LSTM_with_Keras.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "%matplotlib inline\n",
    "from IPython.display import display_latex"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "milk = pd.read_csv('monthly-milk-production.csv',index_col='Month')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Milk Production</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Month</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1962-01-01 01:00:00</th>\n",
       "      <td>589.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1962-02-01 01:00:00</th>\n",
       "      <td>561.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1962-03-01 01:00:00</th>\n",
       "      <td>640.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1962-04-01 01:00:00</th>\n",
       "      <td>656.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1962-05-01 01:00:00</th>\n",
       "      <td>727.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Milk Production\n",
       "Month                               \n",
       "1962-01-01 01:00:00            589.0\n",
       "1962-02-01 01:00:00            561.0\n",
       "1962-03-01 01:00:00            640.0\n",
       "1962-04-01 01:00:00            656.0\n",
       "1962-05-01 01:00:00            727.0"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "milk.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "milk.index = pd.to_datetime(milk.index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "split_date  = pd.Timestamp('1975-01-01')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train = milk[milk.index < split_date]\n",
    "test  = milk[milk.index >= split_date]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((156, 1), (12, 1))"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.shape, test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x2723faf5908>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAEACAYAAAC9Gb03AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2723f95e518>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = train.plot()\n",
    "test.plot(ax=ax)\n",
    "plt.legend(['train', 'test'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Scaling Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import MinMaxScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "scaler = MinMaxScaler()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train_scaled = scaler.fit_transform(train)\n",
    "test_scaled  = scaler.transform(test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "156"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(train_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train = []\n",
    "y_train = []\n",
    "\n",
    "for i in range(12,len(train_scaled)):\n",
    "    X_train.append(train_scaled[i-12:i,0])\n",
    "    y_train.append(train_scaled[i,0])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train,y_train = np.array(X_train), np.array(y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(144, 12)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train = np.reshape(X_train, (X_train.shape[0],X_train.shape[1],1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(144, 12, 1)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import  LSTM\n",
    "from keras.layers import Dense\n",
    "from keras.layers import Dropout"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "rnn_reg = Sequential()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "# first LSTM layer\n",
    "rnn_reg.add(LSTM(units=50,return_sequences =True, input_shape = (X_train.shape[1],1)))\n",
    "rnn_reg.add(Dropout(0.20))\n",
    "\n",
    "# Second LSTM layer\n",
    "rnn_reg.add(LSTM(units=50,return_sequences =True))\n",
    "rnn_reg.add(Dropout(0.20))\n",
    "\n",
    "# Third LSTM layer\n",
    "rnn_reg.add(LSTM(units=50,return_sequences =True))\n",
    "rnn_reg.add(Dropout(0.20))\n",
    "\n",
    "# Fourth LSTM layer\n",
    "rnn_reg.add(LSTM(units=50))\n",
    "rnn_reg.add(Dropout(0.20))\n",
    "\n",
    "#Output layer\n",
    "# third LSTM layer\n",
    "rnn_reg.add(Dense(units=1))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "rnn_reg.compile(optimizer='adam', loss = 'mean_squared_error')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1000\n",
      "144/144 [==============================] - 3s - loss: 0.1495     \n",
      "Epoch 2/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0549     \n",
      "Epoch 3/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0361     \n",
      "Epoch 4/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0322     \n",
      "Epoch 5/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0250     \n",
      "Epoch 6/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0286     \n",
      "Epoch 7/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0256     \n",
      "Epoch 8/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0248     \n",
      "Epoch 9/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0210     \n",
      "Epoch 10/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0252     \n",
      "Epoch 11/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0233     \n",
      "Epoch 12/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0237     \n",
      "Epoch 13/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0205     \n",
      "Epoch 14/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0195     \n",
      "Epoch 15/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0183     \n",
      "Epoch 16/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0178     \n",
      "Epoch 17/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0171     \n",
      "Epoch 18/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0256     \n",
      "Epoch 19/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0174     \n",
      "Epoch 20/1000\n",
      "144/144 [==============================] - ETA: 0s - loss: 0.018 - 0s - loss: 0.0182     \n",
      "Epoch 21/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0163     \n",
      "Epoch 22/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0186     \n",
      "Epoch 23/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0169     \n",
      "Epoch 24/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0159     \n",
      "Epoch 25/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0252     \n",
      "Epoch 26/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0156     \n",
      "Epoch 27/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0148     \n",
      "Epoch 28/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0117     \n",
      "Epoch 29/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0117     \n",
      "Epoch 30/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0142     \n",
      "Epoch 31/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0128     \n",
      "Epoch 32/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0132     \n",
      "Epoch 33/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0169     \n",
      "Epoch 34/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0172     \n",
      "Epoch 35/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0147     \n",
      "Epoch 36/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0141     \n",
      "Epoch 37/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0111     \n",
      "Epoch 38/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0128     \n",
      "Epoch 39/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0128     \n",
      "Epoch 40/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0098     \n",
      "Epoch 41/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0105     \n",
      "Epoch 42/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0118     \n",
      "Epoch 43/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0088     \n",
      "Epoch 44/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0124     \n",
      "Epoch 45/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0093     \n",
      "Epoch 46/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0107     \n",
      "Epoch 47/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0105     \n",
      "Epoch 48/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0093     \n",
      "Epoch 49/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0166     \n",
      "Epoch 50/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0154     \n",
      "Epoch 51/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0120     \n",
      "Epoch 52/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0112     \n",
      "Epoch 53/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0093     \n",
      "Epoch 54/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0084     \n",
      "Epoch 55/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0108     \n",
      "Epoch 56/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0099     \n",
      "Epoch 57/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0086     \n",
      "Epoch 58/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0104     \n",
      "Epoch 59/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0086     \n",
      "Epoch 60/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0081     \n",
      "Epoch 61/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0088     \n",
      "Epoch 62/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0093     \n",
      "Epoch 63/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0095     \n",
      "Epoch 64/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0102     \n",
      "Epoch 65/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0080     \n",
      "Epoch 66/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0095     \n",
      "Epoch 67/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0065     \n",
      "Epoch 68/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0068     \n",
      "Epoch 69/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0093     - ETA: 0s - loss: 0\n",
      "Epoch 70/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0080     \n",
      "Epoch 71/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0083     \n",
      "Epoch 72/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0091     \n",
      "Epoch 73/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0087     \n",
      "Epoch 74/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0060     \n",
      "Epoch 75/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0056     \n",
      "Epoch 76/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0071     \n",
      "Epoch 77/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0069     \n",
      "Epoch 78/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0070     \n",
      "Epoch 79/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0075     \n",
      "Epoch 80/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0073     \n",
      "Epoch 81/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0069     \n",
      "Epoch 82/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0088     \n",
      "Epoch 83/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0111     \n",
      "Epoch 84/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0094     \n",
      "Epoch 85/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0098     - ETA: 0s - loss: 0\n",
      "Epoch 86/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0074     \n",
      "Epoch 87/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0079     \n",
      "Epoch 88/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0063     \n",
      "Epoch 89/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0076     \n",
      "Epoch 90/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0071     \n",
      "Epoch 91/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0089     \n",
      "Epoch 92/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0071     \n",
      "Epoch 93/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0070     \n",
      "Epoch 94/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0061     \n",
      "Epoch 95/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0074     \n",
      "Epoch 96/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0094     \n",
      "Epoch 97/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0069     \n",
      "Epoch 98/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0065     \n",
      "Epoch 99/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0056     \n",
      "Epoch 100/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0071     \n",
      "Epoch 101/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0071     \n",
      "Epoch 102/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "144/144 [==============================] - 0s - loss: 0.0054     \n",
      "Epoch 103/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0053     \n",
      "Epoch 104/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0097     \n",
      "Epoch 105/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0114     \n",
      "Epoch 106/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0091     \n",
      "Epoch 107/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0075     \n",
      "Epoch 108/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0070     \n",
      "Epoch 109/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0071     \n",
      "Epoch 110/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0070     \n",
      "Epoch 111/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0066     \n",
      "Epoch 112/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0071     \n",
      "Epoch 113/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0055     \n",
      "Epoch 114/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0051     \n",
      "Epoch 115/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0057     \n",
      "Epoch 116/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0054     \n",
      "Epoch 117/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0057     \n",
      "Epoch 118/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0047     \n",
      "Epoch 119/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0065     \n",
      "Epoch 120/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0057     \n",
      "Epoch 121/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0049     \n",
      "Epoch 122/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0060     \n",
      "Epoch 123/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0058     \n",
      "Epoch 124/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0054     \n",
      "Epoch 125/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0083     \n",
      "Epoch 126/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0086     \n",
      "Epoch 127/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0083     \n",
      "Epoch 128/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0055     \n",
      "Epoch 129/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0072     - ETA: 0s - loss: 0\n",
      "Epoch 130/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0057     \n",
      "Epoch 131/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0056     \n",
      "Epoch 132/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0045     \n",
      "Epoch 133/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0050     \n",
      "Epoch 134/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0047     \n",
      "Epoch 135/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0054     \n",
      "Epoch 136/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0048     \n",
      "Epoch 137/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0045     \n",
      "Epoch 138/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0049     \n",
      "Epoch 139/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0056     \n",
      "Epoch 140/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0048     \n",
      "Epoch 141/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0045     \n",
      "Epoch 142/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0051     \n",
      "Epoch 143/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0052     \n",
      "Epoch 144/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0070     \n",
      "Epoch 145/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0052     \n",
      "Epoch 146/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0049     \n",
      "Epoch 147/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0052     \n",
      "Epoch 148/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0045     \n",
      "Epoch 149/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0051     \n",
      "Epoch 150/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0046     \n",
      "Epoch 151/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0045     \n",
      "Epoch 152/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0051     \n",
      "Epoch 153/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0049     \n",
      "Epoch 154/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0051     \n",
      "Epoch 155/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0046     \n",
      "Epoch 156/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0045     \n",
      "Epoch 157/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0043     \n",
      "Epoch 158/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0046     \n",
      "Epoch 159/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0052     \n",
      "Epoch 160/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0076     \n",
      "Epoch 161/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0058     \n",
      "Epoch 162/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0047     \n",
      "Epoch 163/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0051     \n",
      "Epoch 164/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0051     \n",
      "Epoch 165/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0042     \n",
      "Epoch 166/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0049     \n",
      "Epoch 167/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0054     \n",
      "Epoch 168/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0048     \n",
      "Epoch 169/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0050     \n",
      "Epoch 170/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0047     \n",
      "Epoch 171/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0045     \n",
      "Epoch 172/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0053     \n",
      "Epoch 173/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0041     \n",
      "Epoch 174/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0049     \n",
      "Epoch 175/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0048     \n",
      "Epoch 176/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0049     \n",
      "Epoch 177/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0045     \n",
      "Epoch 178/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0045     \n",
      "Epoch 179/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0043     \n",
      "Epoch 180/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0048     \n",
      "Epoch 181/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0054     \n",
      "Epoch 182/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0048     \n",
      "Epoch 183/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0034     \n",
      "Epoch 184/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0048     \n",
      "Epoch 185/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0043     \n",
      "Epoch 186/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0048     \n",
      "Epoch 187/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0049     \n",
      "Epoch 188/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0072     \n",
      "Epoch 189/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0046     \n",
      "Epoch 190/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0045     \n",
      "Epoch 191/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0060     \n",
      "Epoch 192/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0038     \n",
      "Epoch 193/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0039     \n",
      "Epoch 194/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0052     \n",
      "Epoch 195/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0040     \n",
      "Epoch 196/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0048     \n",
      "Epoch 197/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0040     \n",
      "Epoch 198/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0047     \n",
      "Epoch 199/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0048     \n",
      "Epoch 200/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0048     \n",
      "Epoch 201/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0051     \n",
      "Epoch 202/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0041     \n",
      "Epoch 203/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "144/144 [==============================] - 0s - loss: 0.0050     \n",
      "Epoch 204/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0045     \n",
      "Epoch 205/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0051     \n",
      "Epoch 206/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0035     \n",
      "Epoch 207/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0040     \n",
      "Epoch 208/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0059     \n",
      "Epoch 209/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0050     \n",
      "Epoch 210/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0041     \n",
      "Epoch 211/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0050     \n",
      "Epoch 212/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0037     \n",
      "Epoch 213/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0034     \n",
      "Epoch 214/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0047     \n",
      "Epoch 215/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0043     \n",
      "Epoch 216/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0042     \n",
      "Epoch 217/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0048     \n",
      "Epoch 218/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0036     \n",
      "Epoch 219/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0051     \n",
      "Epoch 220/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0039     \n",
      "Epoch 221/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0044     \n",
      "Epoch 222/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0043     \n",
      "Epoch 223/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 224/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0039     \n",
      "Epoch 225/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0040     \n",
      "Epoch 226/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0045     \n",
      "Epoch 227/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0044     \n",
      "Epoch 228/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0032     \n",
      "Epoch 229/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0037     \n",
      "Epoch 230/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0034     \n",
      "Epoch 231/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0041     \n",
      "Epoch 232/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0041     \n",
      "Epoch 233/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0034     \n",
      "Epoch 234/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0043     \n",
      "Epoch 235/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0034     \n",
      "Epoch 236/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0043     \n",
      "Epoch 237/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0045     \n",
      "Epoch 238/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0039     \n",
      "Epoch 239/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0054     - ETA: 0s - lo\n",
      "Epoch 240/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0036     \n",
      "Epoch 241/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0034     \n",
      "Epoch 242/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 243/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0036     \n",
      "Epoch 244/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0029     \n",
      "Epoch 245/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0039     \n",
      "Epoch 246/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0035     \n",
      "Epoch 247/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0030     \n",
      "Epoch 248/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0041     \n",
      "Epoch 249/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 250/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0034     \n",
      "Epoch 251/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0044     \n",
      "Epoch 252/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0038     \n",
      "Epoch 253/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0039     \n",
      "Epoch 254/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0036     \n",
      "Epoch 255/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0032     \n",
      "Epoch 256/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0034     \n",
      "Epoch 257/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0037     \n",
      "Epoch 258/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0037     \n",
      "Epoch 259/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0040     \n",
      "Epoch 260/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0036     \n",
      "Epoch 261/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0045     \n",
      "Epoch 262/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0030     \n",
      "Epoch 263/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0039     \n",
      "Epoch 264/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0039     \n",
      "Epoch 265/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 266/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0044     \n",
      "Epoch 267/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0038     \n",
      "Epoch 268/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0036     \n",
      "Epoch 269/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0032     \n",
      "Epoch 270/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0036     \n",
      "Epoch 271/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0035     \n",
      "Epoch 272/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0031     \n",
      "Epoch 273/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0034     \n",
      "Epoch 274/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 275/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0032     \n",
      "Epoch 276/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0031     \n",
      "Epoch 277/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0031     \n",
      "Epoch 278/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0039     \n",
      "Epoch 279/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0037     \n",
      "Epoch 280/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0032     \n",
      "Epoch 281/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0032     \n",
      "Epoch 282/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0036     \n",
      "Epoch 283/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0044     \n",
      "Epoch 284/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0034     \n",
      "Epoch 285/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0034     \n",
      "Epoch 286/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0047     \n",
      "Epoch 287/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0056     \n",
      "Epoch 288/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0046     \n",
      "Epoch 289/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0054     \n",
      "Epoch 290/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0053     \n",
      "Epoch 291/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0041     \n",
      "Epoch 292/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0031     \n",
      "Epoch 293/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0035     \n",
      "Epoch 294/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 295/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0038     \n",
      "Epoch 296/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 297/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 298/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0036     \n",
      "Epoch 299/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 300/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0042     \n",
      "Epoch 301/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0036     \n",
      "Epoch 302/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0036     \n",
      "Epoch 303/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 304/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 305/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0036     \n",
      "Epoch 306/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0035     \n",
      "Epoch 307/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0031     \n",
      "Epoch 308/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 309/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0037     \n",
      "Epoch 310/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 311/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0031     \n",
      "Epoch 312/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0037     \n",
      "Epoch 313/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0031     \n",
      "Epoch 314/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 315/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0031     \n",
      "Epoch 316/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 317/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0030     \n",
      "Epoch 318/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 319/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0034     \n",
      "Epoch 320/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0034     \n",
      "Epoch 321/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0030     \n",
      "Epoch 322/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 323/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0032     \n",
      "Epoch 324/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 325/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0030     \n",
      "Epoch 326/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 327/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0032     \n",
      "Epoch 328/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 329/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0032     \n",
      "Epoch 330/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0034     \n",
      "Epoch 331/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0030     \n",
      "Epoch 332/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0032     \n",
      "Epoch 333/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0034     \n",
      "Epoch 334/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0030     \n",
      "Epoch 335/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0032     \n",
      "Epoch 336/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0029     \n",
      "Epoch 337/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 338/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     - ETA: 0s - loss:\n",
      "Epoch 339/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 340/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 341/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0035     \n",
      "Epoch 342/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0031     \n",
      "Epoch 343/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0029     \n",
      "Epoch 344/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0029     \n",
      "Epoch 345/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0029     \n",
      "Epoch 346/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 347/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 348/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 349/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 350/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 351/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 352/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0034     \n",
      "Epoch 353/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0035     \n",
      "Epoch 354/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 355/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0031     \n",
      "Epoch 356/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 357/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0041     \n",
      "Epoch 358/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 359/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 360/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0029     \n",
      "Epoch 361/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0031     \n",
      "Epoch 362/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 363/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 364/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0031     \n",
      "Epoch 365/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0032     \n",
      "Epoch 366/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0031     \n",
      "Epoch 367/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 368/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0032     \n",
      "Epoch 369/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0029     \n",
      "Epoch 370/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0029     \n",
      "Epoch 371/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0031     \n",
      "Epoch 372/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 373/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0032     \n",
      "Epoch 374/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 375/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0029     \n",
      "Epoch 376/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0032     \n",
      "Epoch 377/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 378/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0029     \n",
      "Epoch 379/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 380/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0031     \n",
      "Epoch 381/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 382/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0029     \n",
      "Epoch 383/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0043     \n",
      "Epoch 384/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0039     \n",
      "Epoch 385/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0036     \n",
      "Epoch 386/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0031     \n",
      "Epoch 387/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0034     \n",
      "Epoch 388/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0030     \n",
      "Epoch 389/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0029     \n",
      "Epoch 390/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0030     \n",
      "Epoch 391/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 392/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0042     \n",
      "Epoch 393/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0031     \n",
      "Epoch 394/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0034     \n",
      "Epoch 395/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0036     \n",
      "Epoch 396/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0039     \n",
      "Epoch 397/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0038     \n",
      "Epoch 398/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0044     \n",
      "Epoch 399/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0040     \n",
      "Epoch 400/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0040     \n",
      "Epoch 401/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 402/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 403/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0031     \n",
      "Epoch 404/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 405/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 406/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 407/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0031     \n",
      "Epoch 408/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0038     \n",
      "Epoch 409/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0031     \n",
      "Epoch 410/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 411/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 412/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0041     \n",
      "Epoch 413/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 414/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 415/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 416/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 417/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 418/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 419/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0032     \n",
      "Epoch 420/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 421/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 422/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0030     \n",
      "Epoch 423/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0029     \n",
      "Epoch 424/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 425/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 426/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0032     \n",
      "Epoch 427/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 428/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 429/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 430/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 431/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0032     \n",
      "Epoch 432/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 433/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 434/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 435/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0032     \n",
      "Epoch 436/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 437/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 438/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 439/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 440/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0034     \n",
      "Epoch 441/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 442/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 443/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 444/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 445/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 446/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 447/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 448/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 449/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 450/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 451/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0029     \n",
      "Epoch 452/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 453/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 454/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 455/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0030     \n",
      "Epoch 456/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0029     \n",
      "Epoch 457/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 458/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 459/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 460/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 461/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 462/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     - ETA: 0s - loss: 0\n",
      "Epoch 463/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 464/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 465/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 466/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 467/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 468/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 469/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0030     \n",
      "Epoch 470/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 471/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 472/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 473/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 474/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 475/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 476/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 477/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 478/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 479/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 480/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 481/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 482/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 483/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 484/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 485/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 486/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     - ETA: 0s - loss: 0.\n",
      "Epoch 487/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 488/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 489/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 490/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 491/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 492/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 493/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 494/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 495/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 496/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 497/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 498/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0030     \n",
      "Epoch 499/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 500/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0030     \n",
      "Epoch 501/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0029     \n",
      "Epoch 502/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 503/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 504/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 505/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 506/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 507/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 508/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 509/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 510/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0032     \n",
      "Epoch 511/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0030     \n",
      "Epoch 512/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 513/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 514/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 515/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 516/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 517/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 518/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 519/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 520/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 521/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 522/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 523/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 524/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 525/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 526/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 527/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 528/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 529/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 530/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 531/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0029     \n",
      "Epoch 532/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 533/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 534/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 535/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 536/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 537/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 538/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 539/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 540/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 541/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 542/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 543/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 544/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 545/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 546/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 547/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 548/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 549/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 550/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 551/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 552/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 553/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 554/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 555/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 556/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 557/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 558/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 559/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 560/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 561/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 562/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 563/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 564/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 565/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 566/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 567/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 568/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 569/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 570/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 571/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 572/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 573/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 574/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 575/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 576/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 577/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 578/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 579/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 580/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 581/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 582/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 583/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 584/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0030     \n",
      "Epoch 585/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0029     \n",
      "Epoch 586/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 587/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 588/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 589/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 590/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 591/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 592/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 593/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 594/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 595/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 596/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 597/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 598/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 599/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 600/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 601/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 602/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 603/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 604/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 605/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 606/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 607/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 608/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 609/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 610/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 611/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 612/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 613/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 614/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0033     \n",
      "Epoch 615/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 616/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 617/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 618/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 619/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 620/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 621/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 622/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 623/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 624/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 625/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 626/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 627/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 628/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 629/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 630/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 631/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 632/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 633/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 634/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 635/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 636/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 637/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 638/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 639/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 640/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 641/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 642/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 643/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 644/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 645/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 646/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 647/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 648/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 649/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 650/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 651/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 652/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 653/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 654/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 655/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     - ETA: 0s - loss: 0.\n",
      "Epoch 656/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 657/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 658/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 659/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 660/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 661/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 662/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 663/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 664/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 665/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 666/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 667/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 668/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 669/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 670/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 671/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 672/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 673/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 674/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0028     \n",
      "Epoch 675/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0030     \n",
      "Epoch 676/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 677/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 678/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0027     \n",
      "Epoch 679/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 680/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 681/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 682/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 683/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 684/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 685/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 686/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 687/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 688/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 689/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 690/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 691/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 692/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 693/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 694/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 695/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 696/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 697/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 698/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 699/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 700/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 701/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 702/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 703/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 704/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 705/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 706/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 707/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 708/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 709/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 710/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 711/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 712/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 713/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 714/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 715/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 716/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 717/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 718/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 719/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 720/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 721/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 722/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 723/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 724/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 725/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 726/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 727/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 728/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 729/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 730/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 731/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 732/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 733/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 734/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 735/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 736/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 737/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 738/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 739/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 740/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 741/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 742/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 743/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 744/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 745/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0014     \n",
      "Epoch 746/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 747/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 748/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0014     \n",
      "Epoch 749/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 750/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 751/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 752/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 753/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 754/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 755/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 756/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 757/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 758/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 759/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 760/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 761/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 762/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 763/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 764/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 765/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 766/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 767/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 768/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 769/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0014     \n",
      "Epoch 770/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 771/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 772/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 773/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 774/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 775/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 776/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 777/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 778/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 779/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0014     \n",
      "Epoch 780/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 781/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 782/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 783/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 784/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 785/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 786/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 787/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 788/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 789/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 790/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 791/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 792/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 793/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     - ETA: 0s - loss: \n",
      "Epoch 794/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 795/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 796/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 797/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 798/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 799/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 800/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 801/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 802/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 803/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 804/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 805/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 806/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0012     \n",
      "Epoch 807/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 808/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 809/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 810/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 811/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 812/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 813/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 814/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 815/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 816/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 817/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 818/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 819/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0014     \n",
      "Epoch 820/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 821/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 822/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 823/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0024     \n",
      "Epoch 824/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 825/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 826/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 827/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 828/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 829/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 830/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     - ETA: 0s - loss: 0\n",
      "Epoch 831/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 832/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 833/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0014      \n",
      "Epoch 834/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 835/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 836/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 837/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 838/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 839/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 840/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 841/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 842/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 843/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 844/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 845/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 846/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0025     \n",
      "Epoch 847/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 848/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 849/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 850/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 851/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 852/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0013     \n",
      "Epoch 853/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 854/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 855/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 856/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 857/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 858/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 859/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0014     \n",
      "Epoch 860/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 861/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 862/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 863/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0013      \n",
      "Epoch 864/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 865/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 866/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 867/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 868/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 869/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 870/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 871/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 872/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 873/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 874/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 875/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0014     \n",
      "Epoch 876/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0014     \n",
      "Epoch 877/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 878/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 879/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 880/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 881/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 882/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 883/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 884/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 885/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 886/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0014     \n",
      "Epoch 887/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 888/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 889/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 890/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 891/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0012     \n",
      "Epoch 892/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 893/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 894/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 895/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 896/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 897/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 898/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 899/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 900/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 901/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 902/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 903/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 904/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 905/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 906/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 907/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 908/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 909/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 910/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 911/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 912/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 913/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 914/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 915/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 916/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 917/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 918/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 919/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 920/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 921/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 922/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 923/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 924/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0021     \n",
      "Epoch 925/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 926/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 927/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 928/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 929/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     - ETA: 0s - loss: 0.\n",
      "Epoch 930/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 931/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0014     \n",
      "Epoch 932/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0012     \n",
      "Epoch 933/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0012     \n",
      "Epoch 934/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 935/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 936/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 937/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 938/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 939/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 940/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 941/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 942/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 943/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 944/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0023     \n",
      "Epoch 945/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 946/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 947/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 948/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 949/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 950/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 951/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0022     \n",
      "Epoch 952/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0014     \n",
      "Epoch 953/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 954/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0014     \n",
      "Epoch 955/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 956/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 957/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 958/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0012     \n",
      "Epoch 959/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 960/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 961/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 962/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 963/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 964/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0012     \n",
      "Epoch 965/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 966/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 967/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 968/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 969/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 970/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0014     \n",
      "Epoch 971/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0014     \n",
      "Epoch 972/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 973/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 974/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 975/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 976/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 977/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 978/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 979/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0014     \n",
      "Epoch 980/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 981/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 982/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 983/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 984/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 985/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 986/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 987/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 988/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0018     \n",
      "Epoch 989/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0019     \n",
      "Epoch 990/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0014     \n",
      "Epoch 991/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 992/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 993/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0014     \n",
      "Epoch 994/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0013     \n",
      "Epoch 995/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0012     \n",
      "Epoch 996/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 997/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0015     \n",
      "Epoch 998/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0020     \n",
      "Epoch 999/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0017     \n",
      "Epoch 1000/1000\n",
      "144/144 [==============================] - 0s - loss: 0.0013     \n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x27245bbe5f8>"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rnn_reg.fit(X_train,y_train,epochs=1000,batch_size=12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.67548077],\n",
       "       [ 0.55048077],\n",
       "       [ 0.81490385],\n",
       "       [ 0.84134615],\n",
       "       [ 0.99278846],\n",
       "       [ 0.92307692],\n",
       "       [ 0.82451923],\n",
       "       [ 0.73317308],\n",
       "       [ 0.63461538],\n",
       "       [ 0.65865385],\n",
       "       [ 0.58653846],\n",
       "       [ 0.69711538]])"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_scaled"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(144, 12, 1)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(12, 1)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_scaled.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#X_train = np.reshape(X_train, (X_train.shape[0],X_train.shape[1],1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_test = np.reshape(test_scaled,(test_scaled.shape[0],test_scaled.shape[1],1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(12, 1, 1)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Milk Production</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Month</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1962-01-01 01:00:00</th>\n",
       "      <td>589.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1962-02-01 01:00:00</th>\n",
       "      <td>561.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1962-03-01 01:00:00</th>\n",
       "      <td>640.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1962-04-01 01:00:00</th>\n",
       "      <td>656.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1962-05-01 01:00:00</th>\n",
       "      <td>727.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1962-06-01 01:00:00</th>\n",
       "      <td>697.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1962-07-01 01:00:00</th>\n",
       "      <td>640.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1962-08-01 01:00:00</th>\n",
       "      <td>599.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1962-09-01 01:00:00</th>\n",
       "      <td>568.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1962-10-01 01:00:00</th>\n",
       "      <td>577.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1962-11-01 01:00:00</th>\n",
       "      <td>553.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1962-12-01 01:00:00</th>\n",
       "      <td>582.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1963-01-01 01:00:00</th>\n",
       "      <td>600.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Milk Production\n",
       "Month                               \n",
       "1962-01-01 01:00:00            589.0\n",
       "1962-02-01 01:00:00            561.0\n",
       "1962-03-01 01:00:00            640.0\n",
       "1962-04-01 01:00:00            656.0\n",
       "1962-05-01 01:00:00            727.0\n",
       "1962-06-01 01:00:00            697.0\n",
       "1962-07-01 01:00:00            640.0\n",
       "1962-08-01 01:00:00            599.0\n",
       "1962-09-01 01:00:00            568.0\n",
       "1962-10-01 01:00:00            577.0\n",
       "1962-11-01 01:00:00            553.0\n",
       "1962-12-01 01:00:00            582.0\n",
       "1963-01-01 01:00:00            600.0"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "milk.head(13)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "inputs = milk[len(milk) - len(test) - 12:].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(24, 1)"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "inputs.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(24, 1)"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "inputs = inputs.reshape(-1,1)\n",
    "inputs.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "inputs = scaler.transform(inputs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.66105769],\n",
       "       [ 0.54086538],\n",
       "       [ 0.80769231],\n",
       "       [ 0.83894231],\n",
       "       [ 1.        ],\n",
       "       [ 0.94711538],\n",
       "       [ 0.85336538],\n",
       "       [ 0.75480769],\n",
       "       [ 0.62980769],\n",
       "       [ 0.62259615],\n",
       "       [ 0.52884615],\n",
       "       [ 0.625     ],\n",
       "       [ 0.67548077],\n",
       "       [ 0.55048077],\n",
       "       [ 0.81490385],\n",
       "       [ 0.84134615],\n",
       "       [ 0.99278846],\n",
       "       [ 0.92307692],\n",
       "       [ 0.82451923],\n",
       "       [ 0.73317308],\n",
       "       [ 0.63461538],\n",
       "       [ 0.65865385],\n",
       "       [ 0.58653846],\n",
       "       [ 0.69711538]])"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "inputs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test = []\n",
    "for i in range(12,len(inputs)):\n",
    "    X_test.append(inputs[i-12:i,0])\n",
    "X_test = np.array(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(12, 12)"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_test = np.reshape(X_test, (X_test.shape[0],X_test.shape[1],1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(12, 12, 1)"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = rnn_reg.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_pred_real_scale = scaler.inverse_transform(y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 828.88232422],\n",
       "       [ 798.5291748 ],\n",
       "       [ 891.72021484],\n",
       "       [ 886.73754883],\n",
       "       [ 957.78613281],\n",
       "       [ 939.9743042 ],\n",
       "       [ 901.35961914],\n",
       "       [ 860.59393311],\n",
       "       [ 827.98675537],\n",
       "       [ 811.3213501 ],\n",
       "       [ 787.04101562],\n",
       "       [ 810.47937012]], dtype=float32)"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred_real_scale"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Milk Production</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Month</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1975-01-01 01:00:00</th>\n",
       "      <td>834.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-02-01 01:00:00</th>\n",
       "      <td>782.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-03-01 01:00:00</th>\n",
       "      <td>892.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-04-01 01:00:00</th>\n",
       "      <td>903.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-05-01 01:00:00</th>\n",
       "      <td>966.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-06-01 01:00:00</th>\n",
       "      <td>937.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-07-01 01:00:00</th>\n",
       "      <td>896.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-08-01 01:00:00</th>\n",
       "      <td>858.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-09-01 01:00:00</th>\n",
       "      <td>817.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-10-01 01:00:00</th>\n",
       "      <td>827.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-11-01 01:00:00</th>\n",
       "      <td>797.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-12-01 01:00:00</th>\n",
       "      <td>843.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Milk Production\n",
       "Month                               \n",
       "1975-01-01 01:00:00            834.0\n",
       "1975-02-01 01:00:00            782.0\n",
       "1975-03-01 01:00:00            892.0\n",
       "1975-04-01 01:00:00            903.0\n",
       "1975-05-01 01:00:00            966.0\n",
       "1975-06-01 01:00:00            937.0\n",
       "1975-07-01 01:00:00            896.0\n",
       "1975-08-01 01:00:00            858.0\n",
       "1975-09-01 01:00:00            817.0\n",
       "1975-10-01 01:00:00            827.0\n",
       "1975-11-01 01:00:00            797.0\n",
       "1975-12-01 01:00:00            843.0"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Milk Production</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Month</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1975-01-01 01:00:00</th>\n",
       "      <td>834.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-02-01 01:00:00</th>\n",
       "      <td>782.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-03-01 01:00:00</th>\n",
       "      <td>892.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-04-01 01:00:00</th>\n",
       "      <td>903.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-05-01 01:00:00</th>\n",
       "      <td>966.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Milk Production\n",
       "Month                               \n",
       "1975-01-01 01:00:00            834.0\n",
       "1975-02-01 01:00:00            782.0\n",
       "1975-03-01 01:00:00            892.0\n",
       "1975-04-01 01:00:00            903.0\n",
       "1975-05-01 01:00:00            966.0"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "real_predicted = test.copy()\n",
    "real_predicted.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "real_predicted['Predicted Milk Production'] = y_pred_real_scale"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Milk Production</th>\n",
       "      <th>Predicted Milk Production</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Month</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1975-01-01 01:00:00</th>\n",
       "      <td>834.0</td>\n",
       "      <td>828.882324</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-02-01 01:00:00</th>\n",
       "      <td>782.0</td>\n",
       "      <td>798.529175</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-03-01 01:00:00</th>\n",
       "      <td>892.0</td>\n",
       "      <td>891.720215</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-04-01 01:00:00</th>\n",
       "      <td>903.0</td>\n",
       "      <td>886.737549</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-05-01 01:00:00</th>\n",
       "      <td>966.0</td>\n",
       "      <td>957.786133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-06-01 01:00:00</th>\n",
       "      <td>937.0</td>\n",
       "      <td>939.974304</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-07-01 01:00:00</th>\n",
       "      <td>896.0</td>\n",
       "      <td>901.359619</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-08-01 01:00:00</th>\n",
       "      <td>858.0</td>\n",
       "      <td>860.593933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-09-01 01:00:00</th>\n",
       "      <td>817.0</td>\n",
       "      <td>827.986755</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-10-01 01:00:00</th>\n",
       "      <td>827.0</td>\n",
       "      <td>811.321350</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-11-01 01:00:00</th>\n",
       "      <td>797.0</td>\n",
       "      <td>787.041016</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1975-12-01 01:00:00</th>\n",
       "      <td>843.0</td>\n",
       "      <td>810.479370</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Milk Production  Predicted Milk Production\n",
       "Month                                                          \n",
       "1975-01-01 01:00:00            834.0                 828.882324\n",
       "1975-02-01 01:00:00            782.0                 798.529175\n",
       "1975-03-01 01:00:00            892.0                 891.720215\n",
       "1975-04-01 01:00:00            903.0                 886.737549\n",
       "1975-05-01 01:00:00            966.0                 957.786133\n",
       "1975-06-01 01:00:00            937.0                 939.974304\n",
       "1975-07-01 01:00:00            896.0                 901.359619\n",
       "1975-08-01 01:00:00            858.0                 860.593933\n",
       "1975-09-01 01:00:00            817.0                 827.986755\n",
       "1975-10-01 01:00:00            827.0                 811.321350\n",
       "1975-11-01 01:00:00            797.0                 787.041016\n",
       "1975-12-01 01:00:00            843.0                 810.479370"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "real_predicted"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x2724a5e0b70>"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeoAAAFqCAYAAADV3RniAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2724a5e0be0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "real_predicted.plot(figsize=(8,6))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.6.1"
  },
  "nav_menu": {},
  "toc": {
   "navigate_menu": true,
   "number_sections": true,
   "sideBar": true,
   "threshold": 6,
   "toc_cell": false,
   "toc_section_display": "block",
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}