regression_mpg_keras.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "\n",
    "import tensorflow as tf\n",
    "\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras import layers\n",
    "from tensorflow.keras.layers.experimental import preprocessing\n",
    "from keras.callbacks import History \n",
    "\n",
    "history = History()\n",
    "\n",
    "# Make numpy printouts easier to read.\n",
    "np.set_printoptions(precision=3, suppress=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "url = 'http://archive.ics.uci.edu/ml/machine-learning-databases/auto-mpg/auto-mpg.data'\n",
    "column_names = ['MPG', 'Cylinders', 'Displacement', 'Horsepower', 'Weight',\n",
    "                'Acceleration', 'Model Year', 'Origin']\n",
    "\n",
    "dataset = pd.read_csv(url, names=column_names,\n",
    "                          na_values='?', comment='\\t',\n",
    "                          sep=' ', skipinitialspace=True)\n",
    "#reading data, naming the columns, replacing \"?\" with NaN and removing the name of the car.\n",
    "\n",
    "#data = pd.read_fwf('./../../Downloads/auto-mpg.data', delimiter = \" \")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "      MPG  Cylinders  Displacement  Horsepower  Weight  Acceleration  \\\n",
      "0    18.0          8         307.0       130.0  3504.0          12.0   \n",
      "1    15.0          8         350.0       165.0  3693.0          11.5   \n",
      "2    18.0          8         318.0       150.0  3436.0          11.0   \n",
      "3    16.0          8         304.0       150.0  3433.0          12.0   \n",
      "4    17.0          8         302.0       140.0  3449.0          10.5   \n",
      "..    ...        ...           ...         ...     ...           ...   \n",
      "393  27.0          4         140.0        86.0  2790.0          15.6   \n",
      "394  44.0          4          97.0        52.0  2130.0          24.6   \n",
      "395  32.0          4         135.0        84.0  2295.0          11.6   \n",
      "396  28.0          4         120.0        79.0  2625.0          18.6   \n",
      "397  31.0          4         119.0        82.0  2720.0          19.4   \n",
      "\n",
      "     Model Year  Origin  \n",
      "0            70       1  \n",
      "1            70       1  \n",
      "2            70       1  \n",
      "3            70       1  \n",
      "4            70       1  \n",
      "..          ...     ...  \n",
      "393          82       1  \n",
      "394          82       2  \n",
      "395          82       1  \n",
      "396          82       1  \n",
      "397          82       1  \n",
      "\n",
      "[398 rows x 8 columns]\n"
     ]
    }
   ],
   "source": [
    "print(dataset)\n",
    "#print(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "      MPG  Cylinders  Displacement  Horsepower  Weight  Acceleration  \\\n",
      "0    18.0          8         307.0       130.0  3504.0          12.0   \n",
      "1    15.0          8         350.0       165.0  3693.0          11.5   \n",
      "2    18.0          8         318.0       150.0  3436.0          11.0   \n",
      "3    16.0          8         304.0       150.0  3433.0          12.0   \n",
      "4    17.0          8         302.0       140.0  3449.0          10.5   \n",
      "..    ...        ...           ...         ...     ...           ...   \n",
      "393  27.0          4         140.0        86.0  2790.0          15.6   \n",
      "394  44.0          4          97.0        52.0  2130.0          24.6   \n",
      "395  32.0          4         135.0        84.0  2295.0          11.6   \n",
      "396  28.0          4         120.0        79.0  2625.0          18.6   \n",
      "397  31.0          4         119.0        82.0  2720.0          19.4   \n",
      "\n",
      "     Model Year  Origin  \n",
      "0            70       1  \n",
      "1            70       1  \n",
      "2            70       1  \n",
      "3            70       1  \n",
      "4            70       1  \n",
      "..          ...     ...  \n",
      "393          82       1  \n",
      "394          82       2  \n",
      "395          82       1  \n",
      "396          82       1  \n",
      "397          82       1  \n",
      "\n",
      "[392 rows x 8 columns]\n"
     ]
    }
   ],
   "source": [
    "dataset.isna().sum()\n",
    "dataset = dataset.dropna()\n",
    "#Cleaning the data by removing the \"NaN\" values\n",
    "print(dataset)"
   ]
  },
  {
   "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>MPG</th>\n",
       "      <th>Cylinders</th>\n",
       "      <th>Displacement</th>\n",
       "      <th>Horsepower</th>\n",
       "      <th>Weight</th>\n",
       "      <th>Acceleration</th>\n",
       "      <th>Model Year</th>\n",
       "      <th>Origin</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>393</th>\n",
       "      <td>27.0</td>\n",
       "      <td>4</td>\n",
       "      <td>140.0</td>\n",
       "      <td>86.0</td>\n",
       "      <td>2790.0</td>\n",
       "      <td>15.6</td>\n",
       "      <td>82</td>\n",
       "      <td>USA</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>394</th>\n",
       "      <td>44.0</td>\n",
       "      <td>4</td>\n",
       "      <td>97.0</td>\n",
       "      <td>52.0</td>\n",
       "      <td>2130.0</td>\n",
       "      <td>24.6</td>\n",
       "      <td>82</td>\n",
       "      <td>Europe</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>395</th>\n",
       "      <td>32.0</td>\n",
       "      <td>4</td>\n",
       "      <td>135.0</td>\n",
       "      <td>84.0</td>\n",
       "      <td>2295.0</td>\n",
       "      <td>11.6</td>\n",
       "      <td>82</td>\n",
       "      <td>USA</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>396</th>\n",
       "      <td>28.0</td>\n",
       "      <td>4</td>\n",
       "      <td>120.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>2625.0</td>\n",
       "      <td>18.6</td>\n",
       "      <td>82</td>\n",
       "      <td>USA</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>397</th>\n",
       "      <td>31.0</td>\n",
       "      <td>4</td>\n",
       "      <td>119.0</td>\n",
       "      <td>82.0</td>\n",
       "      <td>2720.0</td>\n",
       "      <td>19.4</td>\n",
       "      <td>82</td>\n",
       "      <td>USA</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      MPG  Cylinders  Displacement  Horsepower  Weight  Acceleration  \\\n",
       "393  27.0          4         140.0        86.0  2790.0          15.6   \n",
       "394  44.0          4          97.0        52.0  2130.0          24.6   \n",
       "395  32.0          4         135.0        84.0  2295.0          11.6   \n",
       "396  28.0          4         120.0        79.0  2625.0          18.6   \n",
       "397  31.0          4         119.0        82.0  2720.0          19.4   \n",
       "\n",
       "     Model Year  Origin  \n",
       "393          82     USA  \n",
       "394          82  Europe  \n",
       "395          82     USA  \n",
       "396          82     USA  \n",
       "397          82     USA  "
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset['Origin'] = dataset['Origin'].map({1: 'USA', 2: 'Europe', 3: 'Japan'})\n",
    "dataset.tail()"
   ]
  },
  {
   "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>MPG</th>\n",
       "      <th>Cylinders</th>\n",
       "      <th>Displacement</th>\n",
       "      <th>Horsepower</th>\n",
       "      <th>Weight</th>\n",
       "      <th>Acceleration</th>\n",
       "      <th>Model Year</th>\n",
       "      <th>Europe</th>\n",
       "      <th>Japan</th>\n",
       "      <th>USA</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>393</th>\n",
       "      <td>27.0</td>\n",
       "      <td>4</td>\n",
       "      <td>140.0</td>\n",
       "      <td>86.0</td>\n",
       "      <td>2790.0</td>\n",
       "      <td>15.6</td>\n",
       "      <td>82</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>394</th>\n",
       "      <td>44.0</td>\n",
       "      <td>4</td>\n",
       "      <td>97.0</td>\n",
       "      <td>52.0</td>\n",
       "      <td>2130.0</td>\n",
       "      <td>24.6</td>\n",
       "      <td>82</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>395</th>\n",
       "      <td>32.0</td>\n",
       "      <td>4</td>\n",
       "      <td>135.0</td>\n",
       "      <td>84.0</td>\n",
       "      <td>2295.0</td>\n",
       "      <td>11.6</td>\n",
       "      <td>82</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>396</th>\n",
       "      <td>28.0</td>\n",
       "      <td>4</td>\n",
       "      <td>120.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>2625.0</td>\n",
       "      <td>18.6</td>\n",
       "      <td>82</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>397</th>\n",
       "      <td>31.0</td>\n",
       "      <td>4</td>\n",
       "      <td>119.0</td>\n",
       "      <td>82.0</td>\n",
       "      <td>2720.0</td>\n",
       "      <td>19.4</td>\n",
       "      <td>82</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      MPG  Cylinders  Displacement  Horsepower  Weight  Acceleration  \\\n",
       "393  27.0          4         140.0        86.0  2790.0          15.6   \n",
       "394  44.0          4          97.0        52.0  2130.0          24.6   \n",
       "395  32.0          4         135.0        84.0  2295.0          11.6   \n",
       "396  28.0          4         120.0        79.0  2625.0          18.6   \n",
       "397  31.0          4         119.0        82.0  2720.0          19.4   \n",
       "\n",
       "     Model Year  Europe  Japan  USA  \n",
       "393          82       0      0    1  \n",
       "394          82       1      0    0  \n",
       "395          82       0      0    1  \n",
       "396          82       0      0    1  \n",
       "397          82       0      0    1  "
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset = pd.get_dummies(dataset, prefix='', prefix_sep='')\n",
    "dataset.tail()\n",
    "#Changing the categorical parameter into a numerical parameter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "314 78\n"
     ]
    }
   ],
   "source": [
    "train_dataset = dataset.sample(frac = 0.8, random_state = 69)\n",
    "test_dataset = dataset.drop(train_dataset.index)\n",
    "print(len(train_dataset), len(test_dataset))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<seaborn.axisgrid.PairGrid at 0x7f1122d2a160>"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x900 with 30 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Data inspection and plotting \n",
    "#We are investigating the MPG of a vehicle. We plot graphs to observe the relationship among \"features\".\n",
    "sns.pairplot(train_dataset[['MPG', 'Cylinders', 'Displacement', 'Horsepower', 'Weight']], diag_kind='kde')\n",
    "# <3 seaborn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "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>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>MPG</th>\n",
       "      <td>314.0</td>\n",
       "      <td>23.037261</td>\n",
       "      <td>7.847821</td>\n",
       "      <td>9.0</td>\n",
       "      <td>16.500</td>\n",
       "      <td>22.0</td>\n",
       "      <td>29.0</td>\n",
       "      <td>46.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Cylinders</th>\n",
       "      <td>314.0</td>\n",
       "      <td>5.582803</td>\n",
       "      <td>1.725904</td>\n",
       "      <td>3.0</td>\n",
       "      <td>4.000</td>\n",
       "      <td>6.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Displacement</th>\n",
       "      <td>314.0</td>\n",
       "      <td>200.584395</td>\n",
       "      <td>105.512276</td>\n",
       "      <td>68.0</td>\n",
       "      <td>107.000</td>\n",
       "      <td>156.0</td>\n",
       "      <td>302.0</td>\n",
       "      <td>455.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Horsepower</th>\n",
       "      <td>314.0</td>\n",
       "      <td>106.417197</td>\n",
       "      <td>39.620014</td>\n",
       "      <td>46.0</td>\n",
       "      <td>76.000</td>\n",
       "      <td>95.0</td>\n",
       "      <td>131.5</td>\n",
       "      <td>230.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Weight</th>\n",
       "      <td>314.0</td>\n",
       "      <td>3026.098726</td>\n",
       "      <td>857.700403</td>\n",
       "      <td>1755.0</td>\n",
       "      <td>2248.000</td>\n",
       "      <td>2902.5</td>\n",
       "      <td>3670.0</td>\n",
       "      <td>5140.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Acceleration</th>\n",
       "      <td>314.0</td>\n",
       "      <td>15.461783</td>\n",
       "      <td>2.773437</td>\n",
       "      <td>8.0</td>\n",
       "      <td>13.525</td>\n",
       "      <td>15.5</td>\n",
       "      <td>17.0</td>\n",
       "      <td>24.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Model Year</th>\n",
       "      <td>314.0</td>\n",
       "      <td>75.856688</td>\n",
       "      <td>3.683376</td>\n",
       "      <td>70.0</td>\n",
       "      <td>73.000</td>\n",
       "      <td>76.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>82.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Europe</th>\n",
       "      <td>314.0</td>\n",
       "      <td>0.168790</td>\n",
       "      <td>0.375164</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Japan</th>\n",
       "      <td>314.0</td>\n",
       "      <td>0.187898</td>\n",
       "      <td>0.391254</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>USA</th>\n",
       "      <td>314.0</td>\n",
       "      <td>0.643312</td>\n",
       "      <td>0.479786</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              count         mean         std     min       25%     50%  \\\n",
       "MPG           314.0    23.037261    7.847821     9.0    16.500    22.0   \n",
       "Cylinders     314.0     5.582803    1.725904     3.0     4.000     6.0   \n",
       "Displacement  314.0   200.584395  105.512276    68.0   107.000   156.0   \n",
       "Horsepower    314.0   106.417197   39.620014    46.0    76.000    95.0   \n",
       "Weight        314.0  3026.098726  857.700403  1755.0  2248.000  2902.5   \n",
       "Acceleration  314.0    15.461783    2.773437     8.0    13.525    15.5   \n",
       "Model Year    314.0    75.856688    3.683376    70.0    73.000    76.0   \n",
       "Europe        314.0     0.168790    0.375164     0.0     0.000     0.0   \n",
       "Japan         314.0     0.187898    0.391254     0.0     0.000     0.0   \n",
       "USA           314.0     0.643312    0.479786     0.0     0.000     1.0   \n",
       "\n",
       "                 75%     max  \n",
       "MPG             29.0    46.6  \n",
       "Cylinders        8.0     8.0  \n",
       "Displacement   302.0   455.0  \n",
       "Horsepower     131.5   230.0  \n",
       "Weight        3670.0  5140.0  \n",
       "Acceleration    17.0    24.8  \n",
       "Model Year      79.0    82.0  \n",
       "Europe           0.0     1.0  \n",
       "Japan            0.0     1.0  \n",
       "USA              1.0     1.0  "
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_dataset.describe().transpose()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_features = train_dataset.copy()\n",
    "test_features = test_dataset.copy()\n",
    "\n",
    "train_labels = train_features.pop('MPG')\n",
    "test_labels = test_features.pop('MPG')\n",
    "#Removing the MPG label from our dataset as that is what we would like to predict."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "First example: [[   4.  113.   95. 2228.   14.   71.    0.    1.    0.]]\n",
      "Normalized: [[-0.919 -0.831 -0.289 -0.932 -0.528 -1.321 -0.451  2.079 -1.343]]\n"
     ]
    }
   ],
   "source": [
    "#Feature normalization\n",
    "\n",
    "normalizer = preprocessing.Normalization()\n",
    "normalizer.adapt(np.array(train_features))\n",
    "#print(normalizer.mean.numpy())\n",
    "\n",
    "feature_vec = np.array(train_features[:1])\n",
    "\n",
    "print('First example:', feature_vec)\n",
    "print('Normalized:', normalizer(feature_vec).numpy())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_5\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "normalization_13 (Normalizat (None, 1)                 3         \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 1)                 2         \n",
      "=================================================================\n",
      "Total params: 5\n",
      "Trainable params: 2\n",
      "Non-trainable params: 3\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#Predicting MPG from just horsepower (one feature)\n",
    "horsepower_vec = np.array(train_features['Horsepower'])\n",
    "\n",
    "horsepower_normalizer = preprocessing.Normalization(input_shape=[1,])\n",
    "horsepower_normalizer.adapt(horsepower_vec)\n",
    "horsepower_model = tf.keras.Sequential([\n",
    "    horsepower_normalizer,\n",
    "    layers.Dense(units=1, use_bias = 'True')#, activation = 'sigmoid', use_bias = 'True')\n",
    "])\n",
    "\n",
    "horsepower_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:11 out of the last 11 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7f10cc56a4c0> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
      "Epoch 1/1000\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 22.7574 - val_loss: 23.3398\n",
      "Epoch 2/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 23.2176 - val_loss: 23.2597\n",
      "Epoch 3/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 22.3601 - val_loss: 23.1798\n",
      "Epoch 4/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 22.1335 - val_loss: 23.1001\n",
      "Epoch 5/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 23.3146 - val_loss: 23.0200\n",
      "Epoch 6/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 22.3415 - val_loss: 22.9401\n",
      "Epoch 7/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 22.2775 - val_loss: 22.8600\n",
      "Epoch 8/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 22.4099 - val_loss: 22.7801\n",
      "Epoch 9/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 21.8219 - val_loss: 22.7002\n",
      "Epoch 10/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 22.0417 - val_loss: 22.6202\n",
      "Epoch 11/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 22.2325 - val_loss: 22.5401\n",
      "Epoch 12/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 22.0737 - val_loss: 22.4602\n",
      "Epoch 13/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 21.5926 - val_loss: 22.3803\n",
      "Epoch 14/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 22.3900 - val_loss: 22.3002\n",
      "Epoch 15/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 21.9344 - val_loss: 22.2203\n",
      "Epoch 16/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 21.8318 - val_loss: 22.1403\n",
      "Epoch 17/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 21.0867 - val_loss: 22.0605\n",
      "Epoch 18/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 21.0452 - val_loss: 21.9804\n",
      "Epoch 19/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 21.4493 - val_loss: 21.9004\n",
      "Epoch 20/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 21.3993 - val_loss: 21.8205\n",
      "Epoch 21/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 21.3768 - val_loss: 21.7404\n",
      "Epoch 22/1000\n",
      "8/8 [==============================] - 0s 6ms/step - loss: 21.7981 - val_loss: 21.6604\n",
      "Epoch 23/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 20.3888 - val_loss: 21.5806\n",
      "Epoch 24/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 20.2801 - val_loss: 21.5007\n",
      "Epoch 25/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 21.4362 - val_loss: 21.4206\n",
      "Epoch 26/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 20.0376 - val_loss: 21.3408\n",
      "Epoch 27/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 20.5199 - val_loss: 21.2607\n",
      "Epoch 28/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 21.1683 - val_loss: 21.1806\n",
      "Epoch 29/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 20.3091 - val_loss: 21.1007\n",
      "Epoch 30/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 20.4993 - val_loss: 21.0208\n",
      "Epoch 31/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 19.3513 - val_loss: 20.9410\n",
      "Epoch 32/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 20.4844 - val_loss: 20.8608\n",
      "Epoch 33/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 20.6581 - val_loss: 20.7808\n",
      "Epoch 34/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 20.7826 - val_loss: 20.7008\n",
      "Epoch 35/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 21.0859 - val_loss: 20.6208\n",
      "Epoch 36/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 20.6014 - val_loss: 20.5408\n",
      "Epoch 37/1000\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 19.6898 - val_loss: 20.4611\n",
      "Epoch 38/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 19.7053 - val_loss: 20.3811\n",
      "Epoch 39/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 19.7589 - val_loss: 20.3012\n",
      "Epoch 40/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 19.0953 - val_loss: 20.2214\n",
      "Epoch 41/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 19.7958 - val_loss: 20.1412\n",
      "Epoch 42/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 19.4382 - val_loss: 20.0613\n",
      "Epoch 43/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 18.9540 - val_loss: 19.9815\n",
      "Epoch 44/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 19.2283 - val_loss: 19.9014\n",
      "Epoch 45/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 18.9027 - val_loss: 19.8214\n",
      "Epoch 46/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 19.4743 - val_loss: 19.7414\n",
      "Epoch 47/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 19.3753 - val_loss: 19.6614\n",
      "Epoch 48/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 19.1769 - val_loss: 19.5814\n",
      "Epoch 49/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 19.4296 - val_loss: 19.5014\n",
      "Epoch 50/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 19.4632 - val_loss: 19.4214\n",
      "Epoch 51/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 19.0759 - val_loss: 19.3415\n",
      "Epoch 52/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 18.1353 - val_loss: 19.2616\n",
      "Epoch 53/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 18.8328 - val_loss: 19.1817\n",
      "Epoch 54/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 19.0278 - val_loss: 19.1016\n",
      "Epoch 55/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 18.0493 - val_loss: 19.0218\n",
      "Epoch 56/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 18.5516 - val_loss: 18.9418\n",
      "Epoch 57/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 18.3329 - val_loss: 18.8618\n",
      "Epoch 58/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 18.9073 - val_loss: 18.7818\n",
      "Epoch 59/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 18.0361 - val_loss: 18.7019\n",
      "Epoch 60/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 18.2364 - val_loss: 18.6220\n",
      "Epoch 61/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 18.1525 - val_loss: 18.5420\n",
      "Epoch 62/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 18.1813 - val_loss: 18.4620\n",
      "Epoch 63/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 18.4978 - val_loss: 18.3820\n",
      "Epoch 64/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 18.0383 - val_loss: 18.3024\n",
      "Epoch 65/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 18.4441 - val_loss: 18.2231\n",
      "Epoch 66/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 17.5498 - val_loss: 18.1439\n",
      "Epoch 67/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 17.5863 - val_loss: 18.0647\n",
      "Epoch 68/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 17.4096 - val_loss: 17.9855\n",
      "Epoch 69/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 17.5157 - val_loss: 17.9066\n",
      "Epoch 70/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 17.4857 - val_loss: 17.8274\n",
      "Epoch 71/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 17.2091 - val_loss: 17.7483\n",
      "Epoch 72/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 17.5478 - val_loss: 17.6690\n",
      "Epoch 73/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 17.4823 - val_loss: 17.5896\n",
      "Epoch 74/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 16.6384 - val_loss: 17.5103\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 75/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 16.9565 - val_loss: 17.4311\n",
      "Epoch 76/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 16.9508 - val_loss: 17.3519\n",
      "Epoch 77/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 16.9992 - val_loss: 17.2724\n",
      "Epoch 78/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 17.0420 - val_loss: 17.1930\n",
      "Epoch 79/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 16.5610 - val_loss: 17.1136\n",
      "Epoch 80/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 16.9763 - val_loss: 17.0345\n",
      "Epoch 81/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 16.3026 - val_loss: 16.9555\n",
      "Epoch 82/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 16.8918 - val_loss: 16.8764\n",
      "Epoch 83/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 15.7639 - val_loss: 16.7976\n",
      "Epoch 84/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 17.2076 - val_loss: 16.7187\n",
      "Epoch 85/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 16.2642 - val_loss: 16.6397\n",
      "Epoch 86/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 15.6913 - val_loss: 16.5607\n",
      "Epoch 87/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 16.2196 - val_loss: 16.4815\n",
      "Epoch 88/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 15.4314 - val_loss: 16.4027\n",
      "Epoch 89/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 15.6308 - val_loss: 16.3236\n",
      "Epoch 90/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 16.0128 - val_loss: 16.2446\n",
      "Epoch 91/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 15.7999 - val_loss: 16.1652\n",
      "Epoch 92/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 16.2251 - val_loss: 16.0857\n",
      "Epoch 93/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 16.1048 - val_loss: 16.0069\n",
      "Epoch 94/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 15.7875 - val_loss: 15.9276\n",
      "Epoch 95/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 15.3717 - val_loss: 15.8485\n",
      "Epoch 96/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 15.2530 - val_loss: 15.7696\n",
      "Epoch 97/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 15.3408 - val_loss: 15.6904\n",
      "Epoch 98/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 14.9100 - val_loss: 15.6113\n",
      "Epoch 99/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 15.2218 - val_loss: 15.5320\n",
      "Epoch 100/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 15.2960 - val_loss: 15.4526\n",
      "Epoch 101/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 14.7754 - val_loss: 15.3735\n",
      "Epoch 102/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 15.1355 - val_loss: 15.2943\n",
      "Epoch 103/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 14.4550 - val_loss: 15.2149\n",
      "Epoch 104/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 14.6407 - val_loss: 15.1356\n",
      "Epoch 105/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 14.3577 - val_loss: 15.0562\n",
      "Epoch 106/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 14.6381 - val_loss: 14.9767\n",
      "Epoch 107/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 14.4635 - val_loss: 14.8972\n",
      "Epoch 108/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 14.1902 - val_loss: 14.8180\n",
      "Epoch 109/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 14.1870 - val_loss: 14.7386\n",
      "Epoch 110/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 14.0519 - val_loss: 14.6593\n",
      "Epoch 111/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 14.4262 - val_loss: 14.5797\n",
      "Epoch 112/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 14.4324 - val_loss: 14.5001\n",
      "Epoch 113/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 14.3185 - val_loss: 14.4207\n",
      "Epoch 114/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 13.8624 - val_loss: 14.3415\n",
      "Epoch 115/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 14.0489 - val_loss: 14.2617\n",
      "Epoch 116/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 13.4432 - val_loss: 14.1824\n",
      "Epoch 117/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.9277 - val_loss: 14.1030\n",
      "Epoch 118/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 13.5128 - val_loss: 14.0234\n",
      "Epoch 119/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 14.3686 - val_loss: 13.9437\n",
      "Epoch 120/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 13.8999 - val_loss: 13.8642\n",
      "Epoch 121/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 14.1114 - val_loss: 13.7846\n",
      "Epoch 122/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 13.2427 - val_loss: 13.7052\n",
      "Epoch 123/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 13.2574 - val_loss: 13.6255\n",
      "Epoch 124/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 13.4887 - val_loss: 13.5457\n",
      "Epoch 125/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 14.0203 - val_loss: 13.4661\n",
      "Epoch 126/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.7815 - val_loss: 13.3866\n",
      "Epoch 127/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.7019 - val_loss: 13.3076\n",
      "Epoch 128/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 13.0330 - val_loss: 13.2283\n",
      "Epoch 129/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.8097 - val_loss: 13.1492\n",
      "Epoch 130/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.9232 - val_loss: 13.0703\n",
      "Epoch 131/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 13.3224 - val_loss: 12.9907\n",
      "Epoch 132/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 11.7526 - val_loss: 12.9115\n",
      "Epoch 133/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.3847 - val_loss: 12.8319\n",
      "Epoch 134/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.0488 - val_loss: 12.7521\n",
      "Epoch 135/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.1676 - val_loss: 12.6727\n",
      "Epoch 136/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.3057 - val_loss: 12.5933\n",
      "Epoch 137/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.4135 - val_loss: 12.5137\n",
      "Epoch 138/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 11.8562 - val_loss: 12.4340\n",
      "Epoch 139/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.4562 - val_loss: 12.3543\n",
      "Epoch 140/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 11.5732 - val_loss: 12.2757\n",
      "Epoch 141/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 11.7282 - val_loss: 12.1967\n",
      "Epoch 142/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 11.7959 - val_loss: 12.1179\n",
      "Epoch 143/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 12.0368 - val_loss: 12.0394\n",
      "Epoch 144/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 11.1553 - val_loss: 11.9608\n",
      "Epoch 145/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.2209 - val_loss: 11.8819\n",
      "Epoch 146/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 11.3802 - val_loss: 11.8030\n",
      "Epoch 147/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 11.5182 - val_loss: 11.7239\n",
      "Epoch 148/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 11.4762 - val_loss: 11.6449\n",
      "Epoch 149/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 11.2684 - val_loss: 11.5664\n",
      "Epoch 150/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 11.3095 - val_loss: 11.4879\n",
      "Epoch 151/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.8549 - val_loss: 11.4087\n",
      "Epoch 152/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.9689 - val_loss: 11.3298\n",
      "Epoch 153/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.7300 - val_loss: 11.2513\n",
      "Epoch 154/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.9644 - val_loss: 11.1720\n",
      "Epoch 155/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8/8 [==============================] - 0s 3ms/step - loss: 11.0363 - val_loss: 11.0932\n",
      "Epoch 156/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.7416 - val_loss: 11.0141\n",
      "Epoch 157/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.3607 - val_loss: 10.9344\n",
      "Epoch 158/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.9557 - val_loss: 10.8558\n",
      "Epoch 159/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.5437 - val_loss: 10.7769\n",
      "Epoch 160/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.8376 - val_loss: 10.6978\n",
      "Epoch 161/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.3367 - val_loss: 10.6186\n",
      "Epoch 162/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.2105 - val_loss: 10.5398\n",
      "Epoch 163/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.0060 - val_loss: 10.4610\n",
      "Epoch 164/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.0860 - val_loss: 10.3820\n",
      "Epoch 165/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 9.7115 - val_loss: 10.3033\n",
      "Epoch 166/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 9.9259 - val_loss: 10.2240\n",
      "Epoch 167/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 9.6540 - val_loss: 10.1445\n",
      "Epoch 168/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 9.7871 - val_loss: 10.0658\n",
      "Epoch 169/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.0897 - val_loss: 9.9862\n",
      "Epoch 170/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.1202 - val_loss: 9.9073\n",
      "Epoch 171/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 9.4895 - val_loss: 9.8279\n",
      "Epoch 172/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 9.0035 - val_loss: 9.7489\n",
      "Epoch 173/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 9.6521 - val_loss: 9.6694\n",
      "Epoch 174/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 9.1731 - val_loss: 9.5904\n",
      "Epoch 175/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 9.0956 - val_loss: 9.5148\n",
      "Epoch 176/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 9.0978 - val_loss: 9.4425\n",
      "Epoch 177/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 9.3380 - val_loss: 9.3709\n",
      "Epoch 178/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 8.8810 - val_loss: 9.2991\n",
      "Epoch 179/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 9.0771 - val_loss: 9.2270\n",
      "Epoch 180/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 8.9875 - val_loss: 9.1555\n",
      "Epoch 181/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 8.8004 - val_loss: 9.0834\n",
      "Epoch 182/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 9.4405 - val_loss: 9.0113\n",
      "Epoch 183/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 8.8359 - val_loss: 8.9396\n",
      "Epoch 184/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 8.7860 - val_loss: 8.8678\n",
      "Epoch 185/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 9.0288 - val_loss: 8.7959\n",
      "Epoch 186/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 8.9540 - val_loss: 8.7248\n",
      "Epoch 187/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 8.2039 - val_loss: 8.6540\n",
      "Epoch 188/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 7.9666 - val_loss: 8.5840\n",
      "Epoch 189/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 8.4679 - val_loss: 8.5134\n",
      "Epoch 190/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 8.4591 - val_loss: 8.4430\n",
      "Epoch 191/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 8.0533 - val_loss: 8.3736\n",
      "Epoch 192/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 8.0746 - val_loss: 8.3031\n",
      "Epoch 193/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 7.8388 - val_loss: 8.2330\n",
      "Epoch 194/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 8.0017 - val_loss: 8.1627\n",
      "Epoch 195/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 7.6135 - val_loss: 8.0933\n",
      "Epoch 196/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 7.3507 - val_loss: 8.0244\n",
      "Epoch 197/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 7.4147 - val_loss: 7.9557\n",
      "Epoch 198/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 7.3008 - val_loss: 7.8863\n",
      "Epoch 199/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 7.5666 - val_loss: 7.8176\n",
      "Epoch 200/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 7.6207 - val_loss: 7.7494\n",
      "Epoch 201/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.8372 - val_loss: 7.6812\n",
      "Epoch 202/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 7.5219 - val_loss: 7.6129\n",
      "Epoch 203/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.6134 - val_loss: 7.5454\n",
      "Epoch 204/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 7.3169 - val_loss: 7.4776\n",
      "Epoch 205/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 7.3801 - val_loss: 7.4110\n",
      "Epoch 206/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 7.3518 - val_loss: 7.3444\n",
      "Epoch 207/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.8179 - val_loss: 7.2800\n",
      "Epoch 208/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.5427 - val_loss: 7.2188\n",
      "Epoch 209/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 7.0053 - val_loss: 7.1572\n",
      "Epoch 210/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.4827 - val_loss: 7.0957\n",
      "Epoch 211/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.4130 - val_loss: 7.0337\n",
      "Epoch 212/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.6969 - val_loss: 6.9731\n",
      "Epoch 213/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.7696 - val_loss: 6.9130\n",
      "Epoch 214/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.6650 - val_loss: 6.8527\n",
      "Epoch 215/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.4843 - val_loss: 6.7957\n",
      "Epoch 216/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.1483 - val_loss: 6.7396\n",
      "Epoch 217/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.2551 - val_loss: 6.6847\n",
      "Epoch 218/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 6.0688 - val_loss: 6.6295\n",
      "Epoch 219/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.0733 - val_loss: 6.5744\n",
      "Epoch 220/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 5.8026 - val_loss: 6.5212\n",
      "Epoch 221/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 5.7485 - val_loss: 6.4691\n",
      "Epoch 222/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.1279 - val_loss: 6.4154\n",
      "Epoch 223/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.9382 - val_loss: 6.3630\n",
      "Epoch 224/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 6.0462 - val_loss: 6.3146\n",
      "Epoch 225/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.3665 - val_loss: 6.2681\n",
      "Epoch 226/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.0513 - val_loss: 6.2230\n",
      "Epoch 227/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 5.9143 - val_loss: 6.1775\n",
      "Epoch 228/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 5.6729 - val_loss: 6.1344\n",
      "Epoch 229/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 5.5808 - val_loss: 6.0917\n",
      "Epoch 230/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 5.7294 - val_loss: 6.0487\n",
      "Epoch 231/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 5.6994 - val_loss: 6.0096\n",
      "Epoch 232/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.5246 - val_loss: 5.9708\n",
      "Epoch 233/1000\n",
      "8/8 [==============================] - 0s 6ms/step - loss: 5.3039 - val_loss: 5.9341\n",
      "Epoch 234/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.4739 - val_loss: 5.8988\n",
      "Epoch 235/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.2444 - val_loss: 5.8629\n",
      "Epoch 236/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.5744 - val_loss: 5.8275\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 237/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.1373 - val_loss: 5.7926\n",
      "Epoch 238/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.1152 - val_loss: 5.7581\n",
      "Epoch 239/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.5866 - val_loss: 5.7230\n",
      "Epoch 240/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.2080 - val_loss: 5.6879\n",
      "Epoch 241/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.9108 - val_loss: 5.6538\n",
      "Epoch 242/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.6114 - val_loss: 5.6215\n",
      "Epoch 243/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.1195 - val_loss: 5.5910\n",
      "Epoch 244/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 5.1257 - val_loss: 5.5630\n",
      "Epoch 245/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.0997 - val_loss: 5.5352\n",
      "Epoch 246/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.7594 - val_loss: 5.5085\n",
      "Epoch 247/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.0783 - val_loss: 5.4807\n",
      "Epoch 248/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.0592 - val_loss: 5.4547\n",
      "Epoch 249/1000\n",
      "8/8 [==============================] - 0s 6ms/step - loss: 4.7973 - val_loss: 5.4276\n",
      "Epoch 250/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.9421 - val_loss: 5.4007\n",
      "Epoch 251/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.9749 - val_loss: 5.3742\n",
      "Epoch 252/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.7593 - val_loss: 5.3483\n",
      "Epoch 253/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.5834 - val_loss: 5.3217\n",
      "Epoch 254/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.8119 - val_loss: 5.2962\n",
      "Epoch 255/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.7028 - val_loss: 5.2733\n",
      "Epoch 256/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.5958 - val_loss: 5.2515\n",
      "Epoch 257/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.8270 - val_loss: 5.2306\n",
      "Epoch 258/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.4577 - val_loss: 5.2113\n",
      "Epoch 259/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.4456 - val_loss: 5.1929\n",
      "Epoch 260/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.8527 - val_loss: 5.1729\n",
      "Epoch 261/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.8639 - val_loss: 5.1537\n",
      "Epoch 262/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 4.5723 - val_loss: 5.1360\n",
      "Epoch 263/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.4765 - val_loss: 5.1166\n",
      "Epoch 264/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.5334 - val_loss: 5.0980\n",
      "Epoch 265/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.5031 - val_loss: 5.0803\n",
      "Epoch 266/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.6318 - val_loss: 5.0632\n",
      "Epoch 267/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.4901 - val_loss: 5.0446\n",
      "Epoch 268/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.5203 - val_loss: 5.0276\n",
      "Epoch 269/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.4769 - val_loss: 5.0111\n",
      "Epoch 270/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.1211 - val_loss: 4.9956\n",
      "Epoch 271/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.5375 - val_loss: 4.9809\n",
      "Epoch 272/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.3992 - val_loss: 4.9655\n",
      "Epoch 273/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.2608 - val_loss: 4.9523\n",
      "Epoch 274/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.4868 - val_loss: 4.9395\n",
      "Epoch 275/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.3325 - val_loss: 4.9252\n",
      "Epoch 276/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.2796 - val_loss: 4.9135\n",
      "Epoch 277/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.3210 - val_loss: 4.9031\n",
      "Epoch 278/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.1446 - val_loss: 4.8921\n",
      "Epoch 279/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.2710 - val_loss: 4.8824\n",
      "Epoch 280/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.2622 - val_loss: 4.8720\n",
      "Epoch 281/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.3955 - val_loss: 4.8620\n",
      "Epoch 282/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.2203 - val_loss: 4.8525\n",
      "Epoch 283/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9670 - val_loss: 4.8418\n",
      "Epoch 284/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.3644 - val_loss: 4.8331\n",
      "Epoch 285/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8024 - val_loss: 4.8251\n",
      "Epoch 286/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.1380 - val_loss: 4.8196\n",
      "Epoch 287/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.1282 - val_loss: 4.8137\n",
      "Epoch 288/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.0052 - val_loss: 4.8080\n",
      "Epoch 289/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.1574 - val_loss: 4.8013\n",
      "Epoch 290/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.1559 - val_loss: 4.7954\n",
      "Epoch 291/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9781 - val_loss: 4.7896\n",
      "Epoch 292/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9386 - val_loss: 4.7842\n",
      "Epoch 293/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.0783 - val_loss: 4.7797\n",
      "Epoch 294/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.0714 - val_loss: 4.7749\n",
      "Epoch 295/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7838 - val_loss: 4.7696\n",
      "Epoch 296/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.9025 - val_loss: 4.7639\n",
      "Epoch 297/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8072 - val_loss: 4.7576\n",
      "Epoch 298/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.9133 - val_loss: 4.7530\n",
      "Epoch 299/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.1430 - val_loss: 4.7482\n",
      "Epoch 300/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8844 - val_loss: 4.7429\n",
      "Epoch 301/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9897 - val_loss: 4.7381\n",
      "Epoch 302/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8283 - val_loss: 4.7335\n",
      "Epoch 303/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5439 - val_loss: 4.7284\n",
      "Epoch 304/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7689 - val_loss: 4.7263\n",
      "Epoch 305/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8501 - val_loss: 4.7231\n",
      "Epoch 306/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8585 - val_loss: 4.7204\n",
      "Epoch 307/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6442 - val_loss: 4.7193\n",
      "Epoch 308/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8926 - val_loss: 4.7177\n",
      "Epoch 309/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.0112 - val_loss: 4.7162\n",
      "Epoch 310/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.1865 - val_loss: 4.7158\n",
      "Epoch 311/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9461 - val_loss: 4.7137\n",
      "Epoch 312/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7695 - val_loss: 4.7103\n",
      "Epoch 313/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.0700 - val_loss: 4.7091\n",
      "Epoch 314/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7561 - val_loss: 4.7067\n",
      "Epoch 315/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8956 - val_loss: 4.7057\n",
      "Epoch 316/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8817 - val_loss: 4.7045\n",
      "Epoch 317/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.0908 - val_loss: 4.7013\n",
      "Epoch 318/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7386 - val_loss: 4.6977\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 319/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7336 - val_loss: 4.6960\n",
      "Epoch 320/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6384 - val_loss: 4.6944\n",
      "Epoch 321/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7149 - val_loss: 4.6918\n",
      "Epoch 322/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6894 - val_loss: 4.6889\n",
      "Epoch 323/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9107 - val_loss: 4.6860\n",
      "Epoch 324/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6802 - val_loss: 4.6847\n",
      "Epoch 325/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6382 - val_loss: 4.6823\n",
      "Epoch 326/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6816 - val_loss: 4.6804\n",
      "Epoch 327/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7604 - val_loss: 4.6794\n",
      "Epoch 328/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7928 - val_loss: 4.6773\n",
      "Epoch 329/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7249 - val_loss: 4.6757\n",
      "Epoch 330/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.0525 - val_loss: 4.6726\n",
      "Epoch 331/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8506 - val_loss: 4.6706\n",
      "Epoch 332/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6084 - val_loss: 4.6693\n",
      "Epoch 333/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8595 - val_loss: 4.6666\n",
      "Epoch 334/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7121 - val_loss: 4.6653\n",
      "Epoch 335/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6927 - val_loss: 4.6646\n",
      "Epoch 336/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7772 - val_loss: 4.6635\n",
      "Epoch 337/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5328 - val_loss: 4.6623\n",
      "Epoch 338/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9207 - val_loss: 4.6605\n",
      "Epoch 339/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.1248 - val_loss: 4.6587\n",
      "Epoch 340/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7419 - val_loss: 4.6575\n",
      "Epoch 341/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6622 - val_loss: 4.6555\n",
      "Epoch 342/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6935 - val_loss: 4.6551\n",
      "Epoch 343/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8948 - val_loss: 4.6547\n",
      "Epoch 344/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9066 - val_loss: 4.6529\n",
      "Epoch 345/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6522 - val_loss: 4.6505\n",
      "Epoch 346/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6108 - val_loss: 4.6495\n",
      "Epoch 347/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7358 - val_loss: 4.6488\n",
      "Epoch 348/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5596 - val_loss: 4.6476\n",
      "Epoch 349/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6118 - val_loss: 4.6454\n",
      "Epoch 350/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6080 - val_loss: 4.6436\n",
      "Epoch 351/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9063 - val_loss: 4.6421\n",
      "Epoch 352/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6436 - val_loss: 4.6415\n",
      "Epoch 353/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7406 - val_loss: 4.6395\n",
      "Epoch 354/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.4738 - val_loss: 4.6383\n",
      "Epoch 355/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8103 - val_loss: 4.6373\n",
      "Epoch 356/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6253 - val_loss: 4.6373\n",
      "Epoch 357/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6449 - val_loss: 4.6375\n",
      "Epoch 358/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6213 - val_loss: 4.6369\n",
      "Epoch 359/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7475 - val_loss: 4.6356\n",
      "Epoch 360/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5415 - val_loss: 4.6341\n",
      "Epoch 361/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9133 - val_loss: 4.6341\n",
      "Epoch 362/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9622 - val_loss: 4.6345\n",
      "Epoch 363/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5593 - val_loss: 4.6340\n",
      "Epoch 364/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8525 - val_loss: 4.6338\n",
      "Epoch 365/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5255 - val_loss: 4.6334\n",
      "Epoch 366/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6966 - val_loss: 4.6337\n",
      "Epoch 367/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8292 - val_loss: 4.6338\n",
      "Epoch 368/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7718 - val_loss: 4.6328\n",
      "Epoch 369/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7716 - val_loss: 4.6323\n",
      "Epoch 370/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5223 - val_loss: 4.6310\n",
      "Epoch 371/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8579 - val_loss: 4.6310\n",
      "Epoch 372/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8858 - val_loss: 4.6314\n",
      "Epoch 373/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6681 - val_loss: 4.6310\n",
      "Epoch 374/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7817 - val_loss: 4.6306\n",
      "Epoch 375/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7918 - val_loss: 4.6309\n",
      "Epoch 376/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8286 - val_loss: 4.6299\n",
      "Epoch 377/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5787 - val_loss: 4.6299\n",
      "Epoch 378/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5097 - val_loss: 4.6298\n",
      "Epoch 379/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8207 - val_loss: 4.6293\n",
      "Epoch 380/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6284 - val_loss: 4.6302\n",
      "Epoch 381/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5720 - val_loss: 4.6275\n",
      "Epoch 382/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6776 - val_loss: 4.6273\n",
      "Epoch 383/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6404 - val_loss: 4.6274\n",
      "Epoch 384/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7300 - val_loss: 4.6257\n",
      "Epoch 385/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6228 - val_loss: 4.6255\n",
      "Epoch 386/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8567 - val_loss: 4.6265\n",
      "Epoch 387/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8006 - val_loss: 4.6256\n",
      "Epoch 388/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7092 - val_loss: 4.6238\n",
      "Epoch 389/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5234 - val_loss: 4.6240\n",
      "Epoch 390/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5380 - val_loss: 4.6240\n",
      "Epoch 391/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6395 - val_loss: 4.6254\n",
      "Epoch 392/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8953 - val_loss: 4.6248\n",
      "Epoch 393/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7502 - val_loss: 4.6237\n",
      "Epoch 394/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5194 - val_loss: 4.6235\n",
      "Epoch 395/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6492 - val_loss: 4.6228\n",
      "Epoch 396/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7600 - val_loss: 4.6230\n",
      "Epoch 397/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.0635 - val_loss: 4.6239\n",
      "Epoch 398/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4320 - val_loss: 4.6243\n",
      "Epoch 399/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8072 - val_loss: 4.6231\n",
      "Epoch 400/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9467 - val_loss: 4.6227\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 401/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9314 - val_loss: 4.6212\n",
      "Epoch 402/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8168 - val_loss: 4.6224\n",
      "Epoch 403/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6935 - val_loss: 4.6223\n",
      "Epoch 404/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6370 - val_loss: 4.6222\n",
      "Epoch 405/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6576 - val_loss: 4.6217\n",
      "Epoch 406/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6092 - val_loss: 4.6208\n",
      "Epoch 407/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7313 - val_loss: 4.6221\n",
      "Epoch 408/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8975 - val_loss: 4.6227\n",
      "Epoch 409/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5894 - val_loss: 4.6216\n",
      "Epoch 410/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7730 - val_loss: 4.6214\n",
      "Epoch 411/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7158 - val_loss: 4.6219\n",
      "Epoch 412/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8425 - val_loss: 4.6219\n",
      "Epoch 413/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5267 - val_loss: 4.6214\n",
      "Epoch 414/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7268 - val_loss: 4.6202\n",
      "Epoch 415/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8723 - val_loss: 4.6198\n",
      "Epoch 416/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8175 - val_loss: 4.6206\n",
      "Epoch 417/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4944 - val_loss: 4.6221\n",
      "Epoch 418/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7099 - val_loss: 4.6230\n",
      "Epoch 419/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6057 - val_loss: 4.6210\n",
      "Epoch 420/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4747 - val_loss: 4.6200\n",
      "Epoch 421/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6094 - val_loss: 4.6199\n",
      "Epoch 422/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5816 - val_loss: 4.6202\n",
      "Epoch 423/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8031 - val_loss: 4.6196\n",
      "Epoch 424/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8982 - val_loss: 4.6208\n",
      "Epoch 425/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8433 - val_loss: 4.6207\n",
      "Epoch 426/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8439 - val_loss: 4.6206\n",
      "Epoch 427/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6622 - val_loss: 4.6207\n",
      "Epoch 428/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8763 - val_loss: 4.6207\n",
      "Epoch 429/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6529 - val_loss: 4.6206\n",
      "Epoch 430/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.3672 - val_loss: 4.6218\n",
      "Epoch 431/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5076 - val_loss: 4.6216\n",
      "Epoch 432/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6894 - val_loss: 4.6210\n",
      "Epoch 433/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8680 - val_loss: 4.6200\n",
      "Epoch 434/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7133 - val_loss: 4.6193\n",
      "Epoch 435/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6407 - val_loss: 4.6190\n",
      "Epoch 436/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6935 - val_loss: 4.6196\n",
      "Epoch 437/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8094 - val_loss: 4.6209\n",
      "Epoch 438/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6284 - val_loss: 4.6217\n",
      "Epoch 439/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7477 - val_loss: 4.6218\n",
      "Epoch 440/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5693 - val_loss: 4.6218\n",
      "Epoch 441/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6493 - val_loss: 4.6213\n",
      "Epoch 442/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6074 - val_loss: 4.6200\n",
      "Epoch 443/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4580 - val_loss: 4.6195\n",
      "Epoch 444/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4932 - val_loss: 4.6199\n",
      "Epoch 445/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6989 - val_loss: 4.6192\n",
      "Epoch 446/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5764 - val_loss: 4.6196\n",
      "Epoch 447/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8264 - val_loss: 4.6204\n",
      "Epoch 448/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7758 - val_loss: 4.6206\n",
      "Epoch 449/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7092 - val_loss: 4.6199\n",
      "Epoch 450/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6024 - val_loss: 4.6209\n",
      "Epoch 451/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7332 - val_loss: 4.6209\n",
      "Epoch 452/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5006 - val_loss: 4.6199\n",
      "Epoch 453/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8555 - val_loss: 4.6206\n",
      "Epoch 454/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5794 - val_loss: 4.6199\n",
      "Epoch 455/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7828 - val_loss: 4.6202\n",
      "Epoch 456/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6604 - val_loss: 4.6198\n",
      "Epoch 457/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7616 - val_loss: 4.6187\n",
      "Epoch 458/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6585 - val_loss: 4.6204\n",
      "Epoch 459/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5423 - val_loss: 4.6204\n",
      "Epoch 460/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7408 - val_loss: 4.6216\n",
      "Epoch 461/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6118 - val_loss: 4.6219\n",
      "Epoch 462/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6301 - val_loss: 4.6216\n",
      "Epoch 463/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8884 - val_loss: 4.6205\n",
      "Epoch 464/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6243 - val_loss: 4.6200\n",
      "Epoch 465/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7296 - val_loss: 4.6203\n",
      "Epoch 466/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7895 - val_loss: 4.6194\n",
      "Epoch 467/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4784 - val_loss: 4.6193\n",
      "Epoch 468/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5644 - val_loss: 4.6186\n",
      "Epoch 469/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5289 - val_loss: 4.6198\n",
      "Epoch 470/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4862 - val_loss: 4.6196\n",
      "Epoch 471/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6152 - val_loss: 4.6209\n",
      "Epoch 472/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6529 - val_loss: 4.6205\n",
      "Epoch 473/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8344 - val_loss: 4.6185\n",
      "Epoch 474/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6141 - val_loss: 4.6184\n",
      "Epoch 475/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7117 - val_loss: 4.6183\n",
      "Epoch 476/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5344 - val_loss: 4.6190\n",
      "Epoch 477/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4937 - val_loss: 4.6197\n",
      "Epoch 478/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6908 - val_loss: 4.6199\n",
      "Epoch 479/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5988 - val_loss: 4.6198\n",
      "Epoch 480/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7986 - val_loss: 4.6201\n",
      "Epoch 481/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6331 - val_loss: 4.6196\n",
      "Epoch 482/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6919 - val_loss: 4.6184\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 483/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4653 - val_loss: 4.6184\n",
      "Epoch 484/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9858 - val_loss: 4.6190\n",
      "Epoch 485/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7517 - val_loss: 4.6189\n",
      "Epoch 486/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7573 - val_loss: 4.6180\n",
      "Epoch 487/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.3974 - val_loss: 4.6184\n",
      "Epoch 488/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8167 - val_loss: 4.6188\n",
      "Epoch 489/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5314 - val_loss: 4.6191\n",
      "Epoch 490/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6886 - val_loss: 4.6197\n",
      "Epoch 491/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6378 - val_loss: 4.6192\n",
      "Epoch 492/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5626 - val_loss: 4.6190\n",
      "Epoch 493/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6921 - val_loss: 4.6189\n",
      "Epoch 494/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4478 - val_loss: 4.6186\n",
      "Epoch 495/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4394 - val_loss: 4.6197\n",
      "Epoch 496/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7129 - val_loss: 4.6191\n",
      "Epoch 497/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6155 - val_loss: 4.6202\n",
      "Epoch 498/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.4774 - val_loss: 4.6195\n",
      "Epoch 499/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6337 - val_loss: 4.6187\n",
      "Epoch 500/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7155 - val_loss: 4.6191\n",
      "Epoch 501/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8412 - val_loss: 4.6196\n",
      "Epoch 502/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8206 - val_loss: 4.6191\n",
      "Epoch 503/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7645 - val_loss: 4.6190\n",
      "Epoch 504/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7219 - val_loss: 4.6191\n",
      "Epoch 505/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7254 - val_loss: 4.6194\n",
      "Epoch 506/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6422 - val_loss: 4.6195\n",
      "Epoch 507/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7578 - val_loss: 4.6185\n",
      "Epoch 508/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.0569 - val_loss: 4.6191\n",
      "Epoch 509/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5652 - val_loss: 4.6199\n",
      "Epoch 510/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6193 - val_loss: 4.6194\n",
      "Epoch 511/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5713 - val_loss: 4.6189\n",
      "Epoch 512/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7118 - val_loss: 4.6205\n",
      "Epoch 513/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6476 - val_loss: 4.6197\n",
      "Epoch 514/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.3433 - val_loss: 4.6179\n",
      "Epoch 515/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7198 - val_loss: 4.6180\n",
      "Epoch 516/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6880 - val_loss: 4.6182\n",
      "Epoch 517/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8024 - val_loss: 4.6188\n",
      "Epoch 518/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6048 - val_loss: 4.6182\n",
      "Epoch 519/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4840 - val_loss: 4.6183\n",
      "Epoch 520/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6194 - val_loss: 4.6199\n",
      "Epoch 521/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6733 - val_loss: 4.6205\n",
      "Epoch 522/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6220 - val_loss: 4.6196\n",
      "Epoch 523/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4684 - val_loss: 4.6193\n",
      "Epoch 524/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7297 - val_loss: 4.6185\n",
      "Epoch 525/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4748 - val_loss: 4.6187\n",
      "Epoch 526/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7785 - val_loss: 4.6192\n",
      "Epoch 527/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8591 - val_loss: 4.6175\n",
      "Epoch 528/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9476 - val_loss: 4.6180\n",
      "Epoch 529/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6239 - val_loss: 4.6189\n",
      "Epoch 530/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7411 - val_loss: 4.6192\n",
      "Epoch 531/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8227 - val_loss: 4.6196\n",
      "Epoch 532/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6742 - val_loss: 4.6195\n",
      "Epoch 533/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5909 - val_loss: 4.6191\n",
      "Epoch 534/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6114 - val_loss: 4.6189\n",
      "Epoch 535/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6179 - val_loss: 4.6187\n",
      "Epoch 536/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6046 - val_loss: 4.6194\n",
      "Epoch 537/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7040 - val_loss: 4.6192\n",
      "Epoch 538/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7460 - val_loss: 4.6180\n",
      "Epoch 539/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6852 - val_loss: 4.6177\n",
      "Epoch 540/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6788 - val_loss: 4.6178\n",
      "Epoch 541/1000\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 3.8868 - val_loss: 4.6191\n",
      "Epoch 542/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.6972 - val_loss: 4.6194\n",
      "Epoch 543/1000\n",
      "8/8 [==============================] - 0s 6ms/step - loss: 3.7348 - val_loss: 4.6186\n",
      "Epoch 544/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7074 - val_loss: 4.6192\n",
      "Epoch 545/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7558 - val_loss: 4.6194\n",
      "Epoch 546/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.8474 - val_loss: 4.6196\n",
      "Epoch 547/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.6507 - val_loss: 4.6186\n",
      "Epoch 548/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.6813 - val_loss: 4.6186\n",
      "Epoch 549/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8252 - val_loss: 4.6179\n",
      "Epoch 550/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.8439 - val_loss: 4.6190\n",
      "Epoch 551/1000\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 3.8893 - val_loss: 4.6180\n",
      "Epoch 552/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5455 - val_loss: 4.6191\n",
      "Epoch 553/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.6687 - val_loss: 4.6199\n",
      "Epoch 554/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.8158 - val_loss: 4.6203\n",
      "Epoch 555/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.8172 - val_loss: 4.6202\n",
      "Epoch 556/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6374 - val_loss: 4.6196\n",
      "Epoch 557/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.8378 - val_loss: 4.6194\n",
      "Epoch 558/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.5697 - val_loss: 4.6189\n",
      "Epoch 559/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.4343 - val_loss: 4.6190\n",
      "Epoch 560/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7761 - val_loss: 4.6189\n",
      "Epoch 561/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6783 - val_loss: 4.6194\n",
      "Epoch 562/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.9560 - val_loss: 4.6198\n",
      "Epoch 563/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.6872 - val_loss: 4.6188\n",
      "Epoch 564/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.5352 - val_loss: 4.6189\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 565/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7544 - val_loss: 4.6194\n",
      "Epoch 566/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5259 - val_loss: 4.6193\n",
      "Epoch 567/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6535 - val_loss: 4.6194\n",
      "Epoch 568/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.9194 - val_loss: 4.6190\n",
      "Epoch 569/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.4232 - val_loss: 4.6192\n",
      "Epoch 570/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.6329 - val_loss: 4.6192\n",
      "Epoch 571/1000\n",
      "8/8 [==============================] - 0s 6ms/step - loss: 3.7704 - val_loss: 4.6207\n",
      "Epoch 572/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7149 - val_loss: 4.6195\n",
      "Epoch 573/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8143 - val_loss: 4.6188\n",
      "Epoch 574/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7622 - val_loss: 4.6182\n",
      "Epoch 575/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.9477 - val_loss: 4.6182\n",
      "Epoch 576/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6465 - val_loss: 4.6184\n",
      "Epoch 577/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7261 - val_loss: 4.6182\n",
      "Epoch 578/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7755 - val_loss: 4.6180\n",
      "Epoch 579/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5460 - val_loss: 4.6188\n",
      "Epoch 580/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5976 - val_loss: 4.6186\n",
      "Epoch 581/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5994 - val_loss: 4.6192\n",
      "Epoch 582/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7306 - val_loss: 4.6190\n",
      "Epoch 583/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8887 - val_loss: 4.6186\n",
      "Epoch 584/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5522 - val_loss: 4.6185\n",
      "Epoch 585/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4699 - val_loss: 4.6196\n",
      "Epoch 586/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8014 - val_loss: 4.6198\n",
      "Epoch 587/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7374 - val_loss: 4.6198\n",
      "Epoch 588/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6388 - val_loss: 4.6197\n",
      "Epoch 589/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8466 - val_loss: 4.6192\n",
      "Epoch 590/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.9864 - val_loss: 4.6187\n",
      "Epoch 591/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5624 - val_loss: 4.6196\n",
      "Epoch 592/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8710 - val_loss: 4.6181\n",
      "Epoch 593/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8147 - val_loss: 4.6178\n",
      "Epoch 594/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8471 - val_loss: 4.6180\n",
      "Epoch 595/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6120 - val_loss: 4.6185\n",
      "Epoch 596/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8485 - val_loss: 4.6177\n",
      "Epoch 597/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7701 - val_loss: 4.6191\n",
      "Epoch 598/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8624 - val_loss: 4.6187\n",
      "Epoch 599/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.9184 - val_loss: 4.6190\n",
      "Epoch 600/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6027 - val_loss: 4.6188\n",
      "Epoch 601/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7277 - val_loss: 4.6199\n",
      "Epoch 602/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4074 - val_loss: 4.6205\n",
      "Epoch 603/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.4625 - val_loss: 4.6180\n",
      "Epoch 604/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8283 - val_loss: 4.6187\n",
      "Epoch 605/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6136 - val_loss: 4.6190\n",
      "Epoch 606/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5888 - val_loss: 4.6198\n",
      "Epoch 607/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7071 - val_loss: 4.6191\n",
      "Epoch 608/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7331 - val_loss: 4.6188\n",
      "Epoch 609/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.4600 - val_loss: 4.6192\n",
      "Epoch 610/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5841 - val_loss: 4.6192\n",
      "Epoch 611/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5836 - val_loss: 4.6195\n",
      "Epoch 612/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.4909 - val_loss: 4.6186\n",
      "Epoch 613/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5412 - val_loss: 4.6173\n",
      "Epoch 614/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5121 - val_loss: 4.6180\n",
      "Epoch 615/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5325 - val_loss: 4.6175\n",
      "Epoch 616/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6395 - val_loss: 4.6185\n",
      "Epoch 617/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7257 - val_loss: 4.6184\n",
      "Epoch 618/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5729 - val_loss: 4.6200\n",
      "Epoch 619/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4799 - val_loss: 4.6193\n",
      "Epoch 620/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7109 - val_loss: 4.6189\n",
      "Epoch 621/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6984 - val_loss: 4.6187\n",
      "Epoch 622/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8470 - val_loss: 4.6192\n",
      "Epoch 623/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9056 - val_loss: 4.6195\n",
      "Epoch 624/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6731 - val_loss: 4.6180\n",
      "Epoch 625/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4483 - val_loss: 4.6194\n",
      "Epoch 626/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6145 - val_loss: 4.6198\n",
      "Epoch 627/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5319 - val_loss: 4.6202\n",
      "Epoch 628/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7383 - val_loss: 4.6200\n",
      "Epoch 629/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5772 - val_loss: 4.6194\n",
      "Epoch 630/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.9437 - val_loss: 4.6188\n",
      "Epoch 631/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5960 - val_loss: 4.6191\n",
      "Epoch 632/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8359 - val_loss: 4.6183\n",
      "Epoch 633/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6807 - val_loss: 4.6186\n",
      "Epoch 634/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7338 - val_loss: 4.6190\n",
      "Epoch 635/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6680 - val_loss: 4.6200\n",
      "Epoch 636/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8259 - val_loss: 4.6208\n",
      "Epoch 637/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7782 - val_loss: 4.6204\n",
      "Epoch 638/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6927 - val_loss: 4.6197\n",
      "Epoch 639/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7597 - val_loss: 4.6184\n",
      "Epoch 640/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7187 - val_loss: 4.6181\n",
      "Epoch 641/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8423 - val_loss: 4.6181\n",
      "Epoch 642/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8040 - val_loss: 4.6181\n",
      "Epoch 643/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5875 - val_loss: 4.6177\n",
      "Epoch 644/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5233 - val_loss: 4.6175\n",
      "Epoch 645/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.4971 - val_loss: 4.6187\n",
      "Epoch 646/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7810 - val_loss: 4.6200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 647/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.4779 - val_loss: 4.6198\n",
      "Epoch 648/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5094 - val_loss: 4.6189\n",
      "Epoch 649/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.3839 - val_loss: 4.6175\n",
      "Epoch 650/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7784 - val_loss: 4.6180\n",
      "Epoch 651/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9745 - val_loss: 4.6182\n",
      "Epoch 652/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6386 - val_loss: 4.6185\n",
      "Epoch 653/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5090 - val_loss: 4.6190\n",
      "Epoch 654/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5918 - val_loss: 4.6199\n",
      "Epoch 655/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8086 - val_loss: 4.6198\n",
      "Epoch 656/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7091 - val_loss: 4.6197\n",
      "Epoch 657/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6438 - val_loss: 4.6179\n",
      "Epoch 658/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5576 - val_loss: 4.6183\n",
      "Epoch 659/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7226 - val_loss: 4.6195\n",
      "Epoch 660/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5885 - val_loss: 4.6195\n",
      "Epoch 661/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4695 - val_loss: 4.6173\n",
      "Epoch 662/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5666 - val_loss: 4.6174\n",
      "Epoch 663/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6481 - val_loss: 4.6185\n",
      "Epoch 664/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6459 - val_loss: 4.6177\n",
      "Epoch 665/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7394 - val_loss: 4.6174\n",
      "Epoch 666/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9574 - val_loss: 4.6190\n",
      "Epoch 667/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7001 - val_loss: 4.6192\n",
      "Epoch 668/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7780 - val_loss: 4.6194\n",
      "Epoch 669/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8092 - val_loss: 4.6179\n",
      "Epoch 670/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5845 - val_loss: 4.6182\n",
      "Epoch 671/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8046 - val_loss: 4.6181\n",
      "Epoch 672/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5716 - val_loss: 4.6169\n",
      "Epoch 673/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7863 - val_loss: 4.6181\n",
      "Epoch 674/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7011 - val_loss: 4.6192\n",
      "Epoch 675/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8132 - val_loss: 4.6188\n",
      "Epoch 676/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5626 - val_loss: 4.6198\n",
      "Epoch 677/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6543 - val_loss: 4.6195\n",
      "Epoch 678/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6061 - val_loss: 4.6196\n",
      "Epoch 679/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8660 - val_loss: 4.6202\n",
      "Epoch 680/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5899 - val_loss: 4.6191\n",
      "Epoch 681/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5946 - val_loss: 4.6179\n",
      "Epoch 682/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7261 - val_loss: 4.6180\n",
      "Epoch 683/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6762 - val_loss: 4.6178\n",
      "Epoch 684/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6963 - val_loss: 4.6172\n",
      "Epoch 685/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8624 - val_loss: 4.6181\n",
      "Epoch 686/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7959 - val_loss: 4.6192\n",
      "Epoch 687/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8184 - val_loss: 4.6191\n",
      "Epoch 688/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7467 - val_loss: 4.6199\n",
      "Epoch 689/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6049 - val_loss: 4.6208\n",
      "Epoch 690/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6030 - val_loss: 4.6208\n",
      "Epoch 691/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7517 - val_loss: 4.6194\n",
      "Epoch 692/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6307 - val_loss: 4.6195\n",
      "Epoch 693/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6657 - val_loss: 4.6192\n",
      "Epoch 694/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8498 - val_loss: 4.6188\n",
      "Epoch 695/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8045 - val_loss: 4.6191\n",
      "Epoch 696/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5929 - val_loss: 4.6195\n",
      "Epoch 697/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6644 - val_loss: 4.6199\n",
      "Epoch 698/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8318 - val_loss: 4.6193\n",
      "Epoch 699/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9386 - val_loss: 4.6189\n",
      "Epoch 700/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6204 - val_loss: 4.6181\n",
      "Epoch 701/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5325 - val_loss: 4.6183\n",
      "Epoch 702/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6328 - val_loss: 4.6188\n",
      "Epoch 703/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5046 - val_loss: 4.6190\n",
      "Epoch 704/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8710 - val_loss: 4.6194\n",
      "Epoch 705/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5420 - val_loss: 4.6182\n",
      "Epoch 706/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6544 - val_loss: 4.6196\n",
      "Epoch 707/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6433 - val_loss: 4.6189\n",
      "Epoch 708/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5902 - val_loss: 4.6185\n",
      "Epoch 709/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6338 - val_loss: 4.6191\n",
      "Epoch 710/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7562 - val_loss: 4.6191\n",
      "Epoch 711/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6129 - val_loss: 4.6200\n",
      "Epoch 712/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6262 - val_loss: 4.6195\n",
      "Epoch 713/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4273 - val_loss: 4.6187\n",
      "Epoch 714/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6686 - val_loss: 4.6174\n",
      "Epoch 715/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6138 - val_loss: 4.6174\n",
      "Epoch 716/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6147 - val_loss: 4.6183\n",
      "Epoch 717/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7652 - val_loss: 4.6191\n",
      "Epoch 718/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7402 - val_loss: 4.6186\n",
      "Epoch 719/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5390 - val_loss: 4.6190\n",
      "Epoch 720/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7191 - val_loss: 4.6196\n",
      "Epoch 721/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6835 - val_loss: 4.6196\n",
      "Epoch 722/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6145 - val_loss: 4.6197\n",
      "Epoch 723/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6107 - val_loss: 4.6192\n",
      "Epoch 724/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6137 - val_loss: 4.6191\n",
      "Epoch 725/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7983 - val_loss: 4.6184\n",
      "Epoch 726/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6910 - val_loss: 4.6189\n",
      "Epoch 727/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.4774 - val_loss: 4.6190\n",
      "Epoch 728/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6908 - val_loss: 4.6203\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 729/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5538 - val_loss: 4.6201\n",
      "Epoch 730/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7165 - val_loss: 4.6183\n",
      "Epoch 731/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6087 - val_loss: 4.6183\n",
      "Epoch 732/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6553 - val_loss: 4.6189\n",
      "Epoch 733/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.9181 - val_loss: 4.6192\n",
      "Epoch 734/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5393 - val_loss: 4.6200\n",
      "Epoch 735/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5188 - val_loss: 4.6193\n",
      "Epoch 736/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5903 - val_loss: 4.6187\n",
      "Epoch 737/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6091 - val_loss: 4.6194\n",
      "Epoch 738/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.4074 - val_loss: 4.6188\n",
      "Epoch 739/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7043 - val_loss: 4.6199\n",
      "Epoch 740/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6407 - val_loss: 4.6190\n",
      "Epoch 741/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7595 - val_loss: 4.6181\n",
      "Epoch 742/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8471 - val_loss: 4.6197\n",
      "Epoch 743/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5734 - val_loss: 4.6183\n",
      "Epoch 744/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5860 - val_loss: 4.6172\n",
      "Epoch 745/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7037 - val_loss: 4.6171\n",
      "Epoch 746/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4698 - val_loss: 4.6174\n",
      "Epoch 747/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6819 - val_loss: 4.6193\n",
      "Epoch 748/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5041 - val_loss: 4.6190\n",
      "Epoch 749/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7662 - val_loss: 4.6195\n",
      "Epoch 750/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6462 - val_loss: 4.6177\n",
      "Epoch 751/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6406 - val_loss: 4.6180\n",
      "Epoch 752/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4994 - val_loss: 4.6179\n",
      "Epoch 753/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6566 - val_loss: 4.6178\n",
      "Epoch 754/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7144 - val_loss: 4.6185\n",
      "Epoch 755/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8691 - val_loss: 4.6186\n",
      "Epoch 756/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6846 - val_loss: 4.6185\n",
      "Epoch 757/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7604 - val_loss: 4.6187\n",
      "Epoch 758/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7488 - val_loss: 4.6198\n",
      "Epoch 759/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5974 - val_loss: 4.6198\n",
      "Epoch 760/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5774 - val_loss: 4.6196\n",
      "Epoch 761/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6042 - val_loss: 4.6191\n",
      "Epoch 762/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.3970 - val_loss: 4.6197\n",
      "Epoch 763/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7252 - val_loss: 4.6192\n",
      "Epoch 764/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5283 - val_loss: 4.6197\n",
      "Epoch 765/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7221 - val_loss: 4.6201\n",
      "Epoch 766/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5432 - val_loss: 4.6184\n",
      "Epoch 767/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7087 - val_loss: 4.6201\n",
      "Epoch 768/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6716 - val_loss: 4.6190\n",
      "Epoch 769/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7669 - val_loss: 4.6197\n",
      "Epoch 770/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7628 - val_loss: 4.6192\n",
      "Epoch 771/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6991 - val_loss: 4.6186\n",
      "Epoch 772/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6641 - val_loss: 4.6177\n",
      "Epoch 773/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6739 - val_loss: 4.6184\n",
      "Epoch 774/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5406 - val_loss: 4.6186\n",
      "Epoch 775/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8453 - val_loss: 4.6190\n",
      "Epoch 776/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6064 - val_loss: 4.6196\n",
      "Epoch 777/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5724 - val_loss: 4.6186\n",
      "Epoch 778/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6858 - val_loss: 4.6178\n",
      "Epoch 779/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6476 - val_loss: 4.6172\n",
      "Epoch 780/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6272 - val_loss: 4.6178\n",
      "Epoch 781/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6502 - val_loss: 4.6183\n",
      "Epoch 782/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8207 - val_loss: 4.6188\n",
      "Epoch 783/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7069 - val_loss: 4.6188\n",
      "Epoch 784/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6816 - val_loss: 4.6196\n",
      "Epoch 785/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.3113 - val_loss: 4.6199\n",
      "Epoch 786/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5107 - val_loss: 4.6191\n",
      "Epoch 787/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7536 - val_loss: 4.6199\n",
      "Epoch 788/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6745 - val_loss: 4.6197\n",
      "Epoch 789/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7347 - val_loss: 4.6180\n",
      "Epoch 790/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7693 - val_loss: 4.6183\n",
      "Epoch 791/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7740 - val_loss: 4.6191\n",
      "Epoch 792/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7886 - val_loss: 4.6192\n",
      "Epoch 793/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6812 - val_loss: 4.6176\n",
      "Epoch 794/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7462 - val_loss: 4.6196\n",
      "Epoch 795/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.3447 - val_loss: 4.6191\n",
      "Epoch 796/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5844 - val_loss: 4.6181\n",
      "Epoch 797/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6954 - val_loss: 4.6185\n",
      "Epoch 798/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6103 - val_loss: 4.6195\n",
      "Epoch 799/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5876 - val_loss: 4.6188\n",
      "Epoch 800/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9927 - val_loss: 4.6203\n",
      "Epoch 801/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6974 - val_loss: 4.6202\n",
      "Epoch 802/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7660 - val_loss: 4.6185\n",
      "Epoch 803/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8343 - val_loss: 4.6183\n",
      "Epoch 804/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7294 - val_loss: 4.6191\n",
      "Epoch 805/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7246 - val_loss: 4.6187\n",
      "Epoch 806/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6660 - val_loss: 4.6191\n",
      "Epoch 807/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9783 - val_loss: 4.6188\n",
      "Epoch 808/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6710 - val_loss: 4.6197\n",
      "Epoch 809/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6543 - val_loss: 4.6194\n",
      "Epoch 810/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6226 - val_loss: 4.6192\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 811/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5670 - val_loss: 4.6189\n",
      "Epoch 812/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8070 - val_loss: 4.6188\n",
      "Epoch 813/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4953 - val_loss: 4.6185\n",
      "Epoch 814/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7294 - val_loss: 4.6184\n",
      "Epoch 815/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8496 - val_loss: 4.6179\n",
      "Epoch 816/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8932 - val_loss: 4.6184\n",
      "Epoch 817/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6922 - val_loss: 4.6191\n",
      "Epoch 818/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8729 - val_loss: 4.6192\n",
      "Epoch 819/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6241 - val_loss: 4.6200\n",
      "Epoch 820/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7172 - val_loss: 4.6192\n",
      "Epoch 821/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6627 - val_loss: 4.6182\n",
      "Epoch 822/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7330 - val_loss: 4.6188\n",
      "Epoch 823/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8170 - val_loss: 4.6195\n",
      "Epoch 824/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.5747 - val_loss: 4.6206\n",
      "Epoch 825/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6824 - val_loss: 4.6206\n",
      "Epoch 826/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8865 - val_loss: 4.6189\n",
      "Epoch 827/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8870 - val_loss: 4.6188\n",
      "Epoch 828/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6913 - val_loss: 4.6191\n",
      "Epoch 829/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6821 - val_loss: 4.6185\n",
      "Epoch 830/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8962 - val_loss: 4.6183\n",
      "Epoch 831/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6244 - val_loss: 4.6187\n",
      "Epoch 832/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6414 - val_loss: 4.6177\n",
      "Epoch 833/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.3924 - val_loss: 4.6180\n",
      "Epoch 834/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6212 - val_loss: 4.6176\n",
      "Epoch 835/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6842 - val_loss: 4.6182\n",
      "Epoch 836/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5947 - val_loss: 4.6176\n",
      "Epoch 837/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6944 - val_loss: 4.6178\n",
      "Epoch 838/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6579 - val_loss: 4.6183\n",
      "Epoch 839/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7800 - val_loss: 4.6195\n",
      "Epoch 840/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6575 - val_loss: 4.6198\n",
      "Epoch 841/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6226 - val_loss: 4.6199\n",
      "Epoch 842/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6231 - val_loss: 4.6195\n",
      "Epoch 843/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6296 - val_loss: 4.6190\n",
      "Epoch 844/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7673 - val_loss: 4.6202\n",
      "Epoch 845/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7098 - val_loss: 4.6170\n",
      "Epoch 846/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8187 - val_loss: 4.6182\n",
      "Epoch 847/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6859 - val_loss: 4.6162\n",
      "Epoch 848/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6152 - val_loss: 4.6183\n",
      "Epoch 849/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5850 - val_loss: 4.6194\n",
      "Epoch 850/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6853 - val_loss: 4.6182\n",
      "Epoch 851/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7003 - val_loss: 4.6184\n",
      "Epoch 852/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5942 - val_loss: 4.6171\n",
      "Epoch 853/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7691 - val_loss: 4.6173\n",
      "Epoch 854/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5535 - val_loss: 4.6177\n",
      "Epoch 855/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8454 - val_loss: 4.6187\n",
      "Epoch 856/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6860 - val_loss: 4.6197\n",
      "Epoch 857/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5921 - val_loss: 4.6203\n",
      "Epoch 858/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7925 - val_loss: 4.6198\n",
      "Epoch 859/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6950 - val_loss: 4.6188\n",
      "Epoch 860/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8279 - val_loss: 4.6192\n",
      "Epoch 861/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4700 - val_loss: 4.6175\n",
      "Epoch 862/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5567 - val_loss: 4.6173\n",
      "Epoch 863/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6186 - val_loss: 4.6182\n",
      "Epoch 864/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6129 - val_loss: 4.6186\n",
      "Epoch 865/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5806 - val_loss: 4.6189\n",
      "Epoch 866/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4356 - val_loss: 4.6187\n",
      "Epoch 867/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5942 - val_loss: 4.6197\n",
      "Epoch 868/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6771 - val_loss: 4.6205\n",
      "Epoch 869/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5661 - val_loss: 4.6198\n",
      "Epoch 870/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6587 - val_loss: 4.6178\n",
      "Epoch 871/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7083 - val_loss: 4.6172\n",
      "Epoch 872/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.0248 - val_loss: 4.6171\n",
      "Epoch 873/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.3834 - val_loss: 4.6174\n",
      "Epoch 874/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7031 - val_loss: 4.6192\n",
      "Epoch 875/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6164 - val_loss: 4.6189\n",
      "Epoch 876/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8484 - val_loss: 4.6187\n",
      "Epoch 877/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5732 - val_loss: 4.6182\n",
      "Epoch 878/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.9549 - val_loss: 4.6183\n",
      "Epoch 879/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6941 - val_loss: 4.6192\n",
      "Epoch 880/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7844 - val_loss: 4.6192\n",
      "Epoch 881/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9758 - val_loss: 4.6191\n",
      "Epoch 882/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6213 - val_loss: 4.6188\n",
      "Epoch 883/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8638 - val_loss: 4.6195\n",
      "Epoch 884/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5283 - val_loss: 4.6191\n",
      "Epoch 885/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7318 - val_loss: 4.6189\n",
      "Epoch 886/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.4347 - val_loss: 4.6193\n",
      "Epoch 887/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7467 - val_loss: 4.6190\n",
      "Epoch 888/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7334 - val_loss: 4.6189\n",
      "Epoch 889/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7498 - val_loss: 4.6196\n",
      "Epoch 890/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5198 - val_loss: 4.6198\n",
      "Epoch 891/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5272 - val_loss: 4.6201\n",
      "Epoch 892/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7094 - val_loss: 4.6182\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 893/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.3239 - val_loss: 4.6178\n",
      "Epoch 894/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5975 - val_loss: 4.6170\n",
      "Epoch 895/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5988 - val_loss: 4.6187\n",
      "Epoch 896/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6093 - val_loss: 4.6188\n",
      "Epoch 897/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6244 - val_loss: 4.6189\n",
      "Epoch 898/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8277 - val_loss: 4.6198\n",
      "Epoch 899/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6089 - val_loss: 4.6210\n",
      "Epoch 900/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7485 - val_loss: 4.6193\n",
      "Epoch 901/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9184 - val_loss: 4.6194\n",
      "Epoch 902/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8530 - val_loss: 4.6186\n",
      "Epoch 903/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8496 - val_loss: 4.6177\n",
      "Epoch 904/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6777 - val_loss: 4.6179\n",
      "Epoch 905/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6207 - val_loss: 4.6191\n",
      "Epoch 906/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8125 - val_loss: 4.6184\n",
      "Epoch 907/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6649 - val_loss: 4.6187\n",
      "Epoch 908/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4881 - val_loss: 4.6196\n",
      "Epoch 909/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8035 - val_loss: 4.6189\n",
      "Epoch 910/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6948 - val_loss: 4.6191\n",
      "Epoch 911/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6338 - val_loss: 4.6182\n",
      "Epoch 912/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8116 - val_loss: 4.6188\n",
      "Epoch 913/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5481 - val_loss: 4.6176\n",
      "Epoch 914/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4619 - val_loss: 4.6190\n",
      "Epoch 915/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7402 - val_loss: 4.6202\n",
      "Epoch 916/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5262 - val_loss: 4.6204\n",
      "Epoch 917/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7245 - val_loss: 4.6191\n",
      "Epoch 918/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4979 - val_loss: 4.6191\n",
      "Epoch 919/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5192 - val_loss: 4.6183\n",
      "Epoch 920/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7697 - val_loss: 4.6179\n",
      "Epoch 921/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6966 - val_loss: 4.6185\n",
      "Epoch 922/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6573 - val_loss: 4.6192\n",
      "Epoch 923/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7441 - val_loss: 4.6186\n",
      "Epoch 924/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8627 - val_loss: 4.6186\n",
      "Epoch 925/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6206 - val_loss: 4.6179\n",
      "Epoch 926/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6945 - val_loss: 4.6190\n",
      "Epoch 927/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7528 - val_loss: 4.6190\n",
      "Epoch 928/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.1631 - val_loss: 4.6196\n",
      "Epoch 929/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9539 - val_loss: 4.6207\n",
      "Epoch 930/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5745 - val_loss: 4.6201\n",
      "Epoch 931/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7721 - val_loss: 4.6198\n",
      "Epoch 932/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6923 - val_loss: 4.6195\n",
      "Epoch 933/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6777 - val_loss: 4.6176\n",
      "Epoch 934/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5780 - val_loss: 4.6187\n",
      "Epoch 935/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8818 - val_loss: 4.6183\n",
      "Epoch 936/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6036 - val_loss: 4.6190\n",
      "Epoch 937/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6656 - val_loss: 4.6197\n",
      "Epoch 938/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6206 - val_loss: 4.6188\n",
      "Epoch 939/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8157 - val_loss: 4.6178\n",
      "Epoch 940/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6353 - val_loss: 4.6178\n",
      "Epoch 941/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6235 - val_loss: 4.6185\n",
      "Epoch 942/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7410 - val_loss: 4.6187\n",
      "Epoch 943/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7715 - val_loss: 4.6198\n",
      "Epoch 944/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8803 - val_loss: 4.6195\n",
      "Epoch 945/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6336 - val_loss: 4.6202\n",
      "Epoch 946/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6803 - val_loss: 4.6192\n",
      "Epoch 947/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9825 - val_loss: 4.6195\n",
      "Epoch 948/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5563 - val_loss: 4.6200\n",
      "Epoch 949/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9513 - val_loss: 4.6188\n",
      "Epoch 950/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6894 - val_loss: 4.6184\n",
      "Epoch 951/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8219 - val_loss: 4.6186\n",
      "Epoch 952/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9117 - val_loss: 4.6181\n",
      "Epoch 953/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6218 - val_loss: 4.6194\n",
      "Epoch 954/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.4505 - val_loss: 4.6192\n",
      "Epoch 955/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.9996 - val_loss: 4.6179\n",
      "Epoch 956/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6829 - val_loss: 4.6186\n",
      "Epoch 957/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7871 - val_loss: 4.6182\n",
      "Epoch 958/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6400 - val_loss: 4.6178\n",
      "Epoch 959/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5675 - val_loss: 4.6183\n",
      "Epoch 960/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7310 - val_loss: 4.6196\n",
      "Epoch 961/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6286 - val_loss: 4.6193\n",
      "Epoch 962/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5603 - val_loss: 4.6197\n",
      "Epoch 963/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6041 - val_loss: 4.6190\n",
      "Epoch 964/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7269 - val_loss: 4.6203\n",
      "Epoch 965/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.4920 - val_loss: 4.6184\n",
      "Epoch 966/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.4989 - val_loss: 4.6182\n",
      "Epoch 967/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8574 - val_loss: 4.6188\n",
      "Epoch 968/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6325 - val_loss: 4.6193\n",
      "Epoch 969/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6788 - val_loss: 4.6190\n",
      "Epoch 970/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7670 - val_loss: 4.6192\n",
      "Epoch 971/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5562 - val_loss: 4.6185\n",
      "Epoch 972/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7336 - val_loss: 4.6179\n",
      "Epoch 973/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5690 - val_loss: 4.6188\n",
      "Epoch 974/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.6392 - val_loss: 4.6187\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 975/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5832 - val_loss: 4.6193\n",
      "Epoch 976/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8142 - val_loss: 4.6195\n",
      "Epoch 977/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4786 - val_loss: 4.6213\n",
      "Epoch 978/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7697 - val_loss: 4.6192\n",
      "Epoch 979/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8624 - val_loss: 4.6179\n",
      "Epoch 980/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6640 - val_loss: 4.6178\n",
      "Epoch 981/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6628 - val_loss: 4.6184\n",
      "Epoch 982/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7020 - val_loss: 4.6184\n",
      "Epoch 983/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7778 - val_loss: 4.6191\n",
      "Epoch 984/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8036 - val_loss: 4.6189\n",
      "Epoch 985/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7544 - val_loss: 4.6183\n",
      "Epoch 986/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8321 - val_loss: 4.6194\n",
      "Epoch 987/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6174 - val_loss: 4.6184\n",
      "Epoch 988/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8872 - val_loss: 4.6192\n",
      "Epoch 989/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6797 - val_loss: 4.6194\n",
      "Epoch 990/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6891 - val_loss: 4.6195\n",
      "Epoch 991/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6941 - val_loss: 4.6193\n",
      "Epoch 992/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7600 - val_loss: 4.6193\n",
      "Epoch 993/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6800 - val_loss: 4.6175\n",
      "Epoch 994/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7522 - val_loss: 4.6174\n",
      "Epoch 995/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5319 - val_loss: 4.6176\n",
      "Epoch 996/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7236 - val_loss: 4.6196\n",
      "Epoch 997/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6525 - val_loss: 4.6192\n",
      "Epoch 998/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6579 - val_loss: 4.6173\n",
      "Epoch 999/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7674 - val_loss: 4.6177\n",
      "Epoch 1000/1000\n",
      "8/8 [==============================] - ETA: 0s - loss: 4.138 - 0s 4ms/step - loss: 3.8357 - val_loss: 4.6174\n"
     ]
    }
   ],
   "source": [
    "horsepower_model.predict(horsepower_vec[:10])\n",
    "horsepower_model.compile(\n",
    "    optimizer=tf.optimizers.Adam(learning_rate=0.01),\n",
    "    loss = 'mean_absolute_error')\n",
    "#%%time 'MeanSquaredError'\n",
    "history_with_bias_lr = horsepower_model.fit(train_features['Horsepower'], train_labels, validation_split = 0.2, epochs = 1000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "def plot_loss(history):\n",
    "    plt.plot(history.history['loss'], label='loss')\n",
    "    plt.plot(history.history['val_loss'], label='validation_loss')\n",
    "    #plt.ylim([0, 10])\n",
    "    plt.xlabel('Epoch')\n",
    "    plt.ylabel('Error [MPG]')\n",
    "    plt.legend()\n",
    "    plt.grid(True)\n",
    "    \n",
    "#print(history.history.keys())   \n",
    "#print(history.history['loss'])\n",
    "#plot_loss(history)\n",
    "#plot_loss(history_with_bias)\n",
    "plot_loss(history_with_bias_lr)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_results = {}\n",
    "\n",
    "test_results['horsepower_model'] = horsepower_model.evaluate(test_features['Horsepower'], test_labels, verbose = 0)\n",
    "#Running our model on test data and storing results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:11 out of the last 11 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7f10cc103b80> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n"
     ]
    },
    {
     "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": [
    "def plot_horsepower(x, y):\n",
    "    plt.scatter(train_features['Horsepower'], train_labels, label='Data')\n",
    "    plt.plot(x, y, color='k', label='Predictions')\n",
    "    plt.xlabel('Horsepower')\n",
    "    plt.ylabel('MPG')\n",
    "    plt.legend()\n",
    "x = tf.linspace(0.0, 250, 251)\n",
    "y = horsepower_model.predict(x)\n",
    "plot_horsepower(x,y)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1000\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 23.2953 - val_loss: 23.4693\n",
      "Epoch 2/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 22.9722 - val_loss: 23.3993\n",
      "Epoch 3/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 22.1349 - val_loss: 23.3307\n",
      "Epoch 4/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 22.5145 - val_loss: 23.2579\n",
      "Epoch 5/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 22.1350 - val_loss: 23.1904\n",
      "Epoch 6/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 22.4689 - val_loss: 23.1199\n",
      "Epoch 7/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 22.1407 - val_loss: 23.0487\n",
      "Epoch 8/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 22.6600 - val_loss: 22.9773\n",
      "Epoch 9/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 22.2347 - val_loss: 22.9076\n",
      "Epoch 10/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 21.9536 - val_loss: 22.8383\n",
      "Epoch 11/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 22.5073 - val_loss: 22.7657\n",
      "Epoch 12/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 22.1253 - val_loss: 22.6957\n",
      "Epoch 13/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 21.9651 - val_loss: 22.6279\n",
      "Epoch 14/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 21.4460 - val_loss: 22.5613\n",
      "Epoch 15/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 22.0447 - val_loss: 22.4887\n",
      "Epoch 16/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 21.9486 - val_loss: 22.4163\n",
      "Epoch 17/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 20.8446 - val_loss: 22.3466\n",
      "Epoch 18/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 22.0847 - val_loss: 22.2746\n",
      "Epoch 19/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 20.9717 - val_loss: 22.2041\n",
      "Epoch 20/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 21.0501 - val_loss: 22.1352\n",
      "Epoch 21/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 20.9456 - val_loss: 22.0642\n",
      "Epoch 22/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 21.1518 - val_loss: 21.9949\n",
      "Epoch 23/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 20.6665 - val_loss: 21.9264\n",
      "Epoch 24/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 21.0326 - val_loss: 21.8532\n",
      "Epoch 25/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 21.0116 - val_loss: 21.7805\n",
      "Epoch 26/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 20.5400 - val_loss: 21.7113\n",
      "Epoch 27/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 21.0167 - val_loss: 21.6387\n",
      "Epoch 28/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 20.4911 - val_loss: 21.5703\n",
      "Epoch 29/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 20.1106 - val_loss: 21.5001\n",
      "Epoch 30/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 20.5756 - val_loss: 21.4289\n",
      "Epoch 31/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 20.3649 - val_loss: 21.3575\n",
      "Epoch 32/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 20.8068 - val_loss: 21.2862\n",
      "Epoch 33/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 19.7208 - val_loss: 21.2184\n",
      "Epoch 34/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 20.1892 - val_loss: 21.1473\n",
      "Epoch 35/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 19.4938 - val_loss: 21.0749\n",
      "Epoch 36/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 19.7016 - val_loss: 21.0053\n",
      "Epoch 37/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 19.9145 - val_loss: 20.9337\n",
      "Epoch 38/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 19.4146 - val_loss: 20.8630\n",
      "Epoch 39/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 19.2945 - val_loss: 20.7904\n",
      "Epoch 40/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 19.2487 - val_loss: 20.7191\n",
      "Epoch 41/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 19.6133 - val_loss: 20.6452\n",
      "Epoch 42/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 20.3209 - val_loss: 20.5718\n",
      "Epoch 43/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 19.3620 - val_loss: 20.5013\n",
      "Epoch 44/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 20.0556 - val_loss: 20.4315\n",
      "Epoch 45/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 19.9209 - val_loss: 20.3597\n",
      "Epoch 46/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 19.7642 - val_loss: 20.2904\n",
      "Epoch 47/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 18.7073 - val_loss: 20.2211\n",
      "Epoch 48/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 18.3132 - val_loss: 20.1521\n",
      "Epoch 49/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 18.8795 - val_loss: 20.0782\n",
      "Epoch 50/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 18.7983 - val_loss: 20.0057\n",
      "Epoch 51/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 18.4348 - val_loss: 19.9354\n",
      "Epoch 52/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 17.8535 - val_loss: 19.8647\n",
      "Epoch 53/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 18.4724 - val_loss: 19.7930\n",
      "Epoch 54/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 18.4404 - val_loss: 19.7215\n",
      "Epoch 55/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 18.0205 - val_loss: 19.6521\n",
      "Epoch 56/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 18.7479 - val_loss: 19.5798\n",
      "Epoch 57/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 19.0811 - val_loss: 19.5056\n",
      "Epoch 58/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 18.2196 - val_loss: 19.4354\n",
      "Epoch 59/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 17.5015 - val_loss: 19.3650\n",
      "Epoch 60/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 18.7897 - val_loss: 19.2920\n",
      "Epoch 61/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 18.7676 - val_loss: 19.2187\n",
      "Epoch 62/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 17.0195 - val_loss: 19.1511\n",
      "Epoch 63/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 17.8506 - val_loss: 19.0780\n",
      "Epoch 64/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 18.1204 - val_loss: 19.0048\n",
      "Epoch 65/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 17.3168 - val_loss: 18.9338\n",
      "Epoch 66/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 17.6061 - val_loss: 18.8611\n",
      "Epoch 67/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 16.9665 - val_loss: 18.7906\n",
      "Epoch 68/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 17.5857 - val_loss: 18.7161\n",
      "Epoch 69/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 17.8027 - val_loss: 18.6443\n",
      "Epoch 70/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 17.7134 - val_loss: 18.5707\n",
      "Epoch 71/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 18.2113 - val_loss: 18.4971\n",
      "Epoch 72/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 16.7380 - val_loss: 18.4284\n",
      "Epoch 73/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 17.0166 - val_loss: 18.3544\n",
      "Epoch 74/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 17.5439 - val_loss: 18.2819\n",
      "Epoch 75/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 16.7576 - val_loss: 18.2097\n",
      "Epoch 76/1000\n",
      "8/8 [==============================] - 0s 6ms/step - loss: 16.9996 - val_loss: 18.1375\n",
      "Epoch 77/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 15.8704 - val_loss: 18.0668\n",
      "Epoch 78/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 16.7376 - val_loss: 17.9915\n",
      "Epoch 79/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 16.0999 - val_loss: 17.9192\n",
      "Epoch 80/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 17.0741 - val_loss: 17.8430\n",
      "Epoch 81/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 15.3251 - val_loss: 17.7712\n",
      "Epoch 82/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8/8 [==============================] - 0s 4ms/step - loss: 16.4120 - val_loss: 17.6958\n",
      "Epoch 83/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 15.6815 - val_loss: 17.6221\n",
      "Epoch 84/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 15.0712 - val_loss: 17.5498\n",
      "Epoch 85/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 16.1730 - val_loss: 17.4734\n",
      "Epoch 86/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 15.1236 - val_loss: 17.4011\n",
      "Epoch 87/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 15.7868 - val_loss: 17.3265\n",
      "Epoch 88/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 15.4353 - val_loss: 17.2519\n",
      "Epoch 89/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 16.3942 - val_loss: 17.1742\n",
      "Epoch 90/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 16.6001 - val_loss: 17.0982\n",
      "Epoch 91/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 14.7906 - val_loss: 17.0290\n",
      "Epoch 92/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 15.2756 - val_loss: 16.9556\n",
      "Epoch 93/1000\n",
      "8/8 [==============================] - 0s 6ms/step - loss: 16.3742 - val_loss: 16.8790\n",
      "Epoch 94/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 14.7986 - val_loss: 16.8076\n",
      "Epoch 95/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 15.7387 - val_loss: 16.7297\n",
      "Epoch 96/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 15.0211 - val_loss: 16.6590\n",
      "Epoch 97/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 15.7501 - val_loss: 16.5838\n",
      "Epoch 98/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 14.9976 - val_loss: 16.5116\n",
      "Epoch 99/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 14.6094 - val_loss: 16.4399\n",
      "Epoch 100/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 14.9715 - val_loss: 16.3645\n",
      "Epoch 101/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 14.6487 - val_loss: 16.2921\n",
      "Epoch 102/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 14.2127 - val_loss: 16.2190\n",
      "Epoch 103/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 15.2583 - val_loss: 16.1419\n",
      "Epoch 104/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 15.0201 - val_loss: 16.0637\n",
      "Epoch 105/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 14.2338 - val_loss: 15.9915\n",
      "Epoch 106/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 14.9834 - val_loss: 15.9158\n",
      "Epoch 107/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 14.2043 - val_loss: 15.8410\n",
      "Epoch 108/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 14.6543 - val_loss: 15.7672\n",
      "Epoch 109/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 14.0450 - val_loss: 15.6934\n",
      "Epoch 110/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 13.8973 - val_loss: 15.6202\n",
      "Epoch 111/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 13.6479 - val_loss: 15.5464\n",
      "Epoch 112/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 14.0597 - val_loss: 15.4715\n",
      "Epoch 113/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 13.6798 - val_loss: 15.3959\n",
      "Epoch 114/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 13.9536 - val_loss: 15.3211\n",
      "Epoch 115/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 13.4900 - val_loss: 15.2471\n",
      "Epoch 116/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 13.9112 - val_loss: 15.1711\n",
      "Epoch 117/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 13.1384 - val_loss: 15.0988\n",
      "Epoch 118/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.4454 - val_loss: 15.0254\n",
      "Epoch 119/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.5127 - val_loss: 14.9518\n",
      "Epoch 120/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.1011 - val_loss: 14.8763\n",
      "Epoch 121/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 13.0888 - val_loss: 14.8006\n",
      "Epoch 122/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 13.4454 - val_loss: 14.7252\n",
      "Epoch 123/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.9114 - val_loss: 14.6501\n",
      "Epoch 124/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.9127 - val_loss: 14.5762\n",
      "Epoch 125/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 13.1341 - val_loss: 14.4996\n",
      "Epoch 126/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.3089 - val_loss: 14.4262\n",
      "Epoch 127/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 12.5234 - val_loss: 14.3515\n",
      "Epoch 128/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 13.1513 - val_loss: 14.2759\n",
      "Epoch 129/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.1022 - val_loss: 14.2033\n",
      "Epoch 130/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.9626 - val_loss: 14.1250\n",
      "Epoch 131/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.1206 - val_loss: 14.0511\n",
      "Epoch 132/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 11.6339 - val_loss: 13.9732\n",
      "Epoch 133/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.0371 - val_loss: 13.8943\n",
      "Epoch 134/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.2150 - val_loss: 13.8123\n",
      "Epoch 135/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 11.9272 - val_loss: 13.7338\n",
      "Epoch 136/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 11.5880 - val_loss: 13.6530\n",
      "Epoch 137/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 12.1620 - val_loss: 13.5701\n",
      "Epoch 138/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 11.9970 - val_loss: 13.4877\n",
      "Epoch 139/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 11.5904 - val_loss: 13.4102\n",
      "Epoch 140/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.7349 - val_loss: 13.3316\n",
      "Epoch 141/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 11.7605 - val_loss: 13.2491\n",
      "Epoch 142/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 11.6034 - val_loss: 13.1672\n",
      "Epoch 143/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 11.6004 - val_loss: 13.0840\n",
      "Epoch 144/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 11.5613 - val_loss: 13.0031\n",
      "Epoch 145/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 11.1970 - val_loss: 12.9224\n",
      "Epoch 146/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 11.5227 - val_loss: 12.8404\n",
      "Epoch 147/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 11.4312 - val_loss: 12.7602\n",
      "Epoch 148/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.6401 - val_loss: 12.6801\n",
      "Epoch 149/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.8128 - val_loss: 12.5968\n",
      "Epoch 150/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.9402 - val_loss: 12.5146\n",
      "Epoch 151/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.5612 - val_loss: 12.4292\n",
      "Epoch 152/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.9285 - val_loss: 12.3441\n",
      "Epoch 153/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.8441 - val_loss: 12.2606\n",
      "Epoch 154/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 11.2893 - val_loss: 12.1754\n",
      "Epoch 155/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.6119 - val_loss: 12.0935\n",
      "Epoch 156/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.5873 - val_loss: 12.0100\n",
      "Epoch 157/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.2915 - val_loss: 11.9254\n",
      "Epoch 158/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.5848 - val_loss: 11.8409\n",
      "Epoch 159/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.6129 - val_loss: 11.7591\n",
      "Epoch 160/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 9.9246 - val_loss: 11.6801\n",
      "Epoch 161/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.3810 - val_loss: 11.5991\n",
      "Epoch 162/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8/8 [==============================] - 0s 3ms/step - loss: 10.4703 - val_loss: 11.5174\n",
      "Epoch 163/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 9.7392 - val_loss: 11.4366\n",
      "Epoch 164/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.5579 - val_loss: 11.3534\n",
      "Epoch 165/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 9.7996 - val_loss: 11.2735\n",
      "Epoch 166/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 10.4485 - val_loss: 11.1901\n",
      "Epoch 167/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 9.5618 - val_loss: 11.1106\n",
      "Epoch 168/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 9.7527 - val_loss: 11.0271\n",
      "Epoch 169/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 9.3678 - val_loss: 10.9446\n",
      "Epoch 170/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 10.1134 - val_loss: 10.8624\n",
      "Epoch 171/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 9.2502 - val_loss: 10.7809\n",
      "Epoch 172/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 9.1196 - val_loss: 10.6979\n",
      "Epoch 173/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 9.3979 - val_loss: 10.6138\n",
      "Epoch 174/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 9.1315 - val_loss: 10.5304\n",
      "Epoch 175/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 9.4512 - val_loss: 10.4470\n",
      "Epoch 176/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 9.1912 - val_loss: 10.3669\n",
      "Epoch 177/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 9.2386 - val_loss: 10.2855\n",
      "Epoch 178/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 9.6564 - val_loss: 10.2033\n",
      "Epoch 179/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 9.9046 - val_loss: 10.1190\n",
      "Epoch 180/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 8.7891 - val_loss: 10.0385\n",
      "Epoch 181/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 9.3250 - val_loss: 9.9570\n",
      "Epoch 182/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 8.6066 - val_loss: 9.8788\n",
      "Epoch 183/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 8.4770 - val_loss: 9.7943\n",
      "Epoch 184/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 8.4503 - val_loss: 9.7150\n",
      "Epoch 185/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 8.7488 - val_loss: 9.6357\n",
      "Epoch 186/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 8.0499 - val_loss: 9.5562\n",
      "Epoch 187/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 7.7370 - val_loss: 9.4771\n",
      "Epoch 188/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 7.9670 - val_loss: 9.3952\n",
      "Epoch 189/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 7.5015 - val_loss: 9.3142\n",
      "Epoch 190/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 7.5184 - val_loss: 9.2373\n",
      "Epoch 191/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 7.6263 - val_loss: 9.1630\n",
      "Epoch 192/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 7.9076 - val_loss: 9.0796\n",
      "Epoch 193/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 7.4939 - val_loss: 8.9976\n",
      "Epoch 194/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 7.7660 - val_loss: 8.9155\n",
      "Epoch 195/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 7.8720 - val_loss: 8.8362\n",
      "Epoch 196/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 7.8004 - val_loss: 8.7570\n",
      "Epoch 197/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 7.2823 - val_loss: 8.6771\n",
      "Epoch 198/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 7.3422 - val_loss: 8.5981\n",
      "Epoch 199/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 7.4447 - val_loss: 8.5141\n",
      "Epoch 200/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 7.2191 - val_loss: 8.4363\n",
      "Epoch 201/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 7.3107 - val_loss: 8.3555\n",
      "Epoch 202/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 6.7810 - val_loss: 8.2850\n",
      "Epoch 203/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 7.3644 - val_loss: 8.2015\n",
      "Epoch 204/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 6.7225 - val_loss: 8.1226\n",
      "Epoch 205/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 7.5716 - val_loss: 8.0357\n",
      "Epoch 206/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 6.8961 - val_loss: 7.9570\n",
      "Epoch 207/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 7.1536 - val_loss: 7.8768\n",
      "Epoch 208/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.6710 - val_loss: 7.7997\n",
      "Epoch 209/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 6.7381 - val_loss: 7.7241\n",
      "Epoch 210/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 6.1577 - val_loss: 7.6419\n",
      "Epoch 211/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.7750 - val_loss: 7.5582\n",
      "Epoch 212/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 6.4519 - val_loss: 7.4774\n",
      "Epoch 213/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.4833 - val_loss: 7.3981\n",
      "Epoch 214/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.3349 - val_loss: 7.3222\n",
      "Epoch 215/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.7488 - val_loss: 7.2411\n",
      "Epoch 216/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.0363 - val_loss: 7.1639\n",
      "Epoch 217/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 6.3910 - val_loss: 7.0867\n",
      "Epoch 218/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.1797 - val_loss: 7.0078\n",
      "Epoch 219/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.7150 - val_loss: 6.9324\n",
      "Epoch 220/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.9066 - val_loss: 6.8570\n",
      "Epoch 221/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.8922 - val_loss: 6.7821\n",
      "Epoch 222/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 6.1015 - val_loss: 6.7014\n",
      "Epoch 223/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.9983 - val_loss: 6.6200\n",
      "Epoch 224/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 5.9108 - val_loss: 6.5444\n",
      "Epoch 225/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 5.6104 - val_loss: 6.4666\n",
      "Epoch 226/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.7971 - val_loss: 6.3888\n",
      "Epoch 227/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 5.4141 - val_loss: 6.3127\n",
      "Epoch 228/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.4981 - val_loss: 6.2416\n",
      "Epoch 229/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 5.2750 - val_loss: 6.1666\n",
      "Epoch 230/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 5.3418 - val_loss: 6.0968\n",
      "Epoch 231/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.4980 - val_loss: 6.0259\n",
      "Epoch 232/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 5.4458 - val_loss: 5.9551\n",
      "Epoch 233/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.1914 - val_loss: 5.8850\n",
      "Epoch 234/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 5.0289 - val_loss: 5.8141\n",
      "Epoch 235/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.9476 - val_loss: 5.7461\n",
      "Epoch 236/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.7970 - val_loss: 5.6791\n",
      "Epoch 237/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.8953 - val_loss: 5.6128\n",
      "Epoch 238/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 5.1303 - val_loss: 5.5420\n",
      "Epoch 239/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.7737 - val_loss: 5.4682\n",
      "Epoch 240/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.8358 - val_loss: 5.4002\n",
      "Epoch 241/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 5.0287 - val_loss: 5.3331\n",
      "Epoch 242/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.7709 - val_loss: 5.2664\n",
      "Epoch 243/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.9682 - val_loss: 5.1982\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 244/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.2259 - val_loss: 5.1375\n",
      "Epoch 245/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.7146 - val_loss: 5.0768\n",
      "Epoch 246/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.5943 - val_loss: 5.0246\n",
      "Epoch 247/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.6733 - val_loss: 4.9718\n",
      "Epoch 248/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.2681 - val_loss: 4.9108\n",
      "Epoch 249/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.3756 - val_loss: 4.8482\n",
      "Epoch 250/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.0523 - val_loss: 4.7961\n",
      "Epoch 251/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.2825 - val_loss: 4.7383\n",
      "Epoch 252/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.0028 - val_loss: 4.6791\n",
      "Epoch 253/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.0055 - val_loss: 4.6286\n",
      "Epoch 254/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9646 - val_loss: 4.5785\n",
      "Epoch 255/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9769 - val_loss: 4.5156\n",
      "Epoch 256/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6347 - val_loss: 4.4606\n",
      "Epoch 257/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9455 - val_loss: 4.4006\n",
      "Epoch 258/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.8911 - val_loss: 4.3458\n",
      "Epoch 259/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5551 - val_loss: 4.2963\n",
      "Epoch 260/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 4.1045 - val_loss: 4.2425\n",
      "Epoch 261/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5217 - val_loss: 4.1923\n",
      "Epoch 262/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.6757 - val_loss: 4.1383\n",
      "Epoch 263/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.9187 - val_loss: 4.0946\n",
      "Epoch 264/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4983 - val_loss: 4.0433\n",
      "Epoch 265/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.2293 - val_loss: 4.0039\n",
      "Epoch 266/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4078 - val_loss: 3.9640\n",
      "Epoch 267/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.5402 - val_loss: 3.9184\n",
      "Epoch 268/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.3750 - val_loss: 3.8672\n",
      "Epoch 269/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.4595 - val_loss: 3.8241\n",
      "Epoch 270/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.2028 - val_loss: 3.7794\n",
      "Epoch 271/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.3642 - val_loss: 3.7435\n",
      "Epoch 272/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.3119 - val_loss: 3.7094\n",
      "Epoch 273/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1867 - val_loss: 3.6648\n",
      "Epoch 274/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9493 - val_loss: 3.6152\n",
      "Epoch 275/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1393 - val_loss: 3.5774\n",
      "Epoch 276/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0870 - val_loss: 3.5326\n",
      "Epoch 277/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0750 - val_loss: 3.5009\n",
      "Epoch 278/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0318 - val_loss: 3.4715\n",
      "Epoch 279/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0368 - val_loss: 3.4350\n",
      "Epoch 280/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9771 - val_loss: 3.4051\n",
      "Epoch 281/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8833 - val_loss: 3.3842\n",
      "Epoch 282/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0782 - val_loss: 3.3683\n",
      "Epoch 283/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8293 - val_loss: 3.3409\n",
      "Epoch 284/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.7712 - val_loss: 3.3166\n",
      "Epoch 285/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9949 - val_loss: 3.2949\n",
      "Epoch 286/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8886 - val_loss: 3.2758\n",
      "Epoch 287/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.6263 - val_loss: 3.2635\n",
      "Epoch 288/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9366 - val_loss: 3.2477\n",
      "Epoch 289/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0144 - val_loss: 3.2272\n",
      "Epoch 290/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.7453 - val_loss: 3.2058\n",
      "Epoch 291/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.5860 - val_loss: 3.1835\n",
      "Epoch 292/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.6479 - val_loss: 3.1664\n",
      "Epoch 293/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.7076 - val_loss: 3.1535\n",
      "Epoch 294/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.7053 - val_loss: 3.1368\n",
      "Epoch 295/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.6479 - val_loss: 3.1237\n",
      "Epoch 296/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.5002 - val_loss: 3.1149\n",
      "Epoch 297/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.6322 - val_loss: 3.1068\n",
      "Epoch 298/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.5784 - val_loss: 3.1017\n",
      "Epoch 299/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.5972 - val_loss: 3.0930\n",
      "Epoch 300/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.5291 - val_loss: 3.0792\n",
      "Epoch 301/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.6178 - val_loss: 3.0707\n",
      "Epoch 302/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.5562 - val_loss: 3.0726\n",
      "Epoch 303/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.6728 - val_loss: 3.0596\n",
      "Epoch 304/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.7632 - val_loss: 3.0475\n",
      "Epoch 305/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.5971 - val_loss: 3.0368\n",
      "Epoch 306/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.7127 - val_loss: 3.0281\n",
      "Epoch 307/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3672 - val_loss: 3.0118\n",
      "Epoch 308/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4686 - val_loss: 3.0041\n",
      "Epoch 309/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.7219 - val_loss: 3.0003\n",
      "Epoch 310/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.6039 - val_loss: 2.9909\n",
      "Epoch 311/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3475 - val_loss: 2.9923\n",
      "Epoch 312/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3858 - val_loss: 2.9827\n",
      "Epoch 313/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.4059 - val_loss: 2.9787\n",
      "Epoch 314/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2869 - val_loss: 2.9714\n",
      "Epoch 315/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.6127 - val_loss: 2.9705\n",
      "Epoch 316/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3673 - val_loss: 2.9659\n",
      "Epoch 317/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4541 - val_loss: 2.9615\n",
      "Epoch 318/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.5328 - val_loss: 2.9646\n",
      "Epoch 319/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2719 - val_loss: 2.9630\n",
      "Epoch 320/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3692 - val_loss: 2.9553\n",
      "Epoch 321/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4402 - val_loss: 2.9512\n",
      "Epoch 322/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3709 - val_loss: 2.9496\n",
      "Epoch 323/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2640 - val_loss: 2.9470\n",
      "Epoch 324/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.5262 - val_loss: 2.9472\n",
      "Epoch 325/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.5488 - val_loss: 2.9540\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 326/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4323 - val_loss: 2.9513\n",
      "Epoch 327/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2709 - val_loss: 2.9468\n",
      "Epoch 328/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3272 - val_loss: 2.9489\n",
      "Epoch 329/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3328 - val_loss: 2.9474\n",
      "Epoch 330/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0882 - val_loss: 2.9385\n",
      "Epoch 331/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2905 - val_loss: 2.9381\n",
      "Epoch 332/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3520 - val_loss: 2.9411\n",
      "Epoch 333/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4407 - val_loss: 2.9372\n",
      "Epoch 334/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3535 - val_loss: 2.9370\n",
      "Epoch 335/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.5089 - val_loss: 2.9329\n",
      "Epoch 336/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2836 - val_loss: 2.9291\n",
      "Epoch 337/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3279 - val_loss: 2.9301\n",
      "Epoch 338/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2900 - val_loss: 2.9301\n",
      "Epoch 339/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3232 - val_loss: 2.9280\n",
      "Epoch 340/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4396 - val_loss: 2.9349\n",
      "Epoch 341/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1664 - val_loss: 2.9304\n",
      "Epoch 342/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1831 - val_loss: 2.9279\n",
      "Epoch 343/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3434 - val_loss: 2.9292\n",
      "Epoch 344/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4012 - val_loss: 2.9290\n",
      "Epoch 345/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2069 - val_loss: 2.9300\n",
      "Epoch 346/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4795 - val_loss: 2.9324\n",
      "Epoch 347/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2453 - val_loss: 2.9269\n",
      "Epoch 348/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1635 - val_loss: 2.9301\n",
      "Epoch 349/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3848 - val_loss: 2.9314\n",
      "Epoch 350/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3737 - val_loss: 2.9262\n",
      "Epoch 351/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3101 - val_loss: 2.9283\n",
      "Epoch 352/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.0812 - val_loss: 2.9303\n",
      "Epoch 353/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.4242 - val_loss: 2.9291\n",
      "Epoch 354/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2964 - val_loss: 2.9221\n",
      "Epoch 355/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.4513 - val_loss: 2.9263\n",
      "Epoch 356/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2606 - val_loss: 2.9286\n",
      "Epoch 357/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.1273 - val_loss: 2.9257\n",
      "Epoch 358/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.1866 - val_loss: 2.9217\n",
      "Epoch 359/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.5158 - val_loss: 2.9277\n",
      "Epoch 360/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3173 - val_loss: 2.9299\n",
      "Epoch 361/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.6573 - val_loss: 2.9282\n",
      "Epoch 362/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.4389 - val_loss: 2.9223\n",
      "Epoch 363/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.5147 - val_loss: 2.9235\n",
      "Epoch 364/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.5165 - val_loss: 2.9222\n",
      "Epoch 365/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3692 - val_loss: 2.9236\n",
      "Epoch 366/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2426 - val_loss: 2.9245\n",
      "Epoch 367/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.1381 - val_loss: 2.9243\n",
      "Epoch 368/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3394 - val_loss: 2.9180\n",
      "Epoch 369/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2697 - val_loss: 2.9133\n",
      "Epoch 370/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3111 - val_loss: 2.9151\n",
      "Epoch 371/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2561 - val_loss: 2.9189\n",
      "Epoch 372/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2275 - val_loss: 2.9091\n",
      "Epoch 373/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.5443 - val_loss: 2.9075\n",
      "Epoch 374/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3773 - val_loss: 2.9084\n",
      "Epoch 375/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2210 - val_loss: 2.9099\n",
      "Epoch 376/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.4465 - val_loss: 2.9085\n",
      "Epoch 377/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.1979 - val_loss: 2.9038\n",
      "Epoch 378/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2911 - val_loss: 2.9104\n",
      "Epoch 379/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2108 - val_loss: 2.9035\n",
      "Epoch 380/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.4413 - val_loss: 2.9023\n",
      "Epoch 381/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1350 - val_loss: 2.9047\n",
      "Epoch 382/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3346 - val_loss: 2.9129\n",
      "Epoch 383/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1663 - val_loss: 2.9067\n",
      "Epoch 384/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3790 - val_loss: 2.9118\n",
      "Epoch 385/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3412 - val_loss: 2.9152\n",
      "Epoch 386/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3019 - val_loss: 2.9114\n",
      "Epoch 387/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4746 - val_loss: 2.9028\n",
      "Epoch 388/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2167 - val_loss: 2.9146\n",
      "Epoch 389/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3869 - val_loss: 2.9115\n",
      "Epoch 390/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2388 - val_loss: 2.9114\n",
      "Epoch 391/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1476 - val_loss: 2.9094\n",
      "Epoch 392/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3010 - val_loss: 2.9103\n",
      "Epoch 393/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0258 - val_loss: 2.9009\n",
      "Epoch 394/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2869 - val_loss: 2.8988\n",
      "Epoch 395/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4625 - val_loss: 2.9082\n",
      "Epoch 396/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2017 - val_loss: 2.9087\n",
      "Epoch 397/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3168 - val_loss: 2.9081\n",
      "Epoch 398/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3682 - val_loss: 2.9086\n",
      "Epoch 399/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4001 - val_loss: 2.9092\n",
      "Epoch 400/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2053 - val_loss: 2.9041\n",
      "Epoch 401/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4261 - val_loss: 2.9131\n",
      "Epoch 402/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4978 - val_loss: 2.9129\n",
      "Epoch 403/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4444 - val_loss: 2.9156\n",
      "Epoch 404/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.4089 - val_loss: 2.9123\n",
      "Epoch 405/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1840 - val_loss: 2.9077\n",
      "Epoch 406/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2807 - val_loss: 2.9084\n",
      "Epoch 407/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1405 - val_loss: 2.9092\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 408/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3330 - val_loss: 2.9072\n",
      "Epoch 409/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4731 - val_loss: 2.9042\n",
      "Epoch 410/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0498 - val_loss: 2.9031\n",
      "Epoch 411/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3829 - val_loss: 2.9142\n",
      "Epoch 412/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2491 - val_loss: 2.9167\n",
      "Epoch 413/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3051 - val_loss: 2.9104\n",
      "Epoch 414/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4435 - val_loss: 2.9123\n",
      "Epoch 415/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2071 - val_loss: 2.9078\n",
      "Epoch 416/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4010 - val_loss: 2.9115\n",
      "Epoch 417/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3112 - val_loss: 2.9155\n",
      "Epoch 418/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1201 - val_loss: 2.9074\n",
      "Epoch 419/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4117 - val_loss: 2.9087\n",
      "Epoch 420/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1545 - val_loss: 2.9067\n",
      "Epoch 421/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4159 - val_loss: 2.9060\n",
      "Epoch 422/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4451 - val_loss: 2.9148\n",
      "Epoch 423/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2304 - val_loss: 2.9118\n",
      "Epoch 424/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2750 - val_loss: 2.9087\n",
      "Epoch 425/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2661 - val_loss: 2.9095\n",
      "Epoch 426/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4058 - val_loss: 2.9165\n",
      "Epoch 427/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1042 - val_loss: 2.9086\n",
      "Epoch 428/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0535 - val_loss: 2.9111\n",
      "Epoch 429/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0812 - val_loss: 2.9205\n",
      "Epoch 430/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3138 - val_loss: 2.9140\n",
      "Epoch 431/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1370 - val_loss: 2.9075\n",
      "Epoch 432/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0944 - val_loss: 2.9142\n",
      "Epoch 433/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3680 - val_loss: 2.9178\n",
      "Epoch 434/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3136 - val_loss: 2.9137\n",
      "Epoch 435/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2735 - val_loss: 2.9102\n",
      "Epoch 436/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0874 - val_loss: 2.9156\n",
      "Epoch 437/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4427 - val_loss: 2.9135\n",
      "Epoch 438/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1423 - val_loss: 2.9061\n",
      "Epoch 439/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 2.2496 - val_loss: 2.9060\n",
      "Epoch 440/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2736 - val_loss: 2.9139\n",
      "Epoch 441/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3311 - val_loss: 2.9110\n",
      "Epoch 442/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3566 - val_loss: 2.9079\n",
      "Epoch 443/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3123 - val_loss: 2.9100\n",
      "Epoch 444/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0784 - val_loss: 2.9063\n",
      "Epoch 445/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1105 - val_loss: 2.9092\n",
      "Epoch 446/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0268 - val_loss: 2.9130\n",
      "Epoch 447/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2509 - val_loss: 2.9101\n",
      "Epoch 448/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0957 - val_loss: 2.9149\n",
      "Epoch 449/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1093 - val_loss: 2.9079\n",
      "Epoch 450/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2535 - val_loss: 2.9164\n",
      "Epoch 451/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3357 - val_loss: 2.9151\n",
      "Epoch 452/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4715 - val_loss: 2.9151\n",
      "Epoch 453/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2105 - val_loss: 2.9149\n",
      "Epoch 454/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1528 - val_loss: 2.9113\n",
      "Epoch 455/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2312 - val_loss: 2.9090\n",
      "Epoch 456/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2548 - val_loss: 2.9106\n",
      "Epoch 457/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1925 - val_loss: 2.9129\n",
      "Epoch 458/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1564 - val_loss: 2.9148\n",
      "Epoch 459/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2533 - val_loss: 2.9171\n",
      "Epoch 460/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0762 - val_loss: 2.9068\n",
      "Epoch 461/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3127 - val_loss: 2.9131\n",
      "Epoch 462/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2664 - val_loss: 2.9138\n",
      "Epoch 463/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3060 - val_loss: 2.9164\n",
      "Epoch 464/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4195 - val_loss: 2.9110\n",
      "Epoch 465/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4271 - val_loss: 2.9079\n",
      "Epoch 466/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3482 - val_loss: 2.9075\n",
      "Epoch 467/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2516 - val_loss: 2.9111\n",
      "Epoch 468/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2698 - val_loss: 2.9132\n",
      "Epoch 469/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0395 - val_loss: 2.9087\n",
      "Epoch 470/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3638 - val_loss: 2.9112\n",
      "Epoch 471/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.1822 - val_loss: 2.9205\n",
      "Epoch 472/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2425 - val_loss: 2.9144\n",
      "Epoch 473/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1781 - val_loss: 2.9119\n",
      "Epoch 474/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3198 - val_loss: 2.9155\n",
      "Epoch 475/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1959 - val_loss: 2.9135\n",
      "Epoch 476/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3168 - val_loss: 2.9158\n",
      "Epoch 477/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4110 - val_loss: 2.9146\n",
      "Epoch 478/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2453 - val_loss: 2.9171\n",
      "Epoch 479/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2464 - val_loss: 2.9145\n",
      "Epoch 480/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1690 - val_loss: 2.9073\n",
      "Epoch 481/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1356 - val_loss: 2.9071\n",
      "Epoch 482/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2513 - val_loss: 2.9138\n",
      "Epoch 483/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1892 - val_loss: 2.9154\n",
      "Epoch 484/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1613 - val_loss: 2.9108\n",
      "Epoch 485/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2207 - val_loss: 2.9088\n",
      "Epoch 486/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3778 - val_loss: 2.9105\n",
      "Epoch 487/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.1958 - val_loss: 2.9203\n",
      "Epoch 488/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3573 - val_loss: 2.9145\n",
      "Epoch 489/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3265 - val_loss: 2.9083\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 490/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2768 - val_loss: 2.9096\n",
      "Epoch 491/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2961 - val_loss: 2.9139\n",
      "Epoch 492/1000\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2.1836 - val_loss: 2.9163\n",
      "Epoch 493/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2881 - val_loss: 2.9203\n",
      "Epoch 494/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.1504 - val_loss: 2.9126\n",
      "Epoch 495/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2842 - val_loss: 2.9160\n",
      "Epoch 496/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2962 - val_loss: 2.9101\n",
      "Epoch 497/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2482 - val_loss: 2.9114\n",
      "Epoch 498/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.5200 - val_loss: 2.9148\n",
      "Epoch 499/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2357 - val_loss: 2.9144\n",
      "Epoch 500/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4167 - val_loss: 2.9124\n",
      "Epoch 501/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2684 - val_loss: 2.9127\n",
      "Epoch 502/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4003 - val_loss: 2.9208\n",
      "Epoch 503/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2318 - val_loss: 2.9164\n",
      "Epoch 504/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2690 - val_loss: 2.9150\n",
      "Epoch 505/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2331 - val_loss: 2.9173\n",
      "Epoch 506/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2617 - val_loss: 2.9240\n",
      "Epoch 507/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1603 - val_loss: 2.9159\n",
      "Epoch 508/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1509 - val_loss: 2.9192\n",
      "Epoch 509/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2946 - val_loss: 2.9209\n",
      "Epoch 510/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1736 - val_loss: 2.9227\n",
      "Epoch 511/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3135 - val_loss: 2.9176\n",
      "Epoch 512/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1290 - val_loss: 2.9190\n",
      "Epoch 513/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2951 - val_loss: 2.9217\n",
      "Epoch 514/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2614 - val_loss: 2.9173\n",
      "Epoch 515/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2640 - val_loss: 2.9104\n",
      "Epoch 516/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1107 - val_loss: 2.9196\n",
      "Epoch 517/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4199 - val_loss: 2.9239\n",
      "Epoch 518/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1039 - val_loss: 2.9140\n",
      "Epoch 519/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1220 - val_loss: 2.9132\n",
      "Epoch 520/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2426 - val_loss: 2.9160\n",
      "Epoch 521/1000\n",
      "8/8 [==============================] - 0s 6ms/step - loss: 2.1486 - val_loss: 2.9224\n",
      "Epoch 522/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4190 - val_loss: 2.9208\n",
      "Epoch 523/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1653 - val_loss: 2.9244\n",
      "Epoch 524/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2708 - val_loss: 2.9200\n",
      "Epoch 525/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1448 - val_loss: 2.9252\n",
      "Epoch 526/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3456 - val_loss: 2.9243\n",
      "Epoch 527/1000\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2.3972 - val_loss: 2.9207\n",
      "Epoch 528/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2881 - val_loss: 2.9196\n",
      "Epoch 529/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1757 - val_loss: 2.9155\n",
      "Epoch 530/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2787 - val_loss: 2.9253\n",
      "Epoch 531/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2637 - val_loss: 2.9261\n",
      "Epoch 532/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3089 - val_loss: 2.9208\n",
      "Epoch 533/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2264 - val_loss: 2.9171\n",
      "Epoch 534/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2580 - val_loss: 2.9227\n",
      "Epoch 535/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0923 - val_loss: 2.9262\n",
      "Epoch 536/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4180 - val_loss: 2.9253\n",
      "Epoch 537/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4228 - val_loss: 2.9233\n",
      "Epoch 538/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2866 - val_loss: 2.9250\n",
      "Epoch 539/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2342 - val_loss: 2.9236\n",
      "Epoch 540/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0773 - val_loss: 2.9256\n",
      "Epoch 541/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2745 - val_loss: 2.9230\n",
      "Epoch 542/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2928 - val_loss: 2.9246\n",
      "Epoch 543/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 2.3224 - val_loss: 2.9273\n",
      "Epoch 544/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2129 - val_loss: 2.9235\n",
      "Epoch 545/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2065 - val_loss: 2.9209\n",
      "Epoch 546/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3297 - val_loss: 2.9297\n",
      "Epoch 547/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2849 - val_loss: 2.9233\n",
      "Epoch 548/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.1000 - val_loss: 2.9235\n",
      "Epoch 549/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1816 - val_loss: 2.9246\n",
      "Epoch 550/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2663 - val_loss: 2.9264\n",
      "Epoch 551/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2234 - val_loss: 2.9231\n",
      "Epoch 552/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2251 - val_loss: 2.9237\n",
      "Epoch 553/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.1207 - val_loss: 2.9214\n",
      "Epoch 554/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3479 - val_loss: 2.9242\n",
      "Epoch 555/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3501 - val_loss: 2.9249\n",
      "Epoch 556/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.1900 - val_loss: 2.9247\n",
      "Epoch 557/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3972 - val_loss: 2.9295\n",
      "Epoch 558/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3055 - val_loss: 2.9308\n",
      "Epoch 559/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.1680 - val_loss: 2.9234\n",
      "Epoch 560/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2998 - val_loss: 2.9243\n",
      "Epoch 561/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.0898 - val_loss: 2.9156\n",
      "Epoch 562/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3509 - val_loss: 2.9186\n",
      "Epoch 563/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2639 - val_loss: 2.9211\n",
      "Epoch 564/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.0935 - val_loss: 2.9193\n",
      "Epoch 565/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.5546 - val_loss: 2.9273\n",
      "Epoch 566/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.1563 - val_loss: 2.9196\n",
      "Epoch 567/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3713 - val_loss: 2.9164\n",
      "Epoch 568/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0832 - val_loss: 2.9270\n",
      "Epoch 569/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.0414 - val_loss: 2.9275\n",
      "Epoch 570/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3546 - val_loss: 2.9215\n",
      "Epoch 571/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2468 - val_loss: 2.9216\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 572/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3152 - val_loss: 2.9176\n",
      "Epoch 573/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4126 - val_loss: 2.9286\n",
      "Epoch 574/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3509 - val_loss: 2.9366\n",
      "Epoch 575/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2633 - val_loss: 2.9205\n",
      "Epoch 576/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3585 - val_loss: 2.9211\n",
      "Epoch 577/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0990 - val_loss: 2.9211\n",
      "Epoch 578/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3175 - val_loss: 2.9260\n",
      "Epoch 579/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3967 - val_loss: 2.9282\n",
      "Epoch 580/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4552 - val_loss: 2.9263\n",
      "Epoch 581/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3475 - val_loss: 2.9193\n",
      "Epoch 582/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2677 - val_loss: 2.9167\n",
      "Epoch 583/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3324 - val_loss: 2.9192\n",
      "Epoch 584/1000\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 2.1440 - val_loss: 2.9261\n",
      "Epoch 585/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2006 - val_loss: 2.9260\n",
      "Epoch 586/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2404 - val_loss: 2.9273\n",
      "Epoch 587/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3272 - val_loss: 2.9265\n",
      "Epoch 588/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3355 - val_loss: 2.9306\n",
      "Epoch 589/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3592 - val_loss: 2.9301\n",
      "Epoch 590/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3013 - val_loss: 2.9243\n",
      "Epoch 591/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3284 - val_loss: 2.9252\n",
      "Epoch 592/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3033 - val_loss: 2.9187\n",
      "Epoch 593/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3940 - val_loss: 2.9241\n",
      "Epoch 594/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4567 - val_loss: 2.9299\n",
      "Epoch 595/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3929 - val_loss: 2.9285\n",
      "Epoch 596/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2830 - val_loss: 2.9322\n",
      "Epoch 597/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1364 - val_loss: 2.9287\n",
      "Epoch 598/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2819 - val_loss: 2.9163\n",
      "Epoch 599/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2802 - val_loss: 2.9268\n",
      "Epoch 600/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3427 - val_loss: 2.9276\n",
      "Epoch 601/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3041 - val_loss: 2.9293\n",
      "Epoch 602/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1395 - val_loss: 2.9280\n",
      "Epoch 603/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2490 - val_loss: 2.9269\n",
      "Epoch 604/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2858 - val_loss: 2.9295\n",
      "Epoch 605/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3035 - val_loss: 2.9265\n",
      "Epoch 606/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1569 - val_loss: 2.9246\n",
      "Epoch 607/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2404 - val_loss: 2.9290\n",
      "Epoch 608/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3091 - val_loss: 2.9306\n",
      "Epoch 609/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3049 - val_loss: 2.9207\n",
      "Epoch 610/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0787 - val_loss: 2.9273\n",
      "Epoch 611/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2484 - val_loss: 2.9271\n",
      "Epoch 612/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1645 - val_loss: 2.9264\n",
      "Epoch 613/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2970 - val_loss: 2.9265\n",
      "Epoch 614/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.9955 - val_loss: 2.9254\n",
      "Epoch 615/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3131 - val_loss: 2.9257\n",
      "Epoch 616/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1260 - val_loss: 2.9221\n",
      "Epoch 617/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2890 - val_loss: 2.9341\n",
      "Epoch 618/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4357 - val_loss: 2.9361\n",
      "Epoch 619/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3214 - val_loss: 2.9309\n",
      "Epoch 620/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2564 - val_loss: 2.9248\n",
      "Epoch 621/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2793 - val_loss: 2.9303\n",
      "Epoch 622/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1636 - val_loss: 2.9345\n",
      "Epoch 623/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1930 - val_loss: 2.9345\n",
      "Epoch 624/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2013 - val_loss: 2.9230\n",
      "Epoch 625/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2396 - val_loss: 2.9224\n",
      "Epoch 626/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3064 - val_loss: 2.9299\n",
      "Epoch 627/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2294 - val_loss: 2.9338\n",
      "Epoch 628/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2203 - val_loss: 2.9322\n",
      "Epoch 629/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3485 - val_loss: 2.9279\n",
      "Epoch 630/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2035 - val_loss: 2.9259\n",
      "Epoch 631/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2369 - val_loss: 2.9335\n",
      "Epoch 632/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.4323 - val_loss: 2.9349\n",
      "Epoch 633/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1851 - val_loss: 2.9318\n",
      "Epoch 634/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2793 - val_loss: 2.9273\n",
      "Epoch 635/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1809 - val_loss: 2.9307\n",
      "Epoch 636/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4093 - val_loss: 2.9411\n",
      "Epoch 637/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1235 - val_loss: 2.9354\n",
      "Epoch 638/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3536 - val_loss: 2.9320\n",
      "Epoch 639/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1748 - val_loss: 2.9348\n",
      "Epoch 640/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2499 - val_loss: 2.9296\n",
      "Epoch 641/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3698 - val_loss: 2.9311\n",
      "Epoch 642/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.0839 - val_loss: 2.9291\n",
      "Epoch 643/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2075 - val_loss: 2.9288\n",
      "Epoch 644/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3617 - val_loss: 2.9340\n",
      "Epoch 645/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1001 - val_loss: 2.9306\n",
      "Epoch 646/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2784 - val_loss: 2.9256\n",
      "Epoch 647/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1521 - val_loss: 2.9230\n",
      "Epoch 648/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2964 - val_loss: 2.9327\n",
      "Epoch 649/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3186 - val_loss: 2.9351\n",
      "Epoch 650/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2361 - val_loss: 2.9309\n",
      "Epoch 651/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2014 - val_loss: 2.9278\n",
      "Epoch 652/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3430 - val_loss: 2.9318\n",
      "Epoch 653/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0413 - val_loss: 2.9327\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 654/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1962 - val_loss: 2.9327\n",
      "Epoch 655/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3060 - val_loss: 2.9332\n",
      "Epoch 656/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2179 - val_loss: 2.9282\n",
      "Epoch 657/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3155 - val_loss: 2.9277\n",
      "Epoch 658/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2760 - val_loss: 2.9347\n",
      "Epoch 659/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3245 - val_loss: 2.9315\n",
      "Epoch 660/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4296 - val_loss: 2.9320\n",
      "Epoch 661/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1606 - val_loss: 2.9287\n",
      "Epoch 662/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2392 - val_loss: 2.9271\n",
      "Epoch 663/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1617 - val_loss: 2.9359\n",
      "Epoch 664/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2631 - val_loss: 2.9411\n",
      "Epoch 665/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2342 - val_loss: 2.9359\n",
      "Epoch 666/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4228 - val_loss: 2.9304\n",
      "Epoch 667/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3049 - val_loss: 2.9287\n",
      "Epoch 668/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2223 - val_loss: 2.9314\n",
      "Epoch 669/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1238 - val_loss: 2.9364\n",
      "Epoch 670/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2390 - val_loss: 2.9383\n",
      "Epoch 671/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2650 - val_loss: 2.9338\n",
      "Epoch 672/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3829 - val_loss: 2.9412\n",
      "Epoch 673/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3794 - val_loss: 2.9398\n",
      "Epoch 674/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2674 - val_loss: 2.9367\n",
      "Epoch 675/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2153 - val_loss: 2.9406\n",
      "Epoch 676/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3274 - val_loss: 2.9398\n",
      "Epoch 677/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3355 - val_loss: 2.9362\n",
      "Epoch 678/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1557 - val_loss: 2.9322\n",
      "Epoch 679/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2438 - val_loss: 2.9337\n",
      "Epoch 680/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4070 - val_loss: 2.9351\n",
      "Epoch 681/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3205 - val_loss: 2.9420\n",
      "Epoch 682/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1548 - val_loss: 2.9424\n",
      "Epoch 683/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0748 - val_loss: 2.9344\n",
      "Epoch 684/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1951 - val_loss: 2.9297\n",
      "Epoch 685/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3133 - val_loss: 2.9391\n",
      "Epoch 686/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3017 - val_loss: 2.9418\n",
      "Epoch 687/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2865 - val_loss: 2.9434\n",
      "Epoch 688/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3061 - val_loss: 2.9370\n",
      "Epoch 689/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4194 - val_loss: 2.9328\n",
      "Epoch 690/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3498 - val_loss: 2.9328\n",
      "Epoch 691/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2113 - val_loss: 2.9291\n",
      "Epoch 692/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1853 - val_loss: 2.9328\n",
      "Epoch 693/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0547 - val_loss: 2.9376\n",
      "Epoch 694/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2195 - val_loss: 2.9357\n",
      "Epoch 695/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2782 - val_loss: 2.9315\n",
      "Epoch 696/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2185 - val_loss: 2.9359\n",
      "Epoch 697/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2795 - val_loss: 2.9374\n",
      "Epoch 698/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1514 - val_loss: 2.9380\n",
      "Epoch 699/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0923 - val_loss: 2.9367\n",
      "Epoch 700/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2696 - val_loss: 2.9341\n",
      "Epoch 701/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3481 - val_loss: 2.9376\n",
      "Epoch 702/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3080 - val_loss: 2.9365\n",
      "Epoch 703/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1910 - val_loss: 2.9341\n",
      "Epoch 704/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2732 - val_loss: 2.9408\n",
      "Epoch 705/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2987 - val_loss: 2.9360\n",
      "Epoch 706/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3557 - val_loss: 2.9296\n",
      "Epoch 707/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2517 - val_loss: 2.9399\n",
      "Epoch 708/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3124 - val_loss: 2.9387\n",
      "Epoch 709/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1696 - val_loss: 2.9348\n",
      "Epoch 710/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3379 - val_loss: 2.9319\n",
      "Epoch 711/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2170 - val_loss: 2.9302\n",
      "Epoch 712/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.3978 - val_loss: 2.9372\n",
      "Epoch 713/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2411 - val_loss: 2.9390\n",
      "Epoch 714/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1518 - val_loss: 2.9344\n",
      "Epoch 715/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.1514 - val_loss: 2.9391\n",
      "Epoch 716/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2517 - val_loss: 2.9359\n",
      "Epoch 717/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2302 - val_loss: 2.9413\n",
      "Epoch 718/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0995 - val_loss: 2.9394\n",
      "Epoch 719/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1569 - val_loss: 2.9375\n",
      "Epoch 720/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1921 - val_loss: 2.9332\n",
      "Epoch 721/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1750 - val_loss: 2.9383\n",
      "Epoch 722/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1049 - val_loss: 2.9351\n",
      "Epoch 723/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0534 - val_loss: 2.9372\n",
      "Epoch 724/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2363 - val_loss: 2.9383\n",
      "Epoch 725/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2173 - val_loss: 2.9424\n",
      "Epoch 726/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3578 - val_loss: 2.9381\n",
      "Epoch 727/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4004 - val_loss: 2.9362\n",
      "Epoch 728/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2953 - val_loss: 2.9331\n",
      "Epoch 729/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2141 - val_loss: 2.9379\n",
      "Epoch 730/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2612 - val_loss: 2.9306\n",
      "Epoch 731/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2383 - val_loss: 2.9301\n",
      "Epoch 732/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1731 - val_loss: 2.9360\n",
      "Epoch 733/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4490 - val_loss: 2.9361\n",
      "Epoch 734/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1690 - val_loss: 2.9360\n",
      "Epoch 735/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3672 - val_loss: 2.9371\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 736/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1734 - val_loss: 2.9398\n",
      "Epoch 737/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1056 - val_loss: 2.9287\n",
      "Epoch 738/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1541 - val_loss: 2.9313\n",
      "Epoch 739/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1289 - val_loss: 2.9372\n",
      "Epoch 740/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4270 - val_loss: 2.9410\n",
      "Epoch 741/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.5358 - val_loss: 2.9408\n",
      "Epoch 742/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1804 - val_loss: 2.9315\n",
      "Epoch 743/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2942 - val_loss: 2.9302\n",
      "Epoch 744/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4276 - val_loss: 2.9332\n",
      "Epoch 745/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1887 - val_loss: 2.9354\n",
      "Epoch 746/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4599 - val_loss: 2.9351\n",
      "Epoch 747/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2435 - val_loss: 2.9369\n",
      "Epoch 748/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3199 - val_loss: 2.9359\n",
      "Epoch 749/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3048 - val_loss: 2.9392\n",
      "Epoch 750/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2824 - val_loss: 2.9450\n",
      "Epoch 751/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1103 - val_loss: 2.9439\n",
      "Epoch 752/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3608 - val_loss: 2.9422\n",
      "Epoch 753/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2825 - val_loss: 2.9362\n",
      "Epoch 754/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2154 - val_loss: 2.9313\n",
      "Epoch 755/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2206 - val_loss: 2.9356\n",
      "Epoch 756/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1326 - val_loss: 2.9348\n",
      "Epoch 757/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1643 - val_loss: 2.9432\n",
      "Epoch 758/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2851 - val_loss: 2.9414\n",
      "Epoch 759/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.5290 - val_loss: 2.9409\n",
      "Epoch 760/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2978 - val_loss: 2.9378\n",
      "Epoch 761/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2034 - val_loss: 2.9387\n",
      "Epoch 762/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3499 - val_loss: 2.9349\n",
      "Epoch 763/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2903 - val_loss: 2.9346\n",
      "Epoch 764/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3040 - val_loss: 2.9267\n",
      "Epoch 765/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1175 - val_loss: 2.9345\n",
      "Epoch 766/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4370 - val_loss: 2.9369\n",
      "Epoch 767/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4362 - val_loss: 2.9305\n",
      "Epoch 768/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2181 - val_loss: 2.9356\n",
      "Epoch 769/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1743 - val_loss: 2.9442\n",
      "Epoch 770/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1848 - val_loss: 2.9459\n",
      "Epoch 771/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2834 - val_loss: 2.9391\n",
      "Epoch 772/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2835 - val_loss: 2.9387\n",
      "Epoch 773/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2163 - val_loss: 2.9326\n",
      "Epoch 774/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0375 - val_loss: 2.9417\n",
      "Epoch 775/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2717 - val_loss: 2.9370\n",
      "Epoch 776/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3606 - val_loss: 2.9352\n",
      "Epoch 777/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1996 - val_loss: 2.9439\n",
      "Epoch 778/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2979 - val_loss: 2.9419\n",
      "Epoch 779/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2987 - val_loss: 2.9406\n",
      "Epoch 780/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2754 - val_loss: 2.9349\n",
      "Epoch 781/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3609 - val_loss: 2.9374\n",
      "Epoch 782/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3078 - val_loss: 2.9453\n",
      "Epoch 783/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1878 - val_loss: 2.9419\n",
      "Epoch 784/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2341 - val_loss: 2.9333\n",
      "Epoch 785/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2763 - val_loss: 2.9400\n",
      "Epoch 786/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2321 - val_loss: 2.9455\n",
      "Epoch 787/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1758 - val_loss: 2.9430\n",
      "Epoch 788/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1457 - val_loss: 2.9372\n",
      "Epoch 789/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3059 - val_loss: 2.9387\n",
      "Epoch 790/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2850 - val_loss: 2.9378\n",
      "Epoch 791/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3186 - val_loss: 2.9370\n",
      "Epoch 792/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2613 - val_loss: 2.9391\n",
      "Epoch 793/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1068 - val_loss: 2.9480\n",
      "Epoch 794/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2561 - val_loss: 2.9431\n",
      "Epoch 795/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3172 - val_loss: 2.9426\n",
      "Epoch 796/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1703 - val_loss: 2.9413\n",
      "Epoch 797/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2766 - val_loss: 2.9381\n",
      "Epoch 798/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1689 - val_loss: 2.9324\n",
      "Epoch 799/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3610 - val_loss: 2.9406\n",
      "Epoch 800/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1684 - val_loss: 2.9374\n",
      "Epoch 801/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4219 - val_loss: 2.9385\n",
      "Epoch 802/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3124 - val_loss: 2.9432\n",
      "Epoch 803/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1247 - val_loss: 2.9401\n",
      "Epoch 804/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3227 - val_loss: 2.9396\n",
      "Epoch 805/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0001 - val_loss: 2.9343\n",
      "Epoch 806/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3670 - val_loss: 2.9383\n",
      "Epoch 807/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3034 - val_loss: 2.9421\n",
      "Epoch 808/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4712 - val_loss: 2.9400\n",
      "Epoch 809/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2550 - val_loss: 2.9332\n",
      "Epoch 810/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1255 - val_loss: 2.9335\n",
      "Epoch 811/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1871 - val_loss: 2.9348\n",
      "Epoch 812/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1813 - val_loss: 2.9444\n",
      "Epoch 813/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3776 - val_loss: 2.9436\n",
      "Epoch 814/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2647 - val_loss: 2.9409\n",
      "Epoch 815/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1527 - val_loss: 2.9422\n",
      "Epoch 816/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4925 - val_loss: 2.9362\n",
      "Epoch 817/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1647 - val_loss: 2.9408\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 818/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1068 - val_loss: 2.9418\n",
      "Epoch 819/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2032 - val_loss: 2.9386\n",
      "Epoch 820/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2035 - val_loss: 2.9337\n",
      "Epoch 821/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1571 - val_loss: 2.9348\n",
      "Epoch 822/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1939 - val_loss: 2.9486\n",
      "Epoch 823/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4315 - val_loss: 2.9430\n",
      "Epoch 824/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1277 - val_loss: 2.9467\n",
      "Epoch 825/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2150 - val_loss: 2.9408\n",
      "Epoch 826/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1917 - val_loss: 2.9381\n",
      "Epoch 827/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2349 - val_loss: 2.9440\n",
      "Epoch 828/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4984 - val_loss: 2.9499\n",
      "Epoch 829/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2550 - val_loss: 2.9407\n",
      "Epoch 830/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2593 - val_loss: 2.9358\n",
      "Epoch 831/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2483 - val_loss: 2.9395\n",
      "Epoch 832/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1172 - val_loss: 2.9389\n",
      "Epoch 833/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2077 - val_loss: 2.9385\n",
      "Epoch 834/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3617 - val_loss: 2.9352\n",
      "Epoch 835/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4812 - val_loss: 2.9412\n",
      "Epoch 836/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1008 - val_loss: 2.9388\n",
      "Epoch 837/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1896 - val_loss: 2.9388\n",
      "Epoch 838/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3799 - val_loss: 2.9412\n",
      "Epoch 839/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2433 - val_loss: 2.9392\n",
      "Epoch 840/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2604 - val_loss: 2.9351\n",
      "Epoch 841/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2892 - val_loss: 2.9396\n",
      "Epoch 842/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2744 - val_loss: 2.9373\n",
      "Epoch 843/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2264 - val_loss: 2.9377\n",
      "Epoch 844/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4255 - val_loss: 2.9358\n",
      "Epoch 845/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2001 - val_loss: 2.9313\n",
      "Epoch 846/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3406 - val_loss: 2.9356\n",
      "Epoch 847/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3225 - val_loss: 2.9395\n",
      "Epoch 848/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0927 - val_loss: 2.9381\n",
      "Epoch 849/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2907 - val_loss: 2.9413\n",
      "Epoch 850/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3841 - val_loss: 2.9318\n",
      "Epoch 851/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2033 - val_loss: 2.9363\n",
      "Epoch 852/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2329 - val_loss: 2.9411\n",
      "Epoch 853/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2352 - val_loss: 2.9396\n",
      "Epoch 854/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1587 - val_loss: 2.9354\n",
      "Epoch 855/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3592 - val_loss: 2.9358\n",
      "Epoch 856/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3240 - val_loss: 2.9408\n",
      "Epoch 857/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1143 - val_loss: 2.9472\n",
      "Epoch 858/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2223 - val_loss: 2.9355\n",
      "Epoch 859/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2377 - val_loss: 2.9373\n",
      "Epoch 860/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2564 - val_loss: 2.9437\n",
      "Epoch 861/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1914 - val_loss: 2.9448\n",
      "Epoch 862/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2345 - val_loss: 2.9440\n",
      "Epoch 863/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2909 - val_loss: 2.9376\n",
      "Epoch 864/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2897 - val_loss: 2.9374\n",
      "Epoch 865/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3020 - val_loss: 2.9347\n",
      "Epoch 866/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2502 - val_loss: 2.9381\n",
      "Epoch 867/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2833 - val_loss: 2.9457\n",
      "Epoch 868/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2309 - val_loss: 2.9471\n",
      "Epoch 869/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3542 - val_loss: 2.9396\n",
      "Epoch 870/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4394 - val_loss: 2.9415\n",
      "Epoch 871/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.4192 - val_loss: 2.9417\n",
      "Epoch 872/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1891 - val_loss: 2.9434\n",
      "Epoch 873/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3311 - val_loss: 2.9393\n",
      "Epoch 874/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4178 - val_loss: 2.9423\n",
      "Epoch 875/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3749 - val_loss: 2.9405\n",
      "Epoch 876/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3008 - val_loss: 2.9421\n",
      "Epoch 877/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3332 - val_loss: 2.9453\n",
      "Epoch 878/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3534 - val_loss: 2.9326\n",
      "Epoch 879/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1665 - val_loss: 2.9326\n",
      "Epoch 880/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2532 - val_loss: 2.9412\n",
      "Epoch 881/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2570 - val_loss: 2.9389\n",
      "Epoch 882/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4625 - val_loss: 2.9370\n",
      "Epoch 883/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4057 - val_loss: 2.9525\n",
      "Epoch 884/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1998 - val_loss: 2.9456\n",
      "Epoch 885/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2341 - val_loss: 2.9393\n",
      "Epoch 886/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3816 - val_loss: 2.9470\n",
      "Epoch 887/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3639 - val_loss: 2.9443\n",
      "Epoch 888/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2705 - val_loss: 2.9411\n",
      "Epoch 889/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2298 - val_loss: 2.9402\n",
      "Epoch 890/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1021 - val_loss: 2.9478\n",
      "Epoch 891/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1615 - val_loss: 2.9498\n",
      "Epoch 892/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2775 - val_loss: 2.9475\n",
      "Epoch 893/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2272 - val_loss: 2.9415\n",
      "Epoch 894/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1806 - val_loss: 2.9385\n",
      "Epoch 895/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0459 - val_loss: 2.9408\n",
      "Epoch 896/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2286 - val_loss: 2.9427\n",
      "Epoch 897/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0825 - val_loss: 2.9390\n",
      "Epoch 898/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2071 - val_loss: 2.9398\n",
      "Epoch 899/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2517 - val_loss: 2.9422\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 900/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2919 - val_loss: 2.9425\n",
      "Epoch 901/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.5227 - val_loss: 2.9373\n",
      "Epoch 902/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2540 - val_loss: 2.9436\n",
      "Epoch 903/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2541 - val_loss: 2.9386\n",
      "Epoch 904/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3203 - val_loss: 2.9413\n",
      "Epoch 905/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3006 - val_loss: 2.9411\n",
      "Epoch 906/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4383 - val_loss: 2.9368\n",
      "Epoch 907/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3360 - val_loss: 2.9410\n",
      "Epoch 908/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2862 - val_loss: 2.9426\n",
      "Epoch 909/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3998 - val_loss: 2.9355\n",
      "Epoch 910/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3059 - val_loss: 2.9392\n",
      "Epoch 911/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3993 - val_loss: 2.9445\n",
      "Epoch 912/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2656 - val_loss: 2.9472\n",
      "Epoch 913/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3066 - val_loss: 2.9492\n",
      "Epoch 914/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1395 - val_loss: 2.9457\n",
      "Epoch 915/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.5370 - val_loss: 2.9499\n",
      "Epoch 916/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3114 - val_loss: 2.9482\n",
      "Epoch 917/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1599 - val_loss: 2.9421\n",
      "Epoch 918/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2155 - val_loss: 2.9391\n",
      "Epoch 919/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2072 - val_loss: 2.9441\n",
      "Epoch 920/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3407 - val_loss: 2.9462\n",
      "Epoch 921/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2136 - val_loss: 2.9417\n",
      "Epoch 922/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.6054 - val_loss: 2.9425\n",
      "Epoch 923/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2304 - val_loss: 2.9433\n",
      "Epoch 924/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2661 - val_loss: 2.9430\n",
      "Epoch 925/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0488 - val_loss: 2.9400\n",
      "Epoch 926/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0750 - val_loss: 2.9401\n",
      "Epoch 927/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3046 - val_loss: 2.9439\n",
      "Epoch 928/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2725 - val_loss: 2.9469\n",
      "Epoch 929/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0638 - val_loss: 2.9469\n",
      "Epoch 930/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.5064 - val_loss: 2.9443\n",
      "Epoch 931/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1329 - val_loss: 2.9429\n",
      "Epoch 932/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3850 - val_loss: 2.9401\n",
      "Epoch 933/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2139 - val_loss: 2.9445\n",
      "Epoch 934/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4652 - val_loss: 2.9465\n",
      "Epoch 935/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3499 - val_loss: 2.9466\n",
      "Epoch 936/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3123 - val_loss: 2.9416\n",
      "Epoch 937/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1869 - val_loss: 2.9386\n",
      "Epoch 938/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2089 - val_loss: 2.9437\n",
      "Epoch 939/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1768 - val_loss: 2.9543\n",
      "Epoch 940/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2439 - val_loss: 2.9565\n",
      "Epoch 941/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1689 - val_loss: 2.9500\n",
      "Epoch 942/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3991 - val_loss: 2.9456\n",
      "Epoch 943/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2974 - val_loss: 2.9464\n",
      "Epoch 944/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0526 - val_loss: 2.9537\n",
      "Epoch 945/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4777 - val_loss: 2.9592\n",
      "Epoch 946/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0921 - val_loss: 2.9530\n",
      "Epoch 947/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0861 - val_loss: 2.9440\n",
      "Epoch 948/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.5024 - val_loss: 2.9432\n",
      "Epoch 949/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2589 - val_loss: 2.9375\n",
      "Epoch 950/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2428 - val_loss: 2.9404\n",
      "Epoch 951/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3378 - val_loss: 2.9497\n",
      "Epoch 952/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3320 - val_loss: 2.9581\n",
      "Epoch 953/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2720 - val_loss: 2.9516\n",
      "Epoch 954/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2621 - val_loss: 2.9420\n",
      "Epoch 955/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0504 - val_loss: 2.9468\n",
      "Epoch 956/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3579 - val_loss: 2.9463\n",
      "Epoch 957/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3139 - val_loss: 2.9574\n",
      "Epoch 958/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2961 - val_loss: 2.9524\n",
      "Epoch 959/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3582 - val_loss: 2.9503\n",
      "Epoch 960/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2336 - val_loss: 2.9472\n",
      "Epoch 961/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2161 - val_loss: 2.9450\n",
      "Epoch 962/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2510 - val_loss: 2.9418\n",
      "Epoch 963/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0797 - val_loss: 2.9398\n",
      "Epoch 964/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1946 - val_loss: 2.9494\n",
      "Epoch 965/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1898 - val_loss: 2.9548\n",
      "Epoch 966/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2836 - val_loss: 2.9500\n",
      "Epoch 967/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2736 - val_loss: 2.9440\n",
      "Epoch 968/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.0718 - val_loss: 2.9448\n",
      "Epoch 969/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2978 - val_loss: 2.9542\n",
      "Epoch 970/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2333 - val_loss: 2.9536\n",
      "Epoch 971/1000\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2084 - val_loss: 2.9521\n",
      "Epoch 972/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1287 - val_loss: 2.9480\n",
      "Epoch 973/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3414 - val_loss: 2.9516\n",
      "Epoch 974/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2924 - val_loss: 2.9504\n",
      "Epoch 975/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2445 - val_loss: 2.9466\n",
      "Epoch 976/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3632 - val_loss: 2.9474\n",
      "Epoch 977/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3513 - val_loss: 2.9507\n",
      "Epoch 978/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3665 - val_loss: 2.9462\n",
      "Epoch 979/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2659 - val_loss: 2.9510\n",
      "Epoch 980/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3346 - val_loss: 2.9450\n",
      "Epoch 981/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1821 - val_loss: 2.9505\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 982/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2631 - val_loss: 2.9465\n",
      "Epoch 983/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2760 - val_loss: 2.9492\n",
      "Epoch 984/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2538 - val_loss: 2.9490\n",
      "Epoch 985/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3612 - val_loss: 2.9505\n",
      "Epoch 986/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2534 - val_loss: 2.9565\n",
      "Epoch 987/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4147 - val_loss: 2.9502\n",
      "Epoch 988/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3656 - val_loss: 2.9516\n",
      "Epoch 989/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1402 - val_loss: 2.9457\n",
      "Epoch 990/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.4505 - val_loss: 2.9444\n",
      "Epoch 991/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2139 - val_loss: 2.9541\n",
      "Epoch 992/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3095 - val_loss: 2.9548\n",
      "Epoch 993/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3215 - val_loss: 2.9590\n",
      "Epoch 994/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2340 - val_loss: 2.9494\n",
      "Epoch 995/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3739 - val_loss: 2.9494\n",
      "Epoch 996/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.5446 - val_loss: 2.9503\n",
      "Epoch 997/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3101 - val_loss: 2.9447\n",
      "Epoch 998/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2965 - val_loss: 2.9538\n",
      "Epoch 999/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.2820 - val_loss: 2.9597\n",
      "Epoch 1000/1000\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.1850 - val_loss: 2.9604\n"
     ]
    }
   ],
   "source": [
    "#For multiple features\n",
    "linear_model = tf.keras.Sequential([normalizer,layers.Dense(units=1, use_bias = 'True')])\n",
    "#Note that there is no activation in our last layer as we're performing regression!\n",
    "#No 'hidden' layers as it's LINEAR regression (activation functions in hidden layers introduce non linearity)\n",
    "linear_model.compile(\n",
    "    optimizer=tf.optimizers.Adam(learning_rate=0.01),\n",
    "    loss='mean_absolute_error')\n",
    "#%%time\n",
    "history = linear_model.fit(train_features, train_labels, epochs=1000, verbose = 1, validation_split = 0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEGCAYAAAB/+QKOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deXwc5ZXo/d+p6m61Nq/C8go2q9ltbJYZwIhACCEEJoQl3EyCCcQJIQNkJgt5894LJMy8mZkMJNzwgZABksmbsIQlcIGwjKOOcSCADbYxNmBjbLzb8qJd6u3cP55quS1altxWqyX1+X4+/VHtdZ6qVp966qlFVBVjjDGmO6/YARhjjBmcLEEYY4zJyRKEMcaYnCxBGGOMyckShDHGmJxCxQ6gP9XU1OjUqVPzmre1tZXKysr+DWiQszKXBivz8Hcg5V28eHGDqh6Ua9ywShBTp05l0aJFec0bi8Woq6vr34AGOStzabAyD38HUl4RWdfTODvFZIwxJidLEMYYY3KyBGGMMSanYdUGYYwZOIlEgg0bNtDR0VHsUD5m5MiRrFy5sthhDJi+lDcajTJ58mTC4XCfl2sJwhiTlw0bNlBdXc3UqVMRkWKHs5fm5maqq6uLHcaA6a28qsqOHTvYsGED06ZN6/Ny7RSTMSYvHR0djB07dtAlB/NxIsLYsWP3u7ZnCcIYkzdLDkNHPvvKEkQ6DQt+wuidbxU7EmOMGVQsQXgevPK/qWl4vdiRGGP2U1VVVbFDGNYsQQCMOphox9ZiR2GMMYNKySeIdFpZFR8DrZYgjBmqVJXvfOc7HHfccRx//PE8/vjjAGzevJk5c+YwY8YMjjvuOF5++WVSqRRz587tmvbOO+8scvSDV8lf5up5wl93VnI520EVrNHNmP122/95hxWbmvp1mcdMHMEtnz22T9M+8cQTLFmyhKVLl9LQ0MDs2bP51Kc+xe9+9zs+9alP8YMf/IBUKkVbWxtLlixh48aNLF++HIDdu3f3a9zDScFqECIyRUTqRWSFiLwjIjcGw/9dRN4VkWUi8qSIjOph/rUi8raILBGR/J7A10etFZMpoxNaGwq5GmNMgSxcuJArr7wS3/epra3l9NNP54033uDkk0/mwQcf5NZbb+Xtt9+murqaQw89lDVr1vAP//APPP/884wYMaLY4Q9ahaxBJIF/UtU3RaQaWCwiLwEvAd9X1aSI/CvwfeB7PSzjbFUt/K929QRoA5o3Q1XOp94aY/ahr0f6A23OnDksWLCAZ599lrlz5/KP//iPfPnLX2bp0qW88MIL3HvvvTz66KM88MADxQ51UCpYDUJVN6vqm0F3M7ASmKSqL6pqMpjsr8DkQsXQVxWjxgGgbTuKHIkxJh9nnnkmjzzyCKlUiu3bt/PKK69wyimnsG7dOmpra/nqV7/Ktddey5tvvklDQwPpdJrPf/7z3H777bz55pvFDn/QGpA2CBGZCswEXus26ivAIz3MpsCLIqLAL1T1vh6WPQ+YB1BbW0ssFtvv+Ha2dAKw+K8v07K+dNogWlpa8tpeQ5mVuf+MHDmS5ubmfl/u/mpububcc8/lz3/+M8cffzwiwq233kplZSV/+MMfuOuuuwiHw1RWVvKLX/yC999/n2984xuk02kAbrnllkFRjgORSqX6VIaOjo79+y6oakE/QBWwGLik2/AfAE8C0sN8k4K/44ClwJze1jVr1izNR2zxO6q3jND1z/80r/mHqvr6+mKHMOCszP1nxYoVBVluf2hqaip2CAOqr+XNtc+ARdrDb2pBL3MVkTDwOPBbVX0ia/hc4ELgi0GAH6OqG4O/23CJ5JRCxTl+/AQAWndvK9QqjDFmyCnkVUwC3A+sVNU7soafD3wXuEhV23qYtzJo2EZEKoHzgOWFinXimCoatYJE0/ZCrcIYY4acQtYgTge+BHwiuFR1iYhcAPwcqAZeCobdCyAiE0XkuWDeWmChiCwFXgeeVdXnCxVoVVmIXVqN17GzUKswxpghp2CN1Kq6EMjV4vtcjmGo6ibggqB7DXBioWLrTkRolBFUdlqCMMaYjJJ/1EZGm1dBKNFS7DCMMWbQsAQRSEiUUKq92GEYY8ygYQkikPCjRFI528yNMaYkWYIIpPxyynTwvXzdGNM/Mu+O2LRpE5deemnOaerq6li0aN+PfvvpT39KW9ueg8kLLrigXx/4N3fuXB577LF+W96BsAQRSPtRytVOMRkz3E2cOPGAfoC7J4jnnnuOUaNyPnN0yCv5x31naChKRFK0t7dTXl5e7HCMGVr+eDNsebt/lzn+ePj0j3scffPNNzNlyhSuv/56AG699VZCoRD19fXs2LGDVCrF7bffzsUXX7zXfGvXruXCCy9k+fLltLe3c/XVV7N06VKmT59Oe/ueg8TrrruON954g/b2di699FJuu+027rrrLjZt2sTZZ59NTU0N9fX1TJ06lUWLFlFTU8Mdd9zR9eC/a6+9lptuuom1a9fy6U9/mjPOOINXXnmFSZMm8dRTT/Xpd2b+/Pl8+9vfJplMcvLJJ3PPPfdQVlbGzTffzNNPP00oFOK8887jlltu4fe//z233XYbvu8zcuRIFixYkM9W34sliEA65HZWa/NuSxDGDAFXXHEFN910U1eCePTRR3nhhRe44YYbEBE6Ozs57bTTuOiii5Ae3vNyzz33UFFRwcqVK1m2bBknnXRS17h//ud/ZsyYMaRSKc455xyWLVvGDTfcwB133EF9fT01NTV7LWvx4sU8+OCDvPbaa6gqp556KmeddRajR49m1apVPPTQQ/zyl7/k8ssv5/HHH+fv//7v91m+jo4O5s6dy/z58znyyCP58pe/zD333MOXvvQlnnzySd59911EpOv01g9/+ENeeOEFJk2a1G+nvCxBBNJ+FICO1iZgQnGDMWao2ceRfqHMnDmTbdu2sWnTJrZv387o0aMZP3483/rWt4jFYoRCITZu3MjWrVsZP358zmUsWLCAG264AYATTjiBE044oWvco48+yn333UcymWTz5s2sWLFir/HdLVy4kM997nNUVlYCcMkll/Dyyy9z0UUXMW3aNGbMmAHArFmzWLt2ba/le++995g2bRpHHnkkAFdddRV333033/zmN4lGo1xzzTVceOGFXHjhhXR2dnL66aczd+5cLr/8ci655JI+bcPeWBtEQIMaRGdb/74VyxhTOJdddhmPPfYYjzzyCFdccQW//e1v2b59OwsWLGDJkiXU1tbS0bH/F598+OGH/OQnP2H+/PksW7aMz3zmM3ktJ6OsrKyr2/d9ksnkPqbet1AoxOuvv86ll17KM888w/nnnw/Avffey+2338769euZNWsWO3Yc+OsLLEEENORqEPH2of3YX2NKyRVXXMHDDz/MY489xmWXXUZjYyPjxo0jHA5TX1/PunXr9jn/nDlz+N3vfgfA8uXLWbZsGQBNTU1UVlYycuRItm7dyh//+Meueaqrq3M+WvvMM8/kD3/4A21tbbS2tvLkk09y5pln5l22o446irVr17J69WoAfvOb33DWWWfR0tJCY2MjF1xwAXfeeSdLly4F4IMPPuDUU0/lhz/8IQcddBDr16/Pe90ZdoopIEENImkJwpgh49hjj6W5uZlJkyYxYcIEvvjFL/LZz36W0047jVNOOYXp06fvc/7rrruOq6++mqOPPpqjjz6aWbNmAXDiiScyc+ZMpk+fzpQpUzj99NO75pk3bx7nn38+EydOpL6+vmv4SSedxNy5cznlFPfg6WuvvZaZM2f26XRSLtFolAcffJDLLrusq5H661//Ojt37uTiiy+mo6MDVeWOO9yzUL/zne+watUqVJVzzjmHE0888KcVSQ9P2x6SZs+erb1dw9yTpx+6j4ve+w5v/e3dzDxv341Hw0UsFqOurq7YYQwoK3P/WblyJUcffXS/L7c/NDc3U11dXewwBkxfy5trn4nIYlWdnWt6O8UUkLCrQaQ7rQZhjDFgp5i6+CHXiJSM293UxpjCu/766/nLX/6y17Abb7yRq6++ukgRfZwliEA4EgEgZQnCmD5T1R7vMTD7dvfddw/o+vJpTijkG+WmiEi9iKwQkXdE5MZg+BgReUlEVgV/R/cw/1XBNKtE5KpCxZkRCoUBSCcsQRjTF9FolB07duT1w2MGlqqyY8cOotHofs1XyBpEEvgnVX0zeH3oYhF5CZgLzFfVH4vIzcDNwPeyZxSRMcAtwGxAg3mfVtVdhQrWC7kahCUIY/pm8uTJbNiwge3bB9+rejs6Ovb7x3Ao60t5o9EokydP3q/lFvKNcpuBzUF3s4isBCYBFwN1wWS/BmJ0SxDAp4CXVHUnQJBYzgceKli84jaFJjoLtQpjhpVwOMy0adOKHUZOsViMmTNnFjuMAVOo8g5IG4SITAVmAq8BtUHyANiCe/90d5OA7Ls8NgTDci17HjAPoLa2llgslleMLa2tdGiY5t078l7GUNPS0lIyZc2wMpeGUitzocpb8AQhIlXA48BNqtqU3aClqioiB3QCU1XvA+4Ddx9Evtd7x2IxkhKhMhoqmevk7Z6A0mBlHv4KVd6C3gchImFccvitqj4RDN4qIhOC8ROAbTlm3QhMyeqfHAwrqLiEkVS80KsxxpghoZBXMQlwP7BSVe/IGvU0kLkq6SrgqRyzvwCcJyKjg6uczguGFVRSwvhpa4MwxhgobA3idOBLwCdEZEnwuQD4MfBJEVkFnBv0IyKzReQ/AYLG6R8BbwSfH2YarAspKRG8tNUgjDEGCnsV00Kgpztozskx/SLg2qz+B4AHChNdbkkvQshOMRljDGDPYtpLUiL4agnCGGPAEsReUl6YkJ1iMsYYwBLEXtJeBF8TxQ7DGGMGBUsQWVJeGWE7xWSMMYAliL2kvQhhq0EYYwxgCWIvab+MMFaDMMYYsASxF/WtBmGMMRmWILKoX0YESxDGGAOWIPaifoQISZKpdLFDMcaYorMEkUVDUcqIE7cEYYwxliCyiR8mIiniiVSxQzHGmKKzBJHF8917qeMJa4cwxhhLEFkkFCSIuD3y2xhjLEFk2VODsHshjDHGEkQWL6hBJOOWIIwxxhJEFi8UASBhp5iMMaZwLwwSkQeAC4FtqnpcMOwR4KhgklHAblWdkWPetUAzkAKSqjq7UHFm66pBJK0GYYwxBUsQwK+AnwP/lRmgqldkukXkP4DGfcx/tqo2FCy6HPyuGoQlCGOMKeQrRxeIyNRc40REgMuBTxRq/fnIJIikNVIbY0xBaxD7ciawVVVX9TBegRdFRIFfqOp9PS1IROYB8wBqa2uJxWJ5BdTS0sKHDeuZCry3cgUdbR15LWcoaWlpyXt7DVVW5tJQamUuVHmLlSCuBB7ax/gzVHWjiIwDXhKRd1V1Qa4Jg+RxH8Ds2bO1rq4ur4BisRiHjTkK1sLhhx7C6Wfmt5yhJBaLke/2GqqszKWh1MpcqPIO+FVMIhICLgEe6WkaVd0Y/N0GPAmcMhCxZU4xpZJ2J7UxxhTjMtdzgXdVdUOukSJSKSLVmW7gPGD5QASWuYopbVcxGWNM4RKEiDwEvAocJSIbROSaYNQX6HZ6SUQmishzQW8tsFBElgKvA8+q6vOFijNbpgahKatBGGNMIa9iurKH4XNzDNsEXBB0rwFOLFRc++KHXYJIW4Iwxhi7kzpb5hQTliCMMcYSRDbfD2oQ1khtjDGWILJlTjFZG4QxxliC2EsoHJxiSluCMMYYSxBZQqEyANROMRljjCWIbF2N1FaDMMYYSxB7Cd4op6lkkQMxxpjiswSRzbPLXI0xJsMSRDbf3TcodorJGGMsQewlqEGoJQhjjLEEsZegDcJLWxuEMcbs81lMIrKsD8vYrqrn9FM8xeUFm8MaqY0xpteH9fkED9HrgQBP9184RSZCEh9RO8VkjDG9JYivqeq6fU0gIt/ox3iKLoWPWA3CGGP23Qahqgt7W0BfphlKkvh2o5wxxtBLghCRi0Xk+qz+10RkTfC5rJd5HxCRbSKyPGvYrSKyUUSWBJ+cp69E5HwReU9EVovIzftbqAORkhCeWg3CGGN6u4rpu+zdxlAGnAzUAV/vZd5fAefnGH6nqs4IPs91HykiPnA38GngGOBKETmml3X1mxQh0NRArc4YYwat3hJERFXXZ/UvVNUdqvoRULmvGVV1AbAzj5hOAVar6hpVjQMPAxfnsZy8pMRH7DJXY4zptZF6dHaPqn4zq/egPNf5TRH5MrAI+CdV3dVt/CQgOyltAE7taWEiMg+YB1BbW0ssFssrqJaWFmKxGNNVSHa2572coSRT5lJiZS4NpVbmQpW3twTxmoh8VVV/mT1QRL4GvJ7H+u4BfgRo8Pc/gK/ksZwuqnofcB/A7Nmzta6uLq/lxGIx6urq2LIwTMT3OCvP5QwlmTKXEitzaSi1MheqvL0liG8BfxCR/wG8GQybhWuL+Lv9XZmqbs10i8gvgWdyTLYRmJLVPzkYNiDS+HjWBmGMMftOEKq6DfhbEfkEcGww+FlV/VM+KxORCaq6Oej9HLA8x2RvAEeIyDRcYvgC8D/yWV8+0hJCLEEYY0yvj9qI4q5WOhx4G7hftW/XgIrIQ7irnWpEZANwC1AnIjNwp5jWAl8Lpp0I/KeqXqCqSRH5JvAC7k7uB1T1nTzKlhcV357FZIwx9H6K6ddAAngZd9np0cBNfVmwql6ZY/D9PUy7iaxHegSXv37sEtiBkPZ8vJTVIIwxprcEcYyqHg8gIveTX8P0kKLi241yxhhD7/dBdD1zoq+nloa6tISskdoYY+i9BnGiiDQF3QKUB/0CqKqOKGh0RaASwqO92GEYY0zR9XYVkz9QgQwW6vn4VoMwxpher2Ias6/xqprPozQGNZUQPpYgjDGmt1NMDbhHXWTaHyRrnAKHFiKoYnKN1Olih2GMMUXXW4K4Czgb+AvwEO5hfVrwqIpIPZ8QJdEeb4wx+9TbC4NuAmYAvwe+BLwlIv8W3OU8LNkpJmOMcXq7zBV16nHvhrgXuBo4t9CBFUvaC1kjtTHG0HsjdSXuXQxX4B7v/QQwK3gfxPAkIXzSqCoi0vv0xhgzTPXWBrENWIV7ac8qXMP0bBGZDaCqTxQ2vIGnno8vKVJpJeRbgjDGlK7eEsTvcUnhqOCTTXE1iuHFCxEiTTKthEruLhBjjNmjtxvl5g5QHIOH5+OTIj28L9Yyxphe7bORWkQu7G0BfZlmKFEJESJFMm0JwhhT2no7xfTvIrKRvW+Q6+5fyP1muKHJcwkinrIEYYwpbb0liK3AHb1MsyrXQBF5ALgQ2KaqxwXD/h34LBAHPgCuVtXdOeZdCzQDKSCpqrN7iaH/eD4h0rTbKSZjTInrrQ2i7gCW/Svg58B/ZQ17Cfh+8Na4fwW+D3yvh/nPVtWGA1h/fvwwPu4qJmOMKWW93iiXL1VdAOzsNuzFrPdK/BWYXKj1580LEZI0qZQ9j8kYU9oKliD64CvAH3sYp8CLIrJYROYNYEzguUpVKmXPYzLGlLbe2iAQEQ84TVVf6a+VisgPcE+I/W0Pk5yhqhtFZBzwkoi8G9RIci1rHjAPoLa2llgslldMLS0txGIxdIer9Lz66it8UB3Na1lDRabMpcTKXBpKrcwFK6+q9voB3urLdDnmmwos7zZsLvAqUNHHZdwKfLsv086aNUvzVV9fr6qqy3//I9VbRujqDVvyXtZQkSlzKbEyl4ZSK/OBlBdYpD38pvb1FNN8Efm8HODDiUTkfNxD/y5S1bYepqkUkepMN3AesPxA1rtfglNM6WSilwmNMWZ462uC+BrusRtxEWkSkeasd1XnJCIP4WoKR4nIBhG5BndVUzXutNESEbk3mHaiiDwXzFoLLBSRpcDrwLOq+vz+Fy0/YgnCGGOAPrRBAKhq9f4uWFWvzDH4/h6m3QRcEHSvAU7c3/X1l64EkbIEYYwpbX1KEAAichEwJ+iNqerwuXs6i/hhAFJWgzDGlLg+nWISkR8DNwIrgs+NIvL/FTKwYhHf5UxN22WuxpjS1tcaxAXADFVNA4jIr4G3cHdCDyuZBGE1CGNMqdufG+VGZXWP7O9ABotMG4TajXLGmBLX1xrEvwBviUg97smuc4CbCxZVEWVqEGlLEMaYEtfXO6nTwGnAycHg76nqlkIGVix7EoSdYjLGlLZeE4SqpkXku6r6KPD0AMRUVF6QILAEYYwpcX1tg/hvEfm2iEwRkTGZT0EjKxLx3GWudorJGFPq+toGcUXw9/qsYQoc2r/hFJ8XtgRhjDHQ9zaIm1X1kQGIp+g8uw/CGGOAPpxiCu59+M4AxDIo2GWuxhjjWBtEN74fcR2WIIwxJc7aILrxwnaKyRhjoO9Pc51W6EAGC6/rFJNd5mqMKW37PMUkIt/N6r6s27h/KVRQxeSF3FVMVoMwxpS63togvpDV3f3BfOf3cyyDQiZB2I1yxphS11uCkB66c/V/fGaRB0Rkm4gszxo2RkReEpFVwd/RPcx7VTDNKhG5qrd19RfPz9QgUgO1SmOMGZR6SxDaQ3eu/lx+xcdrGjcD81X1CGA+OR76F1whdQtwKnAKcEtPiaS/hTI1CDvFZIwpcb0liBMz76AGTgi6M/3H97ZwVV0A7Ow2+GLg10H3r4G/yzHrp4CXVHWnqu4CXmKATmnZjXLGGOPs8yomVfULsM5aVd0cdG8BanNMMwlYn9W/IRj2MSIyD5gHUFtbSywWyyuolpYWYrEY0rGbs4Ad27bmvayhIlPmUmJlLg2lVuZClbfP76QuBFVVEenLqap9LeM+4D6A2bNna11dXV7LicVi1NXVkW7dBX+FsaNHku+yhopMmUuJlbk0lFqZC1Xe/XmjXH/ZKiITAIK/23JMsxGYktU/ORhWcF7I5UyxU0zGmBJXjATxNJC5Kukq4Kkc07wAnCcio4PG6fOCYYUX3ChnjdTGmFJX0AQhIg8BrwJHicgGEbkG+DHwSRFZBZwb9CMis0XkPwFUdSfwI+CN4PPDYFjheXYVkzHGQIHbIFT1yh5GnZNj2kXAtVn9DwAPFCi0nnmuXV7sPghjTIkrximmwU2EJB6o1SCMMaXNEkQOKXywGoQxpsRZgsghhY9YDcIYU+IsQeSQwrPLXI0xJc8SRA4pfEsQxpiSZwkiB3eKydogjDGlzRJEDimxBGGMMZYgcrBTTMYYYwkip7TVIIwxxhJELil8PEsQxpgSZwkiB1eDsFNMxpjSZgkih7T4eNYGYYwpcZYgckjj42GnmIwxpc0SRA4pCdnTXI0xJc8SRA5psUZqY4wZ8AQhIkeJyJKsT5OI3NRtmjoRacya5n8NZIwqPh7WBmGMKW0FfWFQLqr6HjADQER83Lumn8wx6cuqeuFAxpaRFh9fO4uxamOMGTSKfYrpHOADVV1X5Dj2ohKyU0zGmJJX7ATxBeChHsb9jYgsFZE/isixAxlUWnx8u4rJGFPiRFWLs2KRCLAJOFZVt3YbNwJIq2qLiFwA/ExVj+hhOfOAeQC1tbWzHn744bziaWlpoaqqCoDyhbdTk9zC+rqf57WsoSK7zKXCylwaSq3MB1Les88+e7Gqzs45UlWL8gEuBl7s47RrgZrepps1a5bmq76+vqt78b9/VtfedkzeyxoqsstcKqzMpaHUynwg5QUWaQ+/qcU8xXQlPZxeEpHxIiJB9ym4U2E7BiqwtBfGtzYIY0yJG/CrmABEpBL4JPC1rGFfB1DVe4FLgetEJAm0A18IMt2AcJe5WoIwxpS2oiQIVW0FxnYbdm9W98+BojUAuMtcLUEYY0pbsa9iGpwkZFcxGWNKniWIHNTz8UkXOwxjjCkqSxA5qIQI2aM2jDElzhJEDq4GYaeYjDGlzRJEDiohfLVTTMaY0mYJIhffGqmNMcYSRA4qIUKShiI9hsQYYwYDSxA5qOe7DnurnDGmhFmCyEWC+wfTdiWTMaZ0WYLIxXMJIp1KFDkQY4wpHksQuQSnmJIJSxDGmNJlCSIHtRqEMcZYgsjJEoQxxliCyEX9MgBSifYiR2KMMcVjCSKHjuhBAGjj5iJHYowxxWMJIoeOivEAhNb8d5EjMcaY4ilaghCRtSLytogsEZFFOcaLiNwlIqtFZJmInDRQsbVWTKFJK4i+dT/s+GCgVmuMMYNKsWsQZ6vqDFWdnWPcp4Ejgs884J6BCsoLR7kkfitoGl78fwdqtcYYM6gUO0Hsy8XAf6nzV2CUiEwYiBWHfY/VOpnG2TfCe8/BKjvVZIwpPUV5J3VAgRdFRIFfqOp93cZPAtZn9W8Ihu3Vciwi83A1DGpra4nFYnkF09LS0jXv6s3uERsvdJ7IhZWHEHn0Kyya/TPiZaPzWvZglV3mUmFlLg2lVuaClVdVi/IBJgV/xwFLgTndxj8DnJHVPx+Yva9lzpo1S/NVX1/f1f3C8s16yPee0bc37FbdukL1R7WqvzxHtaUh7+UPRtllLhVW5tJQamU+kPICi7SH39SinWJS1Y3B323Ak8Ap3SbZCEzJ6p8cDCu4SMhtls5kGsYdDZ//JWxeBveeAQt/ComOgQjDGGOKqigJQkQqRaQ60w2cByzvNtnTwJeDq5lOAxpVdUBuTMgkiHgyeKvc0Z+Fq5+DsYfBf98Cd58Cyx8fiFCMMaZoilWDqAUWishS4HXgWVV9XkS+LiJfD6Z5DlgDrAZ+CXxjoIIr66pBZL0PYvJsmPsMfPZnEG+Bx74C///n4Z0noXXHQIVmjDEDpiiN1Kq6Bjgxx/B7s7oVuH4g48ooC7mnuXbVILLNmgsnXAHzfwR/vRtWB1c4jToEmjdDdBS0boOao8APgwjsXAs1h8OISRCpgkQrrPw/cOp1kGyHps1QeRDUHguhMki0Q9MmGDHRzR8dBVW10Lgeao6AVBy8sEtabTshUuGW29kczN8BkUoIR91b8TqboazaLQtg5xoYPW1AtiWpBCDgF/N6CGNMPuy/NoeuU0ypHAkCIFwO5/8LnPM/YcVT0LAKPnrV1SwqxroE0bwFOhv3zLPpLffJ9lqBb+2oOdIlkm3v5Bx9Wtk4iG3bM+24o2H966472ekSUfkoGHuES2iHnuUS0Qd/ctOWjwbxYPNSiI6AqvHw0V9h0kkuucVb4cACCfQAABQ/SURBVM1fu+UffzmkE2658VZ3uq56oksc615xCTJS5eaLjnBJ0gu57blxsXt506iDoaIGGt6D9t0uCaYSLhF6Yage72KuORIa3oexh7tpdqx2y17/OuOaq+CDtHtbYKTK7at4q1tP2QjXnU5CWRWUjwFNufkRN0243CV+Te9pi2prcEk8OsIl5PZd7rswehok2tw0XsjN19Hk4lR1Cb9iLPgRaN7kylAxxi139zoXix92BwfxVpfkk51uO4462E1XNc5t8zGHupjKRwPqypZOQvMW/GQ77P7I9cfboHU7jJzstkEq7uJOJd2+6Gx2sfkRaNni1o248qSTrjuddOM9362nbQc0bgDPc/Mm2tw+KB/tuv2IW+fYw922HHWI244t21ysybhbVsVYN328zX3vQmWu3M1boLLGHUhFR7jhqi5+Tbnvgh+CspHur19GZcuH8H6n+z6Vj3bbfcvbMG46bHjDHcBV1rjlt+9y2y/Z6eJIJdw29sJu+3Q2uf0XqYR02k2TTsHaBTD5ZBDfbbdRB+8Z174T/DK3barHw64P3fdrxCT3P9Oy1W1bTbvtXDZizzaOjnL7q9I98odw1MXfss19H8Llbj9oypWtbATh+O6C/IRYgsgh4ndrg+hJuBxO/MK+p0l0uB2Z7HT/TJp2n+3vuh+Dlu3w3rNwwhdg2wr3hR45xf0j1f+z+xKd+nX3g+dH3N+dH8D7L7razO517ovj+e4Lufsj90PSsgVCUffFrTnK9Ycr3D9q8Ka8aOe2PXE2vB98Cbe4mlC2D/7k/i59aM+wnfu4w7zxo48Pe/vRvfs//PO+t1uBHAOw8j+Ksu5iORNgYbGjGFgnA3zs+QzD1+zIWDjv7/p9uZYgcigLuwTx7LLNXHLS5ANbWDjq/kYq9x4+KXhyyOipMOVk1z1u+p7xIye5I/aMqacfWBwZql2nmmKxGHVnntH1giRE3BFSPDiKLKuGXWvdkVHTJnf04oWDI7sW153scMPjre4oafyJ7uipqtYlo0SbS1zJdmjcCAef5o6edn/kyt4UXJg27hi3rOatLkl6ITdO1R3NNW1yR1mdTe6oqWEVjD/OJdLt78PEGS4Ja9odEY49zB3Jeb5LoKEySCf5aOnLHHz0SS6GkQe7I9VNb7nTdamES5KpTqgc57ZDpBq2r3TbYvQ0l1wzR7Z+xMWzeamLrXqCW78fdrWDccdAR+OeI/XyUW750RFuWOVBrsypOOwOTh+274LWBhd3stN9bxLtEB3pjkb9iNumBwXflc5mV/sIRV38NUcER/Htbn8m2vhw1UqmTZ/h4kon3SnGERPhg3qY/hkXYygaHD3H3bbPvG43UuWWtXONO/Idf5xbZ7gCd/Tf6WKKt7rtOHHmnppPot0tp2KMqx0kO91R/EHTXa0rc0TdsXvPUXtZUEPoaHTl8Hx3INa+C7a96/4vRkxy20LV7aOpc9zRfMNqV4OYNoe1y19nalUCjvgkdAZH5q3b3NF59QQXe0eT20aNG2DMtKB2lnTrHjHR7fPGDW6fjpzivoM7VsGYw1wNTNNuP7c2uFpRKOKGlY1w310v5PrjrW6bVda47axpt98z29YLBWXc7dZfWePiTCfcNE2bYPQh7jvb2uDiam1w/RVjQVOs3Z7kqP75hdiLJYgcRkTDAMx/dxvptOJ5UuSI+pF0K0v3tgHPcz9GGWMPc39HH7L3dNERH192ZtqqoGpcPmrv8RNnur+VNa69BdypqmxjDt3TPXLSx9eRccQn3d9pc3qeJoc1LYdy8N/W7T3whMv3axlDzbpEjGmn1X18xOk3DngsBXPEuXv1rm09nKl1dcWJpQg2x2IFSRCD+VEbRVNZFuK2i9wP2OrtLUWOxhhjisMSRA/OPmocAE8tGZB784wxZtCxBNGDg8dW8NkTJ3L/wg9Z29Ba7HCMMWbAWYLYhx9ccDRh3+Nbjy6hI5HqfQZjjBlGLEHsw/iRUX58yQm89dFu/vHRJSR6ui/CGGOGIUsQvfjMCRP4wQVH89zbW7jiF6/ygTVaG2NKhCWIPvjqnEO584oTWbm5mXPv+DPf/N2brNzcVOywjDGmoOw+iD763MzJnHH4Qdy/8EN+8+panlm2menjq/nM8ROYdchoDqmpJOJ7+J6QViUS8qiKhHLeQ6GqNLYnGFUR2ec6h909GMaYIcUSxH44qLqMmz89nevOOoxHF63nxRVb+I+X3u9x+ojvEQl5iLjHdvieUB0NkUpDQ0snU8dW0NKZoqGlk6MnjKAi4pNMK9GQhyfCq2t2MH5ElJkHjyIUPP5jV2ucaNgnrUpZyGNkeZhUWhGBVNrdBV4R9kkrRMMeZSGfcMglmcb2BGHPY0xlhNbOJO99EOf1jncJ+R4RXxhdGaGtM0Uk5NHQ0smhB1USDfmkVFF1rwCsjPi0xVN4IvgeeCJ0JNOUhTxCnrgbn1VRIOQJb360i2k1VYytjJBKK9GwT0tngrFVZVSVhehIpFi1tYUNu9v565odfP/T0xlRHsYXIZlOE/I8lm9qJBryOfSgSkKeR2N7gvZEirKQR2WZTzypJNNp4sk0lWUhF1MiRU11GYJ7ttbutjjvbWlhe2OKlZubEAHBbRfX7STTypbGDo4cX00qpexsi6OqNLTECfvCwWMqEBFS6TTV0TDJtJJKKVuaOgj7wviRUdrjKV5csZWLZ0wk5HnEU2k27GyjsixEU0eCkeVhRldEUMATd99NMqWEfSGZUjqTaRIp932piPjsakuQVqUzkcbzoDzsc1B1GZ2JNIp76nBDc5zRlWHa4yk+2N7CaYeOJZl2+21XR5qPdrTRlkgyqjyC7wnt8RRlYY+qshAhX2juSLoDG98jGvZ5f2szR4yrpiORoj2RYnRFhJQqaVW2NXVSWeYzrjrKlqYO1u1opbkjyWEHVTF5dDkiEPY8mjoSdCbThH2PUeVh0qqkVEmllR0t8a4yjCgPISI0dSRQhYqIT2cyTXNHgtrqKB3JFGHfoz2RIpF02z0UHDi1JVK0x1NUlvmEPPcd9DyhNaFsa+5gbKX7DjR3JmlqTxAN+0SDJyU0dySJhDyqo6Gux+s0tMRRlHHV0a7/491tcVrjKZo7EowfEWVURYR0Wnlj7U4OH1dFdTTc9fy27trjKdriScZURmiLp9jdnmDCiGjXgZ8G20NE8INh8WSa3W1xV05fCHmCBDe4ptJKZ9L9/4F7XlxF2GdbW2HaR8U9NHV4mD17ti5alN8DWGKxGHV53Hm5vbmTdzY1sqWxg45EikRKae5I4HsebYkkjW0JRKA8HGJbc0fXjt7dFqcyEqI1nmRLYwcTR5WTCmoMbZ1JUqq89dFuJo6MUh7x6dpN4moWaQVFaY+niPgenck0LZ3JrhpMyHP/UKn08Nm/ZngKedKVPPvyc5T9MIBc00d8r+tBm5kKeG//BlVlIdriya7pRlWE8UTY2RrPGW+y2wJHREPEU2kEQcQdlKTTSkciTTyVJuwLiZSbpzLiEw55+CI0tie6luUOHpW2+N5XTPqeBAd9SnsilbMso8qEV35wHhWR/T/mF5HFqjo71zirQRygg6rLqAtuqhtsMkcnmS9gJHMUlnJHda/+5WXmzDmLZDpNWzzVdVTZ1pnqOtLvSKQJ+YIgKEpjW4KKSIhIyCOV1q4vrScSLNfVIuJJd3Tb0NLJ2MoyomG3blX3hU+k0i65hTzGVUfZuLudspBHZzIVHEG7ZafTiidCJBT80yuMKA+TTKeJ+B4dyTQRXwj7XlaChLAn7GiN4wU1kfKwTyKlbFn7PkcfcyzatY1cos10A+xsjVMWcqcLx1RGaGjpZMWmJg4bV9VVI/E9j5aOBGVhn5AnrN/ZxrqdbZw8dQwKlPkeu9tdbS/keYwoD9HckWT9zjYUGFkeJuy7I97O4Ag5c7TtC0RCPm3xJCJCxBeqo2Ea2xN4wY9PQ0u86yi6IuJTHgmxuy3uagxtcWqqylBczejDNas5ZNpheCKUhd1+27irnaqyEL4vxJNu+6RUCXlCU3uSDbvamFpTSXsixeTRFe6gR4TWeIrWTnfkXRbyGF3haiTrd7VxWE0VW5rcE26TaSWZSrttFXw3MgdHkZDHiPIw25s7qSrz2d2WoKUzychyVyPLfA8373YHTtVRV8PK1MZ3tcXd91GVMZUR1u9qI+x7jAlqOe2JFOvWrefwaYd0JZOR5WGqoyE2N3YgCLva4nzY0MqMKaNIptO0dqaChNBJWt304L7HE0eVUxb2WNvQiucJ1WUhkmll6frdHDdpJGlVBJDgBz8a9oj4HiLu6N8PagCjKsKUhTw+bGhFla7t3ZlIM6oyTEc8Rcj32LS7nZDvIbj/lbAvVERCe/4PkmmqytzZhg+2t3DwmAoizZsoD/v9/hsy4AlCRKYA/4V7aZAC96nqz7pNUwc8BXwYDHpCVX84kHEOByLiqqhZ35vKsj27POS5L1wEb+8jj+pMx55qdqEdP3lk7xP1g1j7GuqOnzAg6xosYsl11J15aO8TDiOx2Fbq6grxdKLBKRbb1pWA+1MxahBJ4J9U9c3gtaOLReQlVV3RbbqXVfXCIsRnjDGGIlzmqqqbVfXNoLsZWAns47GdxhhjiqGojdQiMhVYABynqk1Zw+uAx4ENwCbg26qa87VoIjIPmAdQW1s76+GHH84rlpaWFqqqqvKad6iyMpcGK/PwdyDlPfvss3tspEZVi/IBqoDFwCU5xo0AqoLuC4BVfVnmrFmzNF/19fV5zztUWZlLg5V5+DuQ8gKLtIff1KLcSS0iYVwN4beq+kT38arapKotQfdzQFhEagY4TGOMKWkDniDENbXfD6xU1Tt6mGZ8MB0icgouzh0DF6UxxphiXMV0OvAl4G0RWRIM+3+AgwFU9V7gUuA6EUkC7cAXgqqQMcaYATLgCUJVF7LnyQY9TfNz4OcDE5ExxphchtWjNkRkO7Auz9lrgIZ+DGcosDKXBivz8Hcg5T1EVQ/KNWJYJYgDISKLtKdLvYYpK3NpsDIPf4Uqr70PwhhjTE6WIIwxxuRkCWKP+4odQBFYmUuDlXn4K0h5rQ3CGGNMTlaDMMYYk5MlCGOMMTmVfIIQkfNF5D0RWS0iNxc7nv4iIlNEpF5EVojIOyJyYzB8jIi8JCKrgr+jg+EiIncF22GZiJxU3BLkT0R8EXlLRJ4J+qeJyGtB2R4RkUgwvCzoXx2Mn1rMuPMlIqNE5DEReVdEVorI3wz3/Swi3wq+18tF5CERiQ63/SwiD4jINhFZnjVsv/eriFwVTL9KRK7anxhKOkGIiA/cDXwaOAa4UkSOKW5U/SbzYqZjgNOA64Oy3QzMV9UjgPlBP7htcETwmQfcM/Ah95sbce8ZyfhX4E5VPRzYBVwTDL8G2BUMvzOYbij6GfC8qk4HTsSVfdjuZxGZBNwAzFbV4wAf+ALDbz//Cji/27D92q8iMga4BTgVOAW4JZNU+qSnx7yWwgf4G+CFrP7vA98vdlwFKutTwCeB94AJwbAJwHtB9y+AK7Om75puKH2AycE/zieAZ3CPdWkAQt33OfAC8DdBdyiYTopdhv0s70jcq3ml2/Bhu59xLxhbD4wJ9tszwKeG434GpgLL892vwJXAL7KG7zVdb5+SrkGw54uWsYFh+Ha7oEo9E3gNqFXVzcGoLbh3g8Pw2RY/Bb4LpIP+scBuVU0G/dnl6ipzML4xmH4omQZsBx4MTqv9p4hUMoz3s6puBH4CfARsxu23xQzv/Zyxv/v1gPZ3qSeIYU9EqnDv3rhJs97aB6DukGLYXOcsIhcC21R1cbFjGUAh4CTgHlWdCbSy57QDMCz382jgYlxynAhU8vFTMcPeQOzXUk8QG4EpWf2Tg2HDQg8vZtoqIhOC8ROAbcHw4bAtTgcuEpG1wMO400w/A0aJSObJxdnl6ipzMH4kQ++9IxuADar6WtD/GC5hDOf9fC7woapuV9UE8ARu3w/n/Zyxv/v1gPZ3qSeIN4AjgqsfIriGrqeLHFO/EOnxxUxPA5krGa7CtU1khn85uBriNKAxqyo7JKjq91V1sqpOxe3LP6nqF4F63DtG4ONlzmyLS4Pph9SRtqpuAdaLyFHBoHOAFQzj/Yw7tXSaiFQE3/NMmYftfs6yv/v1BeA8ERkd1LzOC4b1TbEbYYr9wb3z+n3gA+AHxY6nH8t1Bq76uQxYEnwuwJ17nQ+sAv4bGBNML7gruj4A3sZdIVL0chxA+euAZ4LuQ4HXgdXA74GyYHg06F8djD+02HHnWdYZwKJgX/8BGD3c9zNwG/AusBz4DVA23PYz8BCujSWBqylek89+Bb4SlH01cPX+xGCP2jDGGJNTqZ9iMsYY0wNLEMYYY3KyBGGMMSYnSxDGGGNysgRhjDEmJ0sQxuwHEUmJyJKsT789AVhEpmY/udOYYgv1PokxJku7qs4odhDGDASrQRjTD0RkrYj8m4i8LSKvi8jhwfCpIvKn4Bn980Xk4GB4rYg8KSJLg8/fBovyReSXwbsOXhSR8qIVypQ8SxDG7J/ybqeYrsga16iqxwM/xz1VFuB/A79W1ROA3wJ3BcPvAv6sqifinp30TjD8COBuVT0W2A18vsDlMaZHdie1MftBRFpUtSrH8LXAJ1R1TfCQxC2qOlZEGnDP708Ewzerao2IbAcmq2pn1jKmAi+pexkMIvI9IKyqtxe+ZMZ8nNUgjOk/2kP3/ujM6k5h7YSmiCxBGNN/rsj6+2rQ/QruybIAXwReDrrnA9dB1zu0Rw5UkMb0lR2dGLN/ykVkSVb/86qaudR1tIgsw9UCrgyG/QPubW/fwb357epg+I3AfSJyDa6mcB3uyZ3GDBrWBmFMPwjaIGarakOxYzGmv9gpJmOMMTlZDcIYY0xOVoMwxhiTkyUIY4wxOVmCMMYYk5MlCGOMMTlZgjDGGJPT/wU28q1y+8E73QAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_loss(history)\n",
    "test_results['linear_model'] = linear_model.evaluate(test_features, test_labels, verbose = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_28\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "normalization_13 (Normalizat (None, 1)                 3         \n",
      "_________________________________________________________________\n",
      "input_7 (InputLayer)         multiple                  0         \n",
      "_________________________________________________________________\n",
      "dense_68 (Dense)             (None, 64)                128       \n",
      "_________________________________________________________________\n",
      "dense_69 (Dense)             (None, 64)                4160      \n",
      "_________________________________________________________________\n",
      "dense_70 (Dense)             (None, 1)                 65        \n",
      "=================================================================\n",
      "Total params: 4,356\n",
      "Trainable params: 4,353\n",
      "Non-trainable params: 3\n",
      "_________________________________________________________________\n",
      "Model: \"sequential_29\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "normalization_5 (Normalizati (None, 9)                 19        \n",
      "_________________________________________________________________\n",
      "input_8 (InputLayer)         multiple                  0         \n",
      "_________________________________________________________________\n",
      "dense_71 (Dense)             (None, 64)                640       \n",
      "_________________________________________________________________\n",
      "dense_72 (Dense)             (None, 64)                4160      \n",
      "_________________________________________________________________\n",
      "dense_73 (Dense)             (None, 1)                 65        \n",
      "=================================================================\n",
      "Total params: 4,884\n",
      "Trainable params: 4,865\n",
      "Non-trainable params: 19\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#A non-linear regressor\n",
    "\n",
    "def create_model(normalization, ip):\n",
    "    model = keras.Sequential([\n",
    "          normalization,\n",
    "          layers.Input(shape=(ip,)),\n",
    "          layers.Dense(units = 64, activation='relu', use_bias = 'True'),\n",
    "          layers.Dense(units = 64, activation='relu', use_bias = 'True'),\n",
    "          layers.Dense(units = 1, use_bias = 'True')\n",
    "    ])\n",
    "\n",
    "    model.compile(loss='mean_absolute_error', optimizer=tf.keras.optimizers.Adam(0.001))\n",
    "    return model\n",
    "    \n",
    "non_linear_horsepower_regressor = create_model(horsepower_normalizer, 1)\n",
    "#Single input non-linear regression\n",
    "non_linear_regressor = create_model(normalizer, 9)\n",
    "#Multi-input non-linear regression\n",
    "non_linear_horsepower_regressor.summary()\n",
    "non_linear_regressor.summary()\n",
    "#Note that the difference in non-trainable parameters arises from the normalization layer."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/500\n",
      "8/8 [==============================] - 0s 13ms/step - loss: 22.7388 - val_loss: 23.0688\n",
      "Epoch 2/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 22.1286 - val_loss: 22.7435\n",
      "Epoch 3/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 22.3640 - val_loss: 22.3811\n",
      "Epoch 4/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 22.5481 - val_loss: 21.9338\n",
      "Epoch 5/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 21.1067 - val_loss: 21.3786\n",
      "Epoch 6/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 21.1107 - val_loss: 20.6799\n",
      "Epoch 7/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 20.0933 - val_loss: 19.7975\n",
      "Epoch 8/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 19.5056 - val_loss: 18.6823\n",
      "Epoch 9/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 18.2722 - val_loss: 17.2872\n",
      "Epoch 10/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 16.5715 - val_loss: 15.6911\n",
      "Epoch 11/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 14.6600 - val_loss: 13.9563\n",
      "Epoch 12/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 13.2604 - val_loss: 12.0259\n",
      "Epoch 13/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 11.1078 - val_loss: 9.9590\n",
      "Epoch 14/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 9.2382 - val_loss: 7.8735\n",
      "Epoch 15/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 7.3454 - val_loss: 6.1642\n",
      "Epoch 16/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.2266 - val_loss: 5.3967\n",
      "Epoch 17/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 4.0517 - val_loss: 5.2701\n",
      "Epoch 18/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.0707 - val_loss: 5.2384\n",
      "Epoch 19/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.9358 - val_loss: 5.2211\n",
      "Epoch 20/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 4.2340 - val_loss: 5.2103\n",
      "Epoch 21/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7987 - val_loss: 5.1562\n",
      "Epoch 22/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7958 - val_loss: 5.0692\n",
      "Epoch 23/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5425 - val_loss: 5.0292\n",
      "Epoch 24/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.8058 - val_loss: 4.9845\n",
      "Epoch 25/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.5141 - val_loss: 4.9465\n",
      "Epoch 26/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.4247 - val_loss: 4.9100\n",
      "Epoch 27/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.3532 - val_loss: 4.8835\n",
      "Epoch 28/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.4905 - val_loss: 4.8472\n",
      "Epoch 29/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.5616 - val_loss: 4.8130\n",
      "Epoch 30/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.3382 - val_loss: 4.7777\n",
      "Epoch 31/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.4359 - val_loss: 4.7563\n",
      "Epoch 32/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.2746 - val_loss: 4.7407\n",
      "Epoch 33/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.2985 - val_loss: 4.7244\n",
      "Epoch 34/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2111 - val_loss: 4.7081\n",
      "Epoch 35/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.2809 - val_loss: 4.6922\n",
      "Epoch 36/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2392 - val_loss: 4.6816\n",
      "Epoch 37/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.0429 - val_loss: 4.6666\n",
      "Epoch 38/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2346 - val_loss: 4.6607\n",
      "Epoch 39/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.2530 - val_loss: 4.6264\n",
      "Epoch 40/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9767 - val_loss: 4.6106\n",
      "Epoch 41/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1752 - val_loss: 4.5945\n",
      "Epoch 42/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.3998 - val_loss: 4.5849\n",
      "Epoch 43/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.3198 - val_loss: 4.5537\n",
      "Epoch 44/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1866 - val_loss: 4.5447\n",
      "Epoch 45/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.3375 - val_loss: 4.5093\n",
      "Epoch 46/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1426 - val_loss: 4.4792\n",
      "Epoch 47/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0588 - val_loss: 4.4642\n",
      "Epoch 48/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.3505 - val_loss: 4.4801\n",
      "Epoch 49/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.2535 - val_loss: 4.4667\n",
      "Epoch 50/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0198 - val_loss: 4.4664\n",
      "Epoch 51/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2773 - val_loss: 4.4410\n",
      "Epoch 52/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2379 - val_loss: 4.4471\n",
      "Epoch 53/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1542 - val_loss: 4.4207\n",
      "Epoch 54/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1312 - val_loss: 4.4624\n",
      "Epoch 55/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2070 - val_loss: 4.4430\n",
      "Epoch 56/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0357 - val_loss: 4.4136\n",
      "Epoch 57/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1714 - val_loss: 4.4139\n",
      "Epoch 58/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1014 - val_loss: 4.4183\n",
      "Epoch 59/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.2539 - val_loss: 4.3786\n",
      "Epoch 60/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8922 - val_loss: 4.4092\n",
      "Epoch 61/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1745 - val_loss: 4.4012\n",
      "Epoch 62/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9966 - val_loss: 4.3747\n",
      "Epoch 63/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1995 - val_loss: 4.3935\n",
      "Epoch 64/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.3028 - val_loss: 4.4026\n",
      "Epoch 65/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0741 - val_loss: 4.4082\n",
      "Epoch 66/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9346 - val_loss: 4.3447\n",
      "Epoch 67/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 2.9921 - val_loss: 4.3907\n",
      "Epoch 68/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2169 - val_loss: 4.4226\n",
      "Epoch 69/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1679 - val_loss: 4.3853\n",
      "Epoch 70/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0118 - val_loss: 4.3635\n",
      "Epoch 71/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.0539 - val_loss: 4.3794\n",
      "Epoch 72/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0157 - val_loss: 4.3730\n",
      "Epoch 73/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8839 - val_loss: 4.3958\n",
      "Epoch 74/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1403 - val_loss: 4.4101\n",
      "Epoch 75/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9853 - val_loss: 4.3462\n",
      "Epoch 76/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9387 - val_loss: 4.3705\n",
      "Epoch 77/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9034 - val_loss: 4.3862\n",
      "Epoch 78/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1364 - val_loss: 4.3702\n",
      "Epoch 79/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.7637 - val_loss: 4.3993\n",
      "Epoch 80/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9938 - val_loss: 4.3771\n",
      "Epoch 81/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1124 - val_loss: 4.3620\n",
      "Epoch 82/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0831 - val_loss: 4.3790\n",
      "Epoch 83/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1260 - val_loss: 4.3536\n",
      "Epoch 84/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0388 - val_loss: 4.3651\n",
      "Epoch 85/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.3413 - val_loss: 4.3758\n",
      "Epoch 86/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9231 - val_loss: 4.3860\n",
      "Epoch 87/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1163 - val_loss: 4.3361\n",
      "Epoch 88/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1192 - val_loss: 4.3755\n",
      "Epoch 89/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9660 - val_loss: 4.3671\n",
      "Epoch 90/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.3188 - val_loss: 4.3587\n",
      "Epoch 91/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1201 - val_loss: 4.3870\n",
      "Epoch 92/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9982 - val_loss: 4.3518\n",
      "Epoch 93/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9871 - val_loss: 4.3549\n",
      "Epoch 94/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9820 - val_loss: 4.3506\n",
      "Epoch 95/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1531 - val_loss: 4.3430\n",
      "Epoch 96/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1306 - val_loss: 4.3545\n",
      "Epoch 97/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.0196 - val_loss: 4.3996\n",
      "Epoch 98/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9636 - val_loss: 4.3641\n",
      "Epoch 99/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0324 - val_loss: 4.3416\n",
      "Epoch 100/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1369 - val_loss: 4.3460\n",
      "Epoch 101/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0783 - val_loss: 4.3566\n",
      "Epoch 102/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1959 - val_loss: 4.3593\n",
      "Epoch 103/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1799 - val_loss: 4.3851\n",
      "Epoch 104/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2272 - val_loss: 4.3453\n",
      "Epoch 105/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1721 - val_loss: 4.3311\n",
      "Epoch 106/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1259 - val_loss: 4.3411\n",
      "Epoch 107/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8990 - val_loss: 4.3622\n",
      "Epoch 108/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1836 - val_loss: 4.3458\n",
      "Epoch 109/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9637 - val_loss: 4.3838\n",
      "Epoch 110/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0268 - val_loss: 4.3681\n",
      "Epoch 111/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9316 - val_loss: 4.3239\n",
      "Epoch 112/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0514 - val_loss: 4.3493\n",
      "Epoch 113/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1084 - val_loss: 4.3670\n",
      "Epoch 114/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2656 - val_loss: 4.3726\n",
      "Epoch 115/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0548 - val_loss: 4.3721\n",
      "Epoch 116/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9644 - val_loss: 4.3525\n",
      "Epoch 117/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8966 - val_loss: 4.3247\n",
      "Epoch 118/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2095 - val_loss: 4.3343\n",
      "Epoch 119/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9126 - val_loss: 4.3557\n",
      "Epoch 120/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0890 - val_loss: 4.3537\n",
      "Epoch 121/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0016 - val_loss: 4.3762\n",
      "Epoch 122/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9626 - val_loss: 4.3373\n",
      "Epoch 123/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0361 - val_loss: 4.3473\n",
      "Epoch 124/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.2126 - val_loss: 4.3480\n",
      "Epoch 125/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9641 - val_loss: 4.3298\n",
      "Epoch 126/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2787 - val_loss: 4.3245\n",
      "Epoch 127/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0954 - val_loss: 4.3667\n",
      "Epoch 128/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1473 - val_loss: 4.3816\n",
      "Epoch 129/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2693 - val_loss: 4.3356\n",
      "Epoch 130/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0705 - val_loss: 4.3075\n",
      "Epoch 131/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8929 - val_loss: 4.3482\n",
      "Epoch 132/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1276 - val_loss: 4.3669\n",
      "Epoch 133/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8861 - val_loss: 4.3702\n",
      "Epoch 134/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0673 - val_loss: 4.3545\n",
      "Epoch 135/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9709 - val_loss: 4.3400\n",
      "Epoch 136/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9956 - val_loss: 4.3263\n",
      "Epoch 137/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1227 - val_loss: 4.3365\n",
      "Epoch 138/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8877 - val_loss: 4.3343\n",
      "Epoch 139/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0744 - val_loss: 4.3379\n",
      "Epoch 140/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8662 - val_loss: 4.3623\n",
      "Epoch 141/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9987 - val_loss: 4.3290\n",
      "Epoch 142/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9981 - val_loss: 4.3345\n",
      "Epoch 143/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0752 - val_loss: 4.3145\n",
      "Epoch 144/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0628 - val_loss: 4.3509\n",
      "Epoch 145/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 2.9193 - val_loss: 4.3595\n",
      "Epoch 146/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2930 - val_loss: 4.3512\n",
      "Epoch 147/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1032 - val_loss: 4.3412\n",
      "Epoch 148/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0032 - val_loss: 4.3383\n",
      "Epoch 149/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0545 - val_loss: 4.3465\n",
      "Epoch 150/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9133 - val_loss: 4.3393\n",
      "Epoch 151/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2008 - val_loss: 4.3688\n",
      "Epoch 152/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.2831 - val_loss: 4.3362\n",
      "Epoch 153/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8118 - val_loss: 4.3367\n",
      "Epoch 154/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1522 - val_loss: 4.3296\n",
      "Epoch 155/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9349 - val_loss: 4.3208\n",
      "Epoch 156/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1799 - val_loss: 4.3508\n",
      "Epoch 157/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0621 - val_loss: 4.3497\n",
      "Epoch 158/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2291 - val_loss: 4.3395\n",
      "Epoch 159/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0603 - val_loss: 4.3304\n",
      "Epoch 160/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0228 - val_loss: 4.3124\n",
      "Epoch 161/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0348 - val_loss: 4.3441\n",
      "Epoch 162/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8034 - val_loss: 4.3312\n",
      "Epoch 163/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0001 - val_loss: 4.3203\n",
      "Epoch 164/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8978 - val_loss: 4.3251\n",
      "Epoch 165/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0235 - val_loss: 4.3484\n",
      "Epoch 166/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9117 - val_loss: 4.3255\n",
      "Epoch 167/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1548 - val_loss: 4.3395\n",
      "Epoch 168/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0695 - val_loss: 4.3862\n",
      "Epoch 169/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.7550 - val_loss: 4.3461\n",
      "Epoch 170/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0290 - val_loss: 4.2737\n",
      "Epoch 171/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0968 - val_loss: 4.2857\n",
      "Epoch 172/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0129 - val_loss: 4.3520\n",
      "Epoch 173/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9022 - val_loss: 4.3383\n",
      "Epoch 174/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0461 - val_loss: 4.3171\n",
      "Epoch 175/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8359 - val_loss: 4.3054\n",
      "Epoch 176/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1172 - val_loss: 4.3264\n",
      "Epoch 177/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1003 - val_loss: 4.3259\n",
      "Epoch 178/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8999 - val_loss: 4.3138\n",
      "Epoch 179/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.7848 - val_loss: 4.2952\n",
      "Epoch 180/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1213 - val_loss: 4.3135\n",
      "Epoch 181/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8076 - val_loss: 4.3123\n",
      "Epoch 182/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9946 - val_loss: 4.3101\n",
      "Epoch 183/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0214 - val_loss: 4.2987\n",
      "Epoch 184/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9400 - val_loss: 4.3197\n",
      "Epoch 185/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9302 - val_loss: 4.3208\n",
      "Epoch 186/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0691 - val_loss: 4.3039\n",
      "Epoch 187/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9461 - val_loss: 4.3029\n",
      "Epoch 188/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0315 - val_loss: 4.3143\n",
      "Epoch 189/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0074 - val_loss: 4.3031\n",
      "Epoch 190/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9570 - val_loss: 4.3130\n",
      "Epoch 191/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0857 - val_loss: 4.3241\n",
      "Epoch 192/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1869 - val_loss: 4.3099\n",
      "Epoch 193/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9606 - val_loss: 4.3203\n",
      "Epoch 194/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1559 - val_loss: 4.3015\n",
      "Epoch 195/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8949 - val_loss: 4.2987\n",
      "Epoch 196/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.2546 - val_loss: 4.3400\n",
      "Epoch 197/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8665 - val_loss: 4.3584\n",
      "Epoch 198/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9228 - val_loss: 4.2950\n",
      "Epoch 199/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0163 - val_loss: 4.2762\n",
      "Epoch 200/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0418 - val_loss: 4.2646\n",
      "Epoch 201/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9168 - val_loss: 4.3239\n",
      "Epoch 202/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9661 - val_loss: 4.3276\n",
      "Epoch 203/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1500 - val_loss: 4.3065\n",
      "Epoch 204/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0472 - val_loss: 4.3198\n",
      "Epoch 205/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9843 - val_loss: 4.3091\n",
      "Epoch 206/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.2316 - val_loss: 4.2929\n",
      "Epoch 207/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9506 - val_loss: 4.3073\n",
      "Epoch 208/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.2448 - val_loss: 4.3023\n",
      "Epoch 209/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1429 - val_loss: 4.2901\n",
      "Epoch 210/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9443 - val_loss: 4.2849\n",
      "Epoch 211/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0340 - val_loss: 4.2937\n",
      "Epoch 212/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1171 - val_loss: 4.2939\n",
      "Epoch 213/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9926 - val_loss: 4.2848\n",
      "Epoch 214/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1796 - val_loss: 4.3161\n",
      "Epoch 215/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1232 - val_loss: 4.2721\n",
      "Epoch 216/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0797 - val_loss: 4.2777\n",
      "Epoch 217/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8784 - val_loss: 4.2983\n",
      "Epoch 218/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8883 - val_loss: 4.3036\n",
      "Epoch 219/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8573 - val_loss: 4.2705\n",
      "Epoch 220/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0410 - val_loss: 4.2501\n",
      "Epoch 221/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0463 - val_loss: 4.2745\n",
      "Epoch 222/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.7831 - val_loss: 4.2938\n",
      "Epoch 223/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1319 - val_loss: 4.2954\n",
      "Epoch 224/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0531 - val_loss: 4.3049\n",
      "Epoch 225/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0190 - val_loss: 4.2674\n",
      "Epoch 226/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0950 - val_loss: 4.2594\n",
      "Epoch 227/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0865 - val_loss: 4.2710\n",
      "Epoch 228/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9465 - val_loss: 4.2990\n",
      "Epoch 229/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1109 - val_loss: 4.2956\n",
      "Epoch 230/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1068 - val_loss: 4.2861\n",
      "Epoch 231/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0537 - val_loss: 4.2700\n",
      "Epoch 232/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9284 - val_loss: 4.2707\n",
      "Epoch 233/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9679 - val_loss: 4.2825\n",
      "Epoch 234/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9331 - val_loss: 4.2765\n",
      "Epoch 235/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1131 - val_loss: 4.2920\n",
      "Epoch 236/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0130 - val_loss: 4.2579\n",
      "Epoch 237/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9678 - val_loss: 4.2783\n",
      "Epoch 238/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0007 - val_loss: 4.3066\n",
      "Epoch 239/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1378 - val_loss: 4.2856\n",
      "Epoch 240/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0806 - val_loss: 4.2727\n",
      "Epoch 241/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9442 - val_loss: 4.2707\n",
      "Epoch 242/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8412 - val_loss: 4.2675\n",
      "Epoch 243/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8814 - val_loss: 4.2849\n",
      "Epoch 244/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0444 - val_loss: 4.2419\n",
      "Epoch 245/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1773 - val_loss: 4.2491\n",
      "Epoch 246/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1870 - val_loss: 4.2732\n",
      "Epoch 247/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1487 - val_loss: 4.2686\n",
      "Epoch 248/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1092 - val_loss: 4.2630\n",
      "Epoch 249/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0053 - val_loss: 4.2759\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 250/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8046 - val_loss: 4.3048\n",
      "Epoch 251/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0005 - val_loss: 4.2504\n",
      "Epoch 252/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9142 - val_loss: 4.2687\n",
      "Epoch 253/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1664 - val_loss: 4.2559\n",
      "Epoch 254/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9906 - val_loss: 4.2690\n",
      "Epoch 255/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1658 - val_loss: 4.2761\n",
      "Epoch 256/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1958 - val_loss: 4.2781\n",
      "Epoch 257/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8869 - val_loss: 4.2709\n",
      "Epoch 258/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0164 - val_loss: 4.3104\n",
      "Epoch 259/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0541 - val_loss: 4.2815\n",
      "Epoch 260/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9140 - val_loss: 4.2717\n",
      "Epoch 261/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9293 - val_loss: 4.2509\n",
      "Epoch 262/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1154 - val_loss: 4.2868\n",
      "Epoch 263/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.6691 - val_loss: 4.2543\n",
      "Epoch 264/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8959 - val_loss: 4.2576\n",
      "Epoch 265/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1644 - val_loss: 4.2858\n",
      "Epoch 266/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0934 - val_loss: 4.2553\n",
      "Epoch 267/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1901 - val_loss: 4.2597\n",
      "Epoch 268/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9583 - val_loss: 4.2357\n",
      "Epoch 269/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0564 - val_loss: 4.2622\n",
      "Epoch 270/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1892 - val_loss: 4.2886\n",
      "Epoch 271/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1274 - val_loss: 4.2661\n",
      "Epoch 272/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9040 - val_loss: 4.2346\n",
      "Epoch 273/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1556 - val_loss: 4.2573\n",
      "Epoch 274/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1499 - val_loss: 4.2730\n",
      "Epoch 275/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0257 - val_loss: 4.2662\n",
      "Epoch 276/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1934 - val_loss: 4.2451\n",
      "Epoch 277/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0464 - val_loss: 4.2462\n",
      "Epoch 278/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9173 - val_loss: 4.2769\n",
      "Epoch 279/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9983 - val_loss: 4.2815\n",
      "Epoch 280/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.6435 - val_loss: 4.2400\n",
      "Epoch 281/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8930 - val_loss: 4.2471\n",
      "Epoch 282/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9612 - val_loss: 4.2561\n",
      "Epoch 283/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0035 - val_loss: 4.2825\n",
      "Epoch 284/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0797 - val_loss: 4.2440\n",
      "Epoch 285/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0177 - val_loss: 4.2628\n",
      "Epoch 286/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9307 - val_loss: 4.2523\n",
      "Epoch 287/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0288 - val_loss: 4.2845\n",
      "Epoch 288/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9825 - val_loss: 4.2797\n",
      "Epoch 289/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9193 - val_loss: 4.2410\n",
      "Epoch 290/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1785 - val_loss: 4.2498\n",
      "Epoch 291/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1825 - val_loss: 4.2708\n",
      "Epoch 292/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9530 - val_loss: 4.2724\n",
      "Epoch 293/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9103 - val_loss: 4.2461\n",
      "Epoch 294/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9247 - val_loss: 4.2752\n",
      "Epoch 295/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9691 - val_loss: 4.2696\n",
      "Epoch 296/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0098 - val_loss: 4.2418\n",
      "Epoch 297/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0971 - val_loss: 4.2479\n",
      "Epoch 298/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0250 - val_loss: 4.2860\n",
      "Epoch 299/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1170 - val_loss: 4.2586\n",
      "Epoch 300/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8780 - val_loss: 4.2720\n",
      "Epoch 301/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9945 - val_loss: 4.2211\n",
      "Epoch 302/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9927 - val_loss: 4.2566\n",
      "Epoch 303/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.2869 - val_loss: 4.2673\n",
      "Epoch 304/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9989 - val_loss: 4.2683\n",
      "Epoch 305/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9427 - val_loss: 4.2641\n",
      "Epoch 306/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1230 - val_loss: 4.2586\n",
      "Epoch 307/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8853 - val_loss: 4.2386\n",
      "Epoch 308/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9969 - val_loss: 4.2278\n",
      "Epoch 309/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9132 - val_loss: 4.2570\n",
      "Epoch 310/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8232 - val_loss: 4.2764\n",
      "Epoch 311/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9252 - val_loss: 4.2454\n",
      "Epoch 312/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0167 - val_loss: 4.2201\n",
      "Epoch 313/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9963 - val_loss: 4.2371\n",
      "Epoch 314/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 2.9242 - val_loss: 4.2502\n",
      "Epoch 315/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9022 - val_loss: 4.2730\n",
      "Epoch 316/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 2.9741 - val_loss: 4.2363\n",
      "Epoch 317/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.1264 - val_loss: 4.2746\n",
      "Epoch 318/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9765 - val_loss: 4.2514\n",
      "Epoch 319/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1989 - val_loss: 4.2629\n",
      "Epoch 320/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8869 - val_loss: 4.2631\n",
      "Epoch 321/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1217 - val_loss: 4.2619\n",
      "Epoch 322/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.7506 - val_loss: 4.2684\n",
      "Epoch 323/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.7366 - val_loss: 4.2584\n",
      "Epoch 324/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0657 - val_loss: 4.2462\n",
      "Epoch 325/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1442 - val_loss: 4.2547\n",
      "Epoch 326/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.7532 - val_loss: 4.2869\n",
      "Epoch 327/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8400 - val_loss: 4.2562\n",
      "Epoch 328/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8562 - val_loss: 4.2386\n",
      "Epoch 329/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 2.9196 - val_loss: 4.2524\n",
      "Epoch 330/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8658 - val_loss: 4.2570\n",
      "Epoch 331/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1607 - val_loss: 4.2568\n",
      "Epoch 332/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1080 - val_loss: 4.2339\n",
      "Epoch 333/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1780 - val_loss: 4.2569\n",
      "Epoch 334/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0683 - val_loss: 4.2742\n",
      "Epoch 335/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0129 - val_loss: 4.2429\n",
      "Epoch 336/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1704 - val_loss: 4.2504\n",
      "Epoch 337/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1762 - val_loss: 4.2056\n",
      "Epoch 338/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9745 - val_loss: 4.2185\n",
      "Epoch 339/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0183 - val_loss: 4.2311\n",
      "Epoch 340/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1406 - val_loss: 4.2597\n",
      "Epoch 341/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9681 - val_loss: 4.2765\n",
      "Epoch 342/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.7547 - val_loss: 4.2462\n",
      "Epoch 343/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8324 - val_loss: 4.2461\n",
      "Epoch 344/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1908 - val_loss: 4.2390\n",
      "Epoch 345/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9495 - val_loss: 4.2618\n",
      "Epoch 346/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0257 - val_loss: 4.2526\n",
      "Epoch 347/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9664 - val_loss: 4.2172\n",
      "Epoch 348/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1521 - val_loss: 4.2594\n",
      "Epoch 349/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8531 - val_loss: 4.2657\n",
      "Epoch 350/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0452 - val_loss: 4.2870\n",
      "Epoch 351/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8947 - val_loss: 4.2307\n",
      "Epoch 352/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8647 - val_loss: 4.2198\n",
      "Epoch 353/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0908 - val_loss: 4.2641\n",
      "Epoch 354/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8951 - val_loss: 4.2281\n",
      "Epoch 355/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0038 - val_loss: 4.2212\n",
      "Epoch 356/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8529 - val_loss: 4.2651\n",
      "Epoch 357/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9281 - val_loss: 4.2339\n",
      "Epoch 358/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0194 - val_loss: 4.2496\n",
      "Epoch 359/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8695 - val_loss: 4.2425\n",
      "Epoch 360/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0293 - val_loss: 4.2541\n",
      "Epoch 361/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1414 - val_loss: 4.2163\n",
      "Epoch 362/500\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2.8875 - val_loss: 4.2603\n",
      "Epoch 363/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0045 - val_loss: 4.2477\n",
      "Epoch 364/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 2.8602 - val_loss: 4.2363\n",
      "Epoch 365/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9318 - val_loss: 4.2214\n",
      "Epoch 366/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9889 - val_loss: 4.2530\n",
      "Epoch 367/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9754 - val_loss: 4.2597\n",
      "Epoch 368/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9242 - val_loss: 4.2352\n",
      "Epoch 369/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2332 - val_loss: 4.2328\n",
      "Epoch 370/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0750 - val_loss: 4.2437\n",
      "Epoch 371/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0146 - val_loss: 4.2618\n",
      "Epoch 372/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0853 - val_loss: 4.2443\n",
      "Epoch 373/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8536 - val_loss: 4.2400\n",
      "Epoch 374/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0825 - val_loss: 4.2408\n",
      "Epoch 375/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9363 - val_loss: 4.2352\n",
      "Epoch 376/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8387 - val_loss: 4.2367\n",
      "Epoch 377/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9372 - val_loss: 4.2703\n",
      "Epoch 378/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1392 - val_loss: 4.2509\n",
      "Epoch 379/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1518 - val_loss: 4.2085\n",
      "Epoch 380/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2354 - val_loss: 4.2070\n",
      "Epoch 381/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1283 - val_loss: 4.2398\n",
      "Epoch 382/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1919 - val_loss: 4.2575\n",
      "Epoch 383/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8504 - val_loss: 4.2123\n",
      "Epoch 384/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9856 - val_loss: 4.2085\n",
      "Epoch 385/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9428 - val_loss: 4.2335\n",
      "Epoch 386/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9809 - val_loss: 4.2369\n",
      "Epoch 387/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1957 - val_loss: 4.2215\n",
      "Epoch 388/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8785 - val_loss: 4.2488\n",
      "Epoch 389/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9365 - val_loss: 4.2404\n",
      "Epoch 390/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8384 - val_loss: 4.2193\n",
      "Epoch 391/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1008 - val_loss: 4.2110\n",
      "Epoch 392/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0855 - val_loss: 4.2450\n",
      "Epoch 393/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1769 - val_loss: 4.2445\n",
      "Epoch 394/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9484 - val_loss: 4.2395\n",
      "Epoch 395/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9490 - val_loss: 4.2533\n",
      "Epoch 396/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0304 - val_loss: 4.2700\n",
      "Epoch 397/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9854 - val_loss: 4.2120\n",
      "Epoch 398/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0764 - val_loss: 4.2309\n",
      "Epoch 399/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.2178 - val_loss: 4.3099\n",
      "Epoch 400/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8956 - val_loss: 4.2524\n",
      "Epoch 401/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8338 - val_loss: 4.2250\n",
      "Epoch 402/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0338 - val_loss: 4.2402\n",
      "Epoch 403/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0687 - val_loss: 4.2523\n",
      "Epoch 404/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9080 - val_loss: 4.2390\n",
      "Epoch 405/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0817 - val_loss: 4.2157\n",
      "Epoch 406/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1151 - val_loss: 4.2476\n",
      "Epoch 407/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0062 - val_loss: 4.2335\n",
      "Epoch 408/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9810 - val_loss: 4.2388\n",
      "Epoch 409/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.2050 - val_loss: 4.2324\n",
      "Epoch 410/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9759 - val_loss: 4.2430\n",
      "Epoch 411/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1282 - val_loss: 4.2410\n",
      "Epoch 412/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0144 - val_loss: 4.2440\n",
      "Epoch 413/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0558 - val_loss: 4.2332\n",
      "Epoch 414/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0030 - val_loss: 4.2385\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 415/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9714 - val_loss: 4.2442\n",
      "Epoch 416/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9161 - val_loss: 4.2281\n",
      "Epoch 417/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0357 - val_loss: 4.2489\n",
      "Epoch 418/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8627 - val_loss: 4.2446\n",
      "Epoch 419/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0240 - val_loss: 4.2238\n",
      "Epoch 420/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0825 - val_loss: 4.2295\n",
      "Epoch 421/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0648 - val_loss: 4.2296\n",
      "Epoch 422/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9607 - val_loss: 4.2538\n",
      "Epoch 423/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9047 - val_loss: 4.2281\n",
      "Epoch 424/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8032 - val_loss: 4.2378\n",
      "Epoch 425/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9935 - val_loss: 4.2390\n",
      "Epoch 426/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9624 - val_loss: 4.2298\n",
      "Epoch 427/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0579 - val_loss: 4.2552\n",
      "Epoch 428/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1318 - val_loss: 4.2193\n",
      "Epoch 429/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9884 - val_loss: 4.2173\n",
      "Epoch 430/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8864 - val_loss: 4.2383\n",
      "Epoch 431/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8799 - val_loss: 4.2320\n",
      "Epoch 432/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9079 - val_loss: 4.2316\n",
      "Epoch 433/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0356 - val_loss: 4.2259\n",
      "Epoch 434/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8847 - val_loss: 4.2425\n",
      "Epoch 435/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0012 - val_loss: 4.2149\n",
      "Epoch 436/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2689 - val_loss: 4.2476\n",
      "Epoch 437/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9453 - val_loss: 4.2353\n",
      "Epoch 438/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9831 - val_loss: 4.2525\n",
      "Epoch 439/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8299 - val_loss: 4.2232\n",
      "Epoch 440/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9642 - val_loss: 4.2504\n",
      "Epoch 441/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0727 - val_loss: 4.2314\n",
      "Epoch 442/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0301 - val_loss: 4.2174\n",
      "Epoch 443/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9667 - val_loss: 4.2505\n",
      "Epoch 444/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8545 - val_loss: 4.2683\n",
      "Epoch 445/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9520 - val_loss: 4.2110\n",
      "Epoch 446/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.6656 - val_loss: 4.1972\n",
      "Epoch 447/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.2338 - val_loss: 4.2084\n",
      "Epoch 448/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0903 - val_loss: 4.2590\n",
      "Epoch 449/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1642 - val_loss: 4.2526\n",
      "Epoch 450/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1322 - val_loss: 4.2320\n",
      "Epoch 451/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9411 - val_loss: 4.2282\n",
      "Epoch 452/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0720 - val_loss: 4.2334\n",
      "Epoch 453/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9037 - val_loss: 4.2426\n",
      "Epoch 454/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8738 - val_loss: 4.2315\n",
      "Epoch 455/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0689 - val_loss: 4.2321\n",
      "Epoch 456/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0448 - val_loss: 4.2079\n",
      "Epoch 457/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9164 - val_loss: 4.2096\n",
      "Epoch 458/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9444 - val_loss: 4.2412\n",
      "Epoch 459/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9175 - val_loss: 4.2279\n",
      "Epoch 460/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0099 - val_loss: 4.2194\n",
      "Epoch 461/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1176 - val_loss: 4.2611\n",
      "Epoch 462/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0386 - val_loss: 4.2418\n",
      "Epoch 463/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8213 - val_loss: 4.2195\n",
      "Epoch 464/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.7902 - val_loss: 4.2155\n",
      "Epoch 465/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8880 - val_loss: 4.2329\n",
      "Epoch 466/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8964 - val_loss: 4.2309\n",
      "Epoch 467/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.7733 - val_loss: 4.2397\n",
      "Epoch 468/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9165 - val_loss: 4.2152\n",
      "Epoch 469/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9004 - val_loss: 4.2132\n",
      "Epoch 470/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0237 - val_loss: 4.2610\n",
      "Epoch 471/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8443 - val_loss: 4.2233\n",
      "Epoch 472/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9246 - val_loss: 4.2148\n",
      "Epoch 473/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1186 - val_loss: 4.2533\n",
      "Epoch 474/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9158 - val_loss: 4.2314\n",
      "Epoch 475/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8904 - val_loss: 4.2205\n",
      "Epoch 476/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.0610 - val_loss: 4.2456\n",
      "Epoch 477/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9044 - val_loss: 4.2281\n",
      "Epoch 478/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1081 - val_loss: 4.2263\n",
      "Epoch 479/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.2519 - val_loss: 4.2446\n",
      "Epoch 480/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8118 - val_loss: 4.2194\n",
      "Epoch 481/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.1496 - val_loss: 4.2435\n",
      "Epoch 482/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9511 - val_loss: 4.2261\n",
      "Epoch 483/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9999 - val_loss: 4.2349\n",
      "Epoch 484/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1040 - val_loss: 4.2339\n",
      "Epoch 485/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9472 - val_loss: 4.2354\n",
      "Epoch 486/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8390 - val_loss: 4.2154\n",
      "Epoch 487/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.8397 - val_loss: 4.2320\n",
      "Epoch 488/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9589 - val_loss: 4.2235\n",
      "Epoch 489/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.9725 - val_loss: 4.2488\n",
      "Epoch 490/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.8948 - val_loss: 4.2182\n",
      "Epoch 491/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2510 - val_loss: 4.1944\n",
      "Epoch 492/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9782 - val_loss: 4.2304\n",
      "Epoch 493/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9355 - val_loss: 4.2447\n",
      "Epoch 494/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.0719 - val_loss: 4.2371\n",
      "Epoch 495/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.0500 - val_loss: 4.2201\n",
      "Epoch 496/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.1436 - val_loss: 4.2243\n",
      "Epoch 497/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8/8 [==============================] - 0s 5ms/step - loss: 3.1167 - val_loss: 4.2216\n",
      "Epoch 498/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 3.1159 - val_loss: 4.2551\n",
      "Epoch 499/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2114 - val_loss: 4.2273\n",
      "Epoch 500/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 2.8806 - val_loss: 4.2239\n"
     ]
    }
   ],
   "source": [
    "#%%time\n",
    "history_horsepower_regression = non_linear_horsepower_regressor.fit(train_features['Horsepower'], train_labels, validation_split = 0.2, verbose = 1, epochs = 500)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "plot_loss(history_horsepower_regression)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "x = tf.linspace(0.0, 250, 251)\n",
    "y = non_linear_horsepower_regressor.predict(x)\n",
    "plot_horsepower(x, y)\n",
    "#We see that introducing non-linearity to our regression model betters the fit!!\n",
    "test_results['non_linear_horsepower_regressor'] = non_linear_horsepower_regressor.evaluate(test_features['Horsepower'], test_labels, verbose = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/500\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 23.4017 - val_loss: 22.8934\n",
      "Epoch 2/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 23.1740 - val_loss: 22.3544\n",
      "Epoch 3/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 21.6336 - val_loss: 21.6723\n",
      "Epoch 4/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 21.0077 - val_loss: 20.7641\n",
      "Epoch 5/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 20.5105 - val_loss: 19.5253\n",
      "Epoch 6/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 18.6709 - val_loss: 17.8715\n",
      "Epoch 7/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 17.2758 - val_loss: 15.8011\n",
      "Epoch 8/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 14.9625 - val_loss: 13.6545\n",
      "Epoch 9/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 13.0935 - val_loss: 11.4440\n",
      "Epoch 10/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 10.8452 - val_loss: 8.7678\n",
      "Epoch 11/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 7.6795 - val_loss: 6.3427\n",
      "Epoch 12/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 5.8968 - val_loss: 5.3803\n",
      "Epoch 13/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 5.2420 - val_loss: 4.7151\n",
      "Epoch 14/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 4.4910 - val_loss: 4.0016\n",
      "Epoch 15/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.7559 - val_loss: 3.5039\n",
      "Epoch 16/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 3.7313 - val_loss: 3.2706\n",
      "Epoch 17/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 3.2956 - val_loss: 3.1049\n",
      "Epoch 18/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.9954 - val_loss: 3.0236\n",
      "Epoch 19/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.6998 - val_loss: 3.0392\n",
      "Epoch 20/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.5858 - val_loss: 3.0557\n",
      "Epoch 21/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.7394 - val_loss: 2.9817\n",
      "Epoch 22/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.6197 - val_loss: 2.9255\n",
      "Epoch 23/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 2.3921 - val_loss: 2.8454\n",
      "Epoch 24/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2625 - val_loss: 2.7759\n",
      "Epoch 25/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 2.1833 - val_loss: 2.6845\n",
      "Epoch 26/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.0991 - val_loss: 2.6522\n",
      "Epoch 27/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.2306 - val_loss: 2.6075\n",
      "Epoch 28/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 2.1075 - val_loss: 2.5444\n",
      "Epoch 29/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.9918 - val_loss: 2.5600\n",
      "Epoch 30/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.9384 - val_loss: 2.5456\n",
      "Epoch 31/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.8743 - val_loss: 2.5193\n",
      "Epoch 32/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.8545 - val_loss: 2.5666\n",
      "Epoch 33/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.9737 - val_loss: 2.5192\n",
      "Epoch 34/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.8714 - val_loss: 2.4957\n",
      "Epoch 35/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.8215 - val_loss: 2.4964\n",
      "Epoch 36/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.7778 - val_loss: 2.4954\n",
      "Epoch 37/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.9133 - val_loss: 2.4782\n",
      "Epoch 38/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.7541 - val_loss: 2.5158\n",
      "Epoch 39/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.8225 - val_loss: 2.4517\n",
      "Epoch 40/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 1.8468 - val_loss: 2.4998\n",
      "Epoch 41/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.9415 - val_loss: 2.4904\n",
      "Epoch 42/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.9644 - val_loss: 2.4623\n",
      "Epoch 43/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.8971 - val_loss: 2.4801\n",
      "Epoch 44/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.7091 - val_loss: 2.4504\n",
      "Epoch 45/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.7621 - val_loss: 2.4540\n",
      "Epoch 46/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.7032 - val_loss: 2.4620\n",
      "Epoch 47/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.5812 - val_loss: 2.4645\n",
      "Epoch 48/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 1.7812 - val_loss: 2.4741\n",
      "Epoch 49/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.7571 - val_loss: 2.4757\n",
      "Epoch 50/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.7446 - val_loss: 2.4730\n",
      "Epoch 51/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.6842 - val_loss: 2.4529\n",
      "Epoch 52/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.8877 - val_loss: 2.4237\n",
      "Epoch 53/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.8846 - val_loss: 2.4423\n",
      "Epoch 54/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.5764 - val_loss: 2.4492\n",
      "Epoch 55/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.8135 - val_loss: 2.5014\n",
      "Epoch 56/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.7727 - val_loss: 2.4486\n",
      "Epoch 57/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.6988 - val_loss: 2.4316\n",
      "Epoch 58/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.8162 - val_loss: 2.4427\n",
      "Epoch 59/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.8040 - val_loss: 2.4406\n",
      "Epoch 60/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.6409 - val_loss: 2.4512\n",
      "Epoch 61/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.7366 - val_loss: 2.4486\n",
      "Epoch 62/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.7633 - val_loss: 2.4140\n",
      "Epoch 63/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.6789 - val_loss: 2.4593\n",
      "Epoch 64/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.8370 - val_loss: 2.4376\n",
      "Epoch 65/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.6585 - val_loss: 2.4008\n",
      "Epoch 66/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.7101 - val_loss: 2.4493\n",
      "Epoch 67/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.7002 - val_loss: 2.3834\n",
      "Epoch 68/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.7269 - val_loss: 2.4159\n",
      "Epoch 69/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.6514 - val_loss: 2.4067\n",
      "Epoch 70/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.6019 - val_loss: 2.4348\n",
      "Epoch 71/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.7630 - val_loss: 2.3932\n",
      "Epoch 72/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.6365 - val_loss: 2.4298\n",
      "Epoch 73/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.7237 - val_loss: 2.4102\n",
      "Epoch 74/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.7150 - val_loss: 2.3898\n",
      "Epoch 75/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.6244 - val_loss: 2.4407\n",
      "Epoch 76/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5987 - val_loss: 2.3546\n",
      "Epoch 77/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.4806 - val_loss: 2.4115\n",
      "Epoch 78/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.6183 - val_loss: 2.4289\n",
      "Epoch 79/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.6104 - val_loss: 2.3803\n",
      "Epoch 80/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.4889 - val_loss: 2.3881\n",
      "Epoch 81/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.6183 - val_loss: 2.3877\n",
      "Epoch 82/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.7206 - val_loss: 2.3864\n",
      "Epoch 83/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.5042 - val_loss: 2.4014\n",
      "Epoch 84/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5409 - val_loss: 2.3631\n",
      "Epoch 85/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.5791 - val_loss: 2.4347\n",
      "Epoch 86/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5223 - val_loss: 2.3619\n",
      "Epoch 87/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5226 - val_loss: 2.4025\n",
      "Epoch 88/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5662 - val_loss: 2.3862\n",
      "Epoch 89/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.6806 - val_loss: 2.3847\n",
      "Epoch 90/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.6937 - val_loss: 2.3927\n",
      "Epoch 91/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4728 - val_loss: 2.3663\n",
      "Epoch 92/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5359 - val_loss: 2.3778\n",
      "Epoch 93/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5414 - val_loss: 2.3751\n",
      "Epoch 94/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5227 - val_loss: 2.3934\n",
      "Epoch 95/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5834 - val_loss: 2.3896\n",
      "Epoch 96/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4838 - val_loss: 2.4366\n",
      "Epoch 97/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5244 - val_loss: 2.3478\n",
      "Epoch 98/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3702 - val_loss: 2.3844\n",
      "Epoch 99/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.9061 - val_loss: 2.3699\n",
      "Epoch 100/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3948 - val_loss: 2.4075\n",
      "Epoch 101/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.6971 - val_loss: 2.3801\n",
      "Epoch 102/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5467 - val_loss: 2.3637\n",
      "Epoch 103/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4555 - val_loss: 2.4004\n",
      "Epoch 104/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.6493 - val_loss: 2.3506\n",
      "Epoch 105/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5705 - val_loss: 2.3762\n",
      "Epoch 106/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5823 - val_loss: 2.3969\n",
      "Epoch 107/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5042 - val_loss: 2.3643\n",
      "Epoch 108/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4908 - val_loss: 2.3660\n",
      "Epoch 109/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4271 - val_loss: 2.3564\n",
      "Epoch 110/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5016 - val_loss: 2.4169\n",
      "Epoch 111/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4469 - val_loss: 2.3326\n",
      "Epoch 112/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4485 - val_loss: 2.3741\n",
      "Epoch 113/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4416 - val_loss: 2.3393\n",
      "Epoch 114/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.6159 - val_loss: 2.3921\n",
      "Epoch 115/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5856 - val_loss: 2.3761\n",
      "Epoch 116/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5066 - val_loss: 2.3662\n",
      "Epoch 117/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2740 - val_loss: 2.4095\n",
      "Epoch 118/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.6144 - val_loss: 2.3702\n",
      "Epoch 119/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4727 - val_loss: 2.3637\n",
      "Epoch 120/500\n",
      "8/8 [==============================] - ETA: 0s - loss: 1.139 - 0s 3ms/step - loss: 1.4654 - val_loss: 2.3564\n",
      "Epoch 121/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5123 - val_loss: 2.3835\n",
      "Epoch 122/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.6086 - val_loss: 2.3754\n",
      "Epoch 123/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4966 - val_loss: 2.3583\n",
      "Epoch 124/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.5495 - val_loss: 2.3417\n",
      "Epoch 125/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4595 - val_loss: 2.3465\n",
      "Epoch 126/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4667 - val_loss: 2.3818\n",
      "Epoch 127/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5563 - val_loss: 2.3759\n",
      "Epoch 128/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4151 - val_loss: 2.3688\n",
      "Epoch 129/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4568 - val_loss: 2.3449\n",
      "Epoch 130/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.4367 - val_loss: 2.3925\n",
      "Epoch 131/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.6115 - val_loss: 2.3600\n",
      "Epoch 132/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5132 - val_loss: 2.3782\n",
      "Epoch 133/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4238 - val_loss: 2.3783\n",
      "Epoch 134/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.5169 - val_loss: 2.3598\n",
      "Epoch 135/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4424 - val_loss: 2.3892\n",
      "Epoch 136/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5112 - val_loss: 2.3725\n",
      "Epoch 137/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3392 - val_loss: 2.3517\n",
      "Epoch 138/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5157 - val_loss: 2.3727\n",
      "Epoch 139/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4643 - val_loss: 2.3662\n",
      "Epoch 140/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4200 - val_loss: 2.3630\n",
      "Epoch 141/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.5012 - val_loss: 2.3611\n",
      "Epoch 142/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4044 - val_loss: 2.3736\n",
      "Epoch 143/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.5206 - val_loss: 2.3591\n",
      "Epoch 144/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5036 - val_loss: 2.3680\n",
      "Epoch 145/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4359 - val_loss: 2.3602\n",
      "Epoch 146/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4720 - val_loss: 2.4028\n",
      "Epoch 147/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.4027 - val_loss: 2.3427\n",
      "Epoch 148/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.5658 - val_loss: 2.3980\n",
      "Epoch 149/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4520 - val_loss: 2.3789\n",
      "Epoch 150/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4159 - val_loss: 2.3777\n",
      "Epoch 151/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.5614 - val_loss: 2.3478\n",
      "Epoch 152/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.5416 - val_loss: 2.3950\n",
      "Epoch 153/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3913 - val_loss: 2.3916\n",
      "Epoch 154/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.4662 - val_loss: 2.4054\n",
      "Epoch 155/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4735 - val_loss: 2.3476\n",
      "Epoch 156/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3065 - val_loss: 2.3980\n",
      "Epoch 157/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4363 - val_loss: 2.3579\n",
      "Epoch 158/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4505 - val_loss: 2.3710\n",
      "Epoch 159/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.3283 - val_loss: 2.3966\n",
      "Epoch 160/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.5536 - val_loss: 2.3587\n",
      "Epoch 161/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.4674 - val_loss: 2.3618\n",
      "Epoch 162/500\n",
      "8/8 [==============================] - ETA: 0s - loss: 1.252 - 0s 4ms/step - loss: 1.3880 - val_loss: 2.3820\n",
      "Epoch 163/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.3509 - val_loss: 2.4036\n",
      "Epoch 164/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.3351 - val_loss: 2.3828\n",
      "Epoch 165/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.3601 - val_loss: 2.3788\n",
      "Epoch 166/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8/8 [==============================] - 0s 4ms/step - loss: 1.5741 - val_loss: 2.3824\n",
      "Epoch 167/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.4556 - val_loss: 2.3749\n",
      "Epoch 168/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4252 - val_loss: 2.3779\n",
      "Epoch 169/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.3118 - val_loss: 2.3653\n",
      "Epoch 170/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.4153 - val_loss: 2.3569\n",
      "Epoch 171/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3414 - val_loss: 2.3582\n",
      "Epoch 172/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.4397 - val_loss: 2.3731\n",
      "Epoch 173/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.4235 - val_loss: 2.3327\n",
      "Epoch 174/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2664 - val_loss: 2.3648\n",
      "Epoch 175/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.3944 - val_loss: 2.3617\n",
      "Epoch 176/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.3874 - val_loss: 2.3691\n",
      "Epoch 177/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.3871 - val_loss: 2.3601\n",
      "Epoch 178/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.4358 - val_loss: 2.3507\n",
      "Epoch 179/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.4215 - val_loss: 2.3398\n",
      "Epoch 180/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.5402 - val_loss: 2.3685\n",
      "Epoch 181/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.4293 - val_loss: 2.3169\n",
      "Epoch 182/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.3728 - val_loss: 2.3869\n",
      "Epoch 183/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.4737 - val_loss: 2.3546\n",
      "Epoch 184/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4592 - val_loss: 2.3530\n",
      "Epoch 185/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3191 - val_loss: 2.3578\n",
      "Epoch 186/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3578 - val_loss: 2.3688\n",
      "Epoch 187/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2962 - val_loss: 2.3431\n",
      "Epoch 188/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3265 - val_loss: 2.3507\n",
      "Epoch 189/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3240 - val_loss: 2.3483\n",
      "Epoch 190/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3417 - val_loss: 2.3482\n",
      "Epoch 191/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.3767 - val_loss: 2.3596\n",
      "Epoch 192/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.3417 - val_loss: 2.3545\n",
      "Epoch 193/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2798 - val_loss: 2.3478\n",
      "Epoch 194/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.3189 - val_loss: 2.3618\n",
      "Epoch 195/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.4514 - val_loss: 2.3317\n",
      "Epoch 196/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.3174 - val_loss: 2.3668\n",
      "Epoch 197/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4256 - val_loss: 2.3661\n",
      "Epoch 198/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2618 - val_loss: 2.3566\n",
      "Epoch 199/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4182 - val_loss: 2.3249\n",
      "Epoch 200/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2910 - val_loss: 2.3730\n",
      "Epoch 201/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3470 - val_loss: 2.3741\n",
      "Epoch 202/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2995 - val_loss: 2.3727\n",
      "Epoch 203/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3136 - val_loss: 2.3796\n",
      "Epoch 204/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3504 - val_loss: 2.3674\n",
      "Epoch 205/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3959 - val_loss: 2.3533\n",
      "Epoch 206/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2621 - val_loss: 2.3472\n",
      "Epoch 207/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.4023 - val_loss: 2.3496\n",
      "Epoch 208/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.3262 - val_loss: 2.3774\n",
      "Epoch 209/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2160 - val_loss: 2.3414\n",
      "Epoch 210/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2542 - val_loss: 2.3494\n",
      "Epoch 211/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.3055 - val_loss: 2.3567\n",
      "Epoch 212/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.4069 - val_loss: 2.3232\n",
      "Epoch 213/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2881 - val_loss: 2.3611\n",
      "Epoch 214/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2809 - val_loss: 2.3701\n",
      "Epoch 215/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.3376 - val_loss: 2.3399\n",
      "Epoch 216/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4020 - val_loss: 2.3787\n",
      "Epoch 217/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2398 - val_loss: 2.3215\n",
      "Epoch 218/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4615 - val_loss: 2.3743\n",
      "Epoch 219/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3446 - val_loss: 2.3282\n",
      "Epoch 220/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3003 - val_loss: 2.3767\n",
      "Epoch 221/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2249 - val_loss: 2.3704\n",
      "Epoch 222/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2838 - val_loss: 2.3500\n",
      "Epoch 223/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2430 - val_loss: 2.3487\n",
      "Epoch 224/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3368 - val_loss: 2.3787\n",
      "Epoch 225/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1466 - val_loss: 2.3367\n",
      "Epoch 226/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2917 - val_loss: 2.3493\n",
      "Epoch 227/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1874 - val_loss: 2.3758\n",
      "Epoch 228/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2777 - val_loss: 2.3606\n",
      "Epoch 229/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3010 - val_loss: 2.3620\n",
      "Epoch 230/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3443 - val_loss: 2.3649\n",
      "Epoch 231/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2371 - val_loss: 2.3704\n",
      "Epoch 232/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3713 - val_loss: 2.3770\n",
      "Epoch 233/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2785 - val_loss: 2.3493\n",
      "Epoch 234/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3582 - val_loss: 2.3616\n",
      "Epoch 235/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2070 - val_loss: 2.3650\n",
      "Epoch 236/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2462 - val_loss: 2.3856\n",
      "Epoch 237/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2286 - val_loss: 2.3450\n",
      "Epoch 238/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2906 - val_loss: 2.3563\n",
      "Epoch 239/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3926 - val_loss: 2.3698\n",
      "Epoch 240/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2434 - val_loss: 2.3861\n",
      "Epoch 241/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3345 - val_loss: 2.3889\n",
      "Epoch 242/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1147 - val_loss: 2.3597\n",
      "Epoch 243/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3269 - val_loss: 2.3533\n",
      "Epoch 244/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2013 - val_loss: 2.3844\n",
      "Epoch 245/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2104 - val_loss: 2.3816\n",
      "Epoch 246/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4126 - val_loss: 2.3170\n",
      "Epoch 247/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1890 - val_loss: 2.3666\n",
      "Epoch 248/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4441 - val_loss: 2.3540\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 249/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4071 - val_loss: 2.3408\n",
      "Epoch 250/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3395 - val_loss: 2.3730\n",
      "Epoch 251/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2871 - val_loss: 2.3427\n",
      "Epoch 252/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3876 - val_loss: 2.4007\n",
      "Epoch 253/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3023 - val_loss: 2.3397\n",
      "Epoch 254/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.3238 - val_loss: 2.3411\n",
      "Epoch 255/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2325 - val_loss: 2.4148\n",
      "Epoch 256/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1775 - val_loss: 2.3802\n",
      "Epoch 257/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2935 - val_loss: 2.3570\n",
      "Epoch 258/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2313 - val_loss: 2.3647\n",
      "Epoch 259/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2382 - val_loss: 2.3683\n",
      "Epoch 260/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2346 - val_loss: 2.3783\n",
      "Epoch 261/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2670 - val_loss: 2.3461\n",
      "Epoch 262/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.4029 - val_loss: 2.4110\n",
      "Epoch 263/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2364 - val_loss: 2.3461\n",
      "Epoch 264/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1898 - val_loss: 2.4198\n",
      "Epoch 265/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.3605 - val_loss: 2.3042\n",
      "Epoch 266/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2719 - val_loss: 2.3581\n",
      "Epoch 267/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2734 - val_loss: 2.3373\n",
      "Epoch 268/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2125 - val_loss: 2.3442\n",
      "Epoch 269/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2770 - val_loss: 2.3467\n",
      "Epoch 270/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2286 - val_loss: 2.3449\n",
      "Epoch 271/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1401 - val_loss: 2.3657\n",
      "Epoch 272/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1376 - val_loss: 2.3336\n",
      "Epoch 273/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3890 - val_loss: 2.3600\n",
      "Epoch 274/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2856 - val_loss: 2.3672\n",
      "Epoch 275/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3068 - val_loss: 2.3758\n",
      "Epoch 276/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2033 - val_loss: 2.3490\n",
      "Epoch 277/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2273 - val_loss: 2.3594\n",
      "Epoch 278/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2335 - val_loss: 2.3678\n",
      "Epoch 279/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1492 - val_loss: 2.3871\n",
      "Epoch 280/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2615 - val_loss: 2.3372\n",
      "Epoch 281/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2489 - val_loss: 2.4299\n",
      "Epoch 282/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2352 - val_loss: 2.3333\n",
      "Epoch 283/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2068 - val_loss: 2.3942\n",
      "Epoch 284/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2119 - val_loss: 2.3217\n",
      "Epoch 285/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.3043 - val_loss: 2.3936\n",
      "Epoch 286/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2511 - val_loss: 2.3149\n",
      "Epoch 287/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1960 - val_loss: 2.3894\n",
      "Epoch 288/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2449 - val_loss: 2.3332\n",
      "Epoch 289/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1581 - val_loss: 2.3222\n",
      "Epoch 290/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2090 - val_loss: 2.4008\n",
      "Epoch 291/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2719 - val_loss: 2.3332\n",
      "Epoch 292/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2486 - val_loss: 2.3629\n",
      "Epoch 293/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2416 - val_loss: 2.3547\n",
      "Epoch 294/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1390 - val_loss: 2.3627\n",
      "Epoch 295/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2220 - val_loss: 2.3280\n",
      "Epoch 296/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2900 - val_loss: 2.3593\n",
      "Epoch 297/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2879 - val_loss: 2.3792\n",
      "Epoch 298/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2484 - val_loss: 2.3936\n",
      "Epoch 299/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0639 - val_loss: 2.3397\n",
      "Epoch 300/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2519 - val_loss: 2.3180\n",
      "Epoch 301/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2405 - val_loss: 2.3837\n",
      "Epoch 302/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1150 - val_loss: 2.3078\n",
      "Epoch 303/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2516 - val_loss: 2.3723\n",
      "Epoch 304/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0496 - val_loss: 2.3378\n",
      "Epoch 305/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.0933 - val_loss: 2.3274\n",
      "Epoch 306/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2278 - val_loss: 2.3424\n",
      "Epoch 307/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3174 - val_loss: 2.3708\n",
      "Epoch 308/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1881 - val_loss: 2.3397\n",
      "Epoch 309/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2662 - val_loss: 2.4179\n",
      "Epoch 310/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2508 - val_loss: 2.3695\n",
      "Epoch 311/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1459 - val_loss: 2.3453\n",
      "Epoch 312/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3557 - val_loss: 2.3600\n",
      "Epoch 313/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2906 - val_loss: 2.3530\n",
      "Epoch 314/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2387 - val_loss: 2.3790\n",
      "Epoch 315/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2689 - val_loss: 2.3247\n",
      "Epoch 316/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0977 - val_loss: 2.3945\n",
      "Epoch 317/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0833 - val_loss: 2.3237\n",
      "Epoch 318/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1554 - val_loss: 2.3735\n",
      "Epoch 319/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1541 - val_loss: 2.3738\n",
      "Epoch 320/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.3397 - val_loss: 2.3314\n",
      "Epoch 321/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.0589 - val_loss: 2.4215\n",
      "Epoch 322/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2514 - val_loss: 2.3621\n",
      "Epoch 323/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2387 - val_loss: 2.3464\n",
      "Epoch 324/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2488 - val_loss: 2.3476\n",
      "Epoch 325/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2164 - val_loss: 2.3500\n",
      "Epoch 326/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.0742 - val_loss: 2.3552\n",
      "Epoch 327/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2037 - val_loss: 2.3633\n",
      "Epoch 328/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2345 - val_loss: 2.3485\n",
      "Epoch 329/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.0738 - val_loss: 2.3658\n",
      "Epoch 330/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1130 - val_loss: 2.3436\n",
      "Epoch 331/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2608 - val_loss: 2.3966\n",
      "Epoch 332/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2456 - val_loss: 2.3109\n",
      "Epoch 333/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2924 - val_loss: 2.3797\n",
      "Epoch 334/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1734 - val_loss: 2.3604\n",
      "Epoch 335/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2730 - val_loss: 2.3573\n",
      "Epoch 336/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2226 - val_loss: 2.3379\n",
      "Epoch 337/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2136 - val_loss: 2.3491\n",
      "Epoch 338/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1647 - val_loss: 2.3740\n",
      "Epoch 339/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1351 - val_loss: 2.3147\n",
      "Epoch 340/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2109 - val_loss: 2.3682\n",
      "Epoch 341/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1189 - val_loss: 2.3419\n",
      "Epoch 342/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1447 - val_loss: 2.3249\n",
      "Epoch 343/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1921 - val_loss: 2.3702\n",
      "Epoch 344/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.0749 - val_loss: 2.3450\n",
      "Epoch 345/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1673 - val_loss: 2.3938\n",
      "Epoch 346/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.0881 - val_loss: 2.3198\n",
      "Epoch 347/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.0937 - val_loss: 2.3680\n",
      "Epoch 348/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2041 - val_loss: 2.3805\n",
      "Epoch 349/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.0592 - val_loss: 2.3484\n",
      "Epoch 350/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.0615 - val_loss: 2.3314\n",
      "Epoch 351/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1816 - val_loss: 2.3334\n",
      "Epoch 352/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0769 - val_loss: 2.3546\n",
      "Epoch 353/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1846 - val_loss: 2.3281\n",
      "Epoch 354/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1409 - val_loss: 2.3483\n",
      "Epoch 355/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2428 - val_loss: 2.3259\n",
      "Epoch 356/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2890 - val_loss: 2.4016\n",
      "Epoch 357/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1273 - val_loss: 2.3634\n",
      "Epoch 358/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1347 - val_loss: 2.3817\n",
      "Epoch 359/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1580 - val_loss: 2.3506\n",
      "Epoch 360/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2924 - val_loss: 2.3697\n",
      "Epoch 361/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0142 - val_loss: 2.3652\n",
      "Epoch 362/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1648 - val_loss: 2.3541\n",
      "Epoch 363/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 0.9790 - val_loss: 2.3395\n",
      "Epoch 364/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2635 - val_loss: 2.3622\n",
      "Epoch 365/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1358 - val_loss: 2.3939\n",
      "Epoch 366/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0464 - val_loss: 2.3779\n",
      "Epoch 367/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 0.9988 - val_loss: 2.3828\n",
      "Epoch 368/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1229 - val_loss: 2.3542\n",
      "Epoch 369/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1045 - val_loss: 2.3766\n",
      "Epoch 370/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1363 - val_loss: 2.3368\n",
      "Epoch 371/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 1.1411 - val_loss: 2.3785\n",
      "Epoch 372/500\n",
      "8/8 [==============================] - 0s 6ms/step - loss: 1.2143 - val_loss: 2.3845\n",
      "Epoch 373/500\n",
      "8/8 [==============================] - 0s 6ms/step - loss: 1.1952 - val_loss: 2.3580\n",
      "Epoch 374/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 1.1071 - val_loss: 2.3628\n",
      "Epoch 375/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1285 - val_loss: 2.3596\n",
      "Epoch 376/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1432 - val_loss: 2.3875\n",
      "Epoch 377/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0973 - val_loss: 2.3434\n",
      "Epoch 378/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0149 - val_loss: 2.3756\n",
      "Epoch 379/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1700 - val_loss: 2.3393\n",
      "Epoch 380/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1283 - val_loss: 2.3735\n",
      "Epoch 381/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1831 - val_loss: 2.3559\n",
      "Epoch 382/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0384 - val_loss: 2.3829\n",
      "Epoch 383/500\n",
      "8/8 [==============================] - 0s 5ms/step - loss: 1.0917 - val_loss: 2.3473\n",
      "Epoch 384/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.0819 - val_loss: 2.3795\n",
      "Epoch 385/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.0203 - val_loss: 2.3567\n",
      "Epoch 386/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2043 - val_loss: 2.3752\n",
      "Epoch 387/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1138 - val_loss: 2.3836\n",
      "Epoch 388/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2151 - val_loss: 2.3500\n",
      "Epoch 389/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1212 - val_loss: 2.4158\n",
      "Epoch 390/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1549 - val_loss: 2.3485\n",
      "Epoch 391/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1445 - val_loss: 2.4120\n",
      "Epoch 392/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2149 - val_loss: 2.3728\n",
      "Epoch 393/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1453 - val_loss: 2.3689\n",
      "Epoch 394/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0947 - val_loss: 2.3876\n",
      "Epoch 395/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1159 - val_loss: 2.3395\n",
      "Epoch 396/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1049 - val_loss: 2.4252\n",
      "Epoch 397/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1532 - val_loss: 2.3748\n",
      "Epoch 398/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1840 - val_loss: 2.4277\n",
      "Epoch 399/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1084 - val_loss: 2.3496\n",
      "Epoch 400/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2417 - val_loss: 2.3946\n",
      "Epoch 401/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1066 - val_loss: 2.4372\n",
      "Epoch 402/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1058 - val_loss: 2.3612\n",
      "Epoch 403/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1617 - val_loss: 2.4030\n",
      "Epoch 404/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1119 - val_loss: 2.3802\n",
      "Epoch 405/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1766 - val_loss: 2.3932\n",
      "Epoch 406/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1380 - val_loss: 2.3650\n",
      "Epoch 407/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1051 - val_loss: 2.4012\n",
      "Epoch 408/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1249 - val_loss: 2.3809\n",
      "Epoch 409/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0631 - val_loss: 2.3739\n",
      "Epoch 410/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1360 - val_loss: 2.4073\n",
      "Epoch 411/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1680 - val_loss: 2.3684\n",
      "Epoch 412/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.0460 - val_loss: 2.4076\n",
      "Epoch 413/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2104 - val_loss: 2.3675\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 414/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.0029 - val_loss: 2.3903\n",
      "Epoch 415/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.0278 - val_loss: 2.3430\n",
      "Epoch 416/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2806 - val_loss: 2.4114\n",
      "Epoch 417/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 0.9935 - val_loss: 2.3790\n",
      "Epoch 418/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1241 - val_loss: 2.3593\n",
      "Epoch 419/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0553 - val_loss: 2.4440\n",
      "Epoch 420/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1058 - val_loss: 2.3435\n",
      "Epoch 421/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2818 - val_loss: 2.4207\n",
      "Epoch 422/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1752 - val_loss: 2.3797\n",
      "Epoch 423/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 0.9834 - val_loss: 2.4317\n",
      "Epoch 424/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0613 - val_loss: 2.3575\n",
      "Epoch 425/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1881 - val_loss: 2.4612\n",
      "Epoch 426/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1227 - val_loss: 2.3526\n",
      "Epoch 427/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1992 - val_loss: 2.4272\n",
      "Epoch 428/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1278 - val_loss: 2.3543\n",
      "Epoch 429/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2193 - val_loss: 2.4001\n",
      "Epoch 430/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1552 - val_loss: 2.4006\n",
      "Epoch 431/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2030 - val_loss: 2.4062\n",
      "Epoch 432/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1136 - val_loss: 2.3691\n",
      "Epoch 433/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1442 - val_loss: 2.4122\n",
      "Epoch 434/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0110 - val_loss: 2.3751\n",
      "Epoch 435/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1771 - val_loss: 2.3977\n",
      "Epoch 436/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2024 - val_loss: 2.3929\n",
      "Epoch 437/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1120 - val_loss: 2.4475\n",
      "Epoch 438/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1997 - val_loss: 2.3509\n",
      "Epoch 439/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0359 - val_loss: 2.3855\n",
      "Epoch 440/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1421 - val_loss: 2.4497\n",
      "Epoch 441/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1857 - val_loss: 2.3753\n",
      "Epoch 442/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1406 - val_loss: 2.4172\n",
      "Epoch 443/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 0.9551 - val_loss: 2.3703\n",
      "Epoch 444/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 0.9709 - val_loss: 2.4010\n",
      "Epoch 445/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1037 - val_loss: 2.3924\n",
      "Epoch 446/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1542 - val_loss: 2.4416\n",
      "Epoch 447/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1117 - val_loss: 2.3493\n",
      "Epoch 448/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1179 - val_loss: 2.4304\n",
      "Epoch 449/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2618 - val_loss: 2.4076\n",
      "Epoch 450/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0571 - val_loss: 2.4252\n",
      "Epoch 451/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1892 - val_loss: 2.3616\n",
      "Epoch 452/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0872 - val_loss: 2.4320\n",
      "Epoch 453/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1480 - val_loss: 2.3587\n",
      "Epoch 454/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1350 - val_loss: 2.3896\n",
      "Epoch 455/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1070 - val_loss: 2.3866\n",
      "Epoch 456/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2032 - val_loss: 2.3899\n",
      "Epoch 457/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1001 - val_loss: 2.4349\n",
      "Epoch 458/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0568 - val_loss: 2.3412\n",
      "Epoch 459/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1200 - val_loss: 2.4700\n",
      "Epoch 460/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1327 - val_loss: 2.4010\n",
      "Epoch 461/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1487 - val_loss: 2.3768\n",
      "Epoch 462/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 0.9940 - val_loss: 2.3968\n",
      "Epoch 463/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.0584 - val_loss: 2.3855\n",
      "Epoch 464/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1464 - val_loss: 2.3451\n",
      "Epoch 465/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0652 - val_loss: 2.4257\n",
      "Epoch 466/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1564 - val_loss: 2.4137\n",
      "Epoch 467/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2587 - val_loss: 2.3739\n",
      "Epoch 468/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1736 - val_loss: 2.4485\n",
      "Epoch 469/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1533 - val_loss: 2.3447\n",
      "Epoch 470/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1349 - val_loss: 2.4195\n",
      "Epoch 471/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0135 - val_loss: 2.3938\n",
      "Epoch 472/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1847 - val_loss: 2.3911\n",
      "Epoch 473/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1246 - val_loss: 2.3876\n",
      "Epoch 474/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.0994 - val_loss: 2.3932\n",
      "Epoch 475/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0873 - val_loss: 2.4257\n",
      "Epoch 476/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2516 - val_loss: 2.3693\n",
      "Epoch 477/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0706 - val_loss: 2.4375\n",
      "Epoch 478/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0602 - val_loss: 2.3710\n",
      "Epoch 479/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0724 - val_loss: 2.4098\n",
      "Epoch 480/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0111 - val_loss: 2.4071\n",
      "Epoch 481/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0498 - val_loss: 2.3866\n",
      "Epoch 482/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.2224 - val_loss: 2.4093\n",
      "Epoch 483/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.2133 - val_loss: 2.3826\n",
      "Epoch 484/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.0695 - val_loss: 2.3894\n",
      "Epoch 485/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.0001 - val_loss: 2.4188\n",
      "Epoch 486/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1204 - val_loss: 2.3920\n",
      "Epoch 487/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1601 - val_loss: 2.4034\n",
      "Epoch 488/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1559 - val_loss: 2.4015\n",
      "Epoch 489/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1315 - val_loss: 2.3931\n",
      "Epoch 490/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1232 - val_loss: 2.4320\n",
      "Epoch 491/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0762 - val_loss: 2.4017\n",
      "Epoch 492/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0354 - val_loss: 2.4137\n",
      "Epoch 493/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.0139 - val_loss: 2.4138\n",
      "Epoch 494/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1803 - val_loss: 2.4137\n",
      "Epoch 495/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.0750 - val_loss: 2.4225\n",
      "Epoch 496/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1082 - val_loss: 2.4038\n",
      "Epoch 497/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1186 - val_loss: 2.3789\n",
      "Epoch 498/500\n",
      "8/8 [==============================] - 0s 4ms/step - loss: 1.1792 - val_loss: 2.4524\n",
      "Epoch 499/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1195 - val_loss: 2.4047\n",
      "Epoch 500/500\n",
      "8/8 [==============================] - 0s 3ms/step - loss: 1.1862 - val_loss: 2.4641\n"
     ]
    }
   ],
   "source": [
    "#%%time\n",
    "history_regression = non_linear_regressor.fit(train_features, train_labels, validation_split = 0.2, verbose = 1, epochs = 500)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "plot_loss(history_regression)\n",
    "test_results['non_linear_regressor'] = non_linear_regressor.evaluate(test_features, test_labels, verbose = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "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>Mean absolute error on predicting the MPG</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>horsepower_model</th>\n",
       "      <td>3.615615</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>linear_model</th>\n",
       "      <td>2.747614</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>non_linear_horsepower_regressor</th>\n",
       "      <td>3.061307</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>non_linear_regressor</th>\n",
       "      <td>1.629809</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                 Mean absolute error on predicting the MPG\n",
       "horsepower_model                                                  3.615615\n",
       "linear_model                                                      2.747614\n",
       "non_linear_horsepower_regressor                                   3.061307\n",
       "non_linear_regressor                                              1.629809"
      ]
     },
     "execution_count": 146,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(test_results, index=['Mean absolute error on predicting the MPG']).T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}