this highlight pen phone '),\n",
       " ('no_op', 'are color this book phone '),\n",
       " ('no_op', 'are is highlight the book cup '),\n",
       " ('no_op', 'how are this book laptop pen phone '),\n",
       " ('no_op', 'color this the pen '),\n",
       " ('no_op', 'there what is laptop bottle '),\n",
       " ('count', 'how many are there laptop '),\n",
       " ('no_op', 'is highlight cup bottle phone '),\n",
       " ('color', 'what color is this bottle '),\n",
       " ('no_op', 'color is the book laptop pen phone '),\n",
       " ('no_op', 'how are there color the bottle '),\n",
       " ('no_op', 'how many are there color is this highlight the laptop '),\n",
       " ('no_op', 'are there color this highlight the book phone '),\n",
       " ('no_op', 'how there this highlight book laptop pen phone '),\n",
       " ('color', 'what color is this cup '),\n",
       " ('no_op', 'are there color this the book laptop bottle '),\n",
       " ('color', 'color pen '),\n",
       " ('no_op', 'there what this highlight the book cup bottle pen '),\n",
       " ('focus', 'highlight laptop '),\n",
       " ('no_op', 'there what color this highlight the book bottle pen phone '),\n",
       " ('no_op', 'what color is the '),\n",
       " ('no_op', 'are highlight the laptop bottle pen '),\n",
       " ('focus', 'highlight bottle '),\n",
       " ('focus', 'color the bottle '),\n",
       " ('no_op', 'is the book cup laptop '),\n",
       " ('no_op', 'how are there highlight book pen phone '),\n",
       " ('no_op', 'is highlight book cup '),\n",
       " ('no_op', 'how many what highlight the cup laptop bottle '),\n",
       " ('no_op', 'how many color this highlight laptop '),\n",
       " ('no_op', 'color is laptop bottle pen phone '),\n",
       " ('no_op', 'is highlight the book cup '),\n",
       " ('no_op', 'what is the cup laptop bottle phone '),\n",
       " ('no_op', 'how there what this highlight cup bottle '),\n",
       " ('color', 'what color is this book '),\n",
       " ('no_op', 'there this book laptop bottle pen '),\n",
       " ('no_op', 'what color highlight laptop bottle pen phone '),\n",
       " ('color', 'color phone '),\n",
       " ('count', 'how many are there phone '),\n",
       " ('color', 'color bottle '),\n",
       " ('no_op', 'how are what is highlight the bottle '),\n",
       " ('no_op', 'there what color is this the bottle pen phone '),\n",
       " ('focus', 'color the pen '),\n",
       " ('no_op', 'there what is this the cup pen phone '),\n",
       " ('no_op', 'is the cup phone '),\n",
       " ('count', 'how many phone '),\n",
       " ('color', 'what color is this phone '),\n",
       " ('color', 'what color is this pen '),\n",
       " ('no_op', 'are there what color this book laptop bottle phone '),\n",
       " ('focus', 'highlight the cup '),\n",
       " ('count', 'how many are there pen '),\n",
       " ('no_op', 'how are there what is this highlight the cup pen phone '),\n",
       " ('no_op', 'there is this cup bottle pen phone '),\n",
       " ('no_op', 'how many are there what color highlight book laptop bottle pen '),\n",
       " ('no_op', 'there what color highlight the pen phone '),\n",
       " ('no_op', 'how what color is highlight the book cup phone '),\n",
       " ('no_op', 'there what color the book bottle pen '),\n",
       " ('no_op', 'there this the book cup laptop bottle phone '),\n",
       " ('color', 'color laptop '),\n",
       " ('count', 'how many pen '),\n",
       " ('no_op', 'the bottle pen phone '),\n",
       " ('no_op', 'the cup pen phone '),\n",
       " ('focus', 'color the cup '),\n",
       " ('no_op', 'there is cup pen '),\n",
       " ('no_op', 'how many are what highlight the book phone ')]"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[(command_classes[np.argmax(predictions[i])], get_phrase(data_x[i])) for i in range(len(predictions))]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save('data/sf_nn.h5')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}