04-BONUS-Multivariate-RNN.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "___\n",
    "\n",
    "<a href='http://www.pieriandata.com'><img src='Pierian_Data_Logo.png'/></a>\n",
    "___\n",
    "<center><em>Copyright Pierian Data</em></center>\n",
    "<center><em>For more information, visit us at <a href='http://www.pieriandata.com'>www.pieriandata.com</a></em></center>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Multivariate Time Series with RNN\n",
    "---\n",
    "----\n",
    "# PLEASE READ THE FOLLOWING BEFORE GOING THROUGH THIS NOTEBOOK!\n",
    "\n",
    "## Interested in Time Series Analysis with Python? [Check out my best-selling course!](https://www.udemy.com/python-for-time-series-data-analysis/?couponCode=TIME_SERIES_PYTHON).\n",
    "\n",
    "### [More Course Coupons Available Here](https://docs.google.com/document/d/1TfB48iJcCAPaQYiOgB70_QgBXyGaU33nYyUZ6lgtLl8/edit)\n",
    "---\n",
    "---\n",
    "Appliance Energy Usage: A Multivariate Time Series Forecasting Example\n",
    "\n",
    "Experimental data used to create regression models of appliances energy use in a low energy building.\n",
    "Data Set Information:\n",
    "\n",
    "The data set is at 10 min for about 4.5 months. The house temperature and humidity conditions were monitored with a ZigBee wireless sensor network. Each wireless node transmitted the temperature and humidity conditions around 3.3 min. Then, the wireless data was averaged for 10 minutes periods. The energy data was logged every 10 minutes with m-bus energy meters. Weather from the nearest airport weather station (Chievres Airport, Belgium) was downloaded from a public data set from Reliable Prognosis (rp5.ru), and merged together with the experimental data sets using the date and time column. Two random variables have been included in the data set for testing the regression models and to filter out non predictive attributes (parameters).\n",
    "Original source of the dataset:\n",
    "\n",
    "http://archive.ics.uci.edu/ml/datasets/Appliances+energy+prediction\n",
    "___\n",
    "___"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# IGNORE THE CONTENT OF THIS CELL\n",
    "# import tensorflow as tf\n",
    "# tf.compat.v1.disable_eager_execution()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data\n",
    "\n",
    "Let's read in the data set:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df = pd.read_csv('../DATA/energydata_complete.csv',index_col='date',\n",
    "                infer_datetime_format=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Appliances</th>\n",
       "      <th>lights</th>\n",
       "      <th>T1</th>\n",
       "      <th>RH_1</th>\n",
       "      <th>T2</th>\n",
       "      <th>RH_2</th>\n",
       "      <th>T3</th>\n",
       "      <th>RH_3</th>\n",
       "      <th>T4</th>\n",
       "      <th>RH_4</th>\n",
       "      <th>...</th>\n",
       "      <th>T9</th>\n",
       "      <th>RH_9</th>\n",
       "      <th>T_out</th>\n",
       "      <th>Press_mm_hg</th>\n",
       "      <th>RH_out</th>\n",
       "      <th>Windspeed</th>\n",
       "      <th>Visibility</th>\n",
       "      <th>Tdewpoint</th>\n",
       "      <th>rv1</th>\n",
       "      <th>rv2</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2016-01-11 17:00:00</th>\n",
       "      <td>60</td>\n",
       "      <td>30</td>\n",
       "      <td>19.89</td>\n",
       "      <td>47.596667</td>\n",
       "      <td>19.2</td>\n",
       "      <td>44.790000</td>\n",
       "      <td>19.79</td>\n",
       "      <td>44.730000</td>\n",
       "      <td>19.000000</td>\n",
       "      <td>45.566667</td>\n",
       "      <td>...</td>\n",
       "      <td>17.033333</td>\n",
       "      <td>45.53</td>\n",
       "      <td>6.600000</td>\n",
       "      <td>733.5</td>\n",
       "      <td>92.0</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>63.000000</td>\n",
       "      <td>5.3</td>\n",
       "      <td>13.275433</td>\n",
       "      <td>13.275433</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-01-11 17:10:00</th>\n",
       "      <td>60</td>\n",
       "      <td>30</td>\n",
       "      <td>19.89</td>\n",
       "      <td>46.693333</td>\n",
       "      <td>19.2</td>\n",
       "      <td>44.722500</td>\n",
       "      <td>19.79</td>\n",
       "      <td>44.790000</td>\n",
       "      <td>19.000000</td>\n",
       "      <td>45.992500</td>\n",
       "      <td>...</td>\n",
       "      <td>17.066667</td>\n",
       "      <td>45.56</td>\n",
       "      <td>6.483333</td>\n",
       "      <td>733.6</td>\n",
       "      <td>92.0</td>\n",
       "      <td>6.666667</td>\n",
       "      <td>59.166667</td>\n",
       "      <td>5.2</td>\n",
       "      <td>18.606195</td>\n",
       "      <td>18.606195</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-01-11 17:20:00</th>\n",
       "      <td>50</td>\n",
       "      <td>30</td>\n",
       "      <td>19.89</td>\n",
       "      <td>46.300000</td>\n",
       "      <td>19.2</td>\n",
       "      <td>44.626667</td>\n",
       "      <td>19.79</td>\n",
       "      <td>44.933333</td>\n",
       "      <td>18.926667</td>\n",
       "      <td>45.890000</td>\n",
       "      <td>...</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>45.50</td>\n",
       "      <td>6.366667</td>\n",
       "      <td>733.7</td>\n",
       "      <td>92.0</td>\n",
       "      <td>6.333333</td>\n",
       "      <td>55.333333</td>\n",
       "      <td>5.1</td>\n",
       "      <td>28.642668</td>\n",
       "      <td>28.642668</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-01-11 17:30:00</th>\n",
       "      <td>50</td>\n",
       "      <td>40</td>\n",
       "      <td>19.89</td>\n",
       "      <td>46.066667</td>\n",
       "      <td>19.2</td>\n",
       "      <td>44.590000</td>\n",
       "      <td>19.79</td>\n",
       "      <td>45.000000</td>\n",
       "      <td>18.890000</td>\n",
       "      <td>45.723333</td>\n",
       "      <td>...</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>45.40</td>\n",
       "      <td>6.250000</td>\n",
       "      <td>733.8</td>\n",
       "      <td>92.0</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>51.500000</td>\n",
       "      <td>5.0</td>\n",
       "      <td>45.410389</td>\n",
       "      <td>45.410389</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-01-11 17:40:00</th>\n",
       "      <td>60</td>\n",
       "      <td>40</td>\n",
       "      <td>19.89</td>\n",
       "      <td>46.333333</td>\n",
       "      <td>19.2</td>\n",
       "      <td>44.530000</td>\n",
       "      <td>19.79</td>\n",
       "      <td>45.000000</td>\n",
       "      <td>18.890000</td>\n",
       "      <td>45.530000</td>\n",
       "      <td>...</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>45.40</td>\n",
       "      <td>6.133333</td>\n",
       "      <td>733.9</td>\n",
       "      <td>92.0</td>\n",
       "      <td>5.666667</td>\n",
       "      <td>47.666667</td>\n",
       "      <td>4.9</td>\n",
       "      <td>10.084097</td>\n",
       "      <td>10.084097</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 28 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Appliances  lights     T1       RH_1    T2       RH_2  \\\n",
       "date                                                                         \n",
       "2016-01-11 17:00:00          60      30  19.89  47.596667  19.2  44.790000   \n",
       "2016-01-11 17:10:00          60      30  19.89  46.693333  19.2  44.722500   \n",
       "2016-01-11 17:20:00          50      30  19.89  46.300000  19.2  44.626667   \n",
       "2016-01-11 17:30:00          50      40  19.89  46.066667  19.2  44.590000   \n",
       "2016-01-11 17:40:00          60      40  19.89  46.333333  19.2  44.530000   \n",
       "\n",
       "                        T3       RH_3         T4       RH_4  ...         T9  \\\n",
       "date                                                         ...              \n",
       "2016-01-11 17:00:00  19.79  44.730000  19.000000  45.566667  ...  17.033333   \n",
       "2016-01-11 17:10:00  19.79  44.790000  19.000000  45.992500  ...  17.066667   \n",
       "2016-01-11 17:20:00  19.79  44.933333  18.926667  45.890000  ...  17.000000   \n",
       "2016-01-11 17:30:00  19.79  45.000000  18.890000  45.723333  ...  17.000000   \n",
       "2016-01-11 17:40:00  19.79  45.000000  18.890000  45.530000  ...  17.000000   \n",
       "\n",
       "                      RH_9     T_out  Press_mm_hg  RH_out  Windspeed  \\\n",
       "date                                                                   \n",
       "2016-01-11 17:00:00  45.53  6.600000        733.5    92.0   7.000000   \n",
       "2016-01-11 17:10:00  45.56  6.483333        733.6    92.0   6.666667   \n",
       "2016-01-11 17:20:00  45.50  6.366667        733.7    92.0   6.333333   \n",
       "2016-01-11 17:30:00  45.40  6.250000        733.8    92.0   6.000000   \n",
       "2016-01-11 17:40:00  45.40  6.133333        733.9    92.0   5.666667   \n",
       "\n",
       "                     Visibility  Tdewpoint        rv1        rv2  \n",
       "date                                                              \n",
       "2016-01-11 17:00:00   63.000000        5.3  13.275433  13.275433  \n",
       "2016-01-11 17:10:00   59.166667        5.2  18.606195  18.606195  \n",
       "2016-01-11 17:20:00   55.333333        5.1  28.642668  28.642668  \n",
       "2016-01-11 17:30:00   51.500000        5.0  45.410389  45.410389  \n",
       "2016-01-11 17:40:00   47.666667        4.9  10.084097  10.084097  \n",
       "\n",
       "[5 rows x 28 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Index: 19735 entries, 2016-01-11 17:00:00 to 2016-05-27 18:00:00\n",
      "Data columns (total 28 columns):\n",
      "Appliances     19735 non-null int64\n",
      "lights         19735 non-null int64\n",
      "T1             19735 non-null float64\n",
      "RH_1           19735 non-null float64\n",
      "T2             19735 non-null float64\n",
      "RH_2           19735 non-null float64\n",
      "T3             19735 non-null float64\n",
      "RH_3           19735 non-null float64\n",
      "T4             19735 non-null float64\n",
      "RH_4           19735 non-null float64\n",
      "T5             19735 non-null float64\n",
      "RH_5           19735 non-null float64\n",
      "T6             19735 non-null float64\n",
      "RH_6           19735 non-null float64\n",
      "T7             19735 non-null float64\n",
      "RH_7           19735 non-null float64\n",
      "T8             19735 non-null float64\n",
      "RH_8           19735 non-null float64\n",
      "T9             19735 non-null float64\n",
      "RH_9           19735 non-null float64\n",
      "T_out          19735 non-null float64\n",
      "Press_mm_hg    19735 non-null float64\n",
      "RH_out         19735 non-null float64\n",
      "Windspeed      19735 non-null float64\n",
      "Visibility     19735 non-null float64\n",
      "Tdewpoint      19735 non-null float64\n",
      "rv1            19735 non-null float64\n",
      "rv2            19735 non-null float64\n",
      "dtypes: float64(26), int64(2)\n",
      "memory usage: 4.4+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x211446cddc8>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df['Windspeed'].plot(figsize=(12,8))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x211467aec48>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df['Appliances'].plot(figsize=(12,8))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train Test Split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "19735"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Appliances</th>\n",
       "      <th>lights</th>\n",
       "      <th>T1</th>\n",
       "      <th>RH_1</th>\n",
       "      <th>T2</th>\n",
       "      <th>RH_2</th>\n",
       "      <th>T3</th>\n",
       "      <th>RH_3</th>\n",
       "      <th>T4</th>\n",
       "      <th>RH_4</th>\n",
       "      <th>...</th>\n",
       "      <th>T9</th>\n",
       "      <th>RH_9</th>\n",
       "      <th>T_out</th>\n",
       "      <th>Press_mm_hg</th>\n",
       "      <th>RH_out</th>\n",
       "      <th>Windspeed</th>\n",
       "      <th>Visibility</th>\n",
       "      <th>Tdewpoint</th>\n",
       "      <th>rv1</th>\n",
       "      <th>rv2</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2016-01-11 17:00:00</th>\n",
       "      <td>60</td>\n",
       "      <td>30</td>\n",
       "      <td>19.89</td>\n",
       "      <td>47.596667</td>\n",
       "      <td>19.2</td>\n",
       "      <td>44.790000</td>\n",
       "      <td>19.79</td>\n",
       "      <td>44.730000</td>\n",
       "      <td>19.000000</td>\n",
       "      <td>45.566667</td>\n",
       "      <td>...</td>\n",
       "      <td>17.033333</td>\n",
       "      <td>45.53</td>\n",
       "      <td>6.600000</td>\n",
       "      <td>733.5</td>\n",
       "      <td>92.0</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>63.000000</td>\n",
       "      <td>5.3</td>\n",
       "      <td>13.275433</td>\n",
       "      <td>13.275433</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-01-11 17:10:00</th>\n",
       "      <td>60</td>\n",
       "      <td>30</td>\n",
       "      <td>19.89</td>\n",
       "      <td>46.693333</td>\n",
       "      <td>19.2</td>\n",
       "      <td>44.722500</td>\n",
       "      <td>19.79</td>\n",
       "      <td>44.790000</td>\n",
       "      <td>19.000000</td>\n",
       "      <td>45.992500</td>\n",
       "      <td>...</td>\n",
       "      <td>17.066667</td>\n",
       "      <td>45.56</td>\n",
       "      <td>6.483333</td>\n",
       "      <td>733.6</td>\n",
       "      <td>92.0</td>\n",
       "      <td>6.666667</td>\n",
       "      <td>59.166667</td>\n",
       "      <td>5.2</td>\n",
       "      <td>18.606195</td>\n",
       "      <td>18.606195</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-01-11 17:20:00</th>\n",
       "      <td>50</td>\n",
       "      <td>30</td>\n",
       "      <td>19.89</td>\n",
       "      <td>46.300000</td>\n",
       "      <td>19.2</td>\n",
       "      <td>44.626667</td>\n",
       "      <td>19.79</td>\n",
       "      <td>44.933333</td>\n",
       "      <td>18.926667</td>\n",
       "      <td>45.890000</td>\n",
       "      <td>...</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>45.50</td>\n",
       "      <td>6.366667</td>\n",
       "      <td>733.7</td>\n",
       "      <td>92.0</td>\n",
       "      <td>6.333333</td>\n",
       "      <td>55.333333</td>\n",
       "      <td>5.1</td>\n",
       "      <td>28.642668</td>\n",
       "      <td>28.642668</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3 rows × 28 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Appliances  lights     T1       RH_1    T2       RH_2  \\\n",
       "date                                                                         \n",
       "2016-01-11 17:00:00          60      30  19.89  47.596667  19.2  44.790000   \n",
       "2016-01-11 17:10:00          60      30  19.89  46.693333  19.2  44.722500   \n",
       "2016-01-11 17:20:00          50      30  19.89  46.300000  19.2  44.626667   \n",
       "\n",
       "                        T3       RH_3         T4       RH_4  ...         T9  \\\n",
       "date                                                         ...              \n",
       "2016-01-11 17:00:00  19.79  44.730000  19.000000  45.566667  ...  17.033333   \n",
       "2016-01-11 17:10:00  19.79  44.790000  19.000000  45.992500  ...  17.066667   \n",
       "2016-01-11 17:20:00  19.79  44.933333  18.926667  45.890000  ...  17.000000   \n",
       "\n",
       "                      RH_9     T_out  Press_mm_hg  RH_out  Windspeed  \\\n",
       "date                                                                   \n",
       "2016-01-11 17:00:00  45.53  6.600000        733.5    92.0   7.000000   \n",
       "2016-01-11 17:10:00  45.56  6.483333        733.6    92.0   6.666667   \n",
       "2016-01-11 17:20:00  45.50  6.366667        733.7    92.0   6.333333   \n",
       "\n",
       "                     Visibility  Tdewpoint        rv1        rv2  \n",
       "date                                                              \n",
       "2016-01-11 17:00:00   63.000000        5.3  13.275433  13.275433  \n",
       "2016-01-11 17:10:00   59.166667        5.2  18.606195  18.606195  \n",
       "2016-01-11 17:20:00   55.333333        5.1  28.642668  28.642668  \n",
       "\n",
       "[3 rows x 28 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Appliances</th>\n",
       "      <th>lights</th>\n",
       "      <th>T1</th>\n",
       "      <th>RH_1</th>\n",
       "      <th>T2</th>\n",
       "      <th>RH_2</th>\n",
       "      <th>T3</th>\n",
       "      <th>RH_3</th>\n",
       "      <th>T4</th>\n",
       "      <th>RH_4</th>\n",
       "      <th>...</th>\n",
       "      <th>T9</th>\n",
       "      <th>RH_9</th>\n",
       "      <th>T_out</th>\n",
       "      <th>Press_mm_hg</th>\n",
       "      <th>RH_out</th>\n",
       "      <th>Windspeed</th>\n",
       "      <th>Visibility</th>\n",
       "      <th>Tdewpoint</th>\n",
       "      <th>rv1</th>\n",
       "      <th>rv2</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:20:00</th>\n",
       "      <td>100</td>\n",
       "      <td>0</td>\n",
       "      <td>25.566667</td>\n",
       "      <td>46.560000</td>\n",
       "      <td>25.890000</td>\n",
       "      <td>42.025714</td>\n",
       "      <td>27.200000</td>\n",
       "      <td>41.163333</td>\n",
       "      <td>24.7</td>\n",
       "      <td>45.590000</td>\n",
       "      <td>...</td>\n",
       "      <td>23.2</td>\n",
       "      <td>46.7900</td>\n",
       "      <td>22.733333</td>\n",
       "      <td>755.2</td>\n",
       "      <td>55.666667</td>\n",
       "      <td>3.333333</td>\n",
       "      <td>23.666667</td>\n",
       "      <td>13.333333</td>\n",
       "      <td>43.096812</td>\n",
       "      <td>43.096812</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:30:00</th>\n",
       "      <td>90</td>\n",
       "      <td>0</td>\n",
       "      <td>25.500000</td>\n",
       "      <td>46.500000</td>\n",
       "      <td>25.754000</td>\n",
       "      <td>42.080000</td>\n",
       "      <td>27.133333</td>\n",
       "      <td>41.223333</td>\n",
       "      <td>24.7</td>\n",
       "      <td>45.590000</td>\n",
       "      <td>...</td>\n",
       "      <td>23.2</td>\n",
       "      <td>46.7900</td>\n",
       "      <td>22.600000</td>\n",
       "      <td>755.2</td>\n",
       "      <td>56.000000</td>\n",
       "      <td>3.500000</td>\n",
       "      <td>24.500000</td>\n",
       "      <td>13.300000</td>\n",
       "      <td>49.282940</td>\n",
       "      <td>49.282940</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:40:00</th>\n",
       "      <td>270</td>\n",
       "      <td>10</td>\n",
       "      <td>25.500000</td>\n",
       "      <td>46.596667</td>\n",
       "      <td>25.628571</td>\n",
       "      <td>42.768571</td>\n",
       "      <td>27.050000</td>\n",
       "      <td>41.690000</td>\n",
       "      <td>24.7</td>\n",
       "      <td>45.730000</td>\n",
       "      <td>...</td>\n",
       "      <td>23.2</td>\n",
       "      <td>46.7900</td>\n",
       "      <td>22.466667</td>\n",
       "      <td>755.2</td>\n",
       "      <td>56.333333</td>\n",
       "      <td>3.666667</td>\n",
       "      <td>25.333333</td>\n",
       "      <td>13.266667</td>\n",
       "      <td>29.199117</td>\n",
       "      <td>29.199117</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:50:00</th>\n",
       "      <td>420</td>\n",
       "      <td>10</td>\n",
       "      <td>25.500000</td>\n",
       "      <td>46.990000</td>\n",
       "      <td>25.414000</td>\n",
       "      <td>43.036000</td>\n",
       "      <td>26.890000</td>\n",
       "      <td>41.290000</td>\n",
       "      <td>24.7</td>\n",
       "      <td>45.790000</td>\n",
       "      <td>...</td>\n",
       "      <td>23.2</td>\n",
       "      <td>46.8175</td>\n",
       "      <td>22.333333</td>\n",
       "      <td>755.2</td>\n",
       "      <td>56.666667</td>\n",
       "      <td>3.833333</td>\n",
       "      <td>26.166667</td>\n",
       "      <td>13.233333</td>\n",
       "      <td>6.322784</td>\n",
       "      <td>6.322784</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 18:00:00</th>\n",
       "      <td>430</td>\n",
       "      <td>10</td>\n",
       "      <td>25.500000</td>\n",
       "      <td>46.600000</td>\n",
       "      <td>25.264286</td>\n",
       "      <td>42.971429</td>\n",
       "      <td>26.823333</td>\n",
       "      <td>41.156667</td>\n",
       "      <td>24.7</td>\n",
       "      <td>45.963333</td>\n",
       "      <td>...</td>\n",
       "      <td>23.2</td>\n",
       "      <td>46.8450</td>\n",
       "      <td>22.200000</td>\n",
       "      <td>755.2</td>\n",
       "      <td>57.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>27.000000</td>\n",
       "      <td>13.200000</td>\n",
       "      <td>34.118851</td>\n",
       "      <td>34.118851</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 28 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Appliances  lights         T1       RH_1         T2  \\\n",
       "date                                                                       \n",
       "2016-05-27 17:20:00         100       0  25.566667  46.560000  25.890000   \n",
       "2016-05-27 17:30:00          90       0  25.500000  46.500000  25.754000   \n",
       "2016-05-27 17:40:00         270      10  25.500000  46.596667  25.628571   \n",
       "2016-05-27 17:50:00         420      10  25.500000  46.990000  25.414000   \n",
       "2016-05-27 18:00:00         430      10  25.500000  46.600000  25.264286   \n",
       "\n",
       "                          RH_2         T3       RH_3    T4       RH_4  ...  \\\n",
       "date                                                                   ...   \n",
       "2016-05-27 17:20:00  42.025714  27.200000  41.163333  24.7  45.590000  ...   \n",
       "2016-05-27 17:30:00  42.080000  27.133333  41.223333  24.7  45.590000  ...   \n",
       "2016-05-27 17:40:00  42.768571  27.050000  41.690000  24.7  45.730000  ...   \n",
       "2016-05-27 17:50:00  43.036000  26.890000  41.290000  24.7  45.790000  ...   \n",
       "2016-05-27 18:00:00  42.971429  26.823333  41.156667  24.7  45.963333  ...   \n",
       "\n",
       "                       T9     RH_9      T_out  Press_mm_hg     RH_out  \\\n",
       "date                                                                    \n",
       "2016-05-27 17:20:00  23.2  46.7900  22.733333        755.2  55.666667   \n",
       "2016-05-27 17:30:00  23.2  46.7900  22.600000        755.2  56.000000   \n",
       "2016-05-27 17:40:00  23.2  46.7900  22.466667        755.2  56.333333   \n",
       "2016-05-27 17:50:00  23.2  46.8175  22.333333        755.2  56.666667   \n",
       "2016-05-27 18:00:00  23.2  46.8450  22.200000        755.2  57.000000   \n",
       "\n",
       "                     Windspeed  Visibility  Tdewpoint        rv1        rv2  \n",
       "date                                                                         \n",
       "2016-05-27 17:20:00   3.333333   23.666667  13.333333  43.096812  43.096812  \n",
       "2016-05-27 17:30:00   3.500000   24.500000  13.300000  49.282940  49.282940  \n",
       "2016-05-27 17:40:00   3.666667   25.333333  13.266667  29.199117  29.199117  \n",
       "2016-05-27 17:50:00   3.833333   26.166667  13.233333   6.322784   6.322784  \n",
       "2016-05-27 18:00:00   4.000000   27.000000  13.200000  34.118851  34.118851  \n",
       "\n",
       "[5 rows x 28 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.tail(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's imagine we want to predict just 24 hours into the future, we don't need 3 months of data for that, so let's save some training time and only select the last months data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Appliances</th>\n",
       "      <th>lights</th>\n",
       "      <th>T1</th>\n",
       "      <th>RH_1</th>\n",
       "      <th>T2</th>\n",
       "      <th>RH_2</th>\n",
       "      <th>T3</th>\n",
       "      <th>RH_3</th>\n",
       "      <th>T4</th>\n",
       "      <th>RH_4</th>\n",
       "      <th>...</th>\n",
       "      <th>T9</th>\n",
       "      <th>RH_9</th>\n",
       "      <th>T_out</th>\n",
       "      <th>Press_mm_hg</th>\n",
       "      <th>RH_out</th>\n",
       "      <th>Windspeed</th>\n",
       "      <th>Visibility</th>\n",
       "      <th>Tdewpoint</th>\n",
       "      <th>rv1</th>\n",
       "      <th>rv2</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2016-05-01 00:00:00</th>\n",
       "      <td>50</td>\n",
       "      <td>0</td>\n",
       "      <td>21.200000</td>\n",
       "      <td>38.200000</td>\n",
       "      <td>18.390000</td>\n",
       "      <td>41.400000</td>\n",
       "      <td>23.200000</td>\n",
       "      <td>36.400000</td>\n",
       "      <td>19.73</td>\n",
       "      <td>37.730000</td>\n",
       "      <td>...</td>\n",
       "      <td>19.166667</td>\n",
       "      <td>38.200000</td>\n",
       "      <td>4.300000</td>\n",
       "      <td>763.700000</td>\n",
       "      <td>86.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>40.000000</td>\n",
       "      <td>2.200000</td>\n",
       "      <td>15.310003</td>\n",
       "      <td>15.310003</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-01 00:10:00</th>\n",
       "      <td>60</td>\n",
       "      <td>0</td>\n",
       "      <td>21.150000</td>\n",
       "      <td>38.172500</td>\n",
       "      <td>18.390000</td>\n",
       "      <td>41.500000</td>\n",
       "      <td>23.133333</td>\n",
       "      <td>36.466667</td>\n",
       "      <td>19.70</td>\n",
       "      <td>37.663333</td>\n",
       "      <td>...</td>\n",
       "      <td>19.133333</td>\n",
       "      <td>38.290000</td>\n",
       "      <td>4.283333</td>\n",
       "      <td>763.716667</td>\n",
       "      <td>86.333333</td>\n",
       "      <td>2.166667</td>\n",
       "      <td>38.166667</td>\n",
       "      <td>2.216667</td>\n",
       "      <td>11.240067</td>\n",
       "      <td>11.240067</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-01 00:20:00</th>\n",
       "      <td>50</td>\n",
       "      <td>0</td>\n",
       "      <td>21.133333</td>\n",
       "      <td>38.090000</td>\n",
       "      <td>18.323333</td>\n",
       "      <td>41.500000</td>\n",
       "      <td>23.200000</td>\n",
       "      <td>36.500000</td>\n",
       "      <td>19.70</td>\n",
       "      <td>37.590000</td>\n",
       "      <td>...</td>\n",
       "      <td>19.133333</td>\n",
       "      <td>38.363333</td>\n",
       "      <td>4.266667</td>\n",
       "      <td>763.733333</td>\n",
       "      <td>86.666667</td>\n",
       "      <td>2.333333</td>\n",
       "      <td>36.333333</td>\n",
       "      <td>2.233333</td>\n",
       "      <td>21.808814</td>\n",
       "      <td>21.808814</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-01 00:30:00</th>\n",
       "      <td>50</td>\n",
       "      <td>0</td>\n",
       "      <td>21.100000</td>\n",
       "      <td>38.090000</td>\n",
       "      <td>18.290000</td>\n",
       "      <td>41.500000</td>\n",
       "      <td>23.200000</td>\n",
       "      <td>36.500000</td>\n",
       "      <td>19.70</td>\n",
       "      <td>37.500000</td>\n",
       "      <td>...</td>\n",
       "      <td>19.166667</td>\n",
       "      <td>38.500000</td>\n",
       "      <td>4.250000</td>\n",
       "      <td>763.750000</td>\n",
       "      <td>87.000000</td>\n",
       "      <td>2.500000</td>\n",
       "      <td>34.500000</td>\n",
       "      <td>2.250000</td>\n",
       "      <td>47.000534</td>\n",
       "      <td>47.000534</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-01 00:40:00</th>\n",
       "      <td>60</td>\n",
       "      <td>0</td>\n",
       "      <td>21.100000</td>\n",
       "      <td>38.030000</td>\n",
       "      <td>18.290000</td>\n",
       "      <td>41.560000</td>\n",
       "      <td>23.200000</td>\n",
       "      <td>36.500000</td>\n",
       "      <td>19.70</td>\n",
       "      <td>37.500000</td>\n",
       "      <td>...</td>\n",
       "      <td>19.166667</td>\n",
       "      <td>38.633333</td>\n",
       "      <td>4.233333</td>\n",
       "      <td>763.766667</td>\n",
       "      <td>87.333333</td>\n",
       "      <td>2.666667</td>\n",
       "      <td>32.666667</td>\n",
       "      <td>2.266667</td>\n",
       "      <td>8.059441</td>\n",
       "      <td>8.059441</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:20:00</th>\n",
       "      <td>100</td>\n",
       "      <td>0</td>\n",
       "      <td>25.566667</td>\n",
       "      <td>46.560000</td>\n",
       "      <td>25.890000</td>\n",
       "      <td>42.025714</td>\n",
       "      <td>27.200000</td>\n",
       "      <td>41.163333</td>\n",
       "      <td>24.70</td>\n",
       "      <td>45.590000</td>\n",
       "      <td>...</td>\n",
       "      <td>23.200000</td>\n",
       "      <td>46.790000</td>\n",
       "      <td>22.733333</td>\n",
       "      <td>755.200000</td>\n",
       "      <td>55.666667</td>\n",
       "      <td>3.333333</td>\n",
       "      <td>23.666667</td>\n",
       "      <td>13.333333</td>\n",
       "      <td>43.096812</td>\n",
       "      <td>43.096812</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:30:00</th>\n",
       "      <td>90</td>\n",
       "      <td>0</td>\n",
       "      <td>25.500000</td>\n",
       "      <td>46.500000</td>\n",
       "      <td>25.754000</td>\n",
       "      <td>42.080000</td>\n",
       "      <td>27.133333</td>\n",
       "      <td>41.223333</td>\n",
       "      <td>24.70</td>\n",
       "      <td>45.590000</td>\n",
       "      <td>...</td>\n",
       "      <td>23.200000</td>\n",
       "      <td>46.790000</td>\n",
       "      <td>22.600000</td>\n",
       "      <td>755.200000</td>\n",
       "      <td>56.000000</td>\n",
       "      <td>3.500000</td>\n",
       "      <td>24.500000</td>\n",
       "      <td>13.300000</td>\n",
       "      <td>49.282940</td>\n",
       "      <td>49.282940</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:40:00</th>\n",
       "      <td>270</td>\n",
       "      <td>10</td>\n",
       "      <td>25.500000</td>\n",
       "      <td>46.596667</td>\n",
       "      <td>25.628571</td>\n",
       "      <td>42.768571</td>\n",
       "      <td>27.050000</td>\n",
       "      <td>41.690000</td>\n",
       "      <td>24.70</td>\n",
       "      <td>45.730000</td>\n",
       "      <td>...</td>\n",
       "      <td>23.200000</td>\n",
       "      <td>46.790000</td>\n",
       "      <td>22.466667</td>\n",
       "      <td>755.200000</td>\n",
       "      <td>56.333333</td>\n",
       "      <td>3.666667</td>\n",
       "      <td>25.333333</td>\n",
       "      <td>13.266667</td>\n",
       "      <td>29.199117</td>\n",
       "      <td>29.199117</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:50:00</th>\n",
       "      <td>420</td>\n",
       "      <td>10</td>\n",
       "      <td>25.500000</td>\n",
       "      <td>46.990000</td>\n",
       "      <td>25.414000</td>\n",
       "      <td>43.036000</td>\n",
       "      <td>26.890000</td>\n",
       "      <td>41.290000</td>\n",
       "      <td>24.70</td>\n",
       "      <td>45.790000</td>\n",
       "      <td>...</td>\n",
       "      <td>23.200000</td>\n",
       "      <td>46.817500</td>\n",
       "      <td>22.333333</td>\n",
       "      <td>755.200000</td>\n",
       "      <td>56.666667</td>\n",
       "      <td>3.833333</td>\n",
       "      <td>26.166667</td>\n",
       "      <td>13.233333</td>\n",
       "      <td>6.322784</td>\n",
       "      <td>6.322784</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 18:00:00</th>\n",
       "      <td>430</td>\n",
       "      <td>10</td>\n",
       "      <td>25.500000</td>\n",
       "      <td>46.600000</td>\n",
       "      <td>25.264286</td>\n",
       "      <td>42.971429</td>\n",
       "      <td>26.823333</td>\n",
       "      <td>41.156667</td>\n",
       "      <td>24.70</td>\n",
       "      <td>45.963333</td>\n",
       "      <td>...</td>\n",
       "      <td>23.200000</td>\n",
       "      <td>46.845000</td>\n",
       "      <td>22.200000</td>\n",
       "      <td>755.200000</td>\n",
       "      <td>57.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>27.000000</td>\n",
       "      <td>13.200000</td>\n",
       "      <td>34.118851</td>\n",
       "      <td>34.118851</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3853 rows × 28 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Appliances  lights         T1       RH_1         T2  \\\n",
       "date                                                                       \n",
       "2016-05-01 00:00:00          50       0  21.200000  38.200000  18.390000   \n",
       "2016-05-01 00:10:00          60       0  21.150000  38.172500  18.390000   \n",
       "2016-05-01 00:20:00          50       0  21.133333  38.090000  18.323333   \n",
       "2016-05-01 00:30:00          50       0  21.100000  38.090000  18.290000   \n",
       "2016-05-01 00:40:00          60       0  21.100000  38.030000  18.290000   \n",
       "...                         ...     ...        ...        ...        ...   \n",
       "2016-05-27 17:20:00         100       0  25.566667  46.560000  25.890000   \n",
       "2016-05-27 17:30:00          90       0  25.500000  46.500000  25.754000   \n",
       "2016-05-27 17:40:00         270      10  25.500000  46.596667  25.628571   \n",
       "2016-05-27 17:50:00         420      10  25.500000  46.990000  25.414000   \n",
       "2016-05-27 18:00:00         430      10  25.500000  46.600000  25.264286   \n",
       "\n",
       "                          RH_2         T3       RH_3     T4       RH_4  ...  \\\n",
       "date                                                                    ...   \n",
       "2016-05-01 00:00:00  41.400000  23.200000  36.400000  19.73  37.730000  ...   \n",
       "2016-05-01 00:10:00  41.500000  23.133333  36.466667  19.70  37.663333  ...   \n",
       "2016-05-01 00:20:00  41.500000  23.200000  36.500000  19.70  37.590000  ...   \n",
       "2016-05-01 00:30:00  41.500000  23.200000  36.500000  19.70  37.500000  ...   \n",
       "2016-05-01 00:40:00  41.560000  23.200000  36.500000  19.70  37.500000  ...   \n",
       "...                        ...        ...        ...    ...        ...  ...   \n",
       "2016-05-27 17:20:00  42.025714  27.200000  41.163333  24.70  45.590000  ...   \n",
       "2016-05-27 17:30:00  42.080000  27.133333  41.223333  24.70  45.590000  ...   \n",
       "2016-05-27 17:40:00  42.768571  27.050000  41.690000  24.70  45.730000  ...   \n",
       "2016-05-27 17:50:00  43.036000  26.890000  41.290000  24.70  45.790000  ...   \n",
       "2016-05-27 18:00:00  42.971429  26.823333  41.156667  24.70  45.963333  ...   \n",
       "\n",
       "                            T9       RH_9      T_out  Press_mm_hg     RH_out  \\\n",
       "date                                                                           \n",
       "2016-05-01 00:00:00  19.166667  38.200000   4.300000   763.700000  86.000000   \n",
       "2016-05-01 00:10:00  19.133333  38.290000   4.283333   763.716667  86.333333   \n",
       "2016-05-01 00:20:00  19.133333  38.363333   4.266667   763.733333  86.666667   \n",
       "2016-05-01 00:30:00  19.166667  38.500000   4.250000   763.750000  87.000000   \n",
       "2016-05-01 00:40:00  19.166667  38.633333   4.233333   763.766667  87.333333   \n",
       "...                        ...        ...        ...          ...        ...   \n",
       "2016-05-27 17:20:00  23.200000  46.790000  22.733333   755.200000  55.666667   \n",
       "2016-05-27 17:30:00  23.200000  46.790000  22.600000   755.200000  56.000000   \n",
       "2016-05-27 17:40:00  23.200000  46.790000  22.466667   755.200000  56.333333   \n",
       "2016-05-27 17:50:00  23.200000  46.817500  22.333333   755.200000  56.666667   \n",
       "2016-05-27 18:00:00  23.200000  46.845000  22.200000   755.200000  57.000000   \n",
       "\n",
       "                     Windspeed  Visibility  Tdewpoint        rv1        rv2  \n",
       "date                                                                         \n",
       "2016-05-01 00:00:00   2.000000   40.000000   2.200000  15.310003  15.310003  \n",
       "2016-05-01 00:10:00   2.166667   38.166667   2.216667  11.240067  11.240067  \n",
       "2016-05-01 00:20:00   2.333333   36.333333   2.233333  21.808814  21.808814  \n",
       "2016-05-01 00:30:00   2.500000   34.500000   2.250000  47.000534  47.000534  \n",
       "2016-05-01 00:40:00   2.666667   32.666667   2.266667   8.059441   8.059441  \n",
       "...                        ...         ...        ...        ...        ...  \n",
       "2016-05-27 17:20:00   3.333333   23.666667  13.333333  43.096812  43.096812  \n",
       "2016-05-27 17:30:00   3.500000   24.500000  13.300000  49.282940  49.282940  \n",
       "2016-05-27 17:40:00   3.666667   25.333333  13.266667  29.199117  29.199117  \n",
       "2016-05-27 17:50:00   3.833333   26.166667  13.233333   6.322784   6.322784  \n",
       "2016-05-27 18:00:00   4.000000   27.000000  13.200000  34.118851  34.118851  \n",
       "\n",
       "[3853 rows x 28 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.loc['2016-05-01':]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df = df.loc['2016-05-01':]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's also round off the data, to one decimal point precision, otherwise this may cause issues with our network (we will also normalize the data anyways, so this level of precision isn't useful to us)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df = df.round(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3853"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "144.0"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# How many rows per day? We know its every 10 min\n",
    "24*60/10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "test_days = 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "test_ind = test_days*144"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "288"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_ind"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Notice the minus sign in our indexing\n",
    "\n",
    "train = df.iloc[:-test_ind]\n",
    "test = df.iloc[-test_ind:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Appliances</th>\n",
       "      <th>lights</th>\n",
       "      <th>T1</th>\n",
       "      <th>RH_1</th>\n",
       "      <th>T2</th>\n",
       "      <th>RH_2</th>\n",
       "      <th>T3</th>\n",
       "      <th>RH_3</th>\n",
       "      <th>T4</th>\n",
       "      <th>RH_4</th>\n",
       "      <th>...</th>\n",
       "      <th>T9</th>\n",
       "      <th>RH_9</th>\n",
       "      <th>T_out</th>\n",
       "      <th>Press_mm_hg</th>\n",
       "      <th>RH_out</th>\n",
       "      <th>Windspeed</th>\n",
       "      <th>Visibility</th>\n",
       "      <th>Tdewpoint</th>\n",
       "      <th>rv1</th>\n",
       "      <th>rv2</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2016-05-01 00:00:00</th>\n",
       "      <td>50</td>\n",
       "      <td>0</td>\n",
       "      <td>21.20</td>\n",
       "      <td>38.20</td>\n",
       "      <td>18.39</td>\n",
       "      <td>41.40</td>\n",
       "      <td>23.20</td>\n",
       "      <td>36.40</td>\n",
       "      <td>19.73</td>\n",
       "      <td>37.73</td>\n",
       "      <td>...</td>\n",
       "      <td>19.17</td>\n",
       "      <td>38.20</td>\n",
       "      <td>4.30</td>\n",
       "      <td>763.70</td>\n",
       "      <td>86.00</td>\n",
       "      <td>2.00</td>\n",
       "      <td>40.00</td>\n",
       "      <td>2.20</td>\n",
       "      <td>15.31</td>\n",
       "      <td>15.31</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-01 00:10:00</th>\n",
       "      <td>60</td>\n",
       "      <td>0</td>\n",
       "      <td>21.15</td>\n",
       "      <td>38.17</td>\n",
       "      <td>18.39</td>\n",
       "      <td>41.50</td>\n",
       "      <td>23.13</td>\n",
       "      <td>36.47</td>\n",
       "      <td>19.70</td>\n",
       "      <td>37.66</td>\n",
       "      <td>...</td>\n",
       "      <td>19.13</td>\n",
       "      <td>38.29</td>\n",
       "      <td>4.28</td>\n",
       "      <td>763.72</td>\n",
       "      <td>86.33</td>\n",
       "      <td>2.17</td>\n",
       "      <td>38.17</td>\n",
       "      <td>2.22</td>\n",
       "      <td>11.24</td>\n",
       "      <td>11.24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-01 00:20:00</th>\n",
       "      <td>50</td>\n",
       "      <td>0</td>\n",
       "      <td>21.13</td>\n",
       "      <td>38.09</td>\n",
       "      <td>18.32</td>\n",
       "      <td>41.50</td>\n",
       "      <td>23.20</td>\n",
       "      <td>36.50</td>\n",
       "      <td>19.70</td>\n",
       "      <td>37.59</td>\n",
       "      <td>...</td>\n",
       "      <td>19.13</td>\n",
       "      <td>38.36</td>\n",
       "      <td>4.27</td>\n",
       "      <td>763.73</td>\n",
       "      <td>86.67</td>\n",
       "      <td>2.33</td>\n",
       "      <td>36.33</td>\n",
       "      <td>2.23</td>\n",
       "      <td>21.81</td>\n",
       "      <td>21.81</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-01 00:30:00</th>\n",
       "      <td>50</td>\n",
       "      <td>0</td>\n",
       "      <td>21.10</td>\n",
       "      <td>38.09</td>\n",
       "      <td>18.29</td>\n",
       "      <td>41.50</td>\n",
       "      <td>23.20</td>\n",
       "      <td>36.50</td>\n",
       "      <td>19.70</td>\n",
       "      <td>37.50</td>\n",
       "      <td>...</td>\n",
       "      <td>19.17</td>\n",
       "      <td>38.50</td>\n",
       "      <td>4.25</td>\n",
       "      <td>763.75</td>\n",
       "      <td>87.00</td>\n",
       "      <td>2.50</td>\n",
       "      <td>34.50</td>\n",
       "      <td>2.25</td>\n",
       "      <td>47.00</td>\n",
       "      <td>47.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-01 00:40:00</th>\n",
       "      <td>60</td>\n",
       "      <td>0</td>\n",
       "      <td>21.10</td>\n",
       "      <td>38.03</td>\n",
       "      <td>18.29</td>\n",
       "      <td>41.56</td>\n",
       "      <td>23.20</td>\n",
       "      <td>36.50</td>\n",
       "      <td>19.70</td>\n",
       "      <td>37.50</td>\n",
       "      <td>...</td>\n",
       "      <td>19.17</td>\n",
       "      <td>38.63</td>\n",
       "      <td>4.23</td>\n",
       "      <td>763.77</td>\n",
       "      <td>87.33</td>\n",
       "      <td>2.67</td>\n",
       "      <td>32.67</td>\n",
       "      <td>2.27</td>\n",
       "      <td>8.06</td>\n",
       "      <td>8.06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-25 17:20:00</th>\n",
       "      <td>120</td>\n",
       "      <td>0</td>\n",
       "      <td>24.50</td>\n",
       "      <td>37.22</td>\n",
       "      <td>24.13</td>\n",
       "      <td>34.30</td>\n",
       "      <td>25.20</td>\n",
       "      <td>37.64</td>\n",
       "      <td>24.36</td>\n",
       "      <td>38.29</td>\n",
       "      <td>...</td>\n",
       "      <td>21.89</td>\n",
       "      <td>37.03</td>\n",
       "      <td>16.17</td>\n",
       "      <td>756.17</td>\n",
       "      <td>52.67</td>\n",
       "      <td>1.33</td>\n",
       "      <td>31.33</td>\n",
       "      <td>6.43</td>\n",
       "      <td>33.46</td>\n",
       "      <td>33.46</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-25 17:30:00</th>\n",
       "      <td>190</td>\n",
       "      <td>0</td>\n",
       "      <td>24.50</td>\n",
       "      <td>37.16</td>\n",
       "      <td>24.10</td>\n",
       "      <td>34.30</td>\n",
       "      <td>25.20</td>\n",
       "      <td>37.55</td>\n",
       "      <td>24.29</td>\n",
       "      <td>38.16</td>\n",
       "      <td>...</td>\n",
       "      <td>21.89</td>\n",
       "      <td>37.20</td>\n",
       "      <td>16.25</td>\n",
       "      <td>756.15</td>\n",
       "      <td>53.50</td>\n",
       "      <td>1.50</td>\n",
       "      <td>33.50</td>\n",
       "      <td>6.75</td>\n",
       "      <td>0.43</td>\n",
       "      <td>0.43</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-25 17:40:00</th>\n",
       "      <td>160</td>\n",
       "      <td>0</td>\n",
       "      <td>24.50</td>\n",
       "      <td>37.43</td>\n",
       "      <td>24.10</td>\n",
       "      <td>34.43</td>\n",
       "      <td>25.14</td>\n",
       "      <td>37.28</td>\n",
       "      <td>24.29</td>\n",
       "      <td>38.00</td>\n",
       "      <td>...</td>\n",
       "      <td>21.89</td>\n",
       "      <td>37.33</td>\n",
       "      <td>16.33</td>\n",
       "      <td>756.13</td>\n",
       "      <td>54.33</td>\n",
       "      <td>1.67</td>\n",
       "      <td>35.67</td>\n",
       "      <td>7.07</td>\n",
       "      <td>16.67</td>\n",
       "      <td>16.67</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-25 17:50:00</th>\n",
       "      <td>90</td>\n",
       "      <td>0</td>\n",
       "      <td>24.50</td>\n",
       "      <td>37.63</td>\n",
       "      <td>24.03</td>\n",
       "      <td>34.43</td>\n",
       "      <td>25.10</td>\n",
       "      <td>36.99</td>\n",
       "      <td>24.29</td>\n",
       "      <td>37.93</td>\n",
       "      <td>...</td>\n",
       "      <td>22.00</td>\n",
       "      <td>37.36</td>\n",
       "      <td>16.42</td>\n",
       "      <td>756.12</td>\n",
       "      <td>55.17</td>\n",
       "      <td>1.83</td>\n",
       "      <td>37.83</td>\n",
       "      <td>7.38</td>\n",
       "      <td>39.36</td>\n",
       "      <td>39.36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-25 18:00:00</th>\n",
       "      <td>100</td>\n",
       "      <td>0</td>\n",
       "      <td>24.50</td>\n",
       "      <td>38.00</td>\n",
       "      <td>24.00</td>\n",
       "      <td>34.40</td>\n",
       "      <td>25.10</td>\n",
       "      <td>36.73</td>\n",
       "      <td>24.29</td>\n",
       "      <td>37.86</td>\n",
       "      <td>...</td>\n",
       "      <td>22.00</td>\n",
       "      <td>37.36</td>\n",
       "      <td>16.50</td>\n",
       "      <td>756.10</td>\n",
       "      <td>56.00</td>\n",
       "      <td>2.00</td>\n",
       "      <td>40.00</td>\n",
       "      <td>7.70</td>\n",
       "      <td>38.63</td>\n",
       "      <td>38.63</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3565 rows × 28 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Appliances  lights     T1   RH_1     T2   RH_2     T3  \\\n",
       "date                                                                         \n",
       "2016-05-01 00:00:00          50       0  21.20  38.20  18.39  41.40  23.20   \n",
       "2016-05-01 00:10:00          60       0  21.15  38.17  18.39  41.50  23.13   \n",
       "2016-05-01 00:20:00          50       0  21.13  38.09  18.32  41.50  23.20   \n",
       "2016-05-01 00:30:00          50       0  21.10  38.09  18.29  41.50  23.20   \n",
       "2016-05-01 00:40:00          60       0  21.10  38.03  18.29  41.56  23.20   \n",
       "...                         ...     ...    ...    ...    ...    ...    ...   \n",
       "2016-05-25 17:20:00         120       0  24.50  37.22  24.13  34.30  25.20   \n",
       "2016-05-25 17:30:00         190       0  24.50  37.16  24.10  34.30  25.20   \n",
       "2016-05-25 17:40:00         160       0  24.50  37.43  24.10  34.43  25.14   \n",
       "2016-05-25 17:50:00          90       0  24.50  37.63  24.03  34.43  25.10   \n",
       "2016-05-25 18:00:00         100       0  24.50  38.00  24.00  34.40  25.10   \n",
       "\n",
       "                      RH_3     T4   RH_4  ...     T9   RH_9  T_out  \\\n",
       "date                                      ...                        \n",
       "2016-05-01 00:00:00  36.40  19.73  37.73  ...  19.17  38.20   4.30   \n",
       "2016-05-01 00:10:00  36.47  19.70  37.66  ...  19.13  38.29   4.28   \n",
       "2016-05-01 00:20:00  36.50  19.70  37.59  ...  19.13  38.36   4.27   \n",
       "2016-05-01 00:30:00  36.50  19.70  37.50  ...  19.17  38.50   4.25   \n",
       "2016-05-01 00:40:00  36.50  19.70  37.50  ...  19.17  38.63   4.23   \n",
       "...                    ...    ...    ...  ...    ...    ...    ...   \n",
       "2016-05-25 17:20:00  37.64  24.36  38.29  ...  21.89  37.03  16.17   \n",
       "2016-05-25 17:30:00  37.55  24.29  38.16  ...  21.89  37.20  16.25   \n",
       "2016-05-25 17:40:00  37.28  24.29  38.00  ...  21.89  37.33  16.33   \n",
       "2016-05-25 17:50:00  36.99  24.29  37.93  ...  22.00  37.36  16.42   \n",
       "2016-05-25 18:00:00  36.73  24.29  37.86  ...  22.00  37.36  16.50   \n",
       "\n",
       "                     Press_mm_hg  RH_out  Windspeed  Visibility  Tdewpoint  \\\n",
       "date                                                                         \n",
       "2016-05-01 00:00:00       763.70   86.00       2.00       40.00       2.20   \n",
       "2016-05-01 00:10:00       763.72   86.33       2.17       38.17       2.22   \n",
       "2016-05-01 00:20:00       763.73   86.67       2.33       36.33       2.23   \n",
       "2016-05-01 00:30:00       763.75   87.00       2.50       34.50       2.25   \n",
       "2016-05-01 00:40:00       763.77   87.33       2.67       32.67       2.27   \n",
       "...                          ...     ...        ...         ...        ...   \n",
       "2016-05-25 17:20:00       756.17   52.67       1.33       31.33       6.43   \n",
       "2016-05-25 17:30:00       756.15   53.50       1.50       33.50       6.75   \n",
       "2016-05-25 17:40:00       756.13   54.33       1.67       35.67       7.07   \n",
       "2016-05-25 17:50:00       756.12   55.17       1.83       37.83       7.38   \n",
       "2016-05-25 18:00:00       756.10   56.00       2.00       40.00       7.70   \n",
       "\n",
       "                       rv1    rv2  \n",
       "date                               \n",
       "2016-05-01 00:00:00  15.31  15.31  \n",
       "2016-05-01 00:10:00  11.24  11.24  \n",
       "2016-05-01 00:20:00  21.81  21.81  \n",
       "2016-05-01 00:30:00  47.00  47.00  \n",
       "2016-05-01 00:40:00   8.06   8.06  \n",
       "...                    ...    ...  \n",
       "2016-05-25 17:20:00  33.46  33.46  \n",
       "2016-05-25 17:30:00   0.43   0.43  \n",
       "2016-05-25 17:40:00  16.67  16.67  \n",
       "2016-05-25 17:50:00  39.36  39.36  \n",
       "2016-05-25 18:00:00  38.63  38.63  \n",
       "\n",
       "[3565 rows x 28 columns]"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Appliances</th>\n",
       "      <th>lights</th>\n",
       "      <th>T1</th>\n",
       "      <th>RH_1</th>\n",
       "      <th>T2</th>\n",
       "      <th>RH_2</th>\n",
       "      <th>T3</th>\n",
       "      <th>RH_3</th>\n",
       "      <th>T4</th>\n",
       "      <th>RH_4</th>\n",
       "      <th>...</th>\n",
       "      <th>T9</th>\n",
       "      <th>RH_9</th>\n",
       "      <th>T_out</th>\n",
       "      <th>Press_mm_hg</th>\n",
       "      <th>RH_out</th>\n",
       "      <th>Windspeed</th>\n",
       "      <th>Visibility</th>\n",
       "      <th>Tdewpoint</th>\n",
       "      <th>rv1</th>\n",
       "      <th>rv2</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2016-05-25 18:10:00</th>\n",
       "      <td>170</td>\n",
       "      <td>0</td>\n",
       "      <td>24.50</td>\n",
       "      <td>37.86</td>\n",
       "      <td>24.00</td>\n",
       "      <td>34.27</td>\n",
       "      <td>25.00</td>\n",
       "      <td>36.70</td>\n",
       "      <td>24.29</td>\n",
       "      <td>37.79</td>\n",
       "      <td>...</td>\n",
       "      <td>22.00</td>\n",
       "      <td>37.23</td>\n",
       "      <td>16.48</td>\n",
       "      <td>756.1</td>\n",
       "      <td>55.83</td>\n",
       "      <td>2.00</td>\n",
       "      <td>38.17</td>\n",
       "      <td>7.62</td>\n",
       "      <td>38.14</td>\n",
       "      <td>38.14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-25 18:20:00</th>\n",
       "      <td>310</td>\n",
       "      <td>0</td>\n",
       "      <td>24.50</td>\n",
       "      <td>37.30</td>\n",
       "      <td>23.86</td>\n",
       "      <td>34.33</td>\n",
       "      <td>24.94</td>\n",
       "      <td>36.67</td>\n",
       "      <td>24.29</td>\n",
       "      <td>37.79</td>\n",
       "      <td>...</td>\n",
       "      <td>22.00</td>\n",
       "      <td>37.36</td>\n",
       "      <td>16.47</td>\n",
       "      <td>756.1</td>\n",
       "      <td>55.67</td>\n",
       "      <td>2.00</td>\n",
       "      <td>36.33</td>\n",
       "      <td>7.53</td>\n",
       "      <td>31.16</td>\n",
       "      <td>31.16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-25 18:30:00</th>\n",
       "      <td>120</td>\n",
       "      <td>0</td>\n",
       "      <td>24.50</td>\n",
       "      <td>36.96</td>\n",
       "      <td>23.73</td>\n",
       "      <td>34.33</td>\n",
       "      <td>24.85</td>\n",
       "      <td>36.50</td>\n",
       "      <td>24.29</td>\n",
       "      <td>37.79</td>\n",
       "      <td>...</td>\n",
       "      <td>22.03</td>\n",
       "      <td>37.39</td>\n",
       "      <td>16.45</td>\n",
       "      <td>756.1</td>\n",
       "      <td>55.50</td>\n",
       "      <td>2.00</td>\n",
       "      <td>34.50</td>\n",
       "      <td>7.45</td>\n",
       "      <td>4.40</td>\n",
       "      <td>4.40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-25 18:40:00</th>\n",
       "      <td>120</td>\n",
       "      <td>0</td>\n",
       "      <td>24.50</td>\n",
       "      <td>37.00</td>\n",
       "      <td>23.70</td>\n",
       "      <td>34.40</td>\n",
       "      <td>24.84</td>\n",
       "      <td>36.45</td>\n",
       "      <td>24.29</td>\n",
       "      <td>37.90</td>\n",
       "      <td>...</td>\n",
       "      <td>22.10</td>\n",
       "      <td>37.72</td>\n",
       "      <td>16.43</td>\n",
       "      <td>756.1</td>\n",
       "      <td>55.33</td>\n",
       "      <td>2.00</td>\n",
       "      <td>32.67</td>\n",
       "      <td>7.37</td>\n",
       "      <td>27.12</td>\n",
       "      <td>27.12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-25 18:50:00</th>\n",
       "      <td>120</td>\n",
       "      <td>0</td>\n",
       "      <td>24.49</td>\n",
       "      <td>37.07</td>\n",
       "      <td>23.68</td>\n",
       "      <td>34.52</td>\n",
       "      <td>24.84</td>\n",
       "      <td>36.49</td>\n",
       "      <td>24.28</td>\n",
       "      <td>37.93</td>\n",
       "      <td>...</td>\n",
       "      <td>22.10</td>\n",
       "      <td>37.81</td>\n",
       "      <td>16.42</td>\n",
       "      <td>756.1</td>\n",
       "      <td>55.17</td>\n",
       "      <td>2.00</td>\n",
       "      <td>30.83</td>\n",
       "      <td>7.28</td>\n",
       "      <td>10.27</td>\n",
       "      <td>10.27</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:20:00</th>\n",
       "      <td>100</td>\n",
       "      <td>0</td>\n",
       "      <td>25.57</td>\n",
       "      <td>46.56</td>\n",
       "      <td>25.89</td>\n",
       "      <td>42.03</td>\n",
       "      <td>27.20</td>\n",
       "      <td>41.16</td>\n",
       "      <td>24.70</td>\n",
       "      <td>45.59</td>\n",
       "      <td>...</td>\n",
       "      <td>23.20</td>\n",
       "      <td>46.79</td>\n",
       "      <td>22.73</td>\n",
       "      <td>755.2</td>\n",
       "      <td>55.67</td>\n",
       "      <td>3.33</td>\n",
       "      <td>23.67</td>\n",
       "      <td>13.33</td>\n",
       "      <td>43.10</td>\n",
       "      <td>43.10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:30:00</th>\n",
       "      <td>90</td>\n",
       "      <td>0</td>\n",
       "      <td>25.50</td>\n",
       "      <td>46.50</td>\n",
       "      <td>25.75</td>\n",
       "      <td>42.08</td>\n",
       "      <td>27.13</td>\n",
       "      <td>41.22</td>\n",
       "      <td>24.70</td>\n",
       "      <td>45.59</td>\n",
       "      <td>...</td>\n",
       "      <td>23.20</td>\n",
       "      <td>46.79</td>\n",
       "      <td>22.60</td>\n",
       "      <td>755.2</td>\n",
       "      <td>56.00</td>\n",
       "      <td>3.50</td>\n",
       "      <td>24.50</td>\n",
       "      <td>13.30</td>\n",
       "      <td>49.28</td>\n",
       "      <td>49.28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:40:00</th>\n",
       "      <td>270</td>\n",
       "      <td>10</td>\n",
       "      <td>25.50</td>\n",
       "      <td>46.60</td>\n",
       "      <td>25.63</td>\n",
       "      <td>42.77</td>\n",
       "      <td>27.05</td>\n",
       "      <td>41.69</td>\n",
       "      <td>24.70</td>\n",
       "      <td>45.73</td>\n",
       "      <td>...</td>\n",
       "      <td>23.20</td>\n",
       "      <td>46.79</td>\n",
       "      <td>22.47</td>\n",
       "      <td>755.2</td>\n",
       "      <td>56.33</td>\n",
       "      <td>3.67</td>\n",
       "      <td>25.33</td>\n",
       "      <td>13.27</td>\n",
       "      <td>29.20</td>\n",
       "      <td>29.20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:50:00</th>\n",
       "      <td>420</td>\n",
       "      <td>10</td>\n",
       "      <td>25.50</td>\n",
       "      <td>46.99</td>\n",
       "      <td>25.41</td>\n",
       "      <td>43.04</td>\n",
       "      <td>26.89</td>\n",
       "      <td>41.29</td>\n",
       "      <td>24.70</td>\n",
       "      <td>45.79</td>\n",
       "      <td>...</td>\n",
       "      <td>23.20</td>\n",
       "      <td>46.82</td>\n",
       "      <td>22.33</td>\n",
       "      <td>755.2</td>\n",
       "      <td>56.67</td>\n",
       "      <td>3.83</td>\n",
       "      <td>26.17</td>\n",
       "      <td>13.23</td>\n",
       "      <td>6.32</td>\n",
       "      <td>6.32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 18:00:00</th>\n",
       "      <td>430</td>\n",
       "      <td>10</td>\n",
       "      <td>25.50</td>\n",
       "      <td>46.60</td>\n",
       "      <td>25.26</td>\n",
       "      <td>42.97</td>\n",
       "      <td>26.82</td>\n",
       "      <td>41.16</td>\n",
       "      <td>24.70</td>\n",
       "      <td>45.96</td>\n",
       "      <td>...</td>\n",
       "      <td>23.20</td>\n",
       "      <td>46.84</td>\n",
       "      <td>22.20</td>\n",
       "      <td>755.2</td>\n",
       "      <td>57.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>27.00</td>\n",
       "      <td>13.20</td>\n",
       "      <td>34.12</td>\n",
       "      <td>34.12</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>288 rows × 28 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Appliances  lights     T1   RH_1     T2   RH_2     T3  \\\n",
       "date                                                                         \n",
       "2016-05-25 18:10:00         170       0  24.50  37.86  24.00  34.27  25.00   \n",
       "2016-05-25 18:20:00         310       0  24.50  37.30  23.86  34.33  24.94   \n",
       "2016-05-25 18:30:00         120       0  24.50  36.96  23.73  34.33  24.85   \n",
       "2016-05-25 18:40:00         120       0  24.50  37.00  23.70  34.40  24.84   \n",
       "2016-05-25 18:50:00         120       0  24.49  37.07  23.68  34.52  24.84   \n",
       "...                         ...     ...    ...    ...    ...    ...    ...   \n",
       "2016-05-27 17:20:00         100       0  25.57  46.56  25.89  42.03  27.20   \n",
       "2016-05-27 17:30:00          90       0  25.50  46.50  25.75  42.08  27.13   \n",
       "2016-05-27 17:40:00         270      10  25.50  46.60  25.63  42.77  27.05   \n",
       "2016-05-27 17:50:00         420      10  25.50  46.99  25.41  43.04  26.89   \n",
       "2016-05-27 18:00:00         430      10  25.50  46.60  25.26  42.97  26.82   \n",
       "\n",
       "                      RH_3     T4   RH_4  ...     T9   RH_9  T_out  \\\n",
       "date                                      ...                        \n",
       "2016-05-25 18:10:00  36.70  24.29  37.79  ...  22.00  37.23  16.48   \n",
       "2016-05-25 18:20:00  36.67  24.29  37.79  ...  22.00  37.36  16.47   \n",
       "2016-05-25 18:30:00  36.50  24.29  37.79  ...  22.03  37.39  16.45   \n",
       "2016-05-25 18:40:00  36.45  24.29  37.90  ...  22.10  37.72  16.43   \n",
       "2016-05-25 18:50:00  36.49  24.28  37.93  ...  22.10  37.81  16.42   \n",
       "...                    ...    ...    ...  ...    ...    ...    ...   \n",
       "2016-05-27 17:20:00  41.16  24.70  45.59  ...  23.20  46.79  22.73   \n",
       "2016-05-27 17:30:00  41.22  24.70  45.59  ...  23.20  46.79  22.60   \n",
       "2016-05-27 17:40:00  41.69  24.70  45.73  ...  23.20  46.79  22.47   \n",
       "2016-05-27 17:50:00  41.29  24.70  45.79  ...  23.20  46.82  22.33   \n",
       "2016-05-27 18:00:00  41.16  24.70  45.96  ...  23.20  46.84  22.20   \n",
       "\n",
       "                     Press_mm_hg  RH_out  Windspeed  Visibility  Tdewpoint  \\\n",
       "date                                                                         \n",
       "2016-05-25 18:10:00        756.1   55.83       2.00       38.17       7.62   \n",
       "2016-05-25 18:20:00        756.1   55.67       2.00       36.33       7.53   \n",
       "2016-05-25 18:30:00        756.1   55.50       2.00       34.50       7.45   \n",
       "2016-05-25 18:40:00        756.1   55.33       2.00       32.67       7.37   \n",
       "2016-05-25 18:50:00        756.1   55.17       2.00       30.83       7.28   \n",
       "...                          ...     ...        ...         ...        ...   \n",
       "2016-05-27 17:20:00        755.2   55.67       3.33       23.67      13.33   \n",
       "2016-05-27 17:30:00        755.2   56.00       3.50       24.50      13.30   \n",
       "2016-05-27 17:40:00        755.2   56.33       3.67       25.33      13.27   \n",
       "2016-05-27 17:50:00        755.2   56.67       3.83       26.17      13.23   \n",
       "2016-05-27 18:00:00        755.2   57.00       4.00       27.00      13.20   \n",
       "\n",
       "                       rv1    rv2  \n",
       "date                               \n",
       "2016-05-25 18:10:00  38.14  38.14  \n",
       "2016-05-25 18:20:00  31.16  31.16  \n",
       "2016-05-25 18:30:00   4.40   4.40  \n",
       "2016-05-25 18:40:00  27.12  27.12  \n",
       "2016-05-25 18:50:00  10.27  10.27  \n",
       "...                    ...    ...  \n",
       "2016-05-27 17:20:00  43.10  43.10  \n",
       "2016-05-27 17:30:00  49.28  49.28  \n",
       "2016-05-27 17:40:00  29.20  29.20  \n",
       "2016-05-27 17:50:00   6.32   6.32  \n",
       "2016-05-27 18:00:00  34.12  34.12  \n",
       "\n",
       "[288 rows x 28 columns]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Scale Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import MinMaxScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "scaler = MinMaxScaler()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MinMaxScaler(copy=True, feature_range=(0, 1))"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# IGNORE WARNING ITS JUST CONVERTING TO FLOATS\n",
    "# WE ONLY FIT TO TRAININ DATA, OTHERWISE WE ARE CHEATING ASSUMING INFO ABOUT TEST SET\n",
    "scaler.fit(train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "scaled_train = scaler.transform(train)\n",
    "scaled_test = scaler.transform(test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Time Series Generator\n",
    "\n",
    "This class takes in a sequence of data-points gathered at\n",
    "equal intervals, along with time series parameters such as\n",
    "stride, length of history, etc., to produce batches for\n",
    "training/validation.\n",
    "\n",
    "#### Arguments\n",
    "    data: Indexable generator (such as list or Numpy array)\n",
    "        containing consecutive data points (timesteps).\n",
    "        The data should be at 2D, and axis 0 is expected\n",
    "        to be the time dimension.\n",
    "    targets: Targets corresponding to timesteps in `data`.\n",
    "        It should have same length as `data`.\n",
    "    length: Length of the output sequences (in number of timesteps).\n",
    "    sampling_rate: Period between successive individual timesteps\n",
    "        within sequences. For rate `r`, timesteps\n",
    "        `data[i]`, `data[i-r]`, ... `data[i - length]`\n",
    "        are used for create a sample sequence.\n",
    "    stride: Period between successive output sequences.\n",
    "        For stride `s`, consecutive output samples would\n",
    "        be centered around `data[i]`, `data[i+s]`, `data[i+2*s]`, etc.\n",
    "    start_index: Data points earlier than `start_index` will not be used\n",
    "        in the output sequences. This is useful to reserve part of the\n",
    "        data for test or validation.\n",
    "    end_index: Data points later than `end_index` will not be used\n",
    "        in the output sequences. This is useful to reserve part of the\n",
    "        data for test or validation.\n",
    "    shuffle: Whether to shuffle output samples,\n",
    "        or instead draw them in chronological order.\n",
    "    reverse: Boolean: if `true`, timesteps in each output sample will be\n",
    "        in reverse chronological order.\n",
    "    batch_size: Number of timeseries samples in each batch\n",
    "        (except maybe the last one)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from tensorflow.keras.preprocessing.sequence import TimeseriesGenerator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# scaled_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# define generator\n",
    "length = 144 # Length of the output sequences (in number of timesteps)\n",
    "batch_size = 1 #Number of timeseries samples in each batch\n",
    "generator = TimeseriesGenerator(scaled_train, scaled_train, length=length, batch_size=batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3565"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(scaled_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3421"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(generator) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# scaled_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# What does the first batch look like?\n",
    "X,y = generator[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Given the Array: \n",
      "[0.03896104 0.         0.13798978 ... 0.14319527 0.75185111 0.75185111]\n",
      "Predict this y: \n",
      " [[0.03896104 0.         0.30834753 0.29439421 0.16038492 0.49182278\n",
      "  0.0140056  0.36627907 0.24142857 0.24364791 0.12650602 0.36276002\n",
      "  0.12       0.28205572 0.06169297 0.15759185 0.34582624 0.39585974\n",
      "  0.09259259 0.39649608 0.18852459 0.96052632 0.59210526 0.1\n",
      "  0.58333333 0.13609467 0.4576746  0.4576746 ]]\n"
     ]
    }
   ],
   "source": [
    "print(f'Given the Array: \\n{X.flatten()}')\n",
    "print(f'Predict this y: \\n {y}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now you will be able to edit the length so that it makes sense for your time series!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Create the Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense,LSTM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3565, 28)"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scaled_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From c:\\users\\marcial\\.conda\\envs\\tf2gpu\\lib\\site-packages\\tensorflow_core\\python\\ops\\resource_variable_ops.py:1630: calling BaseResourceVariable.__init__ (from tensorflow.python.ops.resource_variable_ops) with constraint is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "If using Keras pass *_constraint arguments to layers.\n"
     ]
    }
   ],
   "source": [
    "# define model\n",
    "model = Sequential()\n",
    "\n",
    "# Simple RNN layer\n",
    "model.add(LSTM(100,input_shape=(length,scaled_train.shape[1])))\n",
    "\n",
    "# Final Prediction (one neuron per feature)\n",
    "model.add(Dense(scaled_train.shape[1]))\n",
    "\n",
    "model.compile(optimizer='adam', loss='mse')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "lstm (LSTM)                  (None, 100)               51600     \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 28)                2828      \n",
      "=================================================================\n",
      "Total params: 54,428\n",
      "Trainable params: 54,428\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## EarlyStopping"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from tensorflow.keras.callbacks import EarlyStopping\n",
    "early_stop = EarlyStopping(monitor='val_loss',patience=1)\n",
    "validation_generator = TimeseriesGenerator(scaled_test,scaled_test, \n",
    "                                           length=length, batch_size=batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "3421/3421 [==============================] - 434s 127ms/step - loss: 0.0117 - val_loss: 0.0101\n",
      "Epoch 2/10\n",
      "3421/3421 [==============================] - 453s 132ms/step - loss: 0.0079 - val_loss: 0.0084\n",
      "Epoch 3/10\n",
      "3421/3421 [==============================] - 474s 138ms/step - loss: 0.0075 - val_loss: 0.0084\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x21148aaa0c8>"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit_generator(generator,epochs=10,\n",
    "                    validation_data=validation_generator,\n",
    "                   callbacks=[early_stop])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['loss', 'val_loss'])"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.history.history.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x2116fd7b748>"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "losses = pd.DataFrame(model.history.history)\n",
    "losses.plot()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Evaluate on Test Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "first_eval_batch = scaled_train[-length:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.1038961 , 0.        , 0.72231687, ..., 0.53550296, 0.15909546,\n",
       "        0.15909546],\n",
       "       [0.11688312, 0.        , 0.73424191, ..., 0.52662722, 0.40344207,\n",
       "        0.40344207],\n",
       "       [0.11688312, 0.        , 0.73424191, ..., 0.51775148, 0.20452271,\n",
       "        0.20452271],\n",
       "       ...,\n",
       "       [0.18181818, 0.        , 0.70017036, ..., 0.50118343, 0.33340004,\n",
       "        0.33340004],\n",
       "       [0.09090909, 0.        , 0.70017036, ..., 0.51952663, 0.78747248,\n",
       "        0.78747248],\n",
       "       [0.1038961 , 0.        , 0.70017036, ..., 0.53846154, 0.77286372,\n",
       "        0.77286372]])"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "first_eval_batch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "first_eval_batch = first_eval_batch.reshape((1, length, scaled_train.shape[1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.12666073, 0.08996339, 0.7190974 , 0.36104494, 0.54831064,\n",
       "        0.4291169 , 0.42736438, 0.45519555, 0.7263751 , 0.38756314,\n",
       "        0.2875382 , 0.3892692 , 0.65530246, 0.0064019 , 0.61901623,\n",
       "        0.39769527, 0.54877293, 0.30695575, 0.5420783 , 0.35719806,\n",
       "        0.59826505, 0.5541311 , 0.3933694 , 0.18549034, 0.585099  ,\n",
       "        0.480112  , 0.50588864, 0.5072879 ]], dtype=float32)"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.predict(first_eval_batch)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.19480519, 0.        , 0.70017036, 0.3920434 , 0.53007217,\n",
       "       0.41064526, 0.40616246, 0.41913319, 0.72714286, 0.4115245 ,\n",
       "       0.30722892, 0.36445121, 0.66777778, 0.        , 0.61119082,\n",
       "       0.39840637, 0.51618399, 0.32953105, 0.53703704, 0.34024896,\n",
       "       0.6057377 , 0.52631579, 0.41881579, 0.2       , 0.55283333,\n",
       "       0.53372781, 0.76305783, 0.76305783])"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scaled_test[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now let's put this logic in a for loop to predict into the future for the entire test range.\n",
    "\n",
    "----"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**NOTE: PAY CLOSE ATTENTION HERE TO WHAT IS BEING OUTPUTED AND IN WHAT DIMENSIONS. ADD YOUR OWN PRINT() STATEMENTS TO SEE WHAT IS TRULY GOING ON!!**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "n_features = scaled_train.shape[1]\n",
    "test_predictions = []\n",
    "\n",
    "first_eval_batch = scaled_train[-length:]\n",
    "current_batch = first_eval_batch.reshape((1, length, n_features))\n",
    "\n",
    "for i in range(len(test)):\n",
    "    \n",
    "    # get prediction 1 time stamp ahead ([0] is for grabbing just the number instead of [array])\n",
    "    current_pred = model.predict(current_batch)[0]\n",
    "    \n",
    "    # store prediction\n",
    "    test_predictions.append(current_pred) \n",
    "    \n",
    "    # update batch to now include prediction and drop first value\n",
    "    current_batch = np.append(current_batch[:,1:,:],[[current_pred]],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[array([0.12666073, 0.08996339, 0.7190974 , 0.36104494, 0.54831064,\n",
       "        0.4291169 , 0.42736438, 0.45519555, 0.7263751 , 0.38756314,\n",
       "        0.2875382 , 0.3892692 , 0.65530246, 0.0064019 , 0.61901623,\n",
       "        0.39769527, 0.54877293, 0.30695575, 0.5420783 , 0.35719806,\n",
       "        0.59826505, 0.5541311 , 0.3933694 , 0.18549034, 0.585099  ,\n",
       "        0.480112  , 0.50588864, 0.5072879 ], dtype=float32),\n",
       " array([0.12966822, 0.13813506, 0.7235894 , 0.35265893, 0.55264044,\n",
       "        0.4196531 , 0.4265879 , 0.46761194, 0.7282277 , 0.38263163,\n",
       "        0.28997192, 0.42143172, 0.6442711 , 0.00793535, 0.6257839 ,\n",
       "        0.3920707 , 0.5732043 , 0.29765746, 0.53481907, 0.35216203,\n",
       "        0.5905128 , 0.56288624, 0.3692661 , 0.17617315, 0.5702653 ,\n",
       "        0.45214626, 0.505977  , 0.51701885], dtype=float32),\n",
       " array([0.12946774, 0.16756342, 0.72496855, 0.34270477, 0.55773157,\n",
       "        0.40929344, 0.42363685, 0.4715474 , 0.7293334 , 0.37480983,\n",
       "        0.29166305, 0.45240924, 0.6302188 , 0.00775022, 0.62829643,\n",
       "        0.37968016, 0.596252  , 0.28934565, 0.5257764 , 0.3471005 ,\n",
       "        0.585973  , 0.5709919 , 0.34259155, 0.16723068, 0.5591853 ,\n",
       "        0.41762218, 0.50918907, 0.52199817], dtype=float32),\n",
       " array([0.13045551, 0.18513457, 0.72690207, 0.3345285 , 0.56330025,\n",
       "        0.39902988, 0.42142916, 0.47362253, 0.7306733 , 0.36764717,\n",
       "        0.2956989 , 0.4833321 , 0.6162778 , 0.00670964, 0.62950635,\n",
       "        0.36480853, 0.6190346 , 0.2825483 , 0.51586914, 0.34238243,\n",
       "        0.5828801 , 0.57764125, 0.31579116, 0.15943475, 0.55152464,\n",
       "        0.38274816, 0.5122113 , 0.526143  ], dtype=float32),\n",
       " array([0.13202707, 0.19392566, 0.7297336 , 0.32728466, 0.5689225 ,\n",
       "        0.38930476, 0.42000648, 0.47538933, 0.732516  , 0.36105335,\n",
       "        0.30143273, 0.5141049 , 0.6033684 , 0.00484789, 0.6296671 ,\n",
       "        0.34825486, 0.6418156 , 0.2767924 , 0.5061649 , 0.33775133,\n",
       "        0.5810087 , 0.583131  , 0.28889245, 0.15295249, 0.54668915,\n",
       "        0.3484921 , 0.51486754, 0.5299095 ], dtype=float32),\n",
       " array([0.13397086, 0.19600977, 0.7333121 , 0.32035226, 0.57400733,\n",
       "        0.3801    , 0.41917235, 0.47719362, 0.7347535 , 0.3545412 ,\n",
       "        0.308486  , 0.5446075 , 0.591571  , 0.0023404 , 0.6290882 ,\n",
       "        0.33040777, 0.6646159 , 0.2718531 , 0.49711263, 0.33315903,\n",
       "        0.5801087 , 0.5873113 , 0.26187336, 0.1479849 , 0.5440216 ,\n",
       "        0.31510058, 0.51708347, 0.533343  ], dtype=float32),\n",
       " array([ 1.3614486e-01,  1.9294430e-01,  7.3757005e-01,  3.1353027e-01,\n",
       "         5.7826984e-01,  3.7155005e-01,  4.1881165e-01,  4.7913966e-01,\n",
       "         7.3738956e-01,  3.4788182e-01,  3.1673348e-01,  5.7486320e-01,\n",
       "         5.8100313e-01, -5.7747960e-04,  6.2821513e-01,  3.1167892e-01,\n",
       "         6.8750888e-01,  2.6765460e-01,  4.8893699e-01,  3.2861984e-01,\n",
       "         5.8022743e-01,  5.9021735e-01,  2.3482284e-01,  1.4480096e-01,\n",
       "         5.4298013e-01,  2.8283271e-01,  5.1899719e-01,  5.3656560e-01],\n",
       "       dtype=float32),\n",
       " array([ 0.13845803,  0.18597855,  0.7423643 ,  0.30678737,  0.5814869 ,\n",
       "         0.3637833 ,  0.4188219 ,  0.4811858 ,  0.7403962 ,  0.34099534,\n",
       "         0.3260863 ,  0.6048688 ,  0.5717267 , -0.00374351,  0.6273485 ,\n",
       "         0.2924311 ,  0.71047443,  0.2641412 ,  0.48172572,  0.32409537,\n",
       "         0.58144605,  0.5918515 ,  0.20780051,  0.14357668,  0.5430324 ,\n",
       "         0.2519399 ,  0.52068627,  0.539604  ], dtype=float32),\n",
       " array([ 0.14086394,  0.17616434,  0.74759185,  0.30018637,  0.58349854,\n",
       "         0.3569426 ,  0.41913274,  0.48326373,  0.7437781 ,  0.33391368,\n",
       "         0.33648434,  0.63459617,  0.5638047 , -0.00703206,  0.6267185 ,\n",
       "         0.2730227 ,  0.73347634,  0.26129025,  0.47553846,  0.31953835,\n",
       "         0.58386666,  0.5922446 ,  0.18091547,  0.14443113,  0.54375815,\n",
       "         0.22269374,  0.52222055,  0.54248315], dtype=float32),\n",
       " array([ 0.14334475,  0.16439618,  0.7531766 ,  0.29383016,  0.5841798 ,\n",
       "         0.35116634,  0.41969186,  0.48529732,  0.7475429 ,  0.32673547,\n",
       "         0.34786028,  0.6639852 ,  0.5572804 , -0.01035259,  0.6264817 ,\n",
       "         0.2537884 ,  0.7564604 ,  0.2590903 ,  0.47040993,  0.3148977 ,\n",
       "         0.58756304,  0.59143066,  0.15429881,  0.14742461,  0.5448414 ,\n",
       "         0.1953626 ,  0.523648  ,  0.54521656], dtype=float32),\n",
       " array([ 0.14590284,  0.15143198,  0.75907075,  0.28784287,  0.5834402 ,\n",
       "         0.34658557,  0.4204659 ,  0.4872185 ,  0.75169325,  0.31960508,\n",
       "         0.36013374,  0.6929549 ,  0.55216765, -0.0136324 ,  0.62673664,\n",
       "         0.23503163,  0.77936935,  0.25754106,  0.46635717,  0.31013092,\n",
       "         0.5925683 ,  0.5894497 ,  0.12810512,  0.15256742,  0.54606575,\n",
       "         0.17019859,  0.5250047 ,  0.54781085], dtype=float32),\n",
       " array([ 0.14855269,  0.13790287,  0.76524466,  0.28235912,  0.5812228 ,\n",
       "         0.34331948,  0.4214372 ,  0.48897448,  0.7562175 ,  0.3126959 ,\n",
       "         0.37320638,  0.72141385,  0.54843926, -0.01680276,  0.6275314 ,\n",
       "         0.2170178 ,  0.80215   ,  0.2566524 ,  0.46337608,  0.30520916,\n",
       "         0.5988651 ,  0.5863448 ,  0.10250609,  0.15982428,  0.54729426,\n",
       "         0.14742327,  0.52631706,  0.55026513], dtype=float32),\n",
       " array([ 0.15131585,  0.12432127,  0.77168125,  0.277518  ,  0.5775033 ,\n",
       "         0.341473  ,  0.4226006 ,  0.49053302,  0.76108825,  0.30619738,\n",
       "         0.38696438,  0.7492706 ,  0.54602176, -0.01978786,  0.6288725 ,\n",
       "         0.19997013,  0.82475996,  0.2564445 ,  0.461442  ,  0.30012003,\n",
       "         0.6063842 ,  0.58216155,  0.07768585,  0.16911952,  0.54844964,\n",
       "         0.12721653,  0.5276065 ,  0.5525727 ], dtype=float32),\n",
       " array([ 0.15421602,  0.11108827,  0.77836776,  0.27345794,  0.5722875 ,\n",
       "         0.34113407,  0.42395863,  0.49188587,  0.766262  ,  0.300303  ,\n",
       "         0.40128013,  0.7764405 ,  0.54479396, -0.02249664,  0.6307328 ,\n",
       "         0.18406875,  0.84717023,  0.25694722,  0.4605092 ,  0.29486912,\n",
       "         0.6150064 ,  0.5769456 ,  0.05383512,  0.18034214,  0.5494954 ,\n",
       "         0.10970932,  0.5288904 ,  0.5547223 ], dtype=float32),\n",
       " array([ 0.15727574,  0.09850342,  0.78529066,  0.270313  ,  0.56560653,\n",
       "         0.3423733 ,  0.42551744,  0.4930496 ,  0.7716832 ,  0.29519916,\n",
       "         0.41601825,  0.80285025,  0.5445905 , -0.02481806,  0.6330598 ,\n",
       "         0.16945195,  0.8693684 ,  0.2581988 ,  0.4605149 ,  0.2894798 ,\n",
       "         0.62456757,  0.57074237,  0.0311465 ,  0.19335225,  0.55042064,\n",
       "         0.09498045,  0.53018296,  0.55670065], dtype=float32),\n",
       " array([ 0.16051342,  0.08677597,  0.7924319 ,  0.26820853,  0.5575113 ,\n",
       "         0.345244  ,  0.4272825 ,  0.49406424,  0.77728736,  0.29105535,\n",
       "         0.43104127,  0.82843876,  0.54520714, -0.02661921,  0.63578475,\n",
       "         0.15621996,  0.89135736,  0.2602439 ,  0.4613835 ,  0.28399265,\n",
       "         0.63486826,  0.5635955 ,  0.00980953,  0.20798802,  0.5512298 ,\n",
       "         0.08305752,  0.5314958 ,  0.5584941 ], dtype=float32),\n",
       " array([ 0.16394134,  0.07603765,  0.7997648 ,  0.2672585 ,  0.5480656 ,\n",
       "         0.3497835 ,  0.42925584,  0.49499086,  0.7830064 ,  0.28801677,\n",
       "         0.44621643,  0.8531567 ,  0.5464097 , -0.02774557,  0.6388313 ,\n",
       "         0.1444393 ,  0.91315466,  0.2631315 ,  0.4630327 ,  0.27846476,\n",
       "         0.6456822 ,  0.5555466 , -0.00999276,  0.22407314,  0.55193496,\n",
       "         0.07392048,  0.53283757,  0.56009036], dtype=float32),\n",
       " array([ 0.16756418,  0.06635503,  0.80725324,  0.2675625 ,  0.53733987,\n",
       "         0.35601443,  0.4314337 ,  0.49590594,  0.7887725 ,  0.2861993 ,\n",
       "         0.46142247,  0.87696356,  0.5479418 , -0.02802248,  0.6421229 ,\n",
       "         0.13414851,  0.93479025,  0.2669122 ,  0.4653796 ,  0.27296793,\n",
       "         0.6567667 ,  0.5466357 , -0.02808701,  0.24142352,  0.55255264,\n",
       "         0.0675077 ,  0.53421426,  0.56147933], dtype=float32),\n",
       " array([ 0.17137814,  0.05774223,  0.81485176,  0.26920402,  0.52540565,\n",
       "         0.3639469 ,  0.4338062 ,  0.49689642,  0.7945204 ,  0.2856866 ,\n",
       "         0.47655475,  0.8998244 ,  0.5495336 , -0.02725772,  0.645589  ,\n",
       "         0.12536329,  0.95630354,  0.2716359 ,  0.46834612,  0.26758665,\n",
       "         0.66787195,  0.53690207, -0.04431137,  0.25985357,  0.5531019 ,\n",
       "         0.06372269,  0.5356284 ,  0.562654  ], dtype=float32),\n",
       " array([ 0.17537028,  0.05017202,  0.82250583,  0.2722488 ,  0.51233107,\n",
       "         0.37358022,  0.43635717,  0.4980531 ,  0.80019057,  0.28653085,\n",
       "         0.49153042,  0.9217055 ,  0.55091023, -0.0252445 ,  0.64916986,\n",
       "         0.11808158,  0.9777406 ,  0.27734953,  0.47186413,  0.26241648,\n",
       "         0.6787489 ,  0.5263866 , -0.05851684,  0.27918184,  0.5536044 ,\n",
       "         0.06244112,  0.5370794 ,  0.5636111 ], dtype=float32),\n",
       " array([ 0.1795187 ,  0.04358619,  0.8301533 ,  0.2767442 ,  0.49817774,\n",
       "         0.38490435,  0.43906567,  0.49946484,  0.80572945,  0.28875384,\n",
       "         0.50629085,  0.94257015,  0.55179834, -0.02176393,  0.65281916,\n",
       "         0.11228774,  0.99915004,  0.28409457,  0.4758787 ,  0.25756115,\n",
       "         0.6891556 ,  0.5151329 , -0.07056694,  0.29923445,  0.5540842 ,\n",
       "         0.06351746,  0.53856295,  0.56434983], dtype=float32),\n",
       " array([ 0.1837922 ,  0.03790425,  0.8377273 ,  0.28271806,  0.48300004,\n",
       "         0.3979012 ,  0.4419081 ,  0.50121355,  0.8110904 ,  0.2923506 ,\n",
       "         0.5208036 ,  0.96237487,  0.55193216, -0.01658779,  0.6565051 ,\n",
       "         0.10795577,  1.0205795 ,  0.29190576,  0.48035082,  0.25313073,\n",
       "         0.6988627 ,  0.50319093, -0.08033714,  0.3198482 ,  0.55456835,\n",
       "         0.06679074,  0.5400716 ,  0.56487316], dtype=float32),\n",
       " array([ 0.18815073,  0.03303098,  0.8451588 ,  0.2901782 ,  0.46684444,\n",
       "         0.41254488,  0.44486052,  0.50336874,  0.81623226,  0.29729372,\n",
       "         0.53506184,  0.9810656 ,  0.5510587 , -0.00948143,  0.66021085,\n",
       "         0.10505221,  1.0420712 ,  0.3008085 ,  0.48525855,  0.24923827,\n",
       "         0.7076561 ,  0.49061924, -0.08771364,  0.3408722 ,  0.5550863 ,\n",
       "         0.07208947,  0.5415946 ,  0.5651865 ], dtype=float32),\n",
       " array([ 1.92545161e-01,  2.88623925e-02,  8.52378547e-01,  2.99111366e-01,\n",
       "         4.49751288e-01,  4.28801715e-01,  4.47901309e-01,  5.05984485e-01,\n",
       "         8.21119905e-01,  3.03537518e-01,  5.49083531e-01,  9.98574317e-01,\n",
       "         5.48941791e-01, -2.05606222e-04,  6.63933218e-01,  1.03537664e-01,\n",
       "         1.06365836e+00,  3.10817301e-01,  4.90596503e-01,  2.45997205e-01,\n",
       "         7.15339363e-01,  4.77487803e-01, -9.25911292e-02,  3.62168014e-01,\n",
       "         5.55670023e-01,  7.92347565e-02,  5.43118477e-01,  5.65298140e-01],\n",
       "       dtype=float32),\n",
       " array([ 0.19691756,  0.025291  ,  0.8593209 ,  0.30948168,  0.4317567 ,\n",
       "         0.44662887,  0.45101297,  0.5090957 ,  0.82572174,  0.311022  ,\n",
       "         0.5629073 ,  1.0148165 ,  0.5453657 ,  0.01148088,  0.6676806 ,\n",
       "         0.10336749,  1.0853591 ,  0.32193336,  0.49637505,  0.24351789,\n",
       "         0.7217347 ,  0.46388066, -0.09487098,  0.38360932,  0.5563526 ,\n",
       "         0.08804356,  0.5446267 ,  0.5652186 ], dtype=float32),\n",
       " array([ 0.20120125,  0.02221001,  0.8659256 ,  0.32122892,  0.41289592,\n",
       "         0.4659728 ,  0.45418486,  0.5127167 ,  0.83000946,  0.31967688,\n",
       "         0.5765903 ,  1.0296891 ,  0.5401375 ,  0.02581866,  0.67147195,\n",
       "         0.10449211,  1.1071736 ,  0.33414283,  0.5026183 ,  0.24190482,\n",
       "         0.72668356,  0.4498974 , -0.09445897,  0.40508038,  0.5571678 ,\n",
       "         0.09832998,  0.54610133,  0.5649608 ], dtype=float32),\n",
       " array([ 0.20532095,  0.01951692,  0.87213975,  0.33426714,  0.39320576,\n",
       "         0.4867676 ,  0.45741424,  0.5168395 ,  0.8339569 ,  0.32942408,\n",
       "         0.5902022 ,  1.0430694 ,  0.5330906 ,  0.04304502,  0.6753332 ,\n",
       "         0.10685633,  1.1290779 ,  0.34741423,  0.50936174,  0.24125354,\n",
       "         0.73004514,  0.4356559 , -0.09126308,  0.42647383,  0.5581492 ,\n",
       "         0.1099064 ,  0.5475218 ,  0.56453997], dtype=float32),\n",
       " array([ 0.20919313,  0.01711586,  0.87792075,  0.3484823 ,  0.37272844,\n",
       "         0.50893176,  0.46070707,  0.5214334 ,  0.8375394 ,  0.34017983,\n",
       "         0.603821  ,  1.0548149 ,  0.5240855 ,  0.06339041,  0.6792954 ,\n",
       "         0.11039838,  1.1510208 ,  0.36169615,  0.51664823,  0.24164769,\n",
       "         0.73169607,  0.42129254, -0.08519221,  0.4476879 ,  0.55932987,\n",
       "         0.12258388,  0.54886603,  0.56397384], dtype=float32),\n",
       " array([ 0.21272679,  0.0149201 ,  0.8832355 ,  0.36373138,  0.35151407,\n",
       "         0.5323661 ,  0.4640782 ,  0.52644527,  0.8407335 ,  0.35185552,\n",
       "         0.617527  ,  1.0647652 ,  0.51301277,  0.08707498,  0.68339187,\n",
       "         0.11504802,  1.17292   ,  0.37691614,  0.52452457,  0.24315661,\n",
       "         0.73152876,  0.40696353, -0.07615503,  0.46862373,  0.56074226,\n",
       "         0.13617235,  0.5501105 ,  0.5632827 ], dtype=float32),\n",
       " array([ 0.21582463,  0.01285346,  0.8880617 ,  0.3798414 ,  0.32962334,\n",
       "         0.5569525 ,  0.46755025,  0.5318004 ,  0.8435162 ,  0.36435783,\n",
       "         0.63139695,  1.0727435 ,  0.49979365,  0.11430405,  0.6876544 ,\n",
       "         0.12072505,  1.19466   ,  0.3929789 ,  0.53303677,  0.2458341 ,\n",
       "         0.72944933,  0.39284497, -0.0640591 ,  0.48918188,  0.5624184 ,\n",
       "         0.15048173,  0.55123085,  0.56248957], dtype=float32),\n",
       " array([ 0.21838476,  0.0108514 ,  0.8923867 ,  0.39661103,  0.30712914,\n",
       "         0.58255064,  0.47115162,  0.5374041 ,  0.8458665 ,  0.37758815,\n",
       "         0.645499  ,  1.0785617 ,  0.48438114,  0.14526343,  0.6921111 ,\n",
       "         0.1273372 ,  1.2160916 ,  0.40976593,  0.54222554,  0.2497166 ,\n",
       "         0.7253754 ,  0.3791325 , -0.04881179,  0.5092599 ,  0.5643907 ,\n",
       "         0.16532174,  0.55220175,  0.56161934], dtype=float32),\n",
       " array([ 0.22030292,  0.00886149,  0.89620525,  0.41381174,  0.28411728,\n",
       "         0.6089988 ,  0.4749134 ,  0.5431442 ,  0.8477638 ,  0.39144054,\n",
       "         0.65988684,  1.0820254 ,  0.46676037,  0.180114  ,  0.69678134,\n",
       "         0.13477795,  1.2370319 ,  0.42713606,  0.55212086,  0.2548231 ,\n",
       "         0.7192335 ,  0.3660397 , -0.03032108,  0.52874994,  0.5666914 ,\n",
       "         0.18050385,  0.5529975 ,  0.56069934], dtype=float32),\n",
       " array([ 0.22147559,  0.00684343,  0.8995168 ,  0.43119246,  0.26068732,\n",
       "         0.6361135 ,  0.47886577,  0.5488944 ,  0.84919   ,  0.4058    ,\n",
       "         0.6745948 ,  1.0829408 ,  0.44694924,  0.21898493,  0.7016722 ,\n",
       "         0.14292517,  1.257269  ,  0.44492695,  0.56273776,  0.2611549 ,\n",
       "         0.71095634,  0.35379672, -0.0084978 ,  0.5475375 ,  0.5693537 ,\n",
       "         0.1958423 ,  0.5535933 ,  0.55975854], dtype=float32),\n",
       " array([0.22180331, 0.00476833, 0.902322  , 0.4484846 , 0.23695189,\n",
       "        0.66369045, 0.48303306, 0.5545176 , 0.85012954, 0.4205394 ,\n",
       "        0.68963456, 1.0811231 , 0.42499727, 0.26196602, 0.70677507,\n",
       "        0.15163907, 1.2765659 , 0.462957  , 0.5740707 , 0.2686962 ,\n",
       "        0.70047975, 0.34264728, 0.01674077, 0.5655015 , 0.57241046,\n",
       "        0.21115658, 0.5539648 , 0.55882657], dtype=float32),\n",
       " array([0.22119495, 0.0026179 , 0.9046182 , 0.46540934, 0.21303618,\n",
       "        0.6915058 , 0.48742846, 0.55987084, 0.85056794, 0.43551624,\n",
       "        0.7049913 , 1.0764035 , 0.40098578, 0.3090981 , 0.7120612 ,\n",
       "        0.1607616 , 1.2946682 , 0.48102844, 0.5860911 , 0.27741396,\n",
       "        0.68774265, 0.3328449 , 0.04546661, 0.5825156 , 0.575893  ,\n",
       "        0.22627336, 0.5540897 , 0.5579327 ], dtype=float32),\n",
       " array([2.1957187e-01, 3.8254820e-04, 9.0639418e-01, 4.8168474e-01,\n",
       "        1.8907814e-01, 7.1931839e-01, 4.9204931e-01, 5.6481004e-01,\n",
       "        8.5049456e-01, 4.5056993e-01, 7.2062248e-01, 1.0686381e+00,\n",
       "        3.7502781e-01, 3.6036178e-01, 7.1748018e-01, 1.7011547e-01,\n",
       "        1.3113124e+00, 4.9893004e-01, 5.9874421e-01, 2.8725761e-01,\n",
       "        6.7268598e-01, 3.2464796e-01, 7.7734269e-02, 5.9845132e-01,\n",
       "        5.7982808e-01, 2.4102911e-01, 5.5394900e-01, 5.5710530e-01],\n",
       "       dtype=float32),\n",
       " array([ 0.21687217, -0.00194014,  0.9076264 ,  0.4970339 ,  0.165227  ,\n",
       "         0.7468717 ,  0.49687213,  0.5691948 ,  0.8499    ,  0.46551767,\n",
       "         0.73645854,  1.0577152 ,  0.3472683 ,  0.4156631 ,  0.7229583 ,\n",
       "         0.17950419,  1.3262355 ,  0.5164403 ,  0.61194974,  0.29815674,\n",
       "         0.655254  ,  0.31831375,  0.1135716 ,  0.61318237,  0.58423454,\n",
       "         0.25527197,  0.55352664,  0.5563696 ], dtype=float32),\n",
       " array([ 0.21305548, -0.0043473 ,  0.90827537,  0.5111931 ,  0.14164482,\n",
       "         0.7738954 ,  0.5018493 ,  0.57289404,  0.8487783 ,  0.4801532 ,\n",
       "         0.75240403,  1.0435611 ,  0.31788483,  0.47481826,  0.7284008 ,\n",
       "         0.18871322,  1.3391854 ,  0.5333291 ,  0.6256029 ,  0.31001946,\n",
       "         0.6353995 ,  0.31409144,  0.15297133,  0.6265909 ,  0.5891185 ,\n",
       "         0.26886427,  0.55281216,  0.5557478 ], dtype=float32),\n",
       " array([ 0.20810662, -0.00683237,  0.9082832 ,  0.52391857,  0.11850595,\n",
       "         0.8001071 ,  0.5069075 ,  0.57578915,  0.8471253 ,  0.4942454 ,\n",
       "         0.76834273,  1.0261458 ,  0.2870889 ,  0.53753716,  0.7336955 ,\n",
       "         0.1975109 ,  1.3499299 ,  0.54936033,  0.6395795 ,  0.322727  ,\n",
       "         0.61309004,  0.3122127 ,  0.19587776,  0.6385741 ,  0.5944671 ,\n",
       "         0.2816836 ,  0.55180055,  0.5552568 ], dtype=float32),\n",
       " array([ 0.20203982, -0.00938586,  0.9075741 ,  0.53499395,  0.09599641,\n",
       "         0.825215  ,  0.5119496 ,  0.57777816,  0.8449409 ,  0.5075394 ,\n",
       "         0.7841432 ,  1.005488  ,  0.2551266 ,  0.6034073 ,  0.7387191 ,\n",
       "         0.20565204,  1.3582683 ,  0.56429386,  0.6537425 ,  0.33613002,\n",
       "         0.5883172 ,  0.31288153,  0.24217436,  0.6490508 ,  0.60024416,\n",
       "         0.29362476,  0.5504942 ,  0.5549071 ], dtype=float32),\n",
       " array([ 0.19490157, -0.01199591,  0.90605795,  0.54423547,  0.07431333,\n",
       "         0.84892166,  0.5168567 ,  0.5787772 ,  0.84222937,  0.5197617 ,\n",
       "         0.7996672 ,  0.98165953,  0.22227846,  0.6718838 ,  0.7433473 ,\n",
       "         0.21288183,  1.3640385 ,  0.5778878 ,  0.66795146,  0.35004383,\n",
       "         0.56110805,  0.31626272,  0.291672  ,  0.65796775,  0.60638434,\n",
       "         0.30459967,  0.5489032 ,  0.55470264], dtype=float32),\n",
       " array([ 0.18677235, -0.01464752,  0.9036356 ,  0.551497  ,  0.05365989,\n",
       "         0.87093043,  0.5214962 ,  0.57871944,  0.83900255,  0.5306269 ,\n",
       "         0.81477803,  0.9547896 ,  0.1888567 ,  0.7422853 ,  0.74746436,\n",
       "         0.21894187,  1.3671256 ,  0.5899027 ,  0.6820722 ,  0.36424726,\n",
       "         0.53153646,  0.3224686 ,  0.34409833,  0.66530436,  0.6127889 ,\n",
       "         0.3145386 ,  0.5470466 ,  0.5546396 ], dtype=float32),\n",
       " array([ 0.17776689, -0.01732085,  0.9002086 ,  0.5566732 ,  0.03423911,\n",
       "         0.8909533 ,  0.5257291 ,  0.57755244,  0.83528084,  0.5398499 ,\n",
       "         0.82935035,  0.9250691 ,  0.15519938,  0.8138025 ,  0.7509755 ,\n",
       "         0.22357878,  1.3674688 ,  0.6001081 ,  0.69598746,  0.37848276,\n",
       "         0.49973428,  0.33154622,  0.39909193,  0.671074  ,  0.6193233 ,\n",
       "         0.32338926,  0.544953  ,  0.5547056 ], dtype=float32),\n",
       " array([ 0.16803257, -0.01998944,  0.8956921 ,  0.55970335,  0.01624393,\n",
       "         0.9087236 ,  0.5294218 ,  0.5752327 ,  0.8310982 ,  0.5471609 ,\n",
       "         0.8432772 ,  0.89275604,  0.12166043,  0.8855203 ,  0.7538174 ,\n",
       "         0.22655386,  1.3650665 ,  0.6082897 ,  0.7096057 ,  0.3924638 ,\n",
       "         0.46589935,  0.34346604,  0.45620286,  0.6753223 ,  0.6258172 ,\n",
       "         0.33111694,  0.54266053,  0.55488133], dtype=float32),\n",
       " array([ 1.5774438e-01, -2.2617573e-02,  8.9002699e-01,  5.6057185e-01,\n",
       "        -1.5529990e-04,  9.2400885e-01,  5.3245491e-01,  5.7171929e-01,\n",
       "         8.2650191e-01,  5.5232227e-01,  8.5647672e-01,  8.5817629e-01,\n",
       "         8.8596322e-02,  9.5645380e-01,  7.5596613e-01,  2.2765487e-01,\n",
       "         1.3599783e+00,  6.1426014e-01,  7.2286874e-01,  4.0588444e-01,\n",
       "         4.3029863e-01,  3.5811269e-01,  5.1489943e-01,  6.7812300e-01,\n",
       "         6.3206637e-01,  3.3770353e-01,  5.4021573e-01,  5.5514050e-01],\n",
       "       dtype=float32),\n",
       " array([ 0.14709851, -0.02515929,  0.8831937 ,  0.559309  , -0.01482258,\n",
       "         0.93662727,  0.5347349 ,  0.56696707,  0.8215562 ,  0.55514586,\n",
       "         0.86889654,  0.8217249 ,  0.05635052,  1.0255977 ,  0.7574423 ,\n",
       "         0.22670835,  1.3523271 ,  0.61786896,  0.7357552 ,  0.41843516,\n",
       "         0.39326644,  0.37528312,  0.5745847 ,  0.67957085,  0.6378396 ,\n",
       "         0.34314826,  0.53767323,  0.55545217], dtype=float32),\n",
       " array([ 0.13630322, -0.02755789,  0.8752198 ,  0.5559905 , -0.02766775,\n",
       "         0.9464598 ,  0.5362004 ,  0.560923  ,  0.81634027,  0.5555086 ,\n",
       "         0.8805141 ,  0.78385836,  0.02523666,  1.0919794 ,  0.7583126 ,\n",
       "         0.22359085,  1.3422987 ,  0.6190128 ,  0.74828005,  0.42981946,\n",
       "         0.35519415,  0.3946901 ,  0.63461936,  0.6797733 ,  0.64288574,\n",
       "         0.3474682 ,  0.5350924 ,  0.5557832 ], dtype=float32),\n",
       " array([ 0.12556867, -0.02974747,  0.8661866 ,  0.55073476, -0.0386529 ,\n",
       "         0.9534601 ,  0.53682774,  0.5535255 ,  0.81094736,  0.5533625 ,\n",
       "         0.89133936,  0.7450832 , -0.00447582,  1.1547154 ,  0.75868523,\n",
       "         0.2182383 ,  1.330135  ,  0.6176416 ,  0.76049024,  0.43977064,\n",
       "         0.31651413,  0.41597605,  0.69434965,  0.67884225,  0.6469461 ,\n",
       "         0.35069966,  0.5325351 ,  0.5561014 ], dtype=float32),\n",
       " array([ 0.11509659, -0.03165597,  0.85622466,  0.543698  , -0.0477946 ,\n",
       "         0.9576601 ,  0.5366309 ,  0.5447085 ,  0.80548126,  0.54873884,\n",
       "         0.9014118 ,  0.70593625, -0.03257356,  1.2130593 ,  0.7587008 ,\n",
       "         0.21065071,  1.31613   ,  0.6137636 ,  0.77245617,  0.44806647,\n",
       "         0.27767718,  0.43873328,  0.75313914,  0.6768886 ,  0.6497664 ,\n",
       "         0.35289943,  0.5300613 ,  0.5563778 ], dtype=float32),\n",
       " array([ 0.10507051, -0.03321031,  0.84550816,  0.5350685 , -0.05516034,\n",
       "         0.95916605,  0.535659  ,  0.53440964,  0.8000512 ,  0.5417452 ,\n",
       "         0.9108008 ,  0.66695976, -0.0589054 ,  1.2664375 ,  0.7585199 ,\n",
       "         0.20089269,  1.3006195 ,  0.6074445 ,  0.7842622 ,  0.4545396 ,\n",
       "         0.2391274 ,  0.4625317 ,  0.8104015 ,  0.67401785,  0.6511119 ,\n",
       "         0.3541462 ,  0.5277265 ,  0.5565904 ], dtype=float32),\n",
       " array([ 0.09564819, -0.0343416 ,  0.834242  ,  0.52505857, -0.06086032,\n",
       "         0.9581517 ,  0.5339897 ,  0.5225808 ,  0.7947686 ,  0.5325564 ,\n",
       "         0.91960377,  0.6286744 , -0.08338466,  1.3144721 ,  0.7583097 ,\n",
       "         0.18908823,  1.28397   ,  0.5988036 ,  0.7959945 ,  0.45908403,\n",
       "         0.2012764 ,  0.48694688,  0.86562794,  0.670329  ,  0.6507797 ,\n",
       "         0.35453886,  0.5255769 ,  0.55672467], dtype=float32),\n",
       " array([ 0.08695579, -0.03499099,  0.8226483 ,  0.5138953 , -0.06503694,\n",
       "         0.9548464 ,  0.53172106,  0.5091992 ,  0.7897404 ,  0.52140015,\n",
       "         0.9279431 ,  0.5915527 , -0.10598505,  1.3569837 ,  0.75822854,\n",
       "         0.17541146,  1.2665641 ,  0.588007  ,  0.80773175,  0.4616558 ,\n",
       "         0.16448095,  0.5115878 ,  0.91840965,  0.66591704,  0.64861196,\n",
       "         0.3541941 ,  0.5236482 ,  0.55677533], dtype=float32),\n",
       " array([ 0.07908472, -0.03511452,  0.81095153,  0.501812  , -0.06785248,\n",
       "         0.94951797,  0.5289625 ,  0.49427477,  0.785066  ,  0.5085407 ,\n",
       "         0.93596345,  0.55599844, -0.1267336 ,  1.3939822 ,  0.75841343,\n",
       "         0.16007462,  1.2487878 ,  0.5752591 ,  0.81953585,  0.46227187,\n",
       "         0.12902716,  0.5361177 ,  0.9684499 ,  0.66087604,  0.64450455,\n",
       "         0.35324106,  0.52196395,  0.55674565], dtype=float32),\n",
       " array([ 0.07209188, -0.03468651,  0.7993679 ,  0.48903936, -0.06947798,\n",
       "         0.9424597 ,  0.52582663,  0.4778548 ,  0.7808336 ,  0.4942622 ,\n",
       "         0.9438258 ,  0.52233255, -0.14569962,  1.4256425 ,  0.7589738 ,\n",
       "         0.14331505,  1.2310162 ,  0.5607914 ,  0.83144784,  0.4610048 ,\n",
       "         0.09512272,  0.5602668 ,  1.0155674 ,  0.65530384,  0.63841265,\n",
       "         0.35181528,  0.52053595,  0.5566458 ], dtype=float32),\n",
       " array([ 0.06600137, -0.03370153,  0.7880952 ,  0.47579992, -0.07008363,\n",
       "         0.9339733 ,  0.5224213 ,  0.46002308,  0.77711743,  0.47885424,\n",
       "         0.95169914,  0.490787  , -0.16298327,  1.4522728 ,  0.75998557,\n",
       "         0.12538205,  1.2136002 ,  0.5448528 ,  0.84348595,  0.45797443,\n",
       "         0.06289711,  0.5838358 ,  1.0596895 ,  0.6493061 ,  0.630351  ,\n",
       "         0.35005093,  0.51936513,  0.55649126], dtype=float32),\n",
       " array([ 0.0608086 , -0.03217363,  0.777307  ,  0.4623019 , -0.06983157,\n",
       "         0.9243591 ,  0.5188446 ,  0.44089547,  0.7739763 ,  0.4626    ,\n",
       "         0.959753  ,  0.46150708, -0.17870373,  1.4742776 ,  0.76149255,\n",
       "         0.10652693,  1.1968579 ,  0.5276996 ,  0.8556462 ,  0.45334   ,\n",
       "         0.0324083 ,  0.6066909 ,  1.1008369 ,  0.64299953,  0.6203917 ,\n",
       "         0.3480743 ,  0.51844394,  0.5563005 ], dtype=float32),\n",
       " array([ 0.05648542, -0.03013501,  0.7671515 ,  0.4487372 , -0.0688716 ,\n",
       "         0.9139042 ,  0.51518106,  0.42061272,  0.77145344,  0.44576725,\n",
       "         0.96814835,  0.43455958, -0.19298899,  1.4921212 ,  0.76350915,\n",
       "         0.08699431,  1.1810659 ,  0.5095869 ,  0.8679048 ,  0.4472903 ,\n",
       "         0.00365496,  0.6287531 ,  1.1391038 ,  0.63651335,  0.60865647,\n",
       "         0.34599805,  0.517758  ,  0.55609286], dtype=float32),\n",
       " array([ 0.05298544, -0.02763291,  0.7577496 ,  0.43527836, -0.0673393 ,\n",
       "         0.9028778 ,  0.51149935,  0.3993306 ,  0.76957625,  0.42860347,\n",
       "         0.97702926,  0.4099438 , -0.20596814,  1.5062938 ,  0.76602656,\n",
       "         0.06701655,  1.1664556 ,  0.4907619 ,  0.88022274,  0.44003338,\n",
       "        -0.02341028,  0.6499851 ,  1.174637  ,  0.62998813,  0.5953077 ,\n",
       "         0.34391722,  0.51728857,  0.55588615], dtype=float32),\n",
       " array([ 0.05024908, -0.02472616,  0.74919647,  0.42207772, -0.06535523,\n",
       "         0.8915247 ,  0.50785184,  0.37721187,  0.7683571 ,  0.41133183,\n",
       "         0.98651636,  0.3876042 , -0.2177659 ,  1.5172831 ,  0.7690181 ,\n",
       "         0.04681011,  1.1532117 ,  0.47145876,  0.89254975,  0.43178844,\n",
       "        -0.0488666 ,  0.67037636,  1.2076153 ,  0.62357384,  0.5805384 ,\n",
       "         0.34190568,  0.5170141 ,  0.5556961 ], dtype=float32),\n",
       " array([ 0.04820856, -0.02148176,  0.74156314,  0.40926653, -0.06302583,\n",
       "         0.8800632 ,  0.50427437,  0.35441786,  0.7677952 ,  0.39415   ,\n",
       "         0.9967022 ,  0.3674436 , -0.22850001,  1.5255531 ,  0.7724446 ,\n",
       "         0.02657317,  1.1414717 ,  0.4518947 ,  0.9048273 ,  0.42277682,\n",
       "        -0.0728143 ,  0.6899319 ,  1.2382313 ,  0.6174256 ,  0.56456083,\n",
       "         0.34001526,  0.51691145,  0.5555348 ], dtype=float32),\n",
       " array([ 0.04679134, -0.01797146,  0.7348993 ,  0.39695424, -0.06044515,\n",
       "         0.8686837 ,  0.5007867 ,  0.33110142,  0.76787853,  0.37722832,\n",
       "         1.0076481 ,  0.34933427, -0.23828128,  1.5315292 ,  0.77625984,\n",
       "         0.00648474,  1.1313289 ,  0.43226793,  0.9169935 ,  0.4132154 ,\n",
       "        -0.09536453,  0.7086616 ,  1.2666789 ,  0.6116991 ,  0.54759645,\n",
       "         0.3382761 ,  0.5169572 ,  0.5554106 ], dtype=float32),\n",
       " array([ 0.04592316, -0.01426905,  0.7292353 ,  0.38522828, -0.05769594,\n",
       "         0.85754764,  0.49739465,  0.30740184,  0.768585  ,  0.36071104,\n",
       "         1.0193839 ,  0.33312818, -0.24721378,  1.5355881 ,  0.78041404,\n",
       "        -0.01329574,  1.1228352 ,  0.41275656,  0.92898494,  0.40331227,\n",
       "        -0.11663316,  0.72657365,  1.2931432 ,  0.6065457 ,  0.5298674 ,\n",
       "         0.33669618,  0.5171279 ,  0.5553285 ], dtype=float32),\n",
       " array([ 0.04553034, -0.01044854,  0.72458416,  0.37415332, -0.05485153,\n",
       "         0.84678805,  0.49409047,  0.2834408 ,  0.76988554,  0.34471533,\n",
       "         1.0319084 ,  0.3186651 , -0.25539654,  1.5380545 ,  0.7848566 ,\n",
       "        -0.03262894,  1.1160035 ,  0.3935178 ,  0.9407403 ,  0.39326072,\n",
       "        -0.13673724,  0.74367106,  1.3177947 ,  0.6021068 ,  0.5115877 ,\n",
       "         0.33526444,  0.51740104,  0.55528915], dtype=float32),\n",
       " array([ 0.04554143, -0.00658208,  0.72094434,  0.3637718 , -0.05197693,\n",
       "         0.83650994,  0.49085483,  0.25931984,  0.7717445 ,  0.32933182,\n",
       "         1.0451926 ,  0.30577978, -0.26292616,  1.5392005 ,  0.7895382 ,\n",
       "        -0.05139577,  1.1108136 ,  0.37468827,  0.9522009 ,  0.38323903,\n",
       "        -0.15579215,  0.7599506 ,  1.340789  ,  0.5985097 ,  0.4929585 ,\n",
       "         0.33395186,  0.51775545,  0.55529034], dtype=float32),\n",
       " array([ 0.04588805, -0.00273919,  0.71830124,  0.35410452, -0.04913033,\n",
       "         0.826792  ,  0.48765814,  0.23511863,  0.7741231 ,  0.31462577,\n",
       "         1.0591822 ,  0.29430747, -0.26989755,  1.539248  ,  0.79441214,\n",
       "        -0.06949667,  1.1072152 ,  0.35638362,  0.9633146 ,  0.3734069 ,\n",
       "        -0.1739118 ,  0.7754028 ,  1.3622643 ,  0.5958642 ,  0.47416237,\n",
       "         0.33271548,  0.5181718 ,  0.5553267 ], dtype=float32),\n",
       " array([ 4.6506196e-02,  1.0137875e-03,  7.1662945e-01,  3.4515184e-01,\n",
       "        -4.6363737e-02,  8.1768858e-01,  4.8446283e-01,  2.1089426e-01,\n",
       "         7.7697909e-01,  3.0063772e-01,  1.0738029e+00,  2.8408802e-01,\n",
       "        -2.7640563e-01,  1.5383732e+00,  7.9943591e-01, -8.6851917e-02,\n",
       "         1.1051331e+00,  3.3870038e-01,  9.7403544e-01,  3.6390728e-01,\n",
       "        -1.9120781e-01,  7.9001379e-01,  1.3823463e+00,  5.9425849e-01,\n",
       "         4.5536143e-01,  3.3150041e-01,  5.1863229e-01,  5.5539072e-01],\n",
       "       dtype=float32),\n",
       " array([ 0.04733622,  0.00461433,  0.71589434,  0.3368954 , -0.04372425,\n",
       "         0.80923104,  0.48122418,  0.18668112,  0.780269  ,  0.28738585,\n",
       "         1.0889647 ,  0.27496967, -0.28254578,  1.5367126 ,  0.8045705 ,\n",
       "        -0.1034007 ,  1.1044708 ,  0.32171595,  0.98432654,  0.35486525,\n",
       "        -0.20779108,  0.8037682 ,  1.4011484 ,  0.5937585 ,  0.43669412,\n",
       "         0.33024347,  0.51912177,  0.555472  ], dtype=float32),\n",
       " array([ 0.0483237 ,  0.00800411,  0.71605474,  0.32930082, -0.04125459,\n",
       "         0.8014326 ,  0.47789186,  0.16249144,  0.7839487 ,  0.27486783,\n",
       "         1.1045668 ,  0.26681253, -0.28841403,  1.5343698 ,  0.80978215,\n",
       "        -0.11910041,  1.1051167 ,  0.30549008,  0.9941591 ,  0.3463903 ,\n",
       "        -0.22377129,  0.8166506 ,  1.4187766 ,  0.59440607,  0.41827476,\n",
       "         0.32887718,  0.51962805,  0.5555591 ], dtype=float32),\n",
       " array([ 0.04941938,  0.01112883,  0.7170634 ,  0.3223204 , -0.03899391,\n",
       "         0.7942897 ,  0.47441176,  0.13831636,  0.78797454,  0.26306388,\n",
       "         1.1205022 ,  0.25949   , -0.2941064 ,  1.531421  ,  0.8150418 ,\n",
       "        -0.13392553,  1.1069472 ,  0.29006612,  1.0035148 ,  0.33857745,\n",
       "        -0.23925765,  0.82864887,  1.4353317 ,  0.59621984,  0.40019357,\n",
       "         0.32733226,  0.5201404 ,  0.55563927], dtype=float32),\n",
       " array([ 0.05057928,  0.0139386 ,  0.71886957,  0.31589636, -0.03697852,\n",
       "         0.7877859 ,  0.47072652,  0.11412735,  0.7923028 ,  0.25193954,\n",
       "         1.1366627 ,  0.25288978, -0.29971758,  1.5279223 ,  0.82032543,\n",
       "        -0.14786613,  1.1098323 ,  0.2754728 ,  1.0123854 ,  0.3315093 ,\n",
       "        -0.25435877,  0.83975536,  1.4509112 ,  0.5991958 ,  0.38251713,\n",
       "         0.3255408 ,  0.52065134,  0.5556989 ], dtype=float32),\n",
       " array([ 0.05176503,  0.01638802,  0.72142005,  0.30996412, -0.03524226,\n",
       "         0.7818959 ,  0.4667769 ,  0.08987805,  0.79689103,  0.24144943,\n",
       "         1.1529421 ,  0.24691543, -0.30534035,  1.5239154 ,  0.82561386,\n",
       "        -0.16092631,  1.1136386 ,  0.2617249 ,  1.0207719 ,  0.325258  ,\n",
       "        -0.26918232,  0.84996897,  1.4656128 ,  0.6033097 ,  0.36528993,\n",
       "         0.32343858,  0.52115536,  0.55572444], dtype=float32),\n",
       " array([ 0.05294331,  0.01843631,  0.7246604 ,  0.30445555, -0.03381757,\n",
       "         0.7765875 ,  0.46250284,  0.06550635,  0.8016984 ,  0.23154087,\n",
       "         1.1692402 ,  0.24148591, -0.31106332,  1.5194328 ,  0.8308916 ,\n",
       "        -0.17312196,  1.1182338 ,  0.24882618,  1.0286863 ,  0.31988773,\n",
       "        -0.28383398,  0.85929567,  1.4795346 ,  0.60851926,  0.3485357 ,\n",
       "         0.3209672 ,  0.52164865,  0.5557024 ], dtype=float32),\n",
       " array([ 0.05408607,  0.02004768,  0.7285358 ,  0.29930198, -0.03273542,\n",
       "         0.7718266 ,  0.45784405,  0.04093684,  0.806685  ,  0.222157  ,\n",
       "         1.1854657 ,  0.23653527, -0.31696975,  1.5145012 ,  0.83614844,\n",
       "        -0.18447891,  1.1234896 ,  0.23676981,  1.0361484 ,  0.31545585,\n",
       "        -0.29841673,  0.86775   ,  1.492777  ,  0.6147676 ,  0.33225965,\n",
       "         0.3180761 ,  0.52212936,  0.5556203 ], dtype=float32),\n",
       " array([ 0.05516975,  0.02119134,  0.7329922 ,  0.294437  , -0.03202583,\n",
       "         0.7675772 ,  0.4527412 ,  0.01608299,  0.81181246,  0.21323994,\n",
       "         1.201538  ,  0.23201223, -0.32313615,  1.5091453 ,  0.8413765 ,\n",
       "        -0.1950309 ,  1.1292828 ,  0.22554071,  1.0431873 ,  0.3120153 ,\n",
       "        -0.31302947,  0.87535465,  1.5054401 ,  0.6219869 ,  0.31645003,\n",
       "         0.3147228 ,  0.5225969 ,  0.5554662 ], dtype=float32),\n",
       " array([ 0.05617558,  0.02184201,  0.7379762 ,  0.28979826, -0.03171789,\n",
       "         0.76380587,  0.44713637, -0.00914916,  0.8170437 ,  0.20473343,\n",
       "         1.2173885 ,  0.22787876, -0.32963005,  1.5033885 ,  0.8465704 ,\n",
       "        -0.2048172 ,  1.1354988 ,  0.2151168 ,  1.0498377 ,  0.30961508,\n",
       "        -0.3277657 ,  0.88214123,  1.5176257 ,  0.6301013 ,  0.30108082,\n",
       "         0.31087378,  0.5230517 ,  0.55522996], dtype=float32),\n",
       " array([ 0.0570888 ,  0.02198038,  0.7434349 ,  0.28533   , -0.03183969,\n",
       "         0.76048195,  0.44097444, -0.03485866,  0.8223428 ,  0.19658571,\n",
       "         1.2329626 ,  0.22410867, -0.33650902,  1.4972558 ,  0.8517259 ,\n",
       "        -0.21388054,  1.1420296 ,  0.20547046,  1.0561416 ,  0.3083024 ,\n",
       "        -0.34271175,  0.8881498 ,  1.5294343 ,  0.63903075,  0.286114  ,\n",
       "         0.30650482,  0.52349496,  0.5549027 ], dtype=float32),\n",
       " array([ 0.05789797,  0.02159349,  0.7493165 ,  0.28098375, -0.03241879,\n",
       "         0.7575799 ,  0.4342033 , -0.06114557,  0.82767445,  0.18875097,\n",
       "         1.2482175 ,  0.2206863 , -0.34381908,  1.4907721 ,  0.85683906,\n",
       "        -0.22226468,  1.1487765 ,  0.19657025,  1.062144  ,  0.30812237,\n",
       "        -0.35794532,  0.89342886,  1.5409635 ,  0.6486939 ,  0.27150095,\n",
       "         0.3016009 ,  0.52392805,  0.55447733], dtype=float32),\n",
       " array([ 0.05859528,  0.02067535,  0.7555695 ,  0.27672005, -0.03348142,\n",
       "         0.75507975,  0.4267758 , -0.08810719,  0.8330031 ,  0.18119135,\n",
       "         1.2631245 ,  0.21760523, -0.3515928 ,  1.4839627 ,  0.8619052 ,\n",
       "        -0.23001242,  1.1556466 ,  0.18838169,  1.067893  ,  0.30911845,\n",
       "        -0.373533  ,  0.8980357 ,  1.5523077 ,  0.6590106 ,  0.25718477,\n",
       "         0.29615548,  0.52435297,  0.5539485 ], dtype=float32),\n",
       " array([ 0.05917536,  0.01922789,  0.7621417 ,  0.27250862, -0.03505263,\n",
       "         0.7529676 ,  0.41864976, -0.11583389,  0.83829284,  0.1738777 ,\n",
       "         1.2776654 ,  0.2148668 , -0.35984862,  1.4768523 ,  0.8669162 ,\n",
       "        -0.23716366,  1.1625534 ,  0.18086854,  1.0734371 ,  0.3113332 ,\n",
       "        -0.38952905,  0.902036  ,  1.5635525 ,  0.66990405,  0.24310178,\n",
       "         0.2901713 ,  0.5247717 ,  0.55331284], dtype=float32),\n",
       " array([ 0.05963517,  0.01726131,  0.76897943,  0.26832888, -0.03715573,\n",
       "         0.7512361 ,  0.40979046, -0.14440373,  0.84350604,  0.16679075,\n",
       "         1.2918338 ,  0.21247828, -0.36858875,  1.4694628 ,  0.87186044,\n",
       "        -0.24375293,  1.1694134 ,  0.17399421,  1.0788234 ,  0.31480616,\n",
       "        -0.40597248,  0.90550405,  1.5747738 ,  0.68130225,  0.22918355,\n",
       "         0.28365874,  0.5251859 ,  0.5525687 ], dtype=float32),\n",
       " array([ 0.05997366,  0.01479539,  0.7760269 ,  0.2641703 , -0.0398117 ,\n",
       "         0.74988437,  0.40017095, -0.17387739,  0.8486029 ,  0.15992197,\n",
       "         1.305633  ,  0.2104522 , -0.37779868,  1.4618121 ,  0.876721  ,\n",
       "        -0.24980748,  1.1761459 ,  0.16772208,  1.0840967 ,  0.31957376,\n",
       "        -0.4228856 ,  0.9085226 ,  1.586035  ,  0.6931381 ,  0.21535775,\n",
       "         0.27663672,  0.52559775,  0.55171597], dtype=float32),\n",
       " array([ 0.06019115,  0.01185995,  0.78322446,  0.26003218, -0.04303889,\n",
       "         0.74891746,  0.389774  , -0.20429257,  0.8535405 ,  0.15327308,\n",
       "         1.3190733 ,  0.20880462, -0.38744527,  1.4539126 ,  0.8814753 ,\n",
       "        -0.25534576,  1.1826684 ,  0.16201612,  1.0892957 ,  0.32566756,\n",
       "        -0.44027126,  0.9111829 ,  1.5973847 ,  0.7053491 ,  0.2015498 ,\n",
       "         0.26913127,  0.5260084 ,  0.55075705], dtype=float32),\n",
       " array([ 0.06028908,  0.00849572,  0.79050845,  0.25592333, -0.04685205,\n",
       "         0.7483467 ,  0.3785933 , -0.23565775,  0.8582719 ,  0.14685702,\n",
       "         1.3321723 ,  0.20755474, -0.39747673,  1.4457694 ,  0.886093  ,\n",
       "        -0.26037455,  1.1888971 ,  0.1568419 ,  1.0944533 ,  0.33311218,\n",
       "        -0.45811158,  0.9135848 ,  1.6088524 ,  0.7178772 ,  0.18768418,\n",
       "         0.26117635,  0.5264194 ,  0.5496958 ], dtype=float32),\n",
       " array([ 0.06026945,  0.00475442,  0.797809  ,  0.2518617 , -0.05126205,\n",
       "         0.74818814,  0.36663476, -0.2679467 ,  0.86274517,  0.14069685,\n",
       "         1.3449525 ,  0.20672448, -0.40782237,  1.4373802 ,  0.8905367 ,\n",
       "        -0.26488814,  1.1947441 ,  0.15216647,  1.099593  ,  0.3419236 ,\n",
       "        -0.47636634,  0.9158361 ,  1.620449  ,  0.7306666 ,  0.17368492,\n",
       "         0.2528131 ,  0.5268316 ,  0.5485381 ], dtype=float32),\n",
       " array([ 6.0134448e-02,  6.9888122e-04,  8.0505097e-01,  2.4787402e-01,\n",
       "        -5.6275140e-02,  7.4846208e-01,  3.5391718e-01, -3.0109340e-01,\n",
       "         8.6690503e-01,  1.3482589e-01,  1.3574402e+00,  2.0633672e-01,\n",
       "        -4.1839373e-01,  1.4287355e+00,  8.9476138e-01, -2.6886654e-01,\n",
       "         1.2001177e+00,  1.4795858e-01,  1.1047291e+00,  3.5210693e-01,\n",
       "        -4.9497300e-01,  9.1805053e-01,  1.6321644e+00,  7.4366248e-01,\n",
       "         1.5947697e-01,  2.4408951e-01,  5.2724552e-01,  5.4729187e-01],\n",
       "       dtype=float32),\n",
       " array([ 0.05988563, -0.00359829,  0.8121525 ,  0.24399428, -0.06189204,\n",
       "         0.7491915 ,  0.34047285, -0.334989  ,  0.87069243,  0.1292862 ,\n",
       "         1.3696645 ,  0.20641688, -0.42908603,  1.419818  ,  0.8987164 ,\n",
       "        -0.2722745 ,  1.2049228 ,  0.14418866,  1.1098647 ,  0.36365545,\n",
       "        -0.5138471 ,  0.92034566,  1.643969  ,  0.7568086 ,  0.14498712,\n",
       "         0.23505989,  0.52766126,  0.5459669 ], dtype=float32),\n",
       " array([ 0.05952384, -0.00805592,  0.8190269 ,  0.24026316, -0.06810796,\n",
       "         0.7504014 ,  0.32634705, -0.36947975,  0.87404484,  0.12412754,\n",
       "         1.381657  ,  0.20699157, -0.4397812 ,  1.4106059 ,  0.90234625,\n",
       "        -0.27506122,  1.209062  ,  0.14082882,  1.1149933 ,  0.376549  ,\n",
       "        -0.532885  ,  0.92284   ,  1.6558154 ,  0.77004504,  0.13014424,\n",
       "         0.22578421,  0.5280789 ,  0.5445754 ], dtype=float32),\n",
       " array([ 0.05904802, -0.01258729,  0.8255832 ,  0.23672596, -0.07491229,\n",
       "         0.75211644,  0.3115974 , -0.40436852,  0.8768998 ,  0.11940549,\n",
       "         1.39345   ,  0.20809053, -0.45035368,  1.401074  ,  0.90559363,\n",
       "        -0.27716073,  1.2124385 ,  0.13785231,  1.1200985 ,  0.3907553 ,\n",
       "        -0.55196756,  0.92564815,  1.667641  ,  0.7833072 ,  0.11488023,\n",
       "         0.21632695,  0.5284978 ,  0.5431307 ], dtype=float32),\n",
       " array([ 0.05845466, -0.01710361,  0.83172894,  0.23343156, -0.08228894,\n",
       "         0.75435936,  0.29629055, -0.43942004,  0.87919605,  0.11517961,\n",
       "         1.405077  ,  0.20974684, -0.4606753 ,  1.3911972 ,  0.9084029 ,\n",
       "        -0.27849305,  1.2149605 ,  0.13523376,  1.1251551 ,  0.4062307 ,\n",
       "        -0.5709657 ,  0.928876  ,  1.6793754 ,  0.79652464,  0.09913049,\n",
       "         0.20675561,  0.52891755,  0.5416484 ], dtype=float32),\n",
       " array([ 0.05773724, -0.02151854,  0.83737266,  0.23042987, -0.09021647,\n",
       "         0.75715065,  0.2805008 , -0.47436768,  0.88087684,  0.11151062,\n",
       "         1.4165717 ,  0.21199735, -0.4706229 ,  1.3809546 ,  0.9107228 ,\n",
       "        -0.27896678,  1.2165442 ,  0.1329488 ,  1.130133  ,  0.4229232 ,\n",
       "        -0.5897477 ,  0.9326141 ,  1.6909449 ,  0.80962104,  0.08283501,\n",
       "         0.19713882,  0.5293376 ,  0.54014534], dtype=float32),\n",
       " array([ 0.05688525, -0.02575248,  0.8424273 ,  0.22776994, -0.0986696 ,\n",
       "         0.7605058 ,  0.2643054 , -0.5089254 ,  0.88189137,  0.10845873,\n",
       "         1.427966  ,  0.2148845 , -0.48008502,  1.3703307 ,  0.91250926,\n",
       "        -0.27848226,  1.2171175 ,  0.13097434,  1.1349984 ,  0.4407758 ,\n",
       "        -0.6081861 ,  0.93693274,  1.7022815 ,  0.82251465,  0.06593899,\n",
       "         0.18754442,  0.52975696,  0.5386392 ], dtype=float32),\n",
       " array([ 0.0558839 , -0.02973763,  0.8468154 ,  0.22549795, -0.10762048,\n",
       "         0.764436  ,  0.24778166, -0.5427993 ,  0.8821975 ,  0.10608116,\n",
       "         1.4392899 ,  0.2184561 , -0.4889698 ,  1.3593198 ,  0.9137287 ,\n",
       "        -0.27693507,  1.216625  ,  0.12928866,  1.1397175 ,  0.45973104,\n",
       "        -0.62616515,  0.9418781 ,  1.7133269 ,  0.8351206 ,  0.048394  ,\n",
       "         0.17803717,  0.53017515,  0.5371486 ], dtype=float32),\n",
       " array([ 0.0547141 , -0.03342158,  0.8504702 ,  0.22365518, -0.11703998,\n",
       "         0.7689471 ,  0.23100299, -0.57570237,  0.8817622 ,  0.10443067,\n",
       "         1.4505675 ,  0.22276647, -0.4972098 ,  1.3479263 ,  0.9143577 ,\n",
       "        -0.27422163,  1.2150275 ,  0.12787202,  1.144258  ,  0.4797343 ,\n",
       "        -0.64358646,  0.9474676 ,  1.7240422 ,  0.84735274,  0.0301601 ,\n",
       "         0.16867664,  0.5305907 ,  0.53569263], dtype=float32),\n",
       " array([ 0.05335268, -0.03677032,  0.85334074,  0.22227634, -0.12689947,\n",
       "         0.77404046,  0.21403643, -0.6073666 ,  0.8805615 ,  0.10355425,\n",
       "         1.4618162 ,  0.22787477, -0.5047666 ,  1.3361651 ,  0.91438425,\n",
       "        -0.27024296,  1.2123038 ,  0.12670723,  1.1485914 ,  0.5007371 ,\n",
       "        -0.6603742 ,  0.9536921 ,  1.7344071 ,  0.8591273 ,  0.01120782,\n",
       "         0.159515  ,  0.53100264,  0.5342901 ], dtype=float32),\n",
       " array([ 0.05177294, -0.03976996,  0.8553928 ,  0.22138788, -0.1371724 ,\n",
       "         0.77971303,  0.19694048, -0.63755304,  0.8785795 ,  0.10349241,\n",
       "         1.4730438 ,  0.23384438, -0.5116321 ,  1.3240604 ,  0.9138048 ,\n",
       "        -0.2649098 ,  1.2084516 ,  0.12578028,  1.1526937 ,  0.5226986 ,\n",
       "        -0.676476  ,  0.96051544,  1.7444245 ,  0.87036574, -0.0084798 ,\n",
       "         0.15059574,  0.5314084 ,  0.53296006], dtype=float32),\n",
       " array([ 0.04994628, -0.04242685,  0.8566108 ,  0.22100772, -0.14783509,\n",
       "         0.7859595 ,  0.17976448, -0.66606134,  0.8758065 ,  0.10427916,\n",
       "         1.4842447 ,  0.24074183, -0.5178272 ,  1.3116438 ,  0.91262347,\n",
       "        -0.2581468 ,  1.2034829 ,  0.12508143,  1.156545  ,  0.54558486,\n",
       "        -0.6918634 ,  0.96788096,  1.7541177 ,  0.88099796, -0.02890539,\n",
       "         0.14195281,  0.5318052 ,  0.53172123], dtype=float32),\n",
       " array([ 0.04784371, -0.04476625,  0.8569968 ,  0.22114395, -0.15886773,\n",
       "         0.79277086,  0.16254857, -0.69273293,  0.87223625,  0.10594212,\n",
       "         1.4953986 ,  0.24863283, -0.5233997 ,  1.2989516 ,  0.9108479 ,\n",
       "        -0.24989608,  1.1974238 ,  0.12460535,  1.160129  ,  0.56936896,\n",
       "        -0.70652926,  0.9757155 ,  1.7635264 ,  0.8909664 , -0.05005416,\n",
       "         0.13360909,  0.5321892 ,  0.53059214], dtype=float32),\n",
       " array([ 0.0454376 , -0.04683085,  0.85656965,  0.2217953 , -0.17025392,\n",
       "         0.80013573,  0.14532459, -0.71745396,  0.8678647 ,  0.1085024 ,\n",
       "         1.5064694 ,  0.25758153, -0.52841854,  1.2860218 ,  0.9084887 ,\n",
       "        -0.24011979,  1.1903123 ,  0.1243517 ,  1.1634314 ,  0.59402525,\n",
       "        -0.72048306,  0.98393744,  1.772703  ,  0.9002285 , -0.0718937 ,\n",
       "         0.12557706,  0.53255486,  0.5295911 ], dtype=float32),\n",
       " array([ 0.04270399, -0.04867715,  0.8553635 ,  0.22295104, -0.18198082,\n",
       "         0.8080405 ,  0.1281172 , -0.7401528 ,  0.8626871 ,  0.11197518,\n",
       "         1.517403  ,  0.26764593, -0.5329686 ,  1.272892  ,  0.90555537,\n",
       "        -0.22880134,  1.1821942 ,  0.12432497,  1.1664394 ,  0.619528  ,\n",
       "        -0.7337475 ,  0.99246156,  1.7817049 ,  0.9087585 , -0.09437212,\n",
       "         0.11785862,  0.53289634,  0.52873546], dtype=float32),\n",
       " array([ 0.03962407, -0.0503726 ,  0.85342413,  0.22459179, -0.19403854,\n",
       "         0.8164679 ,  0.11094546, -0.76079786,  0.85669744,  0.11636942,\n",
       "         1.5281283 ,  0.2788762 , -0.5371444 ,  1.259597  ,  0.9020571 ,\n",
       "        -0.21594617,  1.1731243 ,  0.12453441,  1.16914   ,  0.6458457 ,\n",
       "        -0.746352  ,  1.0012054 ,  1.7905893 ,  0.9165501 , -0.11741897,\n",
       "         0.11044541,  0.53320724,  0.5280417 ], dtype=float32),\n",
       " array([ 0.03618594, -0.05199211,  0.8508071 ,  0.22668971, -0.20641862,\n",
       "         0.8253975 ,  0.09382475, -0.77939326,  0.8498875 ,  0.12168749,\n",
       "         1.538558  ,  0.2913114 , -0.5410432 ,  1.246166  ,  0.89800125,\n",
       "        -0.20158151,  1.1631622 ,  0.12499312,  1.1715194 ,  0.6729379 ,\n",
       "        -0.7583289 ,  1.0100924 ,  1.7994068 ,  0.92361647, -0.14094412,\n",
       "         0.10332008,  0.5334805 ,  0.52752507], dtype=float32),\n",
       " array([ 0.03238568, -0.05361454,  0.8475737 ,  0.22921114, -0.21911381,\n",
       "         0.8348036 ,  0.07676737, -0.7959737 ,  0.8422475 ,  0.12792529,\n",
       "         1.5485896 ,  0.3049772 , -0.5447601 ,  1.2326224 ,  0.89339536,\n",
       "        -0.18575427,  1.1523738 ,  0.12571749,  1.1735625 ,  0.7007522 ,\n",
       "        -0.76970834,  1.019054  ,  1.808197  ,  0.92999107, -0.16484067,\n",
       "         0.09645712,  0.5337101 ,  0.52719814], dtype=float32),\n",
       " array([ 0.02822765, -0.05531909,  0.8437883 ,  0.23211576, -0.23211713,\n",
       "         0.84465694,  0.0597834 , -0.81060106,  0.8337659 ,  0.13507128,\n",
       "         1.5581079 ,  0.31988472, -0.54838383,  1.218983  ,  0.8882467 ,\n",
       "        -0.16852912,  1.1408278 ,  0.126725  ,  1.1752533 ,  0.72922194,\n",
       "        -0.7805168 ,  1.02803   ,  1.8169847 ,  0.9357257 , -0.18898696,\n",
       "         0.08982361,  0.5338911 ,  0.52707046], dtype=float32),\n",
       " array([ 0.02372495, -0.05718275,  0.83951586,  0.23535986, -0.245421  ,\n",
       "         0.85492265,  0.04288194, -0.82335865,  0.82443166,  0.14310631,\n",
       "         1.5669898 ,  0.33602843, -0.5519922 ,  1.205257  ,  0.8825658 ,\n",
       "        -0.14998594,  1.1285977 ,  0.12803346,  1.1765759 ,  0.75826657,\n",
       "        -0.79077303,  1.0369662 ,  1.825777  ,  0.9408905 , -0.21324918,\n",
       "         0.08338119,  0.53402025,  0.5271477 ], dtype=float32),\n",
       " array([ 0.01889866, -0.05927712,  0.8348193 ,  0.23889703, -0.2590166 ,\n",
       "         0.86556137,  0.02607131, -0.834349  ,  0.8142355 ,  0.15200368,\n",
       "         1.5751077 ,  0.35338736, -0.5556501 ,  1.1914481 ,  0.87636757,\n",
       "        -0.1302158 ,  1.1157588 ,  0.12965867,  1.1775159 ,  0.78779167,\n",
       "        -0.8004881 ,  1.0458134 ,  1.8345634 ,  0.9455719 , -0.23748472,\n",
       "         0.07708742,  0.5340969 ,  0.5274307 ], dtype=float32),\n",
       " array([ 0.01377721, -0.06166635,  0.829759  ,  0.2426802 , -0.27289355,\n",
       "         0.8765304 ,  0.00935933, -0.84369   ,  0.8031713 ,  0.16172826,\n",
       "         1.5823324 ,  0.37192422, -0.5594082 ,  1.1775519 ,  0.86967355,\n",
       "        -0.10931774,  1.1023877 ,  0.1316131 ,  1.1780616 ,  0.8176924 ,\n",
       "        -0.8096642 ,  1.0545237 ,  1.843316  ,  0.94986945, -0.2615473 ,\n",
       "         0.07089702,  0.5341234 ,  0.52791536], dtype=float32),\n",
       " array([ 0.00839516, -0.06440534,  0.8243896 ,  0.24666318, -0.28703964,\n",
       "         0.8877845 , -0.00724618, -0.8515125 ,  0.79123753,  0.17223838,\n",
       "         1.5885406 ,  0.39158756, -0.5633018 ,  1.163559  ,  0.8625121 ,\n",
       "        -0.08739436,  1.0885606 ,  0.13390368,  1.178206  ,  0.8478557 ,\n",
       "        -0.8182954 ,  1.0630456 ,  1.851992  ,  0.9538932 , -0.28529066,\n",
       "         0.06476395,  0.53410494,  0.5285919 ], dtype=float32),\n",
       " array([ 0.00279184, -0.06753787,  0.8187626 ,  0.2508024 , -0.3014402 ,\n",
       "         0.89927673, -0.02373668, -0.8579583 ,  0.7784385 ,  0.18348573,\n",
       "         1.5936178 ,  0.41231415, -0.56735224,  1.1494546 ,  0.8549212 ,\n",
       "        -0.0645488 ,  1.0743533 ,  0.13652998,  1.1779464 ,  0.87816507,\n",
       "        -0.82636863,  1.0713228 ,  1.8605344 ,  0.957761  , -0.30857375,\n",
       "         0.05864291,  0.5340499 ,  0.5294448 ], dtype=float32),\n",
       " array([-0.00298974, -0.07109631,  0.81292397,  0.25505883, -0.3160775 ,\n",
       "         0.9109608 , -0.04010241, -0.863177  ,  0.7647855 ,  0.19541717,\n",
       "         1.5974633 ,  0.4340315 , -0.5715672 ,  1.1352208 ,  0.84694743,\n",
       "        -0.04088008,  1.0598384 ,  0.13948378,  1.1772889 ,  0.90850496,\n",
       "        -0.83386415,  1.0792897 ,  1.8688779 ,  0.9615962 , -0.33126402,\n",
       "         0.05249107,  0.5339691 ,  0.5304543 ], dtype=float32),\n",
       " array([-0.00890471, -0.075101  ,  0.80691725,  0.25939918, -0.33093056,\n",
       "         0.9227928 , -0.05633108, -0.8673242 ,  0.75029784,  0.20797692,\n",
       "         1.5999929 ,  0.4566617 , -0.5759419 ,  1.1208357 ,  0.83864677,\n",
       "        -0.01648005,  1.0450848 ,  0.14274813,  1.1762458 ,  0.938765  ,\n",
       "        -0.84075695,  1.0868696 ,  1.8769509 ,  0.9655234 , -0.35324237,\n",
       "         0.0462704 ,  0.5338763 ,  0.5315956 ], dtype=float32),\n",
       " array([-0.01490706, -0.07956073,  0.80078346,  0.26379704, -0.3459741 ,\n",
       "         0.9347327 , -0.07240754, -0.8705594 ,  0.7350024 ,  0.22110838,\n",
       "         1.6011431 ,  0.4801239 , -0.5804617 ,  1.1062747 ,  0.83008444,\n",
       "         0.00856976,  1.0301561 ,  0.14629667,  1.174837  ,  0.9688445 ,\n",
       "        -0.84701824,  1.0939735 ,  1.8846788 ,  0.9696669 , -0.37440506,\n",
       "         0.03994823,  0.5337864 ,  0.5328409 ], dtype=float32),\n",
       " array([-0.02095144, -0.0844731 ,  0.79456383,  0.2682342 , -0.36117843,\n",
       "         0.94674516, -0.0883123 , -0.87304246,  0.71893364,  0.23475745,\n",
       "         1.6008719 ,  0.5043394 , -0.5851034 ,  1.0915098 ,  0.8213324 ,\n",
       "         0.03420082,  1.0151104 ,  0.15009516,  1.1730907 ,  0.99865615,\n",
       "        -0.8526165 ,  1.1004989 ,  1.8919891 ,  0.9741477 , -0.39466643,\n",
       "         0.03349948,  0.53371614,  0.53415984], dtype=float32),\n",
       " array([-0.0269931 , -0.08982588,  0.7883022 ,  0.2727015 , -0.37650818,\n",
       "         0.9588015 , -0.10402036, -0.8749313 ,  0.70213413,  0.24887492,\n",
       "         1.5991588 ,  0.5292346 , -0.58983684,  1.076511  ,  0.81247026,\n",
       "         0.06036034,  0.9999996 ,  0.15410194,  1.1710413 ,  1.0281291 ,\n",
       "        -0.8575174 ,  1.106331  ,  1.8988131 ,  0.97908163, -0.41395923,\n",
       "         0.02690787,  0.5336821 ,  0.5355214 ], dtype=float32),\n",
       " array([-0.03298941, -0.09559809,  0.782045  ,  0.27719885, -0.39192238,\n",
       "         0.970881  , -0.1195012 , -0.87637967,  0.68465245,  0.26341924,\n",
       "         1.5960037 ,  0.55474395, -0.5946254 ,  1.0612434 ,  0.80358267,\n",
       "         0.0870129 ,  0.98486805,  0.15826964,  1.1687291 ,  1.0572107 ,\n",
       "        -0.8616853 ,  1.1113431 ,  1.905088  ,  0.9845785 , -0.43223473,\n",
       "         0.02016697,  0.5337019 ,  0.5368937 ], dtype=float32),\n",
       " array([-0.03889973, -0.10176164,  0.7758448 ,  0.28173608, -0.4073732 ,\n",
       "         0.98297   , -0.13471772, -0.8775327 ,  0.6665436 ,  0.27835932,\n",
       "         1.591427  ,  0.58081293, -0.59942824,  1.0456659 ,  0.7947587 ,\n",
       "         0.11414242,  0.9697539 ,  0.16254817,  1.1661997 ,  1.0858682 ,\n",
       "        -0.8650826 ,  1.1153994 ,  1.9107579 ,  0.9907406 , -0.4494618 ,\n",
       "         0.01328076,  0.53379214,  0.5382465 ], dtype=float32),\n",
       " array([-0.04468545, -0.10828312,  0.76976156,  0.2863326 , -0.42280594,\n",
       "         0.995064  , -0.1496261 , -0.8785261 ,  0.6478678 ,  0.2936768 ,\n",
       "         1.5854645 ,  0.60739994, -0.6041991 ,  1.0297297 ,  0.7860898 ,\n",
       "         0.14175291,  0.954688  ,  0.16688618,  1.1635019 ,  1.1140898 ,\n",
       "        -0.86767006,  1.1183558 ,  1.9157753 ,  0.9976614 , -0.4656236 ,\n",
       "         0.00626395,  0.5339689 ,  0.53955114], dtype=float32),\n",
       " array([-5.0310098e-02, -1.1512518e-01,  7.6386267e-01,  2.9101709e-01,\n",
       "        -4.3815872e-01,  1.0071659e+00, -1.6417669e-01, -8.7948239e-01,\n",
       "         6.2868875e-01,  3.0936775e-01,  1.5781651e+00,  6.3447696e-01,\n",
       "        -6.0888779e-01,  1.0133752e+00,  7.7766925e-01,  1.6986978e-01,\n",
       "         9.3969399e-01,  1.7123426e-01,  1.1606872e+00,  1.1418829e+00,\n",
       "        -8.6940706e-01,  1.1200640e+00,  1.9200999e+00,  1.0054275e+00,\n",
       "        -4.8071644e-01, -8.5797161e-04,  5.3424716e-01,  5.4078120e-01],\n",
       "       dtype=float32),\n",
       " array([-0.05573857, -0.12224807,  0.7582237 ,  0.2958271 , -0.4533625 ,\n",
       "         1.019287  , -0.17831382, -0.8805092 ,  0.60907334,  0.32544363,\n",
       "         1.569588  ,  0.66203016, -0.6134386 ,  0.9965291 ,  0.7695906 ,\n",
       "         0.19853991,  0.92478824,  0.17554784,  1.1578083 ,  1.1692748 ,\n",
       "        -0.87025064,  1.1203747 ,  1.9236977 ,  1.0141158 , -0.4947457 ,\n",
       "        -0.00804877,  0.53464085,  0.541913  ], dtype=float32),\n",
       " array([-0.06093682, -0.12961143,  0.7529295 ,  0.3008082 , -0.46834105,\n",
       "         1.0314444 , -0.19197735, -0.88169706,  0.5890902 ,  0.34193224,\n",
       "         1.559797  ,  0.6900585 , -0.6177913 ,  0.97910374,  0.76194674,\n",
       "         0.22783104,  0.9099807 ,  0.17978907,  1.1549174 ,  1.1963094 ,\n",
       "        -0.87015635,  1.119139  ,  1.9265406 ,  1.0237964 , -0.5077222 ,\n",
       "        -0.01526295,  0.5351623 ,  0.5429261 ], dtype=float32),\n",
       " array([-0.0658708 , -0.13717572,  0.748072  ,  0.30601287, -0.4830119 ,\n",
       "         1.0436624 , -0.20510307, -0.8831167 ,  0.5688092 ,  0.3588766 ,\n",
       "         1.5488605 ,  0.71857154, -0.6218806 ,  0.96099424,  0.7548281 ,\n",
       "         0.2578308 ,  0.8952744 ,  0.1839298 ,  1.1520667 ,  1.2230471 ,\n",
       "        -0.86907786,  1.116214  ,  1.9286048 ,  1.0345304 , -0.51965934,\n",
       "        -0.02244604,  0.5358221 ,  0.5438034 ], dtype=float32),\n",
       " array([-0.07050627, -0.1449033 ,  0.74375015,  0.31149906, -0.49728668,\n",
       "         1.0559694 , -0.21762457, -0.8848183 ,  0.5482996 ,  0.37633374,\n",
       "         1.5368471 ,  0.74758494, -0.62563765,  0.94207966,  0.7483213 ,\n",
       "         0.2886443 ,  0.8806661 ,  0.18795376,  1.1493049 ,  1.2495584 ,\n",
       "        -0.86696947,  1.1114659 ,  1.9298702 ,  1.0463701 , -0.5305703 ,\n",
       "        -0.02953618,  0.536628  ,  0.5445311 ], dtype=float32),\n",
       " array([-0.0748072 , -0.15275969,  0.7400658 ,  0.31732804, -0.5110733 ,\n",
       "         1.0683954 , -0.22947443, -0.8868275 ,  0.5276309 ,  0.3943718 ,\n",
       "         1.5238254 ,  0.7771143 , -0.62899184,  0.922226  ,  0.74250776,\n",
       "         0.3203903 ,  0.86614674,  0.1918582 ,  1.1466775 ,  1.2759247 ,\n",
       "        -0.8637874 ,  1.1047754 ,  1.9303207 ,  1.0593582 , -0.54046345,\n",
       "        -0.03646507,  0.5375845 ,  0.5450989 ], dtype=float32),\n",
       " array([-0.07873625, -0.16071504,  0.73712265,  0.3235613 , -0.52427787,\n",
       "         1.0809718 , -0.24058619, -0.8891435 ,  0.50687194,  0.4130654 ,\n",
       "         1.5098658 ,  0.80716753, -0.63187397,  0.9012911 ,  0.737461  ,\n",
       "         0.35319492,  0.85170263,  0.1956559 ,  1.1442236 ,  1.3022295 ,\n",
       "        -0.8594937 ,  1.096045  ,  1.9299461 ,  1.073525  , -0.54934275,\n",
       "        -0.04315915,  0.53869104,  0.5455004 ], dtype=float32),\n",
       " array([-0.0822539 , -0.1687448 ,  0.7350215 ,  0.3302582 , -0.53680706,\n",
       "         1.0937263 , -0.2508966 , -0.89173555,  0.48609146,  0.4324904 ,\n",
       "         1.4950439 ,  0.83773404, -0.6342213 ,  0.87913316,  0.73324335,\n",
       "         0.38718346,  0.8373157 ,  0.19937629,  1.1419762 ,  1.3285546 ,\n",
       "        -0.8540619 ,  1.0852067 ,  1.9287447 ,  1.0888834 , -0.5572061 ,\n",
       "        -0.04954036,  0.5399409 ,  0.5457338 ], dtype=float32),\n",
       " array([-0.08531981, -0.17683053,  0.7338536 ,  0.33747   , -0.5485723 ,\n",
       "         1.1066804 , -0.26034856, -0.89453846,  0.46535876,  0.45271528,\n",
       "         1.4794441 ,  0.86877245, -0.63598454,  0.855625  ,  0.72990286,\n",
       "         0.422469  ,  0.8229654 ,  0.20306675,  1.1399581 ,  1.3549736 ,\n",
       "        -0.8474843 ,  1.0722351 ,  1.9267277 ,  1.1054255 , -0.5640452 ,\n",
       "        -0.05552798,  0.5413187 ,  0.54580224], dtype=float32),\n",
       " array([-0.08789384, -0.1849603 ,  0.73369604,  0.34523606, -0.5594938 ,\n",
       "         1.1198446 , -0.2688941 , -0.89744884,  0.44474536,  0.4737923 ,\n",
       "         1.4631677 ,  0.9001966 , -0.63713557,  0.8306686 ,  0.7274689 ,\n",
       "         0.45913878,  0.8086311 ,  0.2067927 ,  1.1381785 ,  1.381543  ,\n",
       "        -0.83977985,  1.057158  ,  1.9239258 ,  1.1231141 , -0.5698478 ,\n",
       "        -0.06103993,  0.54279846,  0.545715  ], dtype=float32),\n",
       " array([-0.08993896, -0.19312853,  0.7346039 ,  0.35357773, -0.569505  ,\n",
       "         1.1332144 , -0.27649868, -0.9003211 ,  0.42432636,  0.4957471 ,\n",
       "         1.4463388 ,  0.9318629 , -0.6376761 ,  0.8042163 ,  0.7259462 ,\n",
       "         0.49723777,  0.79429334,  0.21063703,  1.1366328 ,  1.4082962 ,\n",
       "        -0.8310042 ,  1.0400723 ,  1.9203956 ,  1.1418743 , -0.5746015 ,\n",
       "        -0.06599442,  0.5443425 ,  0.54548824], dtype=float32),\n",
       " array([-0.09142476, -0.20133488,  0.73660254,  0.36249226, -0.57855785,\n",
       "         1.1467655 , -0.28314534, -0.9029677 ,  0.40418226,  0.51856655,\n",
       "         1.4291103 ,  0.9635595 , -0.637645  ,  0.7762903 ,  0.7253115 ,\n",
       "         0.536752  ,  0.7799369 ,  0.21469723,  1.135298  ,  1.4352345 ,\n",
       "        -0.8212579 ,  1.0211557 ,  1.9162254 ,  1.1615871 , -0.5782976 ,\n",
       "        -0.07031291,  0.54590005,  0.54514533], dtype=float32),\n",
       " array([-0.09233234, -0.2095826 ,  0.73968333,  0.37194872, -0.586626  ,\n",
       "         1.1604503 , -0.28883865, -0.9051606 ,  0.38439974,  0.5421909 ,\n",
       "         1.4116666 ,  0.9950014 , -0.63712347,  0.7470011 ,  0.72551   ,\n",
       "         0.5775932 ,  0.7655536 ,  0.21907961,  1.1341302 ,  1.462322  ,\n",
       "        -0.81069034,  1.0006753 ,  1.9115386 ,  1.1820817 , -0.58093756,\n",
       "        -0.07392483,  0.5474089 ,  0.54471666], dtype=float32),\n",
       " array([-0.09265937, -0.21787502,  0.74379754,  0.38188565, -0.5937072 ,\n",
       "         1.1741965 , -0.29360735, -0.90664244,  0.36507252,  0.56650615,\n",
       "         1.3942181 ,  1.0258372 , -0.63623434,  0.71656024,  0.7264557 ,\n",
       "         0.6195887 ,  0.75114393,  0.22388992,  1.1330664 ,  1.489483  ,\n",
       "        -0.799501  ,  0.978986  ,  1.9064922 ,  1.203135  , -0.582539  ,\n",
       "        -0.07677316,  0.5487986 ,  0.5442383 ], dtype=float32),\n",
       " array([-0.0924254 , -0.22621171,  0.7488578 ,  0.39221227, -0.5998224 ,\n",
       "         1.1879077 , -0.2975053 , -0.9071424 ,  0.34629908,  0.59134376,\n",
       "         1.3769915 ,  1.0556659 , -0.6351344 ,  0.6852822 ,  0.7280348 ,\n",
       "         0.6624785 ,  0.7367195 ,  0.22922221,  1.132023  ,  1.5166044 ,\n",
       "        -0.7879301 ,  0.95651805,  1.9012698 ,  1.2244734 , -0.5831404 ,\n",
       "        -0.07882031,  0.54999673,  0.5437494 ], dtype=float32),\n",
       " array([-0.0916739 , -0.23458451,  0.75474113,  0.40281248, -0.6050141 ,\n",
       "         1.201467  , -0.30060878, -0.9064004 ,  0.32818028,  0.6164857 ,\n",
       "         1.360211  ,  1.0840663 , -0.6340001 ,  0.65357256,  0.730113  ,\n",
       "         0.7059227 ,  0.72230405,  0.23514484,  1.1309016 ,  1.5435402 ,\n",
       "        -0.7762427 ,  0.9337503 ,  1.8960658 ,  1.245785  , -0.5828052 ,\n",
       "        -0.08005581,  0.55093676,  0.54328793], dtype=float32),\n",
       " array([-0.09047322, -0.2429722 ,  0.7612963 ,  0.41355437, -0.60933965,\n",
       "         1.2147453 , -0.30301228, -0.9041937 ,  0.3108148 ,  0.6416762 ,\n",
       "         1.3440764 ,  1.1106344 , -0.63300824,  0.6219033 ,  0.7325472 ,\n",
       "         0.749519  ,  0.70793426,  0.24168815,  1.1295967 ,  1.5701216 ,\n",
       "        -0.7647076 ,  0.91117424,  1.8910693 ,  1.2667351 , -0.58162254,\n",
       "        -0.08050068,  0.5515661 ,  0.54288703], dtype=float32),\n",
       " array([-0.08891243, -0.25133887,  0.7683535 ,  0.424299  , -0.6128655 ,\n",
       "         1.2276083 , -0.30482137, -0.90036404,  0.29429358,  0.66663814,\n",
       "         1.3287429 ,  1.1350198 , -0.63231415,  0.5907773 ,  0.73519605,\n",
       "         0.79282737,  0.6936596 ,  0.24883652,  1.128005  ,  1.596169  ,\n",
       "        -0.7535737 ,  0.88925236,  1.8864427 ,  1.2869854 , -0.5797041 ,\n",
       "        -0.08021017,  0.5518528 ,  0.5425709 ], dtype=float32),\n",
       " array([-0.08709472, -0.25963306,  0.77573526,  0.434911  , -0.61566144,\n",
       "         1.2399261 , -0.30614468, -0.89483553,  0.27869478,  0.6910919 ,\n",
       "         1.3143075 ,  1.1569571 , -0.63203424,  0.56068766,  0.73793054,\n",
       "         0.8353989 ,  0.6795423 ,  0.256527  ,  1.1260347 ,  1.6215025 ,\n",
       "        -0.74304944,  0.86838025,  1.8823059 ,  1.3062159 , -0.5771782 ,\n",
       "        -0.07927042,  0.55178916,  0.542353  ], dtype=float32),\n",
       " array([-0.08512744, -0.26779   ,  0.7832643 ,  0.44526672, -0.6177961 ,\n",
       "         1.251584  , -0.30708656, -0.88762313,  0.26408023,  0.71477306,\n",
       "         1.3008044 ,  1.1762832 , -0.63223565,  0.5320805 ,  0.74064124,\n",
       "         0.87680304,  0.6656534 ,  0.26465523,  1.1236162 ,  1.6459521 ,\n",
       "        -0.7332899 ,  0.8488603 ,  1.8787255 ,  1.3241428 , -0.5741812 ,\n",
       "        -0.07779401,  0.55138946,  0.54223585], dtype=float32),\n",
       " array([-0.0831142 , -0.27573633,  0.7907731 ,  0.45525956, -0.6193354 ,\n",
       "         1.2624842 , -0.3077417 , -0.87883055,  0.25049138,  0.737447  ,\n",
       "         1.2882122 ,  1.1929412 , -0.63293356,  0.5053248 ,  0.74324054,\n",
       "         0.9166494 ,  0.65207183,  0.27308697,  1.1207086 ,  1.669365  ,\n",
       "        -0.72439235,  0.8308899 ,  1.8757167 ,  1.3405333 , -0.5708496 ,\n",
       "        -0.07590977,  0.55068785,  0.54221225], dtype=float32),\n",
       " array([-0.08114634, -0.28339654,  0.79810894,  0.46480286, -0.6203427 ,\n",
       "         1.2725549 , -0.30819306, -0.86863476,  0.23794822,  0.7589183 ,\n",
       "         1.276464  ,  1.2069712 , -0.63409764,  0.4806954 ,  0.7456622 ,\n",
       "         0.954605  ,  0.63887924,  0.281673  ,  1.1173017 ,  1.6916137 ,\n",
       "        -0.7163997 ,  0.8145632 ,  1.8732471 ,  1.3552139 , -0.5673106 ,\n",
       "        -0.0737532 ,  0.5497307 ,  0.5422681 ], dtype=float32),\n",
       " array([-0.07929802, -0.2906993 ,  0.8051388 ,  0.47383004, -0.62087893,\n",
       "         1.2817485 , -0.30850917, -0.8572672 ,  0.22645018,  0.779037  ,\n",
       "         1.2654685 ,  1.2184925 , -0.63566256,  0.45836702,  0.74785954,\n",
       "         0.9904026 ,  0.6261549 ,  0.29026282,  1.1134185 ,  1.7125969 ,\n",
       "        -0.70930886,  0.79988545,  1.8712492 ,  1.3680753 , -0.56367666,\n",
       "        -0.07145611,  0.5485706 ,  0.54238594], dtype=float32),\n",
       " array([-0.07762451, -0.29758298,  0.81175125,  0.48229384, -0.621005  ,\n",
       "         1.2900437 , -0.3087466 , -0.8449901 ,  0.21597636,  0.7977003 ,\n",
       "         1.2551221 ,  1.2276822 , -0.6375433 ,  0.43841955,  0.7498014 ,\n",
       "         1.0238439 ,  0.6139708 ,  0.29871678,  1.1091073 ,  1.7322437 ,\n",
       "        -0.7030824 ,  0.78679156,  1.8696345 ,  1.3790704 , -0.5600414 ,\n",
       "        -0.06913962,  0.5472605 ,  0.54254675], dtype=float32),\n",
       " array([-0.07616214, -0.3039996 ,  0.8178598 ,  0.49016422, -0.62078226,\n",
       "         1.2974433 , -0.30894995, -0.8320762 ,  0.20648944,  0.8148508 ,\n",
       "         1.2453262 ,  1.234755  , -0.63964576,  0.42085096,  0.7514702 ,\n",
       "         1.0547993 ,  0.60238796,  0.30691394,  1.1044378 ,  1.750514  ,\n",
       "        -0.6976598 ,  0.7751695 ,  1.8683066 ,  1.3882117 , -0.55647916,\n",
       "        -0.06690738,  0.5458496 ,  0.542733  ], dtype=float32),\n",
       " array([-0.07493112, -0.30991703,  0.82340133,  0.4974267 , -0.6202724 ,\n",
       "         1.3039705 , -0.3091542 , -0.81878966,  0.19793834,  0.8304715 ,\n",
       "         1.2359954 ,  1.239944  , -0.6418802 ,  0.4055929 ,  0.7528588 ,\n",
       "         1.0832033 ,  0.59145266,  0.31475624,  1.0994945 ,  1.7673951 ,\n",
       "        -0.6929677 ,  0.7648785 ,  1.8671721 ,  1.3955586 , -0.5530458 ,\n",
       "        -0.06484216,  0.54438144,  0.54292953], dtype=float32),\n",
       " array([-0.07393873, -0.31531882,  0.8283368 ,  0.5040798 , -0.6195372 ,\n",
       "         1.3096656 , -0.30938616, -0.8053736 ,  0.19026211,  0.84458184,\n",
       "         1.2270606 ,  1.2434874 , -0.6441665 ,  0.3925286 ,  0.7539694 ,\n",
       "         1.109048  ,  0.5811957 ,  0.32216927,  1.0943677 ,  1.7829018 ,\n",
       "        -0.6889268 ,  0.7557665 ,  1.8661491 ,  1.4012126 , -0.54978085,\n",
       "        -0.06300593,  0.54289186,  0.5431241 ], dtype=float32),\n",
       " array([-0.073182  , -0.3202033 ,  0.8326487 ,  0.51013374, -0.6186362 ,\n",
       "         1.314583  , -0.309666  , -0.7920405 ,  0.18339401,  0.8572302 ,\n",
       "         1.2184725 ,  1.2456172 , -0.64643925,  0.38150924,  0.75481206,\n",
       "         1.1323746 ,  0.5716322 ,  0.32910207,  1.0891484 ,  1.797071  ,\n",
       "        -0.6854585 ,  0.7476822 ,  1.8651701 ,  1.4053043 , -0.5467107 ,\n",
       "        -0.0614384 ,  0.5414104 ,  0.54330766], dtype=float32),\n",
       " array([-0.07265133, -0.3245816 ,  0.8363383 ,  0.51560736, -0.6176262 ,\n",
       "         1.3187855 , -0.31000784, -0.7789693 ,  0.17726426,  0.8684865 ,\n",
       "         1.2102001 ,  1.2465532 , -0.64864874,  0.3723672 ,  0.7554032 ,\n",
       "         1.1532658 ,  0.56276256,  0.33552384,  1.0839225 ,  1.8099568 ,\n",
       "        -0.682488  ,  0.7404833 ,  1.8641841 ,  1.407982  , -0.54385155,\n",
       "        -0.06016081,  0.5399593 ,  0.5434743 ], dtype=float32),\n",
       " array([-0.07233276, -0.3284745 ,  0.83942235,  0.5205271 , -0.61655873,\n",
       "         1.3223405 , -0.31042126, -0.7663024 ,  0.17180336,  0.8784381 ,\n",
       "         1.202225  ,  1.2464979 , -0.65075994,  0.3649272 ,  0.7557642 ,\n",
       "         1.1718365 ,  0.5545746 ,  0.34142116,  1.0787679 ,  1.8216269 ,\n",
       "        -0.6799458 ,  0.7340407 ,  1.8631574 ,  1.4094051 , -0.5412124 ,\n",
       "        -0.05917784,  0.5385552 ,  0.54362047], dtype=float32),\n",
       " array([-0.07220927, -0.33191028,  0.8419303 ,  0.5249248 , -0.6154796 ,\n",
       "         1.3253177 , -0.31091198, -0.754149  ,  0.16694343,  0.88718104,\n",
       "         1.1945423 ,  1.2456347 , -0.6527513 ,  0.35901427,  0.7559202 ,\n",
       "         1.1882248 ,  0.54704636,  0.34679532,  1.0737512 ,  1.832159  ,\n",
       "        -0.6777701 ,  0.7282416 ,  1.8620701 ,  1.4097334 , -0.53879625,\n",
       "        -0.05848094,  0.5372102 ,  0.543745  ], dtype=float32),\n",
       " array([-0.07226297, -0.33492205,  0.84389997,  0.5288356 , -0.61442786,\n",
       "         1.3277851 , -0.31148222, -0.74258703,  0.16262077,  0.89481723,\n",
       "         1.1871518 ,  1.2441268 , -0.654611  ,  0.3544592 ,  0.7558986 ,\n",
       "         1.2025844 ,  0.5401474 ,  0.3516582 ,  1.0689275 ,  1.8416357 ,\n",
       "        -0.67590564,  0.72298956,  1.860914  ,  1.4091234 , -0.5366023 ,\n",
       "        -0.05805136,  0.53593224,  0.5438478 ], dtype=float32),\n",
       " array([-0.07247575, -0.3375456 ,  0.84537476,  0.5322965 , -0.6134352 ,\n",
       "         1.329809  , -0.31213173, -0.73166835,  0.15877604,  0.9014492 ,\n",
       "         1.1800597 ,  1.2421185 , -0.6563357 ,  0.3511031 ,  0.75572723,\n",
       "         1.215078  ,  0.533843  ,  0.3560311 ,  1.0643407 ,  1.8501406 ,\n",
       "        -0.6743044 ,  0.7182034 ,  1.8596905 ,  1.4077231 , -0.5346279 ,\n",
       "        -0.05786326,  0.5347266 ,  0.54393005], dtype=float32),\n",
       " array([-0.07282922, -0.3398177 ,  0.84640163,  0.5353457 , -0.61252636,\n",
       "         1.331449  , -0.31285784, -0.72142154,  0.15535505,  0.9071783 ,\n",
       "         1.1732724 ,  1.2397338 , -0.65792674,  0.34879753,  0.7554338 ,\n",
       "         1.225871  ,  0.5280947 ,  0.3599408 ,  1.0600216 ,  1.8577585 ,\n",
       "        -0.67292434,  0.713816  ,  1.8584075 ,  1.4056689 , -0.5328667 ,\n",
       "        -0.05788656,  0.5335955 ,  0.54399395], dtype=float32),\n",
       " array([-0.07330623, -0.34177446,  0.8470281 ,  0.53802073, -0.61172   ,\n",
       "         1.332762  , -0.31365636, -0.71185684,  0.15230903,  0.91210055,\n",
       "         1.1667972 ,  1.2370797 , -0.65939015,  0.34740874,  0.7550448 ,\n",
       "         1.235127  ,  0.52286303,  0.3634184 ,  1.0559922 ,  1.8645692 ,\n",
       "        -0.67172945,  0.709772  ,  1.8570769 ,  1.4030837 , -0.53131247,\n",
       "        -0.05808837,  0.5325402 ,  0.5440413 ], dtype=float32),\n",
       " array([-0.07388997, -0.34345055,  0.8473011 ,  0.5403582 , -0.6110282 ,\n",
       "         1.3337977 , -0.31452212, -0.7029707 ,  0.14959423,  0.916307  ,\n",
       "         1.1606394 ,  1.2342471 , -0.66073287,  0.3468139 ,  0.75458455,\n",
       "         1.2430035 ,  0.5181084 ,  0.36649707,  1.0522653 ,  1.87065   ,\n",
       "        -0.67068857,  0.7060265 ,  1.8557136 ,  1.4000778 , -0.5299568 ,\n",
       "        -0.05843591,  0.5315594 ,  0.5440747 ], dtype=float32),\n",
       " array([-0.07456499, -0.34487852,  0.8472661 ,  0.54239285, -0.61045796,\n",
       "         1.3346002 , -0.3154488 , -0.6947483 ,  0.14717248,  0.919881  ,\n",
       "         1.1548021 ,  1.2313119 , -0.6619637 ,  0.34690353,  0.7540753 ,\n",
       "         1.2496507 ,  0.5137926 ,  0.36921132,  1.0488459 ,  1.8760734 ,\n",
       "        -0.66977555,  0.7025432 ,  1.8543327 ,  1.3967474 , -0.5287905 ,\n",
       "        -0.05889674,  0.53065205,  0.5440965 ], dtype=float32),\n",
       " array([-0.07531673, -0.3460886 ,  0.8469654 ,  0.54415745, -0.6100122 ,\n",
       "         1.3352079 , -0.31642926, -0.6871662 ,  0.14500956,  0.92289984,\n",
       "         1.1492857 ,  1.2283366 , -0.66309065,  0.34757993,  0.75353605,\n",
       "         1.2552091 ,  0.50987816,  0.3715953 ,  1.0457324 ,  1.8809041 ,\n",
       "        -0.6689676 ,  0.6992923 ,  1.8529495 ,  1.3931757 , -0.5278036 ,\n",
       "        -0.0594404 ,  0.5298157 ,  0.54410875], dtype=float32),\n",
       " array([-0.07613158, -0.3471076 ,  0.8464382 ,  0.5456821 , -0.6096893 ,\n",
       "         1.335655  , -0.3174567 , -0.6801961 ,  0.14307585,  0.92543304,\n",
       "         1.1440884 ,  1.2253724 , -0.6641215 ,  0.34875596,  0.75298345,\n",
       "         1.2598081 ,  0.50633055,  0.3736827 ,  1.0429184 ,  1.8852042 ,\n",
       "        -0.6682452 ,  0.6962502 ,  1.8515782 ,  1.3894341 , -0.52698505,\n",
       "        -0.06003857,  0.5290474 ,  0.54411346], dtype=float32),\n",
       " array([-0.07699736, -0.3479603 ,  0.84572077,  0.5469948 , -0.6094857 ,\n",
       "         1.3359698 , -0.31852347, -0.6738054 ,  0.14134572,  0.9275428 ,\n",
       "         1.1392057 ,  1.2224603 , -0.66506344,  0.3503547 ,  0.7524316 ,\n",
       "         1.2635664 ,  0.50311744,  0.37550545,  1.0403947 ,  1.8890274 ,\n",
       "        -0.66759187,  0.69339687,  1.8502321 ,  1.3855834 , -0.5263241 ,\n",
       "        -0.06066575,  0.52834415,  0.5441122 ], dtype=float32),\n",
       " array([-0.0779025 , -0.34866825,  0.84484553,  0.54812104, -0.609395  ,\n",
       "         1.3361769 , -0.3196227 , -0.6679596 ,  0.13979633,  0.92928505,\n",
       "         1.1346306 ,  1.219632  , -0.66592306,  0.35230854,  0.75189155,\n",
       "         1.2665913 ,  0.50020844,  0.37709343,  1.0381472 ,  1.8924236 ,\n",
       "        -0.6669942 ,  0.6907161 ,  1.8489221 ,  1.3816745 , -0.5258086 ,\n",
       "        -0.0612996 ,  0.5277025 ,  0.5441065 ], dtype=float32),\n",
       " array([-0.07883652, -0.34925073,  0.84384155,  0.54908365, -0.6094093 ,\n",
       "         1.3362968 , -0.3207468 , -0.66262317,  0.13840833,  0.93070924,\n",
       "         1.1303542 ,  1.2169118 , -0.6667061 ,  0.35455766,  0.7513722 ,\n",
       "         1.2689797 ,  0.49757642,  0.37847495,  1.0361614 ,  1.8954362 ,\n",
       "        -0.66643983,  0.6881945 ,  1.8476574 ,  1.3777498 , -0.5254265 ,\n",
       "        -0.06192078,  0.52711874,  0.54409766], dtype=float32),\n",
       " array([-0.07979044, -0.3497243 ,  0.842735  ,  0.5499034 , -0.60951966,\n",
       "         1.3363469 , -0.32188937, -0.65776026,  0.13716479,  0.9318588 ,\n",
       "         1.1263663 ,  1.2143182 , -0.6674169 ,  0.35705003,  0.7508809 ,\n",
       "         1.2708178 ,  0.49519598,  0.37967566,  1.0344211 ,  1.8981047 ,\n",
       "        -0.66591966,  0.685821  ,  1.8464459 ,  1.3738437 , -0.525166  ,\n",
       "        -0.06251235,  0.52658963,  0.5440866 ], dtype=float32),\n",
       " array([-0.08075596, -0.35010347,  0.8415486 ,  0.55059844, -0.6097165 ,\n",
       "         1.3363417 , -0.3230437 , -0.65333605,  0.13605085,  0.9327725 ,\n",
       "         1.1226553 ,  1.2118636 , -0.66805995,  0.35973924,  0.75042254,\n",
       "         1.2721823 ,  0.49304432,  0.38071892,  1.0329087 ,  1.9004632 ,\n",
       "        -0.6654249 ,  0.68358594,  1.8452935 ,  1.3699846 , -0.52501476,\n",
       "        -0.06306072,  0.52611184,  0.54407394], dtype=float32),\n",
       " array([-0.08172547, -0.35040084,  0.84030265,  0.5511855 , -0.60999006,\n",
       "         1.3362931 , -0.32420376, -0.64931613,  0.13505381,  0.9334836 ,\n",
       "         1.1192086 ,  1.2095559 , -0.66863847,  0.3625854 ,  0.7500004 ,\n",
       "         1.2731417 ,  0.4911006 ,  0.38162607,  1.031607  ,  1.9025431 ,\n",
       "        -0.6649489 ,  0.68148047,  1.8442045 ,  1.3661956 , -0.52496105,\n",
       "        -0.06355508,  0.5256814 ,  0.5440605 ], dtype=float32),\n",
       " array([-0.08269269, -0.35062715,  0.83901453,  0.551679  , -0.61033034,\n",
       "         1.3362113 , -0.32536384, -0.6456679 ,  0.13416222,  0.9340215 ,\n",
       "         1.1160133 ,  1.2073997 , -0.6691559 ,  0.36555335,  0.7496174 ,\n",
       "         1.2737564 ,  0.4893453 ,  0.38241583,  1.0304986 ,  1.904371  ,\n",
       "        -0.6644861 ,  0.6794978 ,  1.8431816 ,  1.3624942 , -0.52499306,\n",
       "        -0.06398658,  0.5252957 ,  0.5440465 ], dtype=float32),\n",
       " array([-0.08365194, -0.35079175,  0.8376995 ,  0.55209196, -0.6107275 ,\n",
       "         1.3361037 , -0.32651868, -0.64236045,  0.13336661,  0.93441164,\n",
       "         1.1130553 ,  1.205396  , -0.66961485,  0.36861262,  0.74927443,\n",
       "         1.2740782 ,  0.4877615 ,  0.3831052 ,  1.0295668 ,  1.9059715 ,\n",
       "        -0.66403216,  0.6776311 ,  1.8422263 ,  1.3588945 , -0.52509964,\n",
       "        -0.06434851,  0.5249511 ,  0.5440325 ], dtype=float32),\n",
       " array([-0.08459821, -0.3509026 ,  0.8363705 ,  0.5524357 , -0.6111723 ,\n",
       "         1.3359774 , -0.32766375, -0.63936394,  0.13265823,  0.9346765 ,\n",
       "         1.1103214 ,  1.2035433 , -0.67001784,  0.37173632,  0.7489717 ,\n",
       "         1.2741543 ,  0.4863336 ,  0.38370883,  1.0287943 ,  1.9073654 ,\n",
       "        -0.66358346,  0.67587435,  1.8413388 ,  1.3554071 , -0.5252697 ,\n",
       "        -0.06463677,  0.5246445 ,  0.5440188 ], dtype=float32),\n",
       " array([-0.08552735, -0.35096657,  0.83503854,  0.5527199 , -0.611656  ,\n",
       "         1.335837  , -0.32879475, -0.636651  ,  0.13202949,  0.9348353 ,\n",
       "         1.1077974 ,  1.2018385 , -0.67036694,  0.3749012 ,  0.748709  ,\n",
       "         1.274024  ,  0.4850468 ,  0.38423997,  1.028166  ,  1.9085716 ,\n",
       "        -0.6631369 ,  0.6742222 ,  1.8405188 ,  1.3520389 , -0.5254932 ,\n",
       "        -0.0648485 ,  0.5243732 ,  0.54400533], dtype=float32),\n",
       " array([-0.08643601, -0.3509893 ,  0.8337134 ,  0.5529533 , -0.6121699 ,\n",
       "         1.3356875 , -0.32990786, -0.63419574,  0.13147369,  0.93490434,\n",
       "         1.1054702 ,  1.2002766 , -0.6706649 ,  0.37808752,  0.7484849 ,\n",
       "         1.273723  ,  0.4838881 ,  0.38470975,  1.0276664 ,  1.9096072 ,\n",
       "        -0.6626911 ,  0.6726696 ,  1.839765  ,  1.3487957 , -0.5257605 ,\n",
       "        -0.06498268,  0.52413434,  0.5439922 ], dtype=float32),\n",
       " array([-0.08732131, -0.3509762 ,  0.8324029 ,  0.5531436 , -0.61270696,\n",
       "         1.3355316 , -0.33100015, -0.63197434,  0.1309848 ,  0.9348979 ,\n",
       "         1.1033264 ,  1.198852  , -0.6709135 ,  0.38127765,  0.7482987 ,\n",
       "         1.2732816 ,  0.48284578,  0.38512778,  1.027282  ,  1.910487  ,\n",
       "        -0.6622444 ,  0.67121166,  1.8390754 ,  1.345681  , -0.5260625 ,\n",
       "        -0.06503928,  0.52392536,  0.5439797 ], dtype=float32),\n",
       " array([-0.08818082, -0.35093126,  0.8311132 ,  0.5532972 , -0.61325985,\n",
       "         1.3353721 , -0.33206862, -0.62996477,  0.13055742,  0.9348283 ,\n",
       "         1.1013535 ,  1.1975573 , -0.6711156 ,  0.38445637,  0.7481482 ,\n",
       "         1.2727256 ,  0.48190874,  0.3855025 ,  1.026999  ,  1.9112251 ,\n",
       "        -0.661796  ,  0.66984385,  1.8384473 ,  1.3426952 , -0.5263906 ,\n",
       "        -0.06502012,  0.52374375,  0.5439676 ], dtype=float32),\n",
       " array([-0.08901294, -0.35085833,  0.82985014,  0.55341977, -0.61382234,\n",
       "         1.3352108 , -0.3331109 , -0.62814647,  0.13018663,  0.934706  ,\n",
       "         1.0995383 ,  1.1963854 , -0.67127275,  0.38761103,  0.7480315 ,\n",
       "         1.2720779 ,  0.48106694,  0.38584116,  1.0268055 ,  1.9118323 ,\n",
       "        -0.6613454 ,  0.6685617 ,  1.8378781 ,  1.3398393 , -0.52673763,\n",
       "        -0.06492721,  0.5235871 ,  0.543956  ], dtype=float32),\n",
       " array([-0.08981628, -0.3507609 ,  0.82861793,  0.55351603, -0.6143891 ,\n",
       "         1.3350494 , -0.33412543, -0.6265008 ,  0.12986818,  0.93453956,\n",
       "         1.0978693 ,  1.1953278 , -0.6713881 ,  0.3907303 ,  0.74794656,\n",
       "         1.2713577 ,  0.48031133,  0.3861496 ,  1.0266898 ,  1.9123214 ,\n",
       "        -0.66089284,  0.6673612 ,  1.8373644 ,  1.3371115 , -0.5270964 ,\n",
       "        -0.06476419,  0.52345324,  0.54394484], dtype=float32),\n",
       " array([-0.09058966, -0.35064152,  0.8274195 ,  0.5535902 , -0.61495525,\n",
       "         1.3348882 , -0.33511057, -0.62501043,  0.12959775,  0.9343376 ,\n",
       "         1.0963356 ,  1.1943768 , -0.6714638 ,  0.393805  ,  0.7478909 ,\n",
       "         1.2705815 ,  0.47963345,  0.38643256,  1.0266416 ,  1.9127011 ,\n",
       "        -0.6604385 ,  0.6662378 ,  1.836903  ,  1.3345102 , -0.527461  ,\n",
       "        -0.06453472,  0.52334   ,  0.5439341 ], dtype=float32),\n",
       " array([-0.09133286, -0.35050282,  0.8262579 ,  0.5536455 , -0.61551607,\n",
       "         1.334729  , -0.33606532, -0.62366015,  0.12937176,  0.9341059 ,\n",
       "         1.0949262 ,  1.1935247 , -0.6715022 ,  0.39682695,  0.7478622 ,\n",
       "         1.2697634 ,  0.4790256 ,  0.38669434,  1.0266516 ,  1.9129812 ,\n",
       "        -0.659983  ,  0.6651879 ,  1.8364903 ,  1.3320327 , -0.5278259 ,\n",
       "        -0.06424337,  0.5232455 ,  0.5439237 ], dtype=float32),\n",
       " array([-0.09204537, -0.3503472 ,  0.825135  ,  0.55368525, -0.6160685 ,\n",
       "         1.3345709 , -0.33698887, -0.6224355 ,  0.12918648,  0.93384993,\n",
       "         1.0936308 ,  1.1927634 , -0.6715059 ,  0.39978936,  0.74785835,\n",
       "         1.2689153 ,  0.47848094,  0.38693818,  1.0267115 ,  1.9131706 ,\n",
       "        -0.6595271 ,  0.6642077 ,  1.836123  ,  1.329676  , -0.5281862 ,\n",
       "        -0.06389465,  0.52316797,  0.5439137 ], dtype=float32),\n",
       " array([-0.09272718, -0.3501766 ,  0.8240516 ,  0.5537119 , -0.6166092 ,\n",
       "         1.3344152 , -0.33788046, -0.62132365,  0.12903902,  0.93357456,\n",
       "         1.0924402 ,  1.1920849 , -0.6714777 ,  0.40268704,  0.7478769 ,\n",
       "         1.2680472 ,  0.47799295,  0.38716674,  1.0268124 ,  1.9132767 ,\n",
       "        -0.65907174,  0.66329324,  1.8357977 ,  1.3274362 , -0.52853787,\n",
       "        -0.06349414,  0.5231057 ,  0.543904  ], dtype=float32),\n",
       " array([-0.0933788 , -0.34999293,  0.82300866,  0.5537272 , -0.6171356 ,\n",
       "         1.3342617 , -0.33874038, -0.62031335,  0.12892607,  0.93328357,\n",
       "         1.0913455 ,  1.1914818 , -0.67142   ,  0.40551504,  0.74791545,\n",
       "         1.2671678 ,  0.4775557 ,  0.38738227,  1.0269481 ,  1.913307  ,\n",
       "        -0.65861803,  0.66244113,  1.8355107 ,  1.3253093 , -0.52887803,\n",
       "        -0.06304675,  0.52305704,  0.54389447], dtype=float32),\n",
       " array([-0.09400041, -0.34979776,  0.8220065 ,  0.5537332 , -0.6176459 ,\n",
       "         1.3341101 , -0.3395685 , -0.61939377,  0.12884475,  0.93298054,\n",
       "         1.0903387 ,  1.190947  , -0.6713355 ,  0.40826994,  0.7479722 ,\n",
       "         1.2662841 ,  0.4771642 ,  0.38758627,  1.0271125 ,  1.9132689 ,\n",
       "        -0.6581671 ,  0.66164774,  1.8352588 ,  1.3232912 , -0.52920353,\n",
       "        -0.06255767,  0.52302045,  0.5438853 ], dtype=float32),\n",
       " array([-0.09459282, -0.34959257,  0.82104504,  0.5537314 , -0.6181383 ,\n",
       "         1.3339609 , -0.34036493, -0.61855567,  0.12879246,  0.93266785,\n",
       "         1.0894117 ,  1.1904742 , -0.67122686,  0.41094878,  0.74804455,\n",
       "         1.2654015 ,  0.47681314,  0.38778016,  1.0272996 ,  1.9131687 ,\n",
       "        -0.65772   ,  0.66090965,  1.8350394 ,  1.3213775 , -0.5295121 ,\n",
       "        -0.06203272,  0.5229947 ,  0.5438763 ], dtype=float32),\n",
       " array([-0.09515648, -0.3493791 ,  0.82012385,  0.5537227 , -0.6186119 ,\n",
       "         1.3338133 , -0.34113035, -0.61779106,  0.12876658,  0.93234813,\n",
       "         1.0885578 ,  1.1900562 , -0.6710968 ,  0.41354913,  0.7481306 ,\n",
       "         1.2645249 ,  0.47649807,  0.3879647 ,  1.0275041 ,  1.9130124 ,\n",
       "        -0.6572782 ,  0.6602237 ,  1.8348486 ,  1.3195632 , -0.5298023 ,\n",
       "        -0.06147691,  0.5229787 ,  0.5438674 ], dtype=float32),\n",
       " array([-0.09569264, -0.34915847,  0.8192424 ,  0.55370814, -0.61906564,\n",
       "         1.3336679 , -0.34186515, -0.6170923 ,  0.12876464,  0.9320233 ,\n",
       "         1.0877708 ,  1.1896881 , -0.67094773,  0.4160696 ,  0.7482286 ,\n",
       "         1.2636583 ,  0.4762153 ,  0.38814056,  1.0277226 ,  1.912806  ,\n",
       "        -0.6568426 ,  0.6595867 ,  1.8346844 ,  1.3178446 , -0.5300731 ,\n",
       "        -0.0608955 ,  0.52297086,  0.5438589 ], dtype=float32),\n",
       " array([-0.09620204, -0.34893212,  0.8183994 ,  0.5536884 , -0.61949897,\n",
       "         1.3335243 , -0.34256995, -0.6164525 ,  0.12878448,  0.93169487,\n",
       "         1.0870447 ,  1.189364  , -0.6707819 ,  0.41850936,  0.7483368 ,\n",
       "         1.2628046 ,  0.4759612 ,  0.38830835,  1.027951  ,  1.9125551 ,\n",
       "        -0.6564142 ,  0.6589956 ,  1.8345435 ,  1.3162177 , -0.5303235 ,\n",
       "        -0.06029328,  0.52297044,  0.54385054], dtype=float32),\n",
       " array([-0.09668558, -0.34870106,  0.81759435,  0.55366457, -0.619912  ,\n",
       "         1.3333818 , -0.34324571, -0.6158659 ,  0.12882367,  0.93136466,\n",
       "         1.0863737 ,  1.1890783 , -0.67060226,  0.42086762,  0.7484532 ,\n",
       "         1.2619668 ,  0.4757324 ,  0.38846838,  1.0281857 ,  1.9122645 ,\n",
       "        -0.65599424,  0.6584473 ,  1.8344234 ,  1.3146771 , -0.53055334,\n",
       "        -0.05967519,  0.5229762 ,  0.54384243], dtype=float32),\n",
       " array([-0.09714454, -0.3484665 ,  0.8168258 ,  0.55363667, -0.6203044 ,\n",
       "         1.333241  , -0.34389335, -0.61532754,  0.12888035,  0.9310334 ,\n",
       "         1.0857531 ,  1.1888273 , -0.6704107 ,  0.42314455,  0.74857634,\n",
       "         1.2611463 ,  0.47552603,  0.38862073,  1.0284244 ,  1.9119395 ,\n",
       "        -0.65558374,  0.6579392 ,  1.8343221 ,  1.313219  , -0.53076273,\n",
       "        -0.05904549,  0.5229876 ,  0.5438344 ], dtype=float32),\n",
       " array([-0.09757999, -0.34822944,  0.8160926 ,  0.55360544, -0.62067693,\n",
       "         1.3331013 , -0.34451345, -0.6148329 ,  0.12895258,  0.93070257,\n",
       "         1.0851786 ,  1.188607  , -0.6702092 ,  0.42534038,  0.74870473,\n",
       "         1.2603446 ,  0.4753396 ,  0.3887656 ,  1.0286646 ,  1.9115844 ,\n",
       "        -0.65518314,  0.65746874,  1.8342376 ,  1.3118392 , -0.53095144,\n",
       "        -0.05840828,  0.5230036 ,  0.5438268 ], dtype=float32),\n",
       " array([-0.09799286, -0.34799096,  0.81539345,  0.55357105, -0.62102944,\n",
       "         1.332963  , -0.34510714, -0.61437756,  0.12903824,  0.9303725 ,\n",
       "         1.0846463 ,  1.1884134 , -0.67000014,  0.4274555 ,  0.7488373 ,\n",
       "         1.2595626 ,  0.47517067,  0.38890302,  1.0289041 ,  1.9112034 ,\n",
       "        -0.6547937 ,  0.6570334 ,  1.834168  ,  1.3105341 , -0.53112054,\n",
       "        -0.05776741,  0.5230235 ,  0.54381937], dtype=float32),\n",
       " array([-0.09838422, -0.34775174,  0.81472695,  0.5535339 , -0.62136275,\n",
       "         1.3328258 , -0.34567538, -0.61395824,  0.12913573,  0.9300444 ,\n",
       "         1.0841522 ,  1.1882432 , -0.6697854 ,  0.42949066,  0.74897254,\n",
       "         1.2588015 ,  0.475017  ,  0.38903278,  1.0291414 ,  1.9108009 ,\n",
       "        -0.6544158 ,  0.6566307 ,  1.8341109 ,  1.3092997 , -0.5312706 ,\n",
       "        -0.05712656,  0.5230466 ,  0.54381216], dtype=float32),\n",
       " array([-0.09875537, -0.3475128 ,  0.8140918 ,  0.5534943 , -0.62167734,\n",
       "         1.3326899 , -0.34621912, -0.61357176,  0.12924354,  0.92971873,\n",
       "         1.0836935 ,  1.1880931 , -0.6695667 ,  0.43144742,  0.74910957,\n",
       "         1.2580614 ,  0.4748773 ,  0.38915533,  1.0293751 ,  1.91038   ,\n",
       "        -0.65405005,  0.6562585 ,  1.8340652 ,  1.3081325 , -0.5314027 ,\n",
       "        -0.05648892,  0.5230725 ,  0.5438053 ], dtype=float32),\n",
       " array([-0.09910702, -0.34727508,  0.8134868 ,  0.5534524 , -0.6219742 ,\n",
       "         1.3325552 , -0.34673947, -0.61321545,  0.12936008,  0.9293963 ,\n",
       "         1.0832672 ,  1.1879615 , -0.6693455 ,  0.43332657,  0.74924743,\n",
       "         1.2573429 ,  0.4747494 ,  0.38927037,  1.0296044 ,  1.909945  ,\n",
       "        -0.653697  ,  0.6559147 ,  1.8340296 ,  1.3070291 , -0.5315176 ,\n",
       "        -0.05585731,  0.52310044,  0.54379886], dtype=float32),\n",
       " array([-0.09944045, -0.34703898,  0.81291044,  0.55340844, -0.62225384,\n",
       "         1.3324218 , -0.34723708, -0.6128866 ,  0.1294839 ,  0.92907727,\n",
       "         1.0828707 ,  1.1878455 , -0.6691236 ,  0.43512997,  0.74938524,\n",
       "         1.2566458 ,  0.4746322 ,  0.38937792,  1.0298281 ,  1.9094983 ,\n",
       "        -0.65335673,  0.65559715,  1.8340024 ,  1.305986  , -0.5316165 ,\n",
       "        -0.05523419,  0.52313006,  0.54379267], dtype=float32),\n",
       " array([-0.09975641, -0.34680545,  0.81236166,  0.5533624 , -0.62251717,\n",
       "         1.3322897 , -0.34771284, -0.61258346,  0.12961376,  0.9287622 ,\n",
       "         1.0825016 ,  1.1877433 , -0.6689021 ,  0.43685925,  0.74952203,\n",
       "         1.2559702 ,  0.47452414,  0.3894782 ,  1.0300456 ,  1.9090437 ,\n",
       "        -0.65303   ,  0.6553043 ,  1.8339826 ,  1.3050002 , -0.53170073,\n",
       "        -0.05462193,  0.5231611 ,  0.5437868 ], dtype=float32),\n",
       " array([-0.100056  , -0.34657493,  0.81183875,  0.5533147 , -0.622765  ,\n",
       "         1.332159  , -0.34816793, -0.6123039 ,  0.12974842,  0.9284518 ,\n",
       "         1.0821579 ,  1.1876531 , -0.6686822 ,  0.43851534,  0.7496574 ,\n",
       "         1.2553159 ,  0.47442454,  0.38957113,  1.0302563 ,  1.9085834 ,\n",
       "        -0.65271664,  0.65503424,  1.8339694 ,  1.3040687 , -0.5317716 ,\n",
       "        -0.05402227,  0.52319306,  0.5437814 ], dtype=float32),\n",
       " array([-0.10034003, -0.34634805,  0.81134105,  0.5532654 , -0.622998  ,\n",
       "         1.3320295 , -0.348603  , -0.6120465 ,  0.12988676,  0.92814624,\n",
       "         1.0818378 ,  1.1875736 , -0.668465  ,  0.44010112,  0.7497908 ,\n",
       "         1.2546831 ,  0.47433245,  0.38965696,  1.0304605 ,  1.9081198 ,\n",
       "        -0.6524165 ,  0.6547852 ,  1.8339616 ,  1.3031888 , -0.53183043,\n",
       "        -0.05343687,  0.52322567,  0.54377633], dtype=float32),\n",
       " array([-0.10060924, -0.34612527,  0.8108672 ,  0.55321455, -0.6232173 ,\n",
       "         1.3319018 , -0.34901896, -0.6118097 ,  0.13002771,  0.9278457 ,\n",
       "         1.0815398 ,  1.1875037 , -0.6682513 ,  0.44161785,  0.74992144,\n",
       "         1.2540712 ,  0.47424698,  0.3897357 ,  1.0306572 ,  1.9076554 ,\n",
       "        -0.6521298 ,  0.6545559 ,  1.8339586 ,  1.3023577 , -0.5318781 ,\n",
       "        -0.05286708,  0.52325857,  0.54377156], dtype=float32),\n",
       " array([-0.10086458, -0.345907  ,  0.810416  ,  0.5531625 , -0.6234235 ,\n",
       "         1.3317761 , -0.3494164 , -0.6115919 ,  0.13017043,  0.9275508 ,\n",
       "         1.0812616 ,  1.1874422 , -0.6680421 ,  0.44306758,  0.7500492 ,\n",
       "         1.25348   ,  0.4741677 ,  0.3898075 ,  1.0308471 ,  1.9071923 ,\n",
       "        -0.6518565 ,  0.6543448 ,  1.8339598 ,  1.3015733 , -0.53191596,\n",
       "        -0.05231401,  0.5232916 ,  0.54376733], dtype=float32),\n",
       " array([-0.10110667, -0.3456937 ,  0.8099865 ,  0.5531095 , -0.6236176 ,\n",
       "         1.331652  , -0.34979632, -0.6113921 ,  0.13031398,  0.9272618 ,\n",
       "         1.081003  ,  1.1873881 , -0.667838  ,  0.4444521 ,  0.75017357,\n",
       "         1.2529091 ,  0.47409362,  0.38987264,  1.0310295 ,  1.9067318 ,\n",
       "        -0.6515965 ,  0.6541506 ,  1.8339643 ,  1.300833  , -0.5319454 ,\n",
       "        -0.05177891,  0.5233245 ,  0.5437635 ], dtype=float32),\n",
       " array([-0.10133631, -0.34548566,  0.8095777 ,  0.5530554 , -0.62380016,\n",
       "         1.3315302 , -0.3501592 , -0.61120933,  0.13045773,  0.9269788 ,\n",
       "         1.0807621 ,  1.1873411 , -0.6676395 ,  0.44577378,  0.75029445,\n",
       "         1.2523581 ,  0.47402465,  0.3899313 ,  1.0312046 ,  1.906276  ,\n",
       "        -0.6513494 ,  0.6539722 ,  1.8339721 ,  1.3001349 , -0.5319673 ,\n",
       "        -0.05126175,  0.5233574 ,  0.54376006], dtype=float32),\n",
       " array([-0.10155404, -0.34528318,  0.80918884,  0.5530006 , -0.6239721 ,\n",
       "         1.3314104 , -0.35050583, -0.6110424 ,  0.13060094,  0.926702  ,\n",
       "         1.0805383 ,  1.1873001 , -0.667447  ,  0.44703427,  0.7504115 ,\n",
       "         1.2518268 ,  0.47396028,  0.38998383,  1.0313725 ,  1.9058264 ,\n",
       "        -0.651115  ,  0.65380836,  1.8339827 ,  1.2994765 , -0.53198266,\n",
       "        -0.05076339,  0.52338994,  0.5437571 ], dtype=float32),\n",
       " array([-0.10176051, -0.34508654,  0.80881864,  0.55294514, -0.62413406,\n",
       "         1.3312932 , -0.35083684, -0.6108907 ,  0.13074298,  0.9264319 ,\n",
       "         1.08033   ,  1.1872648 , -0.6672609 ,  0.44823587,  0.75052464,\n",
       "         1.2513145 ,  0.4738999 ,  0.39003056,  1.0315337 ,  1.9053841 ,\n",
       "        -0.6508929 ,  0.6536581 ,  1.8339951 ,  1.298856  , -0.5319923 ,\n",
       "        -0.05028399,  0.52342194,  0.5437544 ], dtype=float32),\n",
       " array([-0.10195642, -0.3448959 ,  0.80846643,  0.55288917, -0.6242867 ,\n",
       "         1.3311785 , -0.3511528 , -0.6107532 ,  0.13088329,  0.9261683 ,\n",
       "         1.0801367 ,  1.1872348 , -0.6670815 ,  0.4493808 ,  0.75063366,\n",
       "         1.2508209 ,  0.47384346,  0.39007148,  1.031688  ,  1.9049499 ,\n",
       "        -0.6506831 ,  0.65352064,  1.8340099 ,  1.2982721 , -0.53199774,\n",
       "        -0.04982373,  0.5234534 ,  0.5437522 ], dtype=float32),\n",
       " array([-0.10214228, -0.34471136,  0.8081318 ,  0.5528329 , -0.6244306 ,\n",
       "         1.3310665 , -0.35145423, -0.6106289 ,  0.13102157,  0.92591155,\n",
       "         1.0799575 ,  1.1872096 , -0.66690886,  0.4504703 ,  0.7507387 ,\n",
       "         1.2503457 ,  0.47379082,  0.39010736,  1.0318358 ,  1.904525  ,\n",
       "        -0.6504846 ,  0.65339476,  1.834026  ,  1.2977226 , -0.5319992 ,\n",
       "        -0.04938231,  0.52348435,  0.54375035], dtype=float32),\n",
       " array([-0.10231837, -0.3445331 ,  0.8078134 ,  0.5527766 , -0.62456626,\n",
       "         1.3309569 , -0.35174158, -0.61051726,  0.13115734,  0.9256619 ,\n",
       "         1.0797919 ,  1.187189  , -0.6667433 ,  0.45150664,  0.7508397 ,\n",
       "         1.2498887 ,  0.4737417 ,  0.39013818,  1.0319774 ,  1.9041098 ,\n",
       "        -0.6502976 ,  0.65327984,  1.8340431 ,  1.297206  , -0.53199744,\n",
       "        -0.04895993,  0.52351445,  0.54374886], dtype=float32),\n",
       " array([-0.1024852 , -0.34436104,  0.80751085,  0.5527202 , -0.62469447,\n",
       "         1.3308505 , -0.35201564, -0.6104176 ,  0.13129027,  0.92541945,\n",
       "         1.0796388 ,  1.1871728 , -0.6665848 ,  0.45249164,  0.7509365 ,\n",
       "         1.2494494 ,  0.47369564,  0.39016432,  1.0321128 ,  1.9037055 ,\n",
       "        -0.6501212 ,  0.653175  ,  1.8340617 ,  1.2967204 , -0.5319933 ,\n",
       "        -0.04855654,  0.523544  ,  0.5437478 ], dtype=float32),\n",
       " array([-0.10264333, -0.34419534,  0.80722356,  0.55266404, -0.6248152 ,\n",
       "         1.3307474 , -0.3522767 , -0.61032915,  0.13142015,  0.92518437,\n",
       "         1.079498  ,  1.1871604 , -0.6664329 ,  0.45342746,  0.75102925,\n",
       "         1.2490271 ,  0.4736526 ,  0.39018622,  1.0322423 ,  1.9033124 ,\n",
       "        -0.64995515,  0.65307945,  1.8340812 ,  1.2962644 , -0.5319873 ,\n",
       "        -0.04817124,  0.5235727 ,  0.5437471 ], dtype=float32),\n",
       " array([-0.10279315, -0.34403598,  0.8069508 ,  0.5526084 , -0.6249295 ,\n",
       "         1.3306466 , -0.35252532, -0.610251  ,  0.13154663,  0.92495656,\n",
       "         1.0793682 ,  1.1871519 , -0.66628814,  0.45431584,  0.75111806,\n",
       "         1.2486218 ,  0.47361296,  0.390204  ,  1.0323663 ,  1.9029309 ,\n",
       "        -0.649799  ,  0.6529925 ,  1.8341012 ,  1.2958369 , -0.5319801 ,\n",
       "        -0.04780406,  0.5236006 ,  0.54374665], dtype=float32),\n",
       " array([-0.10293499, -0.3438829 ,  0.8066921 ,  0.55255306, -0.6250373 ,\n",
       "         1.3305496 , -0.3527619 , -0.61018306,  0.1316696 ,  0.92473614,\n",
       "         1.0792496 ,  1.1871473 , -0.66615   ,  0.45515856,  0.751203  ,\n",
       "         1.2482332 ,  0.473576  ,  0.39021832,  1.0324849 ,  1.9025611 ,\n",
       "        -0.6496522 ,  0.65291387,  1.8341221 ,  1.2954366 , -0.5319718 ,\n",
       "        -0.04745421,  0.52362794,  0.5437465 ], dtype=float32),\n",
       " array([-0.10306904, -0.3437362 ,  0.8064468 ,  0.55249834, -0.6251391 ,\n",
       "         1.3304554 , -0.35298684, -0.6101244 ,  0.13178903,  0.9245231 ,\n",
       "         1.079141  ,  1.1871458 , -0.66601896,  0.45595738,  0.75128394,\n",
       "         1.2478603 ,  0.47354215,  0.3902292 ,  1.032598  ,  1.9022032 ,\n",
       "        -0.6495144 ,  0.6528427 ,  1.8341435 ,  1.295062  , -0.531963  ,\n",
       "        -0.04712147,  0.5236543 ,  0.54374677], dtype=float32),\n",
       " array([-0.10319584, -0.3435956 ,  0.8062142 ,  0.55244416, -0.62523544,\n",
       "         1.3303646 , -0.35320064, -0.61007434,  0.13190453,  0.9243175 ,\n",
       "         1.079042  ,  1.187148  , -0.66589415,  0.45671406,  0.751361  ,\n",
       "         1.2475035 ,  0.47351056,  0.39023706,  1.0327064 ,  1.9018576 ,\n",
       "        -0.64938533,  0.65277857,  1.8341655 ,  1.2947118 , -0.531954  ,\n",
       "        -0.04680511,  0.52368   ,  0.5437471 ], dtype=float32),\n",
       " array([-0.1033157 , -0.34346107,  0.80599403,  0.55239105, -0.6253264 ,\n",
       "         1.3302771 , -0.35340345, -0.6100325 ,  0.13201648,  0.92411935,\n",
       "         1.0789524 ,  1.1871527 , -0.6657758 ,  0.45743042,  0.75143474,\n",
       "         1.2471625 ,  0.4734816 ,  0.39024225,  1.0328101 ,  1.9015244 ,\n",
       "        -0.6492642 ,  0.65272105,  1.8341877 ,  1.2943851 , -0.53194463,\n",
       "        -0.04650458,  0.52370477,  0.54374796], dtype=float32),\n",
       " array([-0.10342885, -0.34333256,  0.8057857 ,  0.5523388 , -0.6254124 ,\n",
       "         1.3301929 , -0.35359594, -0.60999817,  0.1321243 ,  0.9239285 ,\n",
       "         1.0788711 ,  1.1871605 , -0.6656636 ,  0.45810768,  0.75150466,\n",
       "         1.2468354 ,  0.47345495,  0.39024496,  1.032909  ,  1.901203  ,\n",
       "        -0.64915085,  0.6526695 ,  1.8342106 ,  1.2940806 , -0.53193563,\n",
       "        -0.04621942,  0.52372885,  0.5437488 ], dtype=float32),\n",
       " array([-0.10353573, -0.34320992,  0.80558896,  0.55228764, -0.6254937 ,\n",
       "         1.3301122 , -0.35377845, -0.6099711 ,  0.13222821,  0.92374516,\n",
       "         1.0787978 ,  1.1871712 , -0.66555715,  0.45874813,  0.7515712 ,\n",
       "         1.2465237 ,  0.47343087,  0.39024556,  1.0330036 ,  1.900894  ,\n",
       "        -0.64904505,  0.6526236 ,  1.8342335 ,  1.2937971 , -0.53192693,\n",
       "        -0.04594883,  0.5237522 ,  0.54374987], dtype=float32),\n",
       " array([-0.10363644, -0.34309298,  0.80540305,  0.5522376 , -0.6255705 ,\n",
       "         1.3300347 , -0.3539514 , -0.60995024,  0.13232833,  0.92356896,\n",
       "         1.0787328 ,  1.1871843 , -0.66545665,  0.45935276,  0.7516345 ,\n",
       "         1.246226  ,  0.473409  ,  0.3902442 ,  1.033094  ,  1.9005971 ,\n",
       "        -0.6489458 ,  0.65258294,  1.8342565 ,  1.2935336 , -0.5319187 ,\n",
       "        -0.04569251,  0.52377486,  0.5437512 ], dtype=float32),\n",
       " array([-0.10373128, -0.34298164,  0.80522776,  0.55218875, -0.62564284,\n",
       "         1.3299602 , -0.35411486, -0.6099354 ,  0.13242453,  0.9234003 ,\n",
       "         1.0786746 ,  1.1871998 , -0.6653615 ,  0.45992312,  0.75169474,\n",
       "         1.2459419 ,  0.4733892 ,  0.39024112,  1.0331801 ,  1.9003123 ,\n",
       "        -0.64885324,  0.6525471 ,  1.8342795 ,  1.2932893 , -0.53191096,\n",
       "        -0.04544985,  0.5237965 ,  0.54375273], dtype=float32),\n",
       " array([-0.10382047, -0.34287566,  0.8050624 ,  0.5521412 , -0.62571114,\n",
       "         1.3298892 , -0.35426947, -0.60992616,  0.13251694,  0.9232384 ,\n",
       "         1.0786232 ,  1.1872176 , -0.66527176,  0.460461  ,  0.7517519 ,\n",
       "         1.2456717 ,  0.47337174,  0.39023665,  1.0332626 ,  1.9000388 ,\n",
       "        -0.6487666 ,  0.6525158 ,  1.8343024 ,  1.2930628 , -0.53190356,\n",
       "        -0.04522   ,  0.52381754,  0.5437543 ], dtype=float32),\n",
       " array([-0.10390418, -0.34277493,  0.8049068 ,  0.5520948 , -0.62577575,\n",
       "         1.3298212 , -0.3544154 , -0.6099218 ,  0.13260552,  0.92308354,\n",
       "         1.0785785 ,  1.1872377 , -0.66518694,  0.46096742,  0.75180626,\n",
       "         1.2454146 ,  0.4733562 ,  0.39023086,  1.033341  ,  1.8997769 ,\n",
       "        -0.648686  ,  0.6524887 ,  1.8343256 ,  1.2928534 , -0.5318969 ,\n",
       "        -0.04500256,  0.5238379 ,  0.54375607], dtype=float32),\n",
       " array([-0.10398286, -0.34267947,  0.80476034,  0.5520498 , -0.62583673,\n",
       "         1.3297563 , -0.35455292, -0.6099219 ,  0.13269053,  0.9229357 ,\n",
       "         1.078539  ,  1.1872593 , -0.665107  ,  0.46144417,  0.7518577 ,\n",
       "         1.24517   ,  0.47334236,  0.39022437,  1.0334159 ,  1.8995262 ,\n",
       "        -0.64861053,  0.6524655 ,  1.8343487 ,  1.29266   , -0.53189075,\n",
       "        -0.04479681,  0.52385753,  0.54375803], dtype=float32),\n",
       " array([-0.10405659, -0.3425887 ,  0.8046228 ,  0.55200636, -0.6258939 ,\n",
       "         1.3296944 , -0.3546827 , -0.60992646,  0.13277176,  0.92279446,\n",
       "         1.0785055 ,  1.1872827 , -0.66503125,  0.46189228,  0.75190663,\n",
       "         1.244938  ,  0.47333026,  0.39021677,  1.0334873 ,  1.8992866 ,\n",
       "        -0.6485403 ,  0.6524459 ,  1.8343714 ,  1.292482  , -0.5318852 ,\n",
       "        -0.04460233,  0.5238764 ,  0.54375994], dtype=float32),\n",
       " array([-0.10412563, -0.34250265,  0.80449384,  0.5519644 , -0.62594765,\n",
       "         1.329636  , -0.3548048 , -0.6099345 ,  0.13284947,  0.9226599 ,\n",
       "         1.0784763 ,  1.1873072 , -0.66496015,  0.46231312,  0.75195295,\n",
       "         1.2447177 ,  0.47331983,  0.39020836,  1.0335553 ,  1.8990579 ,\n",
       "        -0.64847463,  0.6524294 ,  1.8343941 ,  1.2923183 , -0.53187984,\n",
       "        -0.04441871,  0.5238946 ,  0.54376197], dtype=float32),\n",
       " array([-0.10419034, -0.3424212 ,  0.8043728 ,  0.55192375, -0.6259982 ,\n",
       "         1.3295803 , -0.35491955, -0.60994583,  0.13292368,  0.9225316 ,\n",
       "         1.0784522 ,  1.1873331 , -0.6648931 ,  0.46270826,  0.7519967 ,\n",
       "         1.2445091 ,  0.47331095,  0.39019933,  1.0336201 ,  1.8988395 ,\n",
       "        -0.6484139 ,  0.6524159 ,  1.8344164 ,  1.2921686 , -0.53187513,\n",
       "        -0.04424538,  0.52391213,  0.54376405], dtype=float32),\n",
       " array([-0.10425075, -0.34234416,  0.80425966,  0.55188465, -0.62604576,\n",
       "         1.3295273 , -0.35502747, -0.6099602 ,  0.13299458,  0.9224094 ,\n",
       "         1.0784328 ,  1.1873604 , -0.66482997,  0.46307865,  0.75203836,\n",
       "         1.2443115 ,  0.4733035 ,  0.39018992,  1.0336814 ,  1.8986309 ,\n",
       "        -0.6483573 ,  0.6524053 ,  1.8344389 ,  1.2920316 , -0.5318711 ,\n",
       "        -0.04408171,  0.5239291 ,  0.54376614], dtype=float32),\n",
       " array([-0.10430702, -0.34227136,  0.80415374,  0.55184704, -0.6260903 ,\n",
       "         1.329477  , -0.35512856, -0.60997725,  0.13306224,  0.9222934 ,\n",
       "         1.078417  ,  1.1873885 , -0.66477066,  0.4634256 ,  0.75207776,\n",
       "         1.2441248 ,  0.47329742,  0.39018005,  1.0337399 ,  1.8984323 ,\n",
       "        -0.6483047 ,  0.6523972 ,  1.8344609 ,  1.2919065 , -0.5318672 ,\n",
       "        -0.04392736,  0.5239453 ,  0.5437683 ], dtype=float32),\n",
       " array([-0.1043595 , -0.34220257,  0.804055  ,  0.55181074, -0.62613183,\n",
       "         1.3294295 , -0.3552234 , -0.6099966 ,  0.13312656,  0.9221833 ,\n",
       "         1.0784049 ,  1.1874175 , -0.6647149 ,  0.46374997,  0.7521148 ,\n",
       "         1.2439482 ,  0.47329253,  0.39016995,  1.0337955 ,  1.8982428 ,\n",
       "        -0.64825606,  0.6523915 ,  1.8344824 ,  1.2917926 , -0.5318641 ,\n",
       "        -0.04378185,  0.52396095,  0.5437704 ], dtype=float32),\n",
       " array([-0.10440827, -0.34213758,  0.80396295,  0.551776  , -0.6261708 ,\n",
       "         1.3293844 , -0.35531196, -0.6100179 ,  0.13318792,  0.92207885,\n",
       "         1.0783963 ,  1.1874472 , -0.66466236,  0.46405324,  0.75214994,\n",
       "         1.2437819 ,  0.47328895,  0.3901597 ,  1.0338479 ,  1.8980631 ,\n",
       "        -0.6482106 ,  0.6523879 ,  1.8345034 ,  1.2916893 , -0.53186107,\n",
       "        -0.04364466,  0.5239759 ,  0.5437726 ], dtype=float32),\n",
       " array([-0.10445351, -0.3420764 ,  0.80387723,  0.5517427 , -0.6262073 ,\n",
       "         1.3293419 , -0.355395  , -0.6100407 ,  0.13324633,  0.92197967,\n",
       "         1.0783906 ,  1.187477  , -0.66461307,  0.46433607,  0.75218314,\n",
       "         1.2436252 ,  0.47328645,  0.39014947,  1.0338978 ,  1.8978919 ,\n",
       "        -0.6481685 ,  0.65238625,  1.8345242 ,  1.2915962 , -0.53185797,\n",
       "        -0.04351587,  0.52399015,  0.54377484], dtype=float32),\n",
       " array([-0.10449555, -0.34201857,  0.8037976 ,  0.5517108 , -0.6262414 ,\n",
       "         1.3293018 , -0.35547242, -0.6100651 ,  0.1333018 ,  0.92188585,\n",
       "         1.0783875 ,  1.1875072 , -0.6645668 ,  0.46460003,  0.7522145 ,\n",
       "         1.2434775 ,  0.4732849 ,  0.39013913,  1.0339448 ,  1.8977294 ,\n",
       "        -0.6481297 ,  0.6523861 ,  1.834545  ,  1.291512  , -0.5318554 ,\n",
       "        -0.04339439,  0.52400404,  0.5437769 ], dtype=float32),\n",
       " array([-0.1045345 , -0.34196427,  0.80372363,  0.55168015, -0.6262729 ,\n",
       "         1.3292639 , -0.35554454, -0.6100907 ,  0.13335449,  0.9217973 ,\n",
       "         1.0783871 ,  1.1875376 , -0.66452336,  0.4648456 ,  0.75224394,\n",
       "         1.2433386 ,  0.47328413,  0.3901287 ,  1.0339892 ,  1.8975748 ,\n",
       "        -0.64809376,  0.65238774,  1.834565  ,  1.2914368 , -0.5318531 ,\n",
       "        -0.04328028,  0.5240173 ,  0.54377896], dtype=float32),\n",
       " array([-0.10457037, -0.3419131 ,  0.80365515,  0.55165106, -0.62630224,\n",
       "         1.3292283 , -0.3556117 , -0.61011714,  0.13340452,  0.9217135 ,\n",
       "         1.078389  ,  1.1875684 , -0.6644828 ,  0.46507412,  0.7522721 ,\n",
       "         1.2432082 ,  0.47328424,  0.39011854,  1.0340314 ,  1.8974284 ,\n",
       "        -0.6480607 ,  0.65239084,  1.8345846 ,  1.291369  , -0.53185076,\n",
       "        -0.0431729 ,  0.5240299 ,  0.54378116], dtype=float32),\n",
       " array([-0.1046034 , -0.34186506,  0.8035919 ,  0.55162334, -0.6263295 ,\n",
       "         1.3291949 , -0.355674  , -0.6101444 ,  0.13345203,  0.92163455,\n",
       "         1.0783924 ,  1.1875991 , -0.6644446 ,  0.46528605,  0.7522984 ,\n",
       "         1.243086  ,  0.4732852 ,  0.39010832,  1.0340711 ,  1.8972895 ,\n",
       "        -0.64803   ,  0.6523951 ,  1.8346038 ,  1.2913097 , -0.5318486 ,\n",
       "        -0.04307217,  0.524042  ,  0.5437832 ], dtype=float32),\n",
       " array([-0.10463382, -0.34182003,  0.8035337 ,  0.5515968 , -0.6263545 ,\n",
       "         1.3291634 , -0.3557319 , -0.61017233,  0.1334971 ,  0.92156005,\n",
       "         1.078398  ,  1.1876293 , -0.664409  ,  0.46548256,  0.7523232 ,\n",
       "         1.2429715 ,  0.47328675,  0.39009836,  1.0341084 ,  1.8971581 ,\n",
       "        -0.64800197,  0.65240055,  1.8346224 ,  1.291257  , -0.5318465 ,\n",
       "        -0.04297762,  0.5240536 ,  0.5437853 ], dtype=float32),\n",
       " array([-0.10466181, -0.34177765,  0.80347955,  0.55157167, -0.6263778 ,\n",
       "         1.3291341 , -0.3557856 , -0.61020046,  0.13353965,  0.92148983,\n",
       "         1.078405  ,  1.1876594 , -0.66437566,  0.46566465,  0.7523466 ,\n",
       "         1.2428644 ,  0.47328866,  0.3900884 ,  1.0341436 ,  1.8970339 ,\n",
       "        -0.64797616,  0.652407  ,  1.8346405 ,  1.2912104 , -0.5318444 ,\n",
       "        -0.04288923,  0.52406454,  0.54378724], dtype=float32),\n",
       " array([-0.10468739, -0.34173802,  0.8034304 ,  0.55154777, -0.62639934,\n",
       "         1.3291061 , -0.3558353 , -0.6102287 ,  0.13358004,  0.9214239 ,\n",
       "         1.0784136 ,  1.1876891 , -0.6643444 ,  0.46583298,  0.7523685 ,\n",
       "         1.2427642 ,  0.4732911 ,  0.39007872,  1.0341766 ,  1.8969164 ,\n",
       "        -0.6479525 ,  0.6524143 ,  1.8346578 ,  1.2911701 , -0.53184223,\n",
       "        -0.04280642,  0.52407503,  0.54378927], dtype=float32),\n",
       " array([-0.10471079, -0.3417009 ,  0.8033848 ,  0.55152506, -0.62641907,\n",
       "         1.3290801 , -0.35588124, -0.61025745,  0.13361816,  0.9213617 ,\n",
       "         1.0784229 ,  1.1877183 , -0.6643154 ,  0.46598825,  0.7523891 ,\n",
       "         1.2426707 ,  0.47329414,  0.39006925,  1.0342077 ,  1.8968053 ,\n",
       "        -0.6479308 ,  0.65242225,  1.8346748 ,  1.2911357 , -0.5318401 ,\n",
       "        -0.04272897,  0.524085  ,  0.5437911 ], dtype=float32),\n",
       " array([-0.10473221, -0.34166616,  0.8033431 ,  0.5515034 , -0.626437  ,\n",
       "         1.3290555 , -0.35592353, -0.6102862 ,  0.1336543 ,  0.9213035 ,\n",
       "         1.0784334 ,  1.187747  , -0.6642882 ,  0.46613142,  0.7524083 ,\n",
       "         1.2425835 ,  0.4732973 ,  0.3900601 ,  1.0342364 ,  1.8967001 ,\n",
       "        -0.647911  ,  0.65243095,  1.8346912 ,  1.2911059 , -0.5318378 ,\n",
       "        -0.04265659,  0.52409446,  0.5437929 ], dtype=float32),\n",
       " array([-0.10475174, -0.34163368,  0.80330527,  0.5514829 , -0.62645364,\n",
       "         1.3290329 , -0.35596249, -0.6103144 ,  0.13368836,  0.9212489 ,\n",
       "         1.078445  ,  1.1877751 , -0.6642629 ,  0.46626306,  0.7524264 ,\n",
       "         1.2425026 ,  0.4733011 ,  0.39005113,  1.0342636 ,  1.8966014 ,\n",
       "        -0.6478929 ,  0.65244025,  1.834707  ,  1.2910806 , -0.53183573,\n",
       "        -0.04258908,  0.5241035 ,  0.54379475], dtype=float32),\n",
       " array([-0.10476931, -0.34160334,  0.80327064,  0.55146366, -0.6264687 ,\n",
       "         1.3290116 , -0.35599837, -0.61034256,  0.13372046,  0.9211978 ,\n",
       "         1.0784572 ,  1.1878028 , -0.66423917,  0.4663838 ,  0.7524435 ,\n",
       "         1.2424271 ,  0.47330493,  0.3900423 ,  1.0342892 ,  1.8965087 ,\n",
       "        -0.6478768 ,  0.65245   ,  1.8347224 ,  1.2910602 , -0.5318338 ,\n",
       "        -0.04252599,  0.5241121 ,  0.5437964 ], dtype=float32),\n",
       " array([-0.1047853 , -0.34157503,  0.8032391 ,  0.55144536, -0.62648237,\n",
       "         1.3289918 , -0.35603127, -0.61037046,  0.13375072,  0.9211496 ,\n",
       "         1.07847   ,  1.1878296 , -0.66421723,  0.4664943 ,  0.75245935,\n",
       "         1.2423571 ,  0.4733091 ,  0.3900339 ,  1.0343127 ,  1.8964208 ,\n",
       "        -0.6478619 ,  0.6524601 ,  1.834737  ,  1.2910434 , -0.5318316 ,\n",
       "        -0.04246738,  0.5241202 ,  0.543798  ], dtype=float32),\n",
       " array([-0.10479973, -0.34154865,  0.8032106 ,  0.5514281 , -0.626495  ,\n",
       "         1.3289729 , -0.3560613 , -0.6103978 ,  0.13377923,  0.921105  ,\n",
       "         1.0784832 ,  1.1878558 , -0.66419697,  0.46659562,  0.7524741 ,\n",
       "         1.2422923 ,  0.47331345,  0.3900257 ,  1.0343348 ,  1.8963383 ,\n",
       "        -0.64784837,  0.6524705 ,  1.8347511 ,  1.2910302 , -0.5318296 ,\n",
       "        -0.0424129 ,  0.524128  ,  0.5437997 ], dtype=float32),\n",
       " array([-0.10481256, -0.34152403,  0.80318475,  0.5514118 , -0.62650627,\n",
       "         1.3289557 , -0.35608897, -0.61042464,  0.13380589,  0.92106295,\n",
       "         1.0784969 ,  1.1878812 , -0.6641781 ,  0.46668783,  0.7524879 ,\n",
       "         1.2422323 ,  0.47331792,  0.39001763,  1.0343553 ,  1.8962605 ,\n",
       "        -0.6478364 ,  0.65248114,  1.8347648 ,  1.2910202 , -0.5318274 ,\n",
       "        -0.04236227,  0.5241352 ,  0.54380125], dtype=float32),\n",
       " array([-0.10482418, -0.3415013 ,  0.8031616 ,  0.5513963 , -0.6265165 ,\n",
       "         1.3289399 , -0.35611406, -0.61045116,  0.13383105,  0.92102396,\n",
       "         1.0785106 ,  1.187906  , -0.6641608 ,  0.4667718 ,  0.75250083,\n",
       "         1.2421767 ,  0.47332257,  0.39000994,  1.0343742 ,  1.896188  ,\n",
       "        -0.6478256 ,  0.652492  ,  1.834778  ,  1.2910136 , -0.5318251 ,\n",
       "        -0.0423153 ,  0.5241422 ,  0.5438028 ], dtype=float32),\n",
       " array([-0.10483446, -0.34148008,  0.80314064,  0.55138177, -0.62652576,\n",
       "         1.3289248 , -0.35613683, -0.610477  ,  0.13385475,  0.92098737,\n",
       "         1.0785247 ,  1.1879299 , -0.6641449 ,  0.4668476 ,  0.75251305,\n",
       "         1.2421255 ,  0.47332728,  0.39000252,  1.0343919 ,  1.8961198 ,\n",
       "        -0.64781606,  0.65250295,  1.8347908 ,  1.2910092 , -0.531823  ,\n",
       "        -0.04227199,  0.5241486 ,  0.54380417], dtype=float32),\n",
       " array([-0.10484373, -0.34146062,  0.8031221 ,  0.55136806, -0.6265339 ,\n",
       "         1.3289111 , -0.35615757, -0.61050195,  0.13387692,  0.9209535 ,\n",
       "         1.078539  ,  1.1879531 , -0.6641301 ,  0.46691665,  0.75252426,\n",
       "         1.2420783 ,  0.4733318 ,  0.38999528,  1.0344082 ,  1.8960558 ,\n",
       "        -0.6478077 ,  0.6525139 ,  1.8348025 ,  1.2910074 , -0.53182083,\n",
       "        -0.04223191,  0.52415484,  0.54380566], dtype=float32),\n",
       " array([-0.10485175, -0.34144244,  0.8031055 ,  0.55135524, -0.62654114,\n",
       "         1.3288981 , -0.35617623, -0.6105265 ,  0.13389763,  0.9209219 ,\n",
       "         1.0785531 ,  1.1879753 , -0.6641165 ,  0.46697852,  0.75253445,\n",
       "         1.2420349 ,  0.47333646,  0.38998827,  1.0344232 ,  1.895996  ,\n",
       "        -0.64780027,  0.6525249 ,  1.8348142 ,  1.2910078 , -0.53181857,\n",
       "        -0.042195  ,  0.52416056,  0.5438069 ], dtype=float32),\n",
       " array([-0.10485878, -0.34142575,  0.80309075,  0.55134296, -0.6265476 ,\n",
       "         1.328886  , -0.35619304, -0.6105506 ,  0.13391712,  0.92089236,\n",
       "         1.0785674 ,  1.1879965 , -0.664104  ,  0.46703434,  0.752544  ,\n",
       "         1.241995  ,  0.47334117,  0.3899815 ,  1.034437  ,  1.8959401 ,\n",
       "        -0.6477937 ,  0.652536  ,  1.8348253 ,  1.29101   , -0.5318164 ,\n",
       "        -0.04216103,  0.52416605,  0.54380816], dtype=float32),\n",
       " array([-0.10486492, -0.34141028,  0.8030777 ,  0.55133134, -0.62655336,\n",
       "         1.3288748 , -0.3562083 , -0.6105738 ,  0.13393523,  0.9208652 ,\n",
       "         1.0785815 ,  1.188017  , -0.6640925 ,  0.46708417,  0.75255275,\n",
       "         1.2419586 ,  0.4733457 ,  0.3899748 ,  1.0344498 ,  1.895888  ,\n",
       "        -0.64778817,  0.6525468 ,  1.8348356 ,  1.291014  , -0.53181434,\n",
       "        -0.04212987,  0.5241711 ,  0.54380935], dtype=float32),\n",
       " array([-0.10487028, -0.341396  ,  0.8030664 ,  0.55132085, -0.6265582 ,\n",
       "         1.3288647 , -0.35622182, -0.61059636,  0.13395226,  0.9208399 ,\n",
       "         1.0785953 ,  1.1880367 , -0.66408205,  0.46712843,  0.75256103,\n",
       "         1.2419252 ,  0.4733503 ,  0.38996848,  1.0344617 ,  1.8958396 ,\n",
       "        -0.64778334,  0.65255743,  1.8348457 ,  1.2910196 , -0.53181213,\n",
       "        -0.04210129,  0.52417594,  0.5438105 ], dtype=float32),\n",
       " array([-0.1048748 , -0.34138295,  0.8030566 ,  0.5513108 , -0.62656265,\n",
       "         1.328855  , -0.35623404, -0.61061823,  0.133968  ,  0.92081654,\n",
       "         1.0786092 ,  1.1880558 , -0.66407245,  0.46716782,  0.7525685 ,\n",
       "         1.2418948 ,  0.47335482,  0.38996246,  1.0344723 ,  1.8957943 ,\n",
       "        -0.64777935,  0.65256816,  1.834855  ,  1.2910265 , -0.53181005,\n",
       "        -0.04207518,  0.5241804 ,  0.5438116 ], dtype=float32),\n",
       " array([-0.10487864, -0.341371  ,  0.80304813,  0.55130136, -0.62656635,\n",
       "         1.3288465 , -0.35624477, -0.61063933,  0.13398276,  0.9207951 ,\n",
       "         1.0786232 ,  1.1880739 , -0.66406393,  0.46720228,  0.75257546,\n",
       "         1.2418672 ,  0.4733593 ,  0.38995665,  1.034482  ,  1.8957521 ,\n",
       "        -0.6477759 ,  0.65257865,  1.8348638 ,  1.2910345 , -0.5318078 ,\n",
       "        -0.04205136,  0.52418476,  0.54381275], dtype=float32),\n",
       " array([-0.10488179, -0.3413599 ,  0.803041  ,  0.5512926 , -0.62656945,\n",
       "         1.3288386 , -0.35625422, -0.61065954,  0.13399644,  0.9207753 ,\n",
       "         1.078636  ,  1.1880912 , -0.664056  ,  0.46723247,  0.75258183,\n",
       "         1.241842  ,  0.4733637 ,  0.38995108,  1.034491  ,  1.8957131 ,\n",
       "        -0.6477731 ,  0.652589  ,  1.8348721 ,  1.2910438 , -0.5318056 ,\n",
       "        -0.04202982,  0.5241886 ,  0.5438138 ], dtype=float32),\n",
       " array([-0.10488438, -0.34134996,  0.8030349 ,  0.5512843 , -0.62657213,\n",
       "         1.3288312 , -0.35626248, -0.61067927,  0.13400908,  0.92075694,\n",
       "         1.078649  ,  1.1881076 , -0.6640489 ,  0.46725872,  0.7525875 ,\n",
       "         1.2418195 ,  0.47336787,  0.38994557,  1.034499  ,  1.8956767 ,\n",
       "        -0.64777094,  0.65259904,  1.8348804 ,  1.2910537 , -0.5318036 ,\n",
       "        -0.04201023,  0.5241923 ,  0.5438147 ], dtype=float32),\n",
       " array([-0.10488624, -0.34134072,  0.80303   ,  0.55127656, -0.62657434,\n",
       "         1.3288245 , -0.35626975, -0.6106981 ,  0.13402084,  0.92074025,\n",
       "         1.078662  ,  1.1881233 , -0.6640425 ,  0.46728107,  0.7525928 ,\n",
       "         1.241799  ,  0.47337204,  0.38994032,  1.0345066 ,  1.8956435 ,\n",
       "        -0.64776933,  0.65260893,  1.8348879 ,  1.2910645 , -0.53180146,\n",
       "        -0.04199269,  0.5241956 ,  0.5438157 ], dtype=float32),\n",
       " array([-0.10488784, -0.34133232,  0.8030261 ,  0.5512694 , -0.62657607,\n",
       "         1.3288182 , -0.35627604, -0.6107163 ,  0.13403161,  0.92072487,\n",
       "         1.0786742 ,  1.1881384 , -0.664037  ,  0.4673002 ,  0.75259763,\n",
       "         1.2417803 ,  0.4733761 ,  0.3899353 ,  1.0345131 ,  1.8956127 ,\n",
       "        -0.6477682 ,  0.6526185 ,  1.8348951 ,  1.2910761 , -0.5317994 ,\n",
       "        -0.04197671,  0.5241989 ,  0.54381645], dtype=float32),\n",
       " array([-0.10488888, -0.34132475,  0.8030229 ,  0.5512628 , -0.62657756,\n",
       "         1.3288128 , -0.35628137, -0.6107338 ,  0.13404167,  0.9207107 ,\n",
       "         1.0786866 ,  1.1881528 , -0.6640321 ,  0.46731645,  0.752602  ,\n",
       "         1.2417642 ,  0.4733801 ,  0.38993064,  1.0345191 ,  1.8955843 ,\n",
       "        -0.6477676 ,  0.65262777,  1.8349022 ,  1.2910879 , -0.53179765,\n",
       "        -0.0419625 ,  0.52420187,  0.5438174 ], dtype=float32),\n",
       " array([-0.10488961, -0.34131807,  0.8030206 ,  0.5512566 , -0.6265786 ,\n",
       "         1.3288078 , -0.35628593, -0.6107506 ,  0.13405089,  0.9206978 ,\n",
       "         1.0786985 ,  1.1881665 , -0.66402745,  0.4673299 ,  0.752606  ,\n",
       "         1.2417494 ,  0.4733839 ,  0.389926  ,  1.0345246 ,  1.8955579 ,\n",
       "        -0.64776725,  0.652637  ,  1.8349086 ,  1.2911005 , -0.53179586,\n",
       "        -0.04194982,  0.5242046 ,  0.5438181 ], dtype=float32),\n",
       " array([-0.10488989, -0.34131193,  0.803019  ,  0.55125064, -0.62657934,\n",
       "         1.3288031 , -0.35628968, -0.61076665,  0.13405946,  0.920686  ,\n",
       "         1.0787103 ,  1.1881797 , -0.6640235 ,  0.46734047,  0.75260955,\n",
       "         1.2417363 ,  0.4733876 ,  0.3899216 ,  1.0345292 ,  1.895534  ,\n",
       "        -0.64776736,  0.6526459 ,  1.8349149 ,  1.2911131 , -0.531794  ,\n",
       "        -0.04193842,  0.52420706,  0.5438189 ], dtype=float32),\n",
       " array([-0.10488995, -0.34130642,  0.8030182 ,  0.55124533, -0.6265798 ,\n",
       "         1.3287988 , -0.35629275, -0.610782  ,  0.13406734,  0.9206754 ,\n",
       "         1.0787215 ,  1.188192  , -0.66401994,  0.46734884,  0.7526128 ,\n",
       "         1.2417251 ,  0.4733913 ,  0.38991752,  1.0345335 ,  1.8955117 ,\n",
       "        -0.64776766,  0.65265435,  1.8349205 ,  1.291126  , -0.53179234,\n",
       "        -0.04192843,  0.5242095 ,  0.5438196 ], dtype=float32),\n",
       " array([-0.10488965, -0.34130156,  0.8030178 ,  0.55124027, -0.62658   ,\n",
       "         1.3287952 , -0.35629508, -0.61079687,  0.13407463,  0.92066586,\n",
       "         1.0787323 ,  1.1882036 , -0.66401696,  0.46735498,  0.75261563,\n",
       "         1.2417151 ,  0.47339475,  0.38991356,  1.0345372 ,  1.8954915 ,\n",
       "        -0.6477683 ,  0.6526627 ,  1.8349262 ,  1.2911394 , -0.5317907 ,\n",
       "        -0.04191955,  0.52421165,  0.5438203 ], dtype=float32),\n",
       " array([-0.10488909, -0.34129724,  0.8030182 ,  0.55123556, -0.62658   ,\n",
       "         1.3287915 , -0.35629687, -0.61081094,  0.13408105,  0.9206569 ,\n",
       "         1.0787429 ,  1.1882148 , -0.6640145 ,  0.46735922,  0.7526181 ,\n",
       "         1.2417063 ,  0.4733981 ,  0.38990977,  1.0345407 ,  1.8954729 ,\n",
       "        -0.6477692 ,  0.6526708 ,  1.8349311 ,  1.2911525 , -0.5317892 ,\n",
       "        -0.04191169,  0.52421385,  0.54382086], dtype=float32),\n",
       " array([-0.10488822, -0.34129336,  0.80301875,  0.5512315 , -0.6265797 ,\n",
       "         1.3287886 , -0.35629818, -0.6108243 ,  0.13408709,  0.92064905,\n",
       "         1.078753  ,  1.1882254 , -0.6640122 ,  0.4673616 ,  0.75262064,\n",
       "         1.2416986 ,  0.4734015 ,  0.38990623,  1.0345436 ,  1.8954562 ,\n",
       "        -0.64777035,  0.6526785 ,  1.834936  ,  1.2911661 , -0.5317877 ,\n",
       "        -0.04190517,  0.5242156 ,  0.54382163], dtype=float32),\n",
       " array([-0.10488721, -0.34129006,  0.80301976,  0.5512275 , -0.62657934,\n",
       "         1.328786  , -0.35629898, -0.6108371 ,  0.13409257,  0.9206419 ,\n",
       "         1.0787628 ,  1.1882354 , -0.6640104 ,  0.46736255,  0.7526225 ,\n",
       "         1.2416922 ,  0.47340465,  0.3899028 ,  1.0345461 ,  1.8954407 ,\n",
       "        -0.64777166,  0.6526859 ,  1.8349402 ,  1.2911794 , -0.5317864 ,\n",
       "        -0.04189949,  0.52421737,  0.5438221 ], dtype=float32),\n",
       " array([-0.10488611, -0.34128708,  0.8030213 ,  0.5512238 , -0.62657887,\n",
       "         1.3287836 , -0.35629937, -0.6108494 ,  0.13409753,  0.9206356 ,\n",
       "         1.0787723 ,  1.1882447 , -0.664009  ,  0.46736202,  0.75262415,\n",
       "         1.2416868 ,  0.47340763,  0.3898996 ,  1.0345482 ,  1.895427  ,\n",
       "        -0.6477732 ,  0.652693  ,  1.8349444 ,  1.2911928 , -0.531785  ,\n",
       "        -0.04189485,  0.5242189 ,  0.54382265], dtype=float32),\n",
       " array([-0.10488474, -0.3412845 ,  0.8030232 ,  0.5512204 , -0.62657815,\n",
       "         1.3287815 , -0.35629937, -0.6108612 ,  0.13410197,  0.92062986,\n",
       "         1.0787814 ,  1.1882535 , -0.6640078 ,  0.4673598 ,  0.75262576,\n",
       "         1.2416825 ,  0.47341067,  0.38989648,  1.0345501 ,  1.8954144 ,\n",
       "        -0.64777476,  0.65269995,  1.8349487 ,  1.2912062 , -0.5317837 ,\n",
       "        -0.04189078,  0.5242204 ,  0.54382324], dtype=float32),\n",
       " array([-0.10488322, -0.3412823 ,  0.80302525,  0.5512173 , -0.6265773 ,\n",
       "         1.3287796 , -0.35629895, -0.6108724 ,  0.13410603,  0.92062485,\n",
       "         1.07879   ,  1.1882619 , -0.6640071 ,  0.46735683,  0.75262713,\n",
       "         1.2416788 ,  0.47341335,  0.3898934 ,  1.0345517 ,  1.8954029 ,\n",
       "        -0.6477767 ,  0.65270656,  1.8349521 ,  1.2912191 , -0.5317825 ,\n",
       "        -0.04188764,  0.5242217 ,  0.5438236 ], dtype=float32),\n",
       " array([-0.10488164, -0.34128055,  0.80302763,  0.55121446, -0.6265765 ,\n",
       "         1.328778  , -0.3562983 , -0.61088294,  0.13410974,  0.9206207 ,\n",
       "         1.0787987 ,  1.1882699 , -0.6640066 ,  0.46735284,  0.7526282 ,\n",
       "         1.241676  ,  0.47341615,  0.3898907 ,  1.034553  ,  1.8953928 ,\n",
       "        -0.6477785 ,  0.65271294,  1.8349558 ,  1.2912322 , -0.5317814 ,\n",
       "        -0.04188506,  0.52422297,  0.543824  ], dtype=float32),\n",
       " array([-0.10487998, -0.34127906,  0.8030302 ,  0.5512119 , -0.6265754 ,\n",
       "         1.3287764 , -0.35629734, -0.61089313,  0.13411295,  0.92061675,\n",
       "         1.0788068 ,  1.1882776 , -0.66400594,  0.46734777,  0.75262916,\n",
       "         1.2416738 ,  0.47341877,  0.38988805,  1.034554  ,  1.895384  ,\n",
       "        -0.6477806 ,  0.65271914,  1.8349596 ,  1.291245  , -0.53178036,\n",
       "        -0.04188305,  0.524224  ,  0.54382455], dtype=float32),\n",
       " array([-0.10487818, -0.3412779 ,  0.803033  ,  0.5512094 , -0.62657434,\n",
       "         1.3287752 , -0.35629615, -0.6109026 ,  0.13411584,  0.9206133 ,\n",
       "         1.0788146 ,  1.1882848 , -0.66400576,  0.46734193,  0.7526298 ,\n",
       "         1.2416725 ,  0.47342122,  0.38988551,  1.034555  ,  1.8953757 ,\n",
       "        -0.64778274,  0.6527251 ,  1.8349622 ,  1.2912575 , -0.5317796 ,\n",
       "        -0.04188157,  0.5242251 ,  0.5438249 ], dtype=float32),\n",
       " array([-0.10487632, -0.34127694,  0.80303586,  0.5512073 , -0.62657315,\n",
       "         1.3287743 , -0.35629478, -0.6109117 ,  0.13411847,  0.92061055,\n",
       "         1.078822  ,  1.1882917 , -0.6640058 ,  0.4673355 ,  0.7526304 ,\n",
       "         1.2416714 ,  0.47342354,  0.38988316,  1.0345554 ,  1.8953687 ,\n",
       "        -0.6477847 ,  0.6527308 ,  1.8349651 ,  1.29127   , -0.53177863,\n",
       "        -0.04188078,  0.52422595,  0.5438253 ], dtype=float32),\n",
       " array([-0.10487444, -0.34127626,  0.80303895,  0.55120534, -0.6265719 ,\n",
       "         1.3287737 , -0.3562931 , -0.6109204 ,  0.13412082,  0.92060804,\n",
       "         1.0788293 ,  1.1882979 , -0.664006  ,  0.46732855,  0.7526309 ,\n",
       "         1.2416711 ,  0.47342587,  0.3898811 ,  1.034556  ,  1.8953626 ,\n",
       "        -0.6477869 ,  0.65273625,  1.834968  ,  1.2912822 , -0.5317779 ,\n",
       "        -0.04188036,  0.52422684,  0.5438257 ], dtype=float32),\n",
       " array([-0.10487255, -0.34127566,  0.8030424 ,  0.55120355, -0.6265706 ,\n",
       "         1.3287728 , -0.35629135, -0.61092854,  0.13412282,  0.92060614,\n",
       "         1.0788361 ,  1.1883038 , -0.66400623,  0.46732116,  0.75263137,\n",
       "         1.2416712 ,  0.47342807,  0.3898791 ,  1.0345562 ,  1.8953569 ,\n",
       "        -0.647789  ,  0.6527412 ,  1.8349701 ,  1.2912941 , -0.5317771 ,\n",
       "        -0.04188015,  0.52422756,  0.54382604], dtype=float32),\n",
       " array([-0.10487058, -0.3412753 ,  0.8030456 ,  0.5512019 , -0.6265694 ,\n",
       "         1.3287722 , -0.3562894 , -0.6109364 ,  0.13412455,  0.92060447,\n",
       "         1.0788424 ,  1.1883094 , -0.6640065 ,  0.46731314,  0.75263155,\n",
       "         1.2416718 ,  0.47343028,  0.38987708,  1.0345564 ,  1.8953519 ,\n",
       "        -0.6477912 ,  0.6527462 ,  1.8349725 ,  1.2913059 , -0.53177655,\n",
       "        -0.04188045,  0.5242283 ,  0.5438263 ], dtype=float32),\n",
       " array([-0.10486849, -0.34127522,  0.80304897,  0.5512005 , -0.626568  ,\n",
       "         1.3287722 , -0.3562874 , -0.61094356,  0.134126  ,  0.9206033 ,\n",
       "         1.078849  ,  1.1883148 , -0.6640071 ,  0.4673047 ,  0.7526318 ,\n",
       "         1.2416729 ,  0.47343218,  0.38987535,  1.0345565 ,  1.8953482 ,\n",
       "        -0.64779335,  0.6527508 ,  1.8349745 ,  1.2913172 , -0.5317758 ,\n",
       "        -0.04188103,  0.5242289 ,  0.54382664], dtype=float32),\n",
       " array([-0.10486642, -0.3412752 ,  0.80305225,  0.55119926, -0.6265665 ,\n",
       "         1.328772  , -0.35628513, -0.61095023,  0.1341274 ,  0.9206022 ,\n",
       "         1.0788549 ,  1.1883197 , -0.6640077 ,  0.46729594,  0.75263184,\n",
       "         1.2416742 ,  0.47343397,  0.38987368,  1.0345564 ,  1.8953449 ,\n",
       "        -0.6477956 ,  0.6527552 ,  1.8349767 ,  1.2913281 , -0.531775  ,\n",
       "        -0.04188202,  0.5242294 ,  0.54382694], dtype=float32),\n",
       " array([-0.10486443, -0.34127536,  0.80305564,  0.5511981 , -0.62656516,\n",
       "         1.3287717 , -0.35628283, -0.6109568 ,  0.13412845,  0.9206015 ,\n",
       "         1.0788606 ,  1.1883242 , -0.66400844,  0.46728703,  0.7526318 ,\n",
       "         1.2416761 ,  0.47343582,  0.38987204,  1.0345558 ,  1.8953421 ,\n",
       "        -0.6477978 ,  0.6527595 ,  1.8349785 ,  1.2913387 , -0.5317743 ,\n",
       "        -0.04188332,  0.52422976,  0.54382724], dtype=float32),\n",
       " array([-0.1048624 , -0.34127575,  0.8030593 ,  0.5511971 , -0.6265637 ,\n",
       "         1.328772  , -0.3562806 , -0.6109626 ,  0.13412938,  0.920601  ,\n",
       "         1.0788658 ,  1.1883284 , -0.66400915,  0.46727803,  0.75263184,\n",
       "         1.241678  ,  0.4734376 ,  0.38987052,  1.0345556 ,  1.8953397 ,\n",
       "        -0.64779997,  0.65276337,  1.8349802 ,  1.2913488 , -0.5317737 ,\n",
       "        -0.04188469,  0.52423024,  0.5438275 ], dtype=float32),\n",
       " array([-0.10486046, -0.34127614,  0.8030627 ,  0.55119616, -0.6265624 ,\n",
       "         1.3287718 , -0.3562782 , -0.6109686 ,  0.13413005,  0.9206008 ,\n",
       "         1.0788708 ,  1.1883323 , -0.6640101 ,  0.4672686 ,  0.75263166,\n",
       "         1.2416798 ,  0.4734391 ,  0.38986892,  1.0345551 ,  1.8953377 ,\n",
       "        -0.64780223,  0.65276706,  1.8349817 ,  1.2913587 , -0.5317732 ,\n",
       "        -0.04188643,  0.5242308 ,  0.5438276 ], dtype=float32),\n",
       " array([-0.10485845, -0.3412767 ,  0.80306613,  0.55119544, -0.62656075,\n",
       "         1.328772  , -0.35627577, -0.61097383,  0.13413067,  0.9206008 ,\n",
       "         1.0788757 ,  1.188336  , -0.66401106,  0.4672591 ,  0.75263137,\n",
       "         1.241682  ,  0.4734406 ,  0.38986766,  1.0345544 ,  1.895336  ,\n",
       "        -0.6478043 ,  0.6527707 ,  1.834983  ,  1.2913681 , -0.5317728 ,\n",
       "        -0.04188846,  0.52423096,  0.54382783], dtype=float32),\n",
       " array([-0.10485645, -0.34127727,  0.8030694 ,  0.5511947 , -0.6265594 ,\n",
       "         1.3287722 , -0.3562734 , -0.6109788 ,  0.13413104,  0.9206008 ,\n",
       "         1.0788801 ,  1.1883395 , -0.6640121 ,  0.4672497 ,  0.75263107,\n",
       "         1.2416846 ,  0.47344208,  0.38986635,  1.0345538 ,  1.8953353 ,\n",
       "        -0.64780647,  0.65277416,  1.8349844 ,  1.2913774 , -0.5317724 ,\n",
       "        -0.04189041,  0.52423126,  0.54382807], dtype=float32),\n",
       " array([-0.1048544 , -0.341278  ,  0.80307287,  0.55119413, -0.626558  ,\n",
       "         1.3287723 , -0.35627106, -0.61098343,  0.1341313 ,  0.92060125,\n",
       "         1.0788847 ,  1.1883427 , -0.6640132 ,  0.46724048,  0.7526307 ,\n",
       "         1.2416872 ,  0.4734434 ,  0.38986516,  1.0345533 ,  1.8953348 ,\n",
       "        -0.6478086 ,  0.65277743,  1.8349855 ,  1.2913865 , -0.5317719 ,\n",
       "        -0.04189266,  0.5242314 ,  0.54382825], dtype=float32),\n",
       " array([-0.10485253, -0.34127873,  0.80307627,  0.55119365, -0.6265567 ,\n",
       "         1.3287728 , -0.35626867, -0.61098796,  0.13413157,  0.9206017 ,\n",
       "         1.0788888 ,  1.1883457 , -0.66401416,  0.46723118,  0.7526305 ,\n",
       "         1.2416899 ,  0.47344464,  0.38986403,  1.0345525 ,  1.8953347 ,\n",
       "        -0.6478105 ,  0.6527804 ,  1.8349866 ,  1.2913951 , -0.53177154,\n",
       "        -0.04189499,  0.5242316 ,  0.5438284 ], dtype=float32),\n",
       " array([-0.10485073, -0.34127957,  0.80307955,  0.5511933 , -0.62655526,\n",
       "         1.3287729 , -0.35626623, -0.6109921 ,  0.1341317 ,  0.92060256,\n",
       "         1.0788924 ,  1.1883484 , -0.66401535,  0.4672222 ,  0.75263005,\n",
       "         1.2416927 ,  0.4734459 ,  0.38986304,  1.0345519 ,  1.8953345 ,\n",
       "        -0.6478124 ,  0.6527832 ,  1.8349875 ,  1.2914035 , -0.53177124,\n",
       "        -0.04189725,  0.5242319 ,  0.54382855], dtype=float32),\n",
       " array([-0.10484894, -0.34128037,  0.8030827 ,  0.5511928 , -0.626554  ,\n",
       "         1.3287736 , -0.3562638 , -0.610996  ,  0.13413163,  0.92060316,\n",
       "         1.078896  ,  1.1883512 , -0.66401654,  0.46721306,  0.7526297 ,\n",
       "         1.2416958 ,  0.47344708,  0.38986218,  1.0345509 ,  1.8953348 ,\n",
       "        -0.6478144 ,  0.65278596,  1.8349885 ,  1.2914114 , -0.5317708 ,\n",
       "        -0.04189959,  0.524232  ,  0.54382867], dtype=float32),\n",
       " array([-0.10484717, -0.34128132,  0.8030859 ,  0.55119264, -0.62655276,\n",
       "         1.3287739 , -0.35626143, -0.6109997 ,  0.13413152,  0.9206039 ,\n",
       "         1.0788997 ,  1.1883535 , -0.66401756,  0.46720383,  0.7526292 ,\n",
       "         1.2416989 ,  0.47344822,  0.3898614 ,  1.0345503 ,  1.8953352 ,\n",
       "        -0.64781624,  0.6527884 ,  1.8349895 ,  1.2914193 , -0.5317707 ,\n",
       "        -0.04190202,  0.5242321 ,  0.54382885], dtype=float32),\n",
       " array([-0.10484537, -0.34128228,  0.8030889 ,  0.55119246, -0.62655145,\n",
       "         1.3287742 , -0.35625905, -0.6110031 ,  0.13413134,  0.9206048 ,\n",
       "         1.0789031 ,  1.1883556 , -0.66401863,  0.46719506,  0.7526287 ,\n",
       "         1.2417016 ,  0.47344917,  0.3898605 ,  1.0345494 ,  1.8953359 ,\n",
       "        -0.64781827,  0.6527909 ,  1.8349903 ,  1.2914265 , -0.5317703 ,\n",
       "        -0.04190441,  0.5242322 ,  0.54382896], dtype=float32),\n",
       " array([-0.10484361, -0.34128317,  0.80309194,  0.55119216, -0.62655026,\n",
       "         1.3287749 , -0.3562568 , -0.6110064 ,  0.13413107,  0.9206058 ,\n",
       "         1.0789062 ,  1.1883575 , -0.6640197 ,  0.46718636,  0.7526282 ,\n",
       "         1.2417047 ,  0.47345006,  0.38985983,  1.0345486 ,  1.8953367 ,\n",
       "        -0.64782006,  0.6527932 ,  1.8349909 ,  1.2914332 , -0.53177   ,\n",
       "        -0.04190702,  0.52423227,  0.543829  ], dtype=float32),\n",
       " array([-0.10484204, -0.3412842 ,  0.8030948 ,  0.551192  , -0.6265491 ,\n",
       "         1.3287754 , -0.3562545 , -0.61100936,  0.13413078,  0.92060673,\n",
       "         1.078909  ,  1.1883594 , -0.6640209 ,  0.46717778,  0.75262773,\n",
       "         1.2417077 ,  0.47345102,  0.389859  ,  1.0345476 ,  1.8953378 ,\n",
       "        -0.6478219 ,  0.6527955 ,  1.8349916 ,  1.29144   , -0.5317697 ,\n",
       "        -0.04190951,  0.52423215,  0.54382926], dtype=float32),\n",
       " array([-0.10484046, -0.34128514,  0.8030974 ,  0.5511918 , -0.6265479 ,\n",
       "         1.328776  , -0.35625222, -0.61101204,  0.13413043,  0.92060757,\n",
       "         1.0789119 ,  1.1883612 , -0.6640222 ,  0.46716955,  0.752627  ,\n",
       "         1.2417108 ,  0.4734518 ,  0.3898583 ,  1.0345467 ,  1.895339  ,\n",
       "        -0.64782363,  0.6527975 ,  1.834992  ,  1.2914463 , -0.53176934,\n",
       "        -0.04191218,  0.5242321 ,  0.54382926], dtype=float32),\n",
       " array([-0.104839  , -0.34128612,  0.80310017,  0.5511918 , -0.62654686,\n",
       "         1.3287764 , -0.35625002, -0.6110147 ,  0.13413005,  0.9206089 ,\n",
       "         1.0789143 ,  1.1883626 , -0.6640232 ,  0.46716148,  0.75262654,\n",
       "         1.2417136 ,  0.4734525 ,  0.38985765,  1.0345457 ,  1.8953403 ,\n",
       "        -0.64782536,  0.6527995 ,  1.8349926 ,  1.2914524 , -0.5317693 ,\n",
       "        -0.04191484,  0.5242321 ,  0.5438294 ], dtype=float32),\n",
       " array([-0.10483745, -0.34128714,  0.80310285,  0.55119175, -0.62654585,\n",
       "         1.3287771 , -0.35624796, -0.6110171 ,  0.13412951,  0.92060995,\n",
       "         1.0789167 ,  1.1883643 , -0.6640243 ,  0.46715376,  0.75262594,\n",
       "         1.2417167 ,  0.47345328,  0.38985717,  1.034545  ,  1.895342  ,\n",
       "        -0.6478269 ,  0.6528014 ,  1.8349931 ,  1.2914582 , -0.53176934,\n",
       "        -0.04191726,  0.52423227,  0.5438294 ], dtype=float32),\n",
       " array([-0.10483607, -0.34128818,  0.80310553,  0.55119187, -0.6265447 ,\n",
       "         1.3287776 , -0.35624585, -0.61101925,  0.1341292 ,  0.920611  ,\n",
       "         1.078919  ,  1.1883656 , -0.6640252 ,  0.46714625,  0.75262547,\n",
       "         1.2417196 ,  0.473454  ,  0.3898568 ,  1.0345442 ,  1.8953434 ,\n",
       "        -0.6478283 ,  0.65280306,  1.8349936 ,  1.291464  , -0.5317693 ,\n",
       "        -0.0419196 ,  0.5242321 ,  0.54382956], dtype=float32),\n",
       " array([-0.10483467, -0.3412892 ,  0.80310816,  0.55119187, -0.62654376,\n",
       "         1.3287779 , -0.35624385, -0.6110216 ,  0.13412862,  0.92061234,\n",
       "         1.0789212 ,  1.1883669 , -0.66402626,  0.4671387 ,  0.75262487,\n",
       "         1.2417226 ,  0.4734547 ,  0.38985628,  1.0345432 ,  1.8953451 ,\n",
       "        -0.6478297 ,  0.65280455,  1.8349938 ,  1.291469  , -0.53176916,\n",
       "        -0.04192197,  0.5242321 ,  0.5438295 ], dtype=float32),\n",
       " array([-0.10483333, -0.34129018,  0.8031106 ,  0.55119205, -0.6265426 ,\n",
       "         1.3287787 , -0.35624197, -0.6110234 ,  0.13412815,  0.92061365,\n",
       "         1.0789231 ,  1.1883682 , -0.6640273 ,  0.46713173,  0.7526245 ,\n",
       "         1.2417256 ,  0.47345567,  0.38985604,  1.0345426 ,  1.8953466 ,\n",
       "        -0.647831  ,  0.6528061 ,  1.8349944 ,  1.2914741 , -0.5317691 ,\n",
       "        -0.04192422,  0.52423203,  0.5438297 ], dtype=float32)]"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.19480519, 0.        , 0.70017036, ..., 0.53372781, 0.76305783,\n",
       "        0.76305783],\n",
       "       [0.37662338, 0.        , 0.70017036, ..., 0.52840237, 0.62337402,\n",
       "        0.62337402],\n",
       "       [0.12987013, 0.        , 0.70017036, ..., 0.52366864, 0.08785271,\n",
       "        0.08785271],\n",
       "       ...,\n",
       "       [0.32467532, 0.33333333, 0.87052811, ..., 0.86804734, 0.58415049,\n",
       "        0.58415049],\n",
       "       [0.51948052, 0.33333333, 0.87052811, ..., 0.86568047, 0.12627577,\n",
       "        0.12627577],\n",
       "       [0.53246753, 0.33333333, 0.87052811, ..., 0.86390533, 0.68260957,\n",
       "        0.68260957]])"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scaled_test"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Inverse Transformations and Compare"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "true_predictions = scaler.inverse_transform(test_predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[117.5287655 ,   2.69890174,  24.6111016 , ...,   6.71389257,\n",
       "         25.28925538,  25.35917734],\n",
       "       [119.84453008,   4.14405182,  24.6374699 , ...,   6.24127183,\n",
       "         25.29366944,  25.84543217],\n",
       "       [119.6901603 ,   5.02690271,  24.6455654 , ...,   5.65781482,\n",
       "         25.45417779,  26.09424841],\n",
       "       ...,\n",
       "       [-60.72377317, -10.23864537,  25.10422948, ...,  -2.10844118,\n",
       "         26.20587751,  27.18516313],\n",
       "       [-60.72269462, -10.23867577,  25.10424487, ...,  -2.10848122,\n",
       "         26.20587751,  27.18516015],\n",
       "       [-60.72166197, -10.23870528,  25.10425922, ...,  -2.10851925,\n",
       "         26.20587453,  27.18516908]])"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "true_predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Appliances</th>\n",
       "      <th>lights</th>\n",
       "      <th>T1</th>\n",
       "      <th>RH_1</th>\n",
       "      <th>T2</th>\n",
       "      <th>RH_2</th>\n",
       "      <th>T3</th>\n",
       "      <th>RH_3</th>\n",
       "      <th>T4</th>\n",
       "      <th>RH_4</th>\n",
       "      <th>...</th>\n",
       "      <th>T9</th>\n",
       "      <th>RH_9</th>\n",
       "      <th>T_out</th>\n",
       "      <th>Press_mm_hg</th>\n",
       "      <th>RH_out</th>\n",
       "      <th>Windspeed</th>\n",
       "      <th>Visibility</th>\n",
       "      <th>Tdewpoint</th>\n",
       "      <th>rv1</th>\n",
       "      <th>rv2</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2016-05-25 18:10:00</th>\n",
       "      <td>170</td>\n",
       "      <td>0</td>\n",
       "      <td>24.50</td>\n",
       "      <td>37.86</td>\n",
       "      <td>24.00</td>\n",
       "      <td>34.27</td>\n",
       "      <td>25.00</td>\n",
       "      <td>36.70</td>\n",
       "      <td>24.29</td>\n",
       "      <td>37.79</td>\n",
       "      <td>...</td>\n",
       "      <td>22.00</td>\n",
       "      <td>37.23</td>\n",
       "      <td>16.48</td>\n",
       "      <td>756.1</td>\n",
       "      <td>55.83</td>\n",
       "      <td>2.00</td>\n",
       "      <td>38.17</td>\n",
       "      <td>7.62</td>\n",
       "      <td>38.14</td>\n",
       "      <td>38.14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-25 18:20:00</th>\n",
       "      <td>310</td>\n",
       "      <td>0</td>\n",
       "      <td>24.50</td>\n",
       "      <td>37.30</td>\n",
       "      <td>23.86</td>\n",
       "      <td>34.33</td>\n",
       "      <td>24.94</td>\n",
       "      <td>36.67</td>\n",
       "      <td>24.29</td>\n",
       "      <td>37.79</td>\n",
       "      <td>...</td>\n",
       "      <td>22.00</td>\n",
       "      <td>37.36</td>\n",
       "      <td>16.47</td>\n",
       "      <td>756.1</td>\n",
       "      <td>55.67</td>\n",
       "      <td>2.00</td>\n",
       "      <td>36.33</td>\n",
       "      <td>7.53</td>\n",
       "      <td>31.16</td>\n",
       "      <td>31.16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-25 18:30:00</th>\n",
       "      <td>120</td>\n",
       "      <td>0</td>\n",
       "      <td>24.50</td>\n",
       "      <td>36.96</td>\n",
       "      <td>23.73</td>\n",
       "      <td>34.33</td>\n",
       "      <td>24.85</td>\n",
       "      <td>36.50</td>\n",
       "      <td>24.29</td>\n",
       "      <td>37.79</td>\n",
       "      <td>...</td>\n",
       "      <td>22.03</td>\n",
       "      <td>37.39</td>\n",
       "      <td>16.45</td>\n",
       "      <td>756.1</td>\n",
       "      <td>55.50</td>\n",
       "      <td>2.00</td>\n",
       "      <td>34.50</td>\n",
       "      <td>7.45</td>\n",
       "      <td>4.40</td>\n",
       "      <td>4.40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-25 18:40:00</th>\n",
       "      <td>120</td>\n",
       "      <td>0</td>\n",
       "      <td>24.50</td>\n",
       "      <td>37.00</td>\n",
       "      <td>23.70</td>\n",
       "      <td>34.40</td>\n",
       "      <td>24.84</td>\n",
       "      <td>36.45</td>\n",
       "      <td>24.29</td>\n",
       "      <td>37.90</td>\n",
       "      <td>...</td>\n",
       "      <td>22.10</td>\n",
       "      <td>37.72</td>\n",
       "      <td>16.43</td>\n",
       "      <td>756.1</td>\n",
       "      <td>55.33</td>\n",
       "      <td>2.00</td>\n",
       "      <td>32.67</td>\n",
       "      <td>7.37</td>\n",
       "      <td>27.12</td>\n",
       "      <td>27.12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-25 18:50:00</th>\n",
       "      <td>120</td>\n",
       "      <td>0</td>\n",
       "      <td>24.49</td>\n",
       "      <td>37.07</td>\n",
       "      <td>23.68</td>\n",
       "      <td>34.52</td>\n",
       "      <td>24.84</td>\n",
       "      <td>36.49</td>\n",
       "      <td>24.28</td>\n",
       "      <td>37.93</td>\n",
       "      <td>...</td>\n",
       "      <td>22.10</td>\n",
       "      <td>37.81</td>\n",
       "      <td>16.42</td>\n",
       "      <td>756.1</td>\n",
       "      <td>55.17</td>\n",
       "      <td>2.00</td>\n",
       "      <td>30.83</td>\n",
       "      <td>7.28</td>\n",
       "      <td>10.27</td>\n",
       "      <td>10.27</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:20:00</th>\n",
       "      <td>100</td>\n",
       "      <td>0</td>\n",
       "      <td>25.57</td>\n",
       "      <td>46.56</td>\n",
       "      <td>25.89</td>\n",
       "      <td>42.03</td>\n",
       "      <td>27.20</td>\n",
       "      <td>41.16</td>\n",
       "      <td>24.70</td>\n",
       "      <td>45.59</td>\n",
       "      <td>...</td>\n",
       "      <td>23.20</td>\n",
       "      <td>46.79</td>\n",
       "      <td>22.73</td>\n",
       "      <td>755.2</td>\n",
       "      <td>55.67</td>\n",
       "      <td>3.33</td>\n",
       "      <td>23.67</td>\n",
       "      <td>13.33</td>\n",
       "      <td>43.10</td>\n",
       "      <td>43.10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:30:00</th>\n",
       "      <td>90</td>\n",
       "      <td>0</td>\n",
       "      <td>25.50</td>\n",
       "      <td>46.50</td>\n",
       "      <td>25.75</td>\n",
       "      <td>42.08</td>\n",
       "      <td>27.13</td>\n",
       "      <td>41.22</td>\n",
       "      <td>24.70</td>\n",
       "      <td>45.59</td>\n",
       "      <td>...</td>\n",
       "      <td>23.20</td>\n",
       "      <td>46.79</td>\n",
       "      <td>22.60</td>\n",
       "      <td>755.2</td>\n",
       "      <td>56.00</td>\n",
       "      <td>3.50</td>\n",
       "      <td>24.50</td>\n",
       "      <td>13.30</td>\n",
       "      <td>49.28</td>\n",
       "      <td>49.28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:40:00</th>\n",
       "      <td>270</td>\n",
       "      <td>10</td>\n",
       "      <td>25.50</td>\n",
       "      <td>46.60</td>\n",
       "      <td>25.63</td>\n",
       "      <td>42.77</td>\n",
       "      <td>27.05</td>\n",
       "      <td>41.69</td>\n",
       "      <td>24.70</td>\n",
       "      <td>45.73</td>\n",
       "      <td>...</td>\n",
       "      <td>23.20</td>\n",
       "      <td>46.79</td>\n",
       "      <td>22.47</td>\n",
       "      <td>755.2</td>\n",
       "      <td>56.33</td>\n",
       "      <td>3.67</td>\n",
       "      <td>25.33</td>\n",
       "      <td>13.27</td>\n",
       "      <td>29.20</td>\n",
       "      <td>29.20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 17:50:00</th>\n",
       "      <td>420</td>\n",
       "      <td>10</td>\n",
       "      <td>25.50</td>\n",
       "      <td>46.99</td>\n",
       "      <td>25.41</td>\n",
       "      <td>43.04</td>\n",
       "      <td>26.89</td>\n",
       "      <td>41.29</td>\n",
       "      <td>24.70</td>\n",
       "      <td>45.79</td>\n",
       "      <td>...</td>\n",
       "      <td>23.20</td>\n",
       "      <td>46.82</td>\n",
       "      <td>22.33</td>\n",
       "      <td>755.2</td>\n",
       "      <td>56.67</td>\n",
       "      <td>3.83</td>\n",
       "      <td>26.17</td>\n",
       "      <td>13.23</td>\n",
       "      <td>6.32</td>\n",
       "      <td>6.32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05-27 18:00:00</th>\n",
       "      <td>430</td>\n",
       "      <td>10</td>\n",
       "      <td>25.50</td>\n",
       "      <td>46.60</td>\n",
       "      <td>25.26</td>\n",
       "      <td>42.97</td>\n",
       "      <td>26.82</td>\n",
       "      <td>41.16</td>\n",
       "      <td>24.70</td>\n",
       "      <td>45.96</td>\n",
       "      <td>...</td>\n",
       "      <td>23.20</td>\n",
       "      <td>46.84</td>\n",
       "      <td>22.20</td>\n",
       "      <td>755.2</td>\n",
       "      <td>57.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>27.00</td>\n",
       "      <td>13.20</td>\n",
       "      <td>34.12</td>\n",
       "      <td>34.12</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>288 rows × 28 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Appliances  lights     T1   RH_1     T2   RH_2     T3  \\\n",
       "date                                                                         \n",
       "2016-05-25 18:10:00         170       0  24.50  37.86  24.00  34.27  25.00   \n",
       "2016-05-25 18:20:00         310       0  24.50  37.30  23.86  34.33  24.94   \n",
       "2016-05-25 18:30:00         120       0  24.50  36.96  23.73  34.33  24.85   \n",
       "2016-05-25 18:40:00         120       0  24.50  37.00  23.70  34.40  24.84   \n",
       "2016-05-25 18:50:00         120       0  24.49  37.07  23.68  34.52  24.84   \n",
       "...                         ...     ...    ...    ...    ...    ...    ...   \n",
       "2016-05-27 17:20:00         100       0  25.57  46.56  25.89  42.03  27.20   \n",
       "2016-05-27 17:30:00          90       0  25.50  46.50  25.75  42.08  27.13   \n",
       "2016-05-27 17:40:00         270      10  25.50  46.60  25.63  42.77  27.05   \n",
       "2016-05-27 17:50:00         420      10  25.50  46.99  25.41  43.04  26.89   \n",
       "2016-05-27 18:00:00         430      10  25.50  46.60  25.26  42.97  26.82   \n",
       "\n",
       "                      RH_3     T4   RH_4  ...     T9   RH_9  T_out  \\\n",
       "date                                      ...                        \n",
       "2016-05-25 18:10:00  36.70  24.29  37.79  ...  22.00  37.23  16.48   \n",
       "2016-05-25 18:20:00  36.67  24.29  37.79  ...  22.00  37.36  16.47   \n",
       "2016-05-25 18:30:00  36.50  24.29  37.79  ...  22.03  37.39  16.45   \n",
       "2016-05-25 18:40:00  36.45  24.29  37.90  ...  22.10  37.72  16.43   \n",
       "2016-05-25 18:50:00  36.49  24.28  37.93  ...  22.10  37.81  16.42   \n",
       "...                    ...    ...    ...  ...    ...    ...    ...   \n",
       "2016-05-27 17:20:00  41.16  24.70  45.59  ...  23.20  46.79  22.73   \n",
       "2016-05-27 17:30:00  41.22  24.70  45.59  ...  23.20  46.79  22.60   \n",
       "2016-05-27 17:40:00  41.69  24.70  45.73  ...  23.20  46.79  22.47   \n",
       "2016-05-27 17:50:00  41.29  24.70  45.79  ...  23.20  46.82  22.33   \n",
       "2016-05-27 18:00:00  41.16  24.70  45.96  ...  23.20  46.84  22.20   \n",
       "\n",
       "                     Press_mm_hg  RH_out  Windspeed  Visibility  Tdewpoint  \\\n",
       "date                                                                         \n",
       "2016-05-25 18:10:00        756.1   55.83       2.00       38.17       7.62   \n",
       "2016-05-25 18:20:00        756.1   55.67       2.00       36.33       7.53   \n",
       "2016-05-25 18:30:00        756.1   55.50       2.00       34.50       7.45   \n",
       "2016-05-25 18:40:00        756.1   55.33       2.00       32.67       7.37   \n",
       "2016-05-25 18:50:00        756.1   55.17       2.00       30.83       7.28   \n",
       "...                          ...     ...        ...         ...        ...   \n",
       "2016-05-27 17:20:00        755.2   55.67       3.33       23.67      13.33   \n",
       "2016-05-27 17:30:00        755.2   56.00       3.50       24.50      13.30   \n",
       "2016-05-27 17:40:00        755.2   56.33       3.67       25.33      13.27   \n",
       "2016-05-27 17:50:00        755.2   56.67       3.83       26.17      13.23   \n",
       "2016-05-27 18:00:00        755.2   57.00       4.00       27.00      13.20   \n",
       "\n",
       "                       rv1    rv2  \n",
       "date                               \n",
       "2016-05-25 18:10:00  38.14  38.14  \n",
       "2016-05-25 18:20:00  31.16  31.16  \n",
       "2016-05-25 18:30:00   4.40   4.40  \n",
       "2016-05-25 18:40:00  27.12  27.12  \n",
       "2016-05-25 18:50:00  10.27  10.27  \n",
       "...                    ...    ...  \n",
       "2016-05-27 17:20:00  43.10  43.10  \n",
       "2016-05-27 17:30:00  49.28  49.28  \n",
       "2016-05-27 17:40:00  29.20  29.20  \n",
       "2016-05-27 17:50:00   6.32   6.32  \n",
       "2016-05-27 18:00:00  34.12  34.12  \n",
       "\n",
       "[288 rows x 28 columns]"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "true_predictions = pd.DataFrame(data=true_predictions,columns=test.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Appliances</th>\n",
       "      <th>lights</th>\n",
       "      <th>T1</th>\n",
       "      <th>RH_1</th>\n",
       "      <th>T2</th>\n",
       "      <th>RH_2</th>\n",
       "      <th>T3</th>\n",
       "      <th>RH_3</th>\n",
       "      <th>T4</th>\n",
       "      <th>RH_4</th>\n",
       "      <th>...</th>\n",
       "      <th>T9</th>\n",
       "      <th>RH_9</th>\n",
       "      <th>T_out</th>\n",
       "      <th>Press_mm_hg</th>\n",
       "      <th>RH_out</th>\n",
       "      <th>Windspeed</th>\n",
       "      <th>Visibility</th>\n",
       "      <th>Tdewpoint</th>\n",
       "      <th>rv1</th>\n",
       "      <th>rv2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>117.528765</td>\n",
       "      <td>2.698902</td>\n",
       "      <td>24.611102</td>\n",
       "      <td>37.002893</td>\n",
       "      <td>24.227434</td>\n",
       "      <td>34.891202</td>\n",
       "      <td>25.151382</td>\n",
       "      <td>37.382300</td>\n",
       "      <td>24.284626</td>\n",
       "      <td>37.261892</td>\n",
       "      <td>...</td>\n",
       "      <td>22.027223</td>\n",
       "      <td>37.597626</td>\n",
       "      <td>16.297667</td>\n",
       "      <td>756.734189</td>\n",
       "      <td>53.896075</td>\n",
       "      <td>1.854903</td>\n",
       "      <td>40.105939</td>\n",
       "      <td>6.713893</td>\n",
       "      <td>25.289255</td>\n",
       "      <td>25.359177</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>119.844530</td>\n",
       "      <td>4.144052</td>\n",
       "      <td>24.637470</td>\n",
       "      <td>36.771019</td>\n",
       "      <td>24.281426</td>\n",
       "      <td>34.572933</td>\n",
       "      <td>25.145838</td>\n",
       "      <td>37.617218</td>\n",
       "      <td>24.297594</td>\n",
       "      <td>37.153201</td>\n",
       "      <td>...</td>\n",
       "      <td>21.988023</td>\n",
       "      <td>37.488395</td>\n",
       "      <td>16.108513</td>\n",
       "      <td>756.933806</td>\n",
       "      <td>52.064223</td>\n",
       "      <td>1.761732</td>\n",
       "      <td>39.215918</td>\n",
       "      <td>6.241272</td>\n",
       "      <td>25.293669</td>\n",
       "      <td>25.845432</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>119.690160</td>\n",
       "      <td>5.026903</td>\n",
       "      <td>24.645565</td>\n",
       "      <td>36.495787</td>\n",
       "      <td>24.344913</td>\n",
       "      <td>34.224538</td>\n",
       "      <td>25.124767</td>\n",
       "      <td>37.691677</td>\n",
       "      <td>24.305334</td>\n",
       "      <td>36.980809</td>\n",
       "      <td>...</td>\n",
       "      <td>21.939192</td>\n",
       "      <td>37.378610</td>\n",
       "      <td>15.997742</td>\n",
       "      <td>757.118615</td>\n",
       "      <td>50.036958</td>\n",
       "      <td>1.672307</td>\n",
       "      <td>38.551120</td>\n",
       "      <td>5.657815</td>\n",
       "      <td>25.454178</td>\n",
       "      <td>26.094248</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>120.450742</td>\n",
       "      <td>5.554037</td>\n",
       "      <td>24.656915</td>\n",
       "      <td>36.269713</td>\n",
       "      <td>24.414354</td>\n",
       "      <td>33.879375</td>\n",
       "      <td>25.109004</td>\n",
       "      <td>37.730938</td>\n",
       "      <td>24.314713</td>\n",
       "      <td>36.822944</td>\n",
       "      <td>...</td>\n",
       "      <td>21.885693</td>\n",
       "      <td>37.276275</td>\n",
       "      <td>15.922274</td>\n",
       "      <td>757.270220</td>\n",
       "      <td>48.000128</td>\n",
       "      <td>1.594348</td>\n",
       "      <td>38.091478</td>\n",
       "      <td>5.068444</td>\n",
       "      <td>25.605200</td>\n",
       "      <td>26.301366</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>121.660848</td>\n",
       "      <td>5.817770</td>\n",
       "      <td>24.673536</td>\n",
       "      <td>36.069421</td>\n",
       "      <td>24.484464</td>\n",
       "      <td>33.552319</td>\n",
       "      <td>25.098846</td>\n",
       "      <td>37.764366</td>\n",
       "      <td>24.327612</td>\n",
       "      <td>36.677616</td>\n",
       "      <td>...</td>\n",
       "      <td>21.833291</td>\n",
       "      <td>37.175826</td>\n",
       "      <td>15.876612</td>\n",
       "      <td>757.395387</td>\n",
       "      <td>45.955826</td>\n",
       "      <td>1.529525</td>\n",
       "      <td>37.801349</td>\n",
       "      <td>4.489516</td>\n",
       "      <td>25.737931</td>\n",
       "      <td>26.489577</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>283</th>\n",
       "      <td>-60.726034</td>\n",
       "      <td>-10.238584</td>\n",
       "      <td>25.104198</td>\n",
       "      <td>42.260453</td>\n",
       "      <td>9.576961</td>\n",
       "      <td>65.146749</td>\n",
       "      <td>19.556375</td>\n",
       "      <td>17.209601</td>\n",
       "      <td>20.138910</td>\n",
       "      <td>49.010220</td>\n",
       "      <td>...</td>\n",
       "      <td>24.686547</td>\n",
       "      <td>70.959932</td>\n",
       "      <td>-14.106939</td>\n",
       "      <td>758.983828</td>\n",
       "      <td>163.459441</td>\n",
       "      <td>12.914524</td>\n",
       "      <td>-26.906157</td>\n",
       "      <td>-2.108361</td>\n",
       "      <td>26.205878</td>\n",
       "      <td>27.185154</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>284</th>\n",
       "      <td>-60.724840</td>\n",
       "      <td>-10.238614</td>\n",
       "      <td>25.104214</td>\n",
       "      <td>42.260452</td>\n",
       "      <td>9.576973</td>\n",
       "      <td>65.146773</td>\n",
       "      <td>19.556390</td>\n",
       "      <td>17.209556</td>\n",
       "      <td>20.138907</td>\n",
       "      <td>49.010243</td>\n",
       "      <td>...</td>\n",
       "      <td>24.686543</td>\n",
       "      <td>70.959968</td>\n",
       "      <td>-14.106977</td>\n",
       "      <td>758.983872</td>\n",
       "      <td>163.459477</td>\n",
       "      <td>12.914582</td>\n",
       "      <td>-26.906160</td>\n",
       "      <td>-2.108402</td>\n",
       "      <td>26.205886</td>\n",
       "      <td>27.185154</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>285</th>\n",
       "      <td>-60.723773</td>\n",
       "      <td>-10.238645</td>\n",
       "      <td>25.104229</td>\n",
       "      <td>42.260455</td>\n",
       "      <td>9.576987</td>\n",
       "      <td>65.146789</td>\n",
       "      <td>19.556405</td>\n",
       "      <td>17.209516</td>\n",
       "      <td>20.138904</td>\n",
       "      <td>49.010267</td>\n",
       "      <td>...</td>\n",
       "      <td>24.686539</td>\n",
       "      <td>70.959999</td>\n",
       "      <td>-14.107010</td>\n",
       "      <td>758.983910</td>\n",
       "      <td>163.459514</td>\n",
       "      <td>12.914640</td>\n",
       "      <td>-26.906157</td>\n",
       "      <td>-2.108441</td>\n",
       "      <td>26.205878</td>\n",
       "      <td>27.185163</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>286</th>\n",
       "      <td>-60.722695</td>\n",
       "      <td>-10.238676</td>\n",
       "      <td>25.104245</td>\n",
       "      <td>42.260455</td>\n",
       "      <td>9.576999</td>\n",
       "      <td>65.146801</td>\n",
       "      <td>19.556419</td>\n",
       "      <td>17.209472</td>\n",
       "      <td>20.138900</td>\n",
       "      <td>49.010296</td>\n",
       "      <td>...</td>\n",
       "      <td>24.686533</td>\n",
       "      <td>70.960035</td>\n",
       "      <td>-14.107045</td>\n",
       "      <td>758.983944</td>\n",
       "      <td>163.459532</td>\n",
       "      <td>12.914690</td>\n",
       "      <td>-26.906149</td>\n",
       "      <td>-2.108481</td>\n",
       "      <td>26.205878</td>\n",
       "      <td>27.185160</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>287</th>\n",
       "      <td>-60.721662</td>\n",
       "      <td>-10.238705</td>\n",
       "      <td>25.104259</td>\n",
       "      <td>42.260460</td>\n",
       "      <td>9.577013</td>\n",
       "      <td>65.146829</td>\n",
       "      <td>19.556432</td>\n",
       "      <td>17.209437</td>\n",
       "      <td>20.138897</td>\n",
       "      <td>49.010325</td>\n",
       "      <td>...</td>\n",
       "      <td>24.686530</td>\n",
       "      <td>70.960069</td>\n",
       "      <td>-14.107077</td>\n",
       "      <td>758.983979</td>\n",
       "      <td>163.459577</td>\n",
       "      <td>12.914741</td>\n",
       "      <td>-26.906146</td>\n",
       "      <td>-2.108519</td>\n",
       "      <td>26.205875</td>\n",
       "      <td>27.185169</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>288 rows × 28 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     Appliances     lights         T1       RH_1         T2       RH_2  \\\n",
       "0    117.528765   2.698902  24.611102  37.002893  24.227434  34.891202   \n",
       "1    119.844530   4.144052  24.637470  36.771019  24.281426  34.572933   \n",
       "2    119.690160   5.026903  24.645565  36.495787  24.344913  34.224538   \n",
       "3    120.450742   5.554037  24.656915  36.269713  24.414354  33.879375   \n",
       "4    121.660848   5.817770  24.673536  36.069421  24.484464  33.552319   \n",
       "..          ...        ...        ...        ...        ...        ...   \n",
       "283  -60.726034 -10.238584  25.104198  42.260453   9.576961  65.146749   \n",
       "284  -60.724840 -10.238614  25.104214  42.260452   9.576973  65.146773   \n",
       "285  -60.723773 -10.238645  25.104229  42.260455   9.576987  65.146789   \n",
       "286  -60.722695 -10.238676  25.104245  42.260455   9.576999  65.146801   \n",
       "287  -60.721662 -10.238705  25.104259  42.260460   9.577013  65.146829   \n",
       "\n",
       "            T3       RH_3         T4       RH_4  ...         T9       RH_9  \\\n",
       "0    25.151382  37.382300  24.284626  37.261892  ...  22.027223  37.597626   \n",
       "1    25.145838  37.617218  24.297594  37.153201  ...  21.988023  37.488395   \n",
       "2    25.124767  37.691677  24.305334  36.980809  ...  21.939192  37.378610   \n",
       "3    25.109004  37.730938  24.314713  36.822944  ...  21.885693  37.276275   \n",
       "4    25.098846  37.764366  24.327612  36.677616  ...  21.833291  37.175826   \n",
       "..         ...        ...        ...        ...  ...        ...        ...   \n",
       "283  19.556375  17.209601  20.138910  49.010220  ...  24.686547  70.959932   \n",
       "284  19.556390  17.209556  20.138907  49.010243  ...  24.686543  70.959968   \n",
       "285  19.556405  17.209516  20.138904  49.010267  ...  24.686539  70.959999   \n",
       "286  19.556419  17.209472  20.138900  49.010296  ...  24.686533  70.960035   \n",
       "287  19.556432  17.209437  20.138897  49.010325  ...  24.686530  70.960069   \n",
       "\n",
       "         T_out  Press_mm_hg      RH_out  Windspeed  Visibility  Tdewpoint  \\\n",
       "0    16.297667   756.734189   53.896075   1.854903   40.105939   6.713893   \n",
       "1    16.108513   756.933806   52.064223   1.761732   39.215918   6.241272   \n",
       "2    15.997742   757.118615   50.036958   1.672307   38.551120   5.657815   \n",
       "3    15.922274   757.270220   48.000128   1.594348   38.091478   5.068444   \n",
       "4    15.876612   757.395387   45.955826   1.529525   37.801349   4.489516   \n",
       "..         ...          ...         ...        ...         ...        ...   \n",
       "283 -14.106939   758.983828  163.459441  12.914524  -26.906157  -2.108361   \n",
       "284 -14.106977   758.983872  163.459477  12.914582  -26.906160  -2.108402   \n",
       "285 -14.107010   758.983910  163.459514  12.914640  -26.906157  -2.108441   \n",
       "286 -14.107045   758.983944  163.459532  12.914690  -26.906149  -2.108481   \n",
       "287 -14.107077   758.983979  163.459577  12.914741  -26.906146  -2.108519   \n",
       "\n",
       "           rv1        rv2  \n",
       "0    25.289255  25.359177  \n",
       "1    25.293669  25.845432  \n",
       "2    25.454178  26.094248  \n",
       "3    25.605200  26.301366  \n",
       "4    25.737931  26.489577  \n",
       "..         ...        ...  \n",
       "283  26.205878  27.185154  \n",
       "284  26.205886  27.185154  \n",
       "285  26.205878  27.185163  \n",
       "286  26.205878  27.185160  \n",
       "287  26.205875  27.185169  \n",
       "\n",
       "[288 rows x 28 columns]"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "true_predictions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# Great Job!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## EXTRA:\n",
    "### Our Saved Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from tensorflow.keras.models import load_model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "model.save(\"multivariate.h5\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.7.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}