HOD model using WP,VPF,CIA,CIC and machine learning-reversed.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load the autoreload extension\n",
    "%reload_ext autoreload\n",
    "# Set extension to reload modules every time before executing code\n",
    "%autoreload 2\n",
    "\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import numpy as np\n",
    "import os\n",
    "%matplotlib inline\n",
    "\n",
    "# Import random forest regressor and multioutput regressor\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.multioutput import MultiOutputRegressor\n",
    "from sklearn.metrics import confusion_matrix,classification_report,accuracy_score,mean_squared_error\n",
    "from sklearn.preprocessing import MinMaxScaler,StandardScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "path = '/mnt/data4/Abhishek/WP/'\n",
    "data = np.load('machine_learning_data.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def extractIndex(filename):\n",
    "    return int(filename.split('_')[-1][:4])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "f = np.loadtxt('/home/ajana/github/HODProject/parameters.txt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "random_data = []\n",
    "\n",
    "for filename in os.listdir(path):\n",
    "    if filename.startswith('wp_galaxies'):\n",
    "        pos = extractIndex(filename)\n",
    "        wp,_ = np.load(os.path.join(path,filename))\n",
    "        #_temp = np.concatenate((wp,f[pos]))\n",
    "        random_data.append(wp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "full_data = np.vstack((random_data))\n",
    "full_data = np.hstack((full_data,data))\n",
    "del random_data\n",
    "np.random.shuffle(full_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Order is wp (29) VPF (30) CIC(15) CIA(15)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_data = full_data[:,:-5]\n",
    "X_data = full_data[:,-5:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_data = y_data[:,59:74]  #CIC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5000, 15)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f3f6e07be90>]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(np.arange(y_data.shape[1]),y_data[0],'*')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X_data, y_data, test_size=0.1,random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((4500, 5), (500, 5))"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape, X_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "#fit Random Forest Regressor\n",
    "\n",
    "reg_rf = RandomForestRegressor(\n",
    "    n_estimators=3000,\n",
    "    criterion='mse',\n",
    "    max_depth=None,\n",
    "    min_samples_split=2,\n",
    "    min_samples_leaf=1,\n",
    "    min_weight_fraction_leaf=0.0,\n",
    "    max_features=None,\n",
    "    max_leaf_nodes=None,\n",
    "    min_impurity_decrease=0.0,\n",
    "    min_impurity_split=None,\n",
    "    bootstrap=True,\n",
    "    oob_score=False,\n",
    "    n_jobs=-1,\n",
    "    random_state=0,\n",
    "    verbose=0,\n",
    "    warm_start=False,\n",
    ")\n",
    "\n",
    "reg_rf.fit(X_train,y_train)\n",
    "y_rf = reg_rf.predict(X_test)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "err = mean_squared_error(y_rf,y_test,multioutput='raw_values',squared=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.0205699  0.00817134 0.00392991 0.00569654 0.00214891 0.00430081\n",
      " 0.00401833 0.00380941 0.00321383 0.0040248  0.00429106 0.00291648\n",
      " 0.00307698 0.00378532 0.00490818]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([0.01476237, 0.12880985, 0.14495341, 0.15333116, 0.14358751,\n",
       "       0.14242188, 0.13491406, 0.13007509, 0.14169343, 0.15341653,\n",
       "       0.17292664, 0.18846694, 0.20327123, 0.21254273, 0.18302224])"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print (np.mean(((y_rf - y_test)/y_test), axis=0))\n",
    "np.std(y_rf,axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_corr(pred,test):\n",
    "    sns.jointplot(pred.iloc[:,0],test.iloc[:,0],kind = \"kde\")\n",
    "    sns.jointplot(pred.iloc[:,1],test.iloc[:,1],kind = 'kde',color='purple')\n",
    "    sns.jointplot(pred.iloc[:,2],test.iloc[:,2],kind = 'kde',color='blue')\n",
    "    sns.jointplot(pred.iloc[:,3],test.iloc[:,3],kind = 'kde',color='green')\n",
    "    sns.jointplot(pred.iloc[:,4],test.iloc[:,4],kind = 'kde',color='red')\n",
    "    plt.tight_layout\n",
    "    \n",
    "def plot_error(pred,test,dims=(12,8)):\n",
    "    fig,ax = plt.subplots(figsize=dims)\n",
    "    sns.distplot(pred.iloc[:,0]-test.iloc[:,0],color='blue',ax=ax,kde_kws={\"label\": \"M_cut\"})\n",
    "    sns.distplot(pred.iloc[:,1]-test.iloc[:,1],color='cyan',ax=ax,kde_kws={\"label\": \"M_1\"})\n",
    "    sns.distplot(pred.iloc[:,2]-test.iloc[:,2],color='purple',ax=ax,kde_kws={\"label\": \"Sigma\"})\n",
    "    sns.distplot(pred.iloc[:,3]-test.iloc[:,3],color='yellow',ax=ax,kde_kws={\"label\": \"Kappa\"})\n",
    "    sns.distplot(pred.iloc[:,4]-test.iloc[:,4],color='red',ax=ax,kde_kws={\"label\": \"Alpha\"})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'WP')"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(np.arange(y_data.shape[1]),y_test[160,:],'*')\n",
    "plt.plot(np.arange(y_data.shape[1]),y_rf[160])\n",
    "plt.xlabel('WP')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'CIC_predicted')"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,8))\n",
    "plt.errorbar(np.mean(y_rf,axis=0),np.mean(y_test,axis=0),np.std(y_rf,axis=0),ecolor='r',fmt='.')\n",
    "plt.errorbar(y_rf,y_test,ecolor='r',fmt='.')\n",
    "plt.xlabel('CIC_true')\n",
    "plt.ylabel('CIC_predicted')\n",
    "#plt.savefig('VPF_correlation')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([8.21665032e-11, 1.34454901e-10, 2.66977651e-10, 4.49789141e-10,\n",
       "       8.00866927e-10, 1.34122003e-09, 2.21506431e-09, 3.74734422e-09,\n",
       "       6.60899762e-09, 1.02169768e-08, 1.70307012e-08, 2.88489818e-08,\n",
       "       5.04068499e-08, 8.83092820e-08, 1.62414882e-07, 2.65172904e-07,\n",
       "       4.07514913e-07, 6.36517267e-07, 1.07348532e-06, 1.95946365e-06,\n",
       "       2.68126326e-06, 3.67424005e-06, 4.05559818e-06, 3.80352607e-06,\n",
       "       2.61755213e-06, 1.36378253e-06, 4.26518247e-07, 6.13398700e-08,\n",
       "       4.72856640e-09, 1.39221227e-10])"
      ]
     },
     "execution_count": 156,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "err"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 229,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.02923531 0.00329706 0.00571132 0.00441526 0.00588189 0.00605791\n",
      " 0.00563519 0.00626721 0.00510051 0.00593019 0.00499478 0.00306184\n",
      " 0.00399232 0.00224409 0.00466897]\n"
     ]
    }
   ],
   "source": [
    "print (np.mean(((y_rf - y_test)/y_test), axis=0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 221,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.02243472,  0.01776045, -0.05481712, -0.00261736, -0.03085204,\n",
       "       -0.002358  , -0.01373077, -0.03321914,  0.00404207, -0.00347799,\n",
       "       -0.0046748 , -0.00054485, -0.05162875,  0.01019179,  0.01620123])"
      ]
     },
     "execution_count": 221,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(y_rf[15] -  y_test[15,:])/y_test[15,:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<seaborn.axisgrid.JointGrid at 0x7f4e22288bd0>"
      ]
     },
     "execution_count": 157,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.jointplot(y_test,y_rf,kind='scatter')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import mean_squared_error\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras import layers\n",
    "from tensorflow import keras\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 194,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import  MinMaxScaler\n",
    "sc= MinMaxScaler()\n",
    "X_data= sc.fit_transform(X_data)\n",
    "y_data= y_data.reshape(-1,y_data.shape[1])\n",
    "y_data=sc.fit_transform(y_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "metadata": {},
   "outputs": [],
   "source": [
    "in_dim = X_train.shape[1]\n",
    "out_dim = y_train.shape[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "model.add(Dense(128, input_dim=in_dim, activation=\"relu\"))\n",
    "model.add(Dense(64, activation=\"relu\"))\n",
    "model.add(Dense(32, activation=\"relu\"))\n",
    "model.add(Dense(out_dim))\n",
    "model.compile(loss=\"mse\", optimizer=\"adam\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "metadata": {},
   "outputs": [],
   "source": [
    "def build_model(in_dim,out_dim):\n",
    "    model = keras.Sequential([\n",
    "        layers.Dense(256, activation='relu', input_dim=in_dim),\n",
    "        layers.Dense(128, activation='relu'),\n",
    "        layers.Dense(64, activation='relu'),\n",
    "        layers.Dense(out_dim)\n",
    "      ])\n",
    "\n",
    "    optimizer = tf.keras.optimizers.RMSprop(0.01)\n",
    "\n",
    "    model.compile(loss='mse',\n",
    "                    optimizer=\"adam\",\n",
    "                    metrics=['mae', 'mse'])\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_3\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_12 (Dense)             (None, 256)               1536      \n",
      "_________________________________________________________________\n",
      "dense_13 (Dense)             (None, 128)               32896     \n",
      "_________________________________________________________________\n",
      "dense_14 (Dense)             (None, 64)                8256      \n",
      "_________________________________________________________________\n",
      "dense_15 (Dense)             (None, 29)                1885      \n",
      "=================================================================\n",
      "Total params: 44,573\n",
      "Trainable params: 44,573\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "mod = build_model(in_dim,out_dim)\n",
    "mod.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 3600 samples, validate on 900 samples\n",
      "Epoch 1/300\n",
      "3600/3600 [==============================] - 1s 188us/sample - loss: 0.0570 - mean_absolute_error: 0.0991 - mean_squared_error: 0.0570 - val_loss: 5.9745e-04 - val_mean_absolute_error: 0.0158 - val_mean_squared_error: 5.9745e-04\n",
      "Epoch 2/300\n",
      "3600/3600 [==============================] - 0s 129us/sample - loss: 5.5709e-04 - mean_absolute_error: 0.0152 - mean_squared_error: 5.5709e-04 - val_loss: 5.3708e-04 - val_mean_absolute_error: 0.0146 - val_mean_squared_error: 5.3708e-04\n",
      "Epoch 3/300\n",
      "3600/3600 [==============================] - 0s 113us/sample - loss: 4.9850e-04 - mean_absolute_error: 0.0141 - mean_squared_error: 4.9851e-04 - val_loss: 4.8743e-04 - val_mean_absolute_error: 0.0136 - val_mean_squared_error: 4.8743e-04\n",
      "Epoch 4/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 4.4565e-04 - mean_absolute_error: 0.0131 - mean_squared_error: 4.4565e-04 - val_loss: 5.0516e-04 - val_mean_absolute_error: 0.0151 - val_mean_squared_error: 5.0516e-04\n",
      "Epoch 5/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 4.0828e-04 - mean_absolute_error: 0.0125 - mean_squared_error: 4.0828e-04 - val_loss: 3.8613e-04 - val_mean_absolute_error: 0.0120 - val_mean_squared_error: 3.8613e-04\n",
      "Epoch 6/300\n",
      "3600/3600 [==============================] - 0s 122us/sample - loss: 3.8257e-04 - mean_absolute_error: 0.0123 - mean_squared_error: 3.8257e-04 - val_loss: 3.5481e-04 - val_mean_absolute_error: 0.0117 - val_mean_squared_error: 3.5481e-04\n",
      "Epoch 7/300\n",
      "3600/3600 [==============================] - 0s 128us/sample - loss: 3.4056e-04 - mean_absolute_error: 0.0116 - mean_squared_error: 3.4056e-04 - val_loss: 3.6761e-04 - val_mean_absolute_error: 0.0123 - val_mean_squared_error: 3.6761e-04\n",
      "Epoch 8/300\n",
      "3600/3600 [==============================] - 0s 129us/sample - loss: 3.1058e-04 - mean_absolute_error: 0.0113 - mean_squared_error: 3.1058e-04 - val_loss: 3.0835e-04 - val_mean_absolute_error: 0.0115 - val_mean_squared_error: 3.0835e-04\n",
      "Epoch 9/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 2.8201e-04 - mean_absolute_error: 0.0109 - mean_squared_error: 2.8201e-04 - val_loss: 2.7369e-04 - val_mean_absolute_error: 0.0106 - val_mean_squared_error: 2.7369e-04\n",
      "Epoch 10/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 2.5712e-04 - mean_absolute_error: 0.0106 - mean_squared_error: 2.5712e-04 - val_loss: 2.4632e-04 - val_mean_absolute_error: 0.0104 - val_mean_squared_error: 2.4632e-04\n",
      "Epoch 11/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 2.2385e-04 - mean_absolute_error: 0.0099 - mean_squared_error: 2.2385e-04 - val_loss: 2.1916e-04 - val_mean_absolute_error: 0.0098 - val_mean_squared_error: 2.1916e-04\n",
      "Epoch 12/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 2.1222e-04 - mean_absolute_error: 0.0101 - mean_squared_error: 2.1222e-04 - val_loss: 1.9202e-04 - val_mean_absolute_error: 0.0095 - val_mean_squared_error: 1.9202e-04\n",
      "Epoch 13/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 1.9247e-04 - mean_absolute_error: 0.0099 - mean_squared_error: 1.9247e-04 - val_loss: 1.7371e-04 - val_mean_absolute_error: 0.0090 - val_mean_squared_error: 1.7371e-04\n",
      "Epoch 14/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 2.1648e-04 - mean_absolute_error: 0.0107 - mean_squared_error: 2.1648e-04 - val_loss: 2.6297e-04 - val_mean_absolute_error: 0.0129 - val_mean_squared_error: 2.6297e-04\n",
      "Epoch 15/300\n",
      "3600/3600 [==============================] - 0s 113us/sample - loss: 1.6844e-04 - mean_absolute_error: 0.0096 - mean_squared_error: 1.6844e-04 - val_loss: 1.2724e-04 - val_mean_absolute_error: 0.0077 - val_mean_squared_error: 1.2724e-04\n",
      "Epoch 16/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 1.4355e-04 - mean_absolute_error: 0.0089 - mean_squared_error: 1.4355e-04 - val_loss: 1.5797e-04 - val_mean_absolute_error: 0.0096 - val_mean_squared_error: 1.5797e-04\n",
      "Epoch 17/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 1.5831e-04 - mean_absolute_error: 0.0093 - mean_squared_error: 1.5831e-04 - val_loss: 1.7560e-04 - val_mean_absolute_error: 0.0099 - val_mean_squared_error: 1.7560e-04\n",
      "Epoch 18/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 1.2238e-04 - mean_absolute_error: 0.0083 - mean_squared_error: 1.2238e-04 - val_loss: 1.5609e-04 - val_mean_absolute_error: 0.0099 - val_mean_squared_error: 1.5609e-04\n",
      "Epoch 19/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 1.4060e-04 - mean_absolute_error: 0.0090 - mean_squared_error: 1.4060e-04 - val_loss: 1.3106e-04 - val_mean_absolute_error: 0.0084 - val_mean_squared_error: 1.3106e-04\n",
      "Epoch 20/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 1.3140e-04 - mean_absolute_error: 0.0084 - mean_squared_error: 1.3140e-04 - val_loss: 1.4199e-04 - val_mean_absolute_error: 0.0093 - val_mean_squared_error: 1.4199e-04\n",
      "Epoch 21/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 1.0062e-04 - mean_absolute_error: 0.0074 - mean_squared_error: 1.0062e-04 - val_loss: 1.0253e-04 - val_mean_absolute_error: 0.0078 - val_mean_squared_error: 1.0253e-04\n",
      "Epoch 22/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 1.0722e-04 - mean_absolute_error: 0.0079 - mean_squared_error: 1.0722e-04 - val_loss: 5.8731e-05 - val_mean_absolute_error: 0.0051 - val_mean_squared_error: 5.8731e-05\n",
      "Epoch 23/300\n",
      "3600/3600 [==============================] - 0s 129us/sample - loss: 1.1887e-04 - mean_absolute_error: 0.0078 - mean_squared_error: 1.1887e-04 - val_loss: 2.4077e-04 - val_mean_absolute_error: 0.0123 - val_mean_squared_error: 2.4077e-04\n",
      "Epoch 24/300\n",
      "3600/3600 [==============================] - 0s 121us/sample - loss: 9.8528e-05 - mean_absolute_error: 0.0072 - mean_squared_error: 9.8528e-05 - val_loss: 9.6749e-05 - val_mean_absolute_error: 0.0068 - val_mean_squared_error: 9.6749e-05\n",
      "Epoch 25/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 1.0159e-04 - mean_absolute_error: 0.0075 - mean_squared_error: 1.0159e-04 - val_loss: 8.9005e-05 - val_mean_absolute_error: 0.0070 - val_mean_squared_error: 8.9005e-05\n",
      "Epoch 26/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 1.5603e-04 - mean_absolute_error: 0.0094 - mean_squared_error: 1.5603e-04 - val_loss: 8.8993e-05 - val_mean_absolute_error: 0.0072 - val_mean_squared_error: 8.8993e-05\n",
      "Epoch 27/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 8.3429e-05 - mean_absolute_error: 0.0066 - mean_squared_error: 8.3429e-05 - val_loss: 1.1638e-04 - val_mean_absolute_error: 0.0080 - val_mean_squared_error: 1.1638e-04\n",
      "Epoch 28/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 9.3737e-05 - mean_absolute_error: 0.0071 - mean_squared_error: 9.3737e-05 - val_loss: 1.3813e-04 - val_mean_absolute_error: 0.0092 - val_mean_squared_error: 1.3813e-04\n",
      "Epoch 29/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 9.0723e-05 - mean_absolute_error: 0.0071 - mean_squared_error: 9.0723e-05 - val_loss: 1.1955e-04 - val_mean_absolute_error: 0.0082 - val_mean_squared_error: 1.1955e-04\n",
      "Epoch 30/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 8.7742e-05 - mean_absolute_error: 0.0070 - mean_squared_error: 8.7742e-05 - val_loss: 1.2240e-04 - val_mean_absolute_error: 0.0089 - val_mean_squared_error: 1.2240e-04\n",
      "Epoch 31/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 9.0607e-05 - mean_absolute_error: 0.0068 - mean_squared_error: 9.0607e-05 - val_loss: 6.9179e-05 - val_mean_absolute_error: 0.0066 - val_mean_squared_error: 6.9179e-05\n",
      "Epoch 32/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 6.8401e-05 - mean_absolute_error: 0.0060 - mean_squared_error: 6.8401e-05 - val_loss: 9.3136e-05 - val_mean_absolute_error: 0.0072 - val_mean_squared_error: 9.3136e-05\n",
      "Epoch 33/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 8.9652e-05 - mean_absolute_error: 0.0069 - mean_squared_error: 8.9652e-05 - val_loss: 2.6071e-04 - val_mean_absolute_error: 0.0130 - val_mean_squared_error: 2.6071e-04\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 34/300\n",
      "3600/3600 [==============================] - 0s 134us/sample - loss: 9.1515e-05 - mean_absolute_error: 0.0071 - mean_squared_error: 9.1515e-05 - val_loss: 9.1673e-05 - val_mean_absolute_error: 0.0077 - val_mean_squared_error: 9.1673e-05\n",
      "Epoch 35/300\n",
      "3600/3600 [==============================] - 0s 132us/sample - loss: 8.7034e-05 - mean_absolute_error: 0.0069 - mean_squared_error: 8.7035e-05 - val_loss: 6.2633e-05 - val_mean_absolute_error: 0.0059 - val_mean_squared_error: 6.2633e-05\n",
      "Epoch 36/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 9.4592e-05 - mean_absolute_error: 0.0073 - mean_squared_error: 9.4592e-05 - val_loss: 7.3565e-05 - val_mean_absolute_error: 0.0067 - val_mean_squared_error: 7.3565e-05\n",
      "Epoch 37/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 7.1829e-05 - mean_absolute_error: 0.0063 - mean_squared_error: 7.1829e-05 - val_loss: 9.5112e-05 - val_mean_absolute_error: 0.0069 - val_mean_squared_error: 9.5112e-05\n",
      "Epoch 38/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 9.1143e-05 - mean_absolute_error: 0.0071 - mean_squared_error: 9.1143e-05 - val_loss: 8.3481e-05 - val_mean_absolute_error: 0.0069 - val_mean_squared_error: 8.3481e-05\n",
      "Epoch 39/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 9.1061e-05 - mean_absolute_error: 0.0073 - mean_squared_error: 9.1061e-05 - val_loss: 6.8745e-05 - val_mean_absolute_error: 0.0064 - val_mean_squared_error: 6.8745e-05\n",
      "Epoch 40/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 7.6882e-05 - mean_absolute_error: 0.0065 - mean_squared_error: 7.6882e-05 - val_loss: 1.0770e-04 - val_mean_absolute_error: 0.0084 - val_mean_squared_error: 1.0770e-04\n",
      "Epoch 41/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 9.5251e-05 - mean_absolute_error: 0.0073 - mean_squared_error: 9.5251e-05 - val_loss: 2.1993e-04 - val_mean_absolute_error: 0.0116 - val_mean_squared_error: 2.1993e-04\n",
      "Epoch 42/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 8.6407e-05 - mean_absolute_error: 0.0069 - mean_squared_error: 8.6407e-05 - val_loss: 3.5253e-05 - val_mean_absolute_error: 0.0039 - val_mean_squared_error: 3.5253e-05\n",
      "Epoch 43/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 6.6480e-05 - mean_absolute_error: 0.0060 - mean_squared_error: 6.6480e-05 - val_loss: 6.0608e-05 - val_mean_absolute_error: 0.0060 - val_mean_squared_error: 6.0608e-05\n",
      "Epoch 44/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 1.0120e-04 - mean_absolute_error: 0.0072 - mean_squared_error: 1.0120e-04 - val_loss: 5.7152e-05 - val_mean_absolute_error: 0.0058 - val_mean_squared_error: 5.7152e-05\n",
      "Epoch 45/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 7.9711e-05 - mean_absolute_error: 0.0067 - mean_squared_error: 7.9711e-05 - val_loss: 7.4747e-05 - val_mean_absolute_error: 0.0064 - val_mean_squared_error: 7.4747e-05\n",
      "Epoch 46/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 5.4652e-04 - mean_absolute_error: 0.0135 - mean_squared_error: 5.4652e-04 - val_loss: 1.0702e-04 - val_mean_absolute_error: 0.0084 - val_mean_squared_error: 1.0702e-04\n",
      "Epoch 47/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 6.4829e-05 - mean_absolute_error: 0.0058 - mean_squared_error: 6.4828e-05 - val_loss: 6.2222e-05 - val_mean_absolute_error: 0.0053 - val_mean_squared_error: 6.2222e-05\n",
      "Epoch 48/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 4.6170e-05 - mean_absolute_error: 0.0048 - mean_squared_error: 4.6170e-05 - val_loss: 5.3372e-05 - val_mean_absolute_error: 0.0053 - val_mean_squared_error: 5.3372e-05\n",
      "Epoch 49/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 4.6107e-05 - mean_absolute_error: 0.0048 - mean_squared_error: 4.6107e-05 - val_loss: 4.9833e-05 - val_mean_absolute_error: 0.0048 - val_mean_squared_error: 4.9833e-05\n",
      "Epoch 50/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 4.8674e-05 - mean_absolute_error: 0.0049 - mean_squared_error: 4.8674e-05 - val_loss: 4.1393e-05 - val_mean_absolute_error: 0.0043 - val_mean_squared_error: 4.1393e-05\n",
      "Epoch 51/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 5.1508e-05 - mean_absolute_error: 0.0051 - mean_squared_error: 5.1508e-05 - val_loss: 7.3209e-05 - val_mean_absolute_error: 0.0066 - val_mean_squared_error: 7.3208e-05\n",
      "Epoch 52/300\n",
      "3600/3600 [==============================] - 0s 106us/sample - loss: 5.7103e-05 - mean_absolute_error: 0.0055 - mean_squared_error: 5.7103e-05 - val_loss: 7.1119e-05 - val_mean_absolute_error: 0.0065 - val_mean_squared_error: 7.1119e-05\n",
      "Epoch 53/300\n",
      "3600/3600 [==============================] - 0s 113us/sample - loss: 5.7240e-05 - mean_absolute_error: 0.0055 - mean_squared_error: 5.7240e-05 - val_loss: 5.5090e-05 - val_mean_absolute_error: 0.0055 - val_mean_squared_error: 5.5090e-05\n",
      "Epoch 54/300\n",
      "3600/3600 [==============================] - 0s 137us/sample - loss: 5.5000e-05 - mean_absolute_error: 0.0054 - mean_squared_error: 5.5000e-05 - val_loss: 7.3501e-05 - val_mean_absolute_error: 0.0058 - val_mean_squared_error: 7.3501e-05\n",
      "Epoch 55/300\n",
      "3600/3600 [==============================] - 0s 117us/sample - loss: 5.4180e-05 - mean_absolute_error: 0.0054 - mean_squared_error: 5.4180e-05 - val_loss: 7.8574e-05 - val_mean_absolute_error: 0.0071 - val_mean_squared_error: 7.8574e-05\n",
      "Epoch 56/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 6.8897e-05 - mean_absolute_error: 0.0063 - mean_squared_error: 6.8897e-05 - val_loss: 6.8813e-05 - val_mean_absolute_error: 0.0061 - val_mean_squared_error: 6.8813e-05\n",
      "Epoch 57/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 7.2949e-05 - mean_absolute_error: 0.0064 - mean_squared_error: 7.2949e-05 - val_loss: 8.9318e-05 - val_mean_absolute_error: 0.0077 - val_mean_squared_error: 8.9318e-05\n",
      "Epoch 58/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 6.2588e-05 - mean_absolute_error: 0.0059 - mean_squared_error: 6.2588e-05 - val_loss: 1.8845e-04 - val_mean_absolute_error: 0.0091 - val_mean_squared_error: 1.8845e-04\n",
      "Epoch 59/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 6.8706e-05 - mean_absolute_error: 0.0062 - mean_squared_error: 6.8706e-05 - val_loss: 1.4299e-04 - val_mean_absolute_error: 0.0092 - val_mean_squared_error: 1.4299e-04\n",
      "Epoch 60/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 6.8169e-05 - mean_absolute_error: 0.0062 - mean_squared_error: 6.8169e-05 - val_loss: 5.1282e-05 - val_mean_absolute_error: 0.0053 - val_mean_squared_error: 5.1282e-05\n",
      "Epoch 61/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 6.5413e-05 - mean_absolute_error: 0.0060 - mean_squared_error: 6.5413e-05 - val_loss: 1.2830e-04 - val_mean_absolute_error: 0.0094 - val_mean_squared_error: 1.2830e-04\n",
      "Epoch 62/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 6.7918e-05 - mean_absolute_error: 0.0061 - mean_squared_error: 6.7918e-05 - val_loss: 5.2623e-05 - val_mean_absolute_error: 0.0055 - val_mean_squared_error: 5.2623e-05\n",
      "Epoch 63/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 7.2801e-05 - mean_absolute_error: 0.0064 - mean_squared_error: 7.2801e-05 - val_loss: 9.8709e-05 - val_mean_absolute_error: 0.0077 - val_mean_squared_error: 9.8709e-05\n",
      "Epoch 64/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 5.9879e-05 - mean_absolute_error: 0.0058 - mean_squared_error: 5.9879e-05 - val_loss: 8.2466e-05 - val_mean_absolute_error: 0.0068 - val_mean_squared_error: 8.2466e-05\n",
      "Epoch 65/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 6.7652e-05 - mean_absolute_error: 0.0063 - mean_squared_error: 6.7652e-05 - val_loss: 6.1134e-05 - val_mean_absolute_error: 0.0058 - val_mean_squared_error: 6.1134e-05\n",
      "Epoch 66/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 5.8594e-05 - mean_absolute_error: 0.0057 - mean_squared_error: 5.8594e-05 - val_loss: 6.3904e-05 - val_mean_absolute_error: 0.0058 - val_mean_squared_error: 6.3904e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 67/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 6.6732e-05 - mean_absolute_error: 0.0061 - mean_squared_error: 6.6732e-05 - val_loss: 5.6065e-05 - val_mean_absolute_error: 0.0055 - val_mean_squared_error: 5.6065e-05\n",
      "Epoch 68/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 4.8646e-05 - mean_absolute_error: 0.0051 - mean_squared_error: 4.8646e-05 - val_loss: 6.1712e-05 - val_mean_absolute_error: 0.0061 - val_mean_squared_error: 6.1712e-05\n",
      "Epoch 69/300\n",
      "3600/3600 [==============================] - 0s 104us/sample - loss: 7.2393e-05 - mean_absolute_error: 0.0064 - mean_squared_error: 7.2393e-05 - val_loss: 5.6916e-05 - val_mean_absolute_error: 0.0058 - val_mean_squared_error: 5.6916e-05\n",
      "Epoch 70/300\n",
      "3600/3600 [==============================] - 0s 105us/sample - loss: 5.5772e-05 - mean_absolute_error: 0.0055 - mean_squared_error: 5.5772e-05 - val_loss: 7.7727e-05 - val_mean_absolute_error: 0.0070 - val_mean_squared_error: 7.7727e-05\n",
      "Epoch 71/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 5.2717e-05 - mean_absolute_error: 0.0054 - mean_squared_error: 5.2717e-05 - val_loss: 1.0214e-04 - val_mean_absolute_error: 0.0075 - val_mean_squared_error: 1.0214e-04\n",
      "Epoch 72/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 5.3407e-05 - mean_absolute_error: 0.0054 - mean_squared_error: 5.3407e-05 - val_loss: 6.8715e-05 - val_mean_absolute_error: 0.0064 - val_mean_squared_error: 6.8715e-05\n",
      "Epoch 73/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 6.1082e-05 - mean_absolute_error: 0.0058 - mean_squared_error: 6.1082e-05 - val_loss: 9.3521e-05 - val_mean_absolute_error: 0.0072 - val_mean_squared_error: 9.3521e-05\n",
      "Epoch 74/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 7.4667e-05 - mean_absolute_error: 0.0064 - mean_squared_error: 7.4667e-05 - val_loss: 1.2833e-04 - val_mean_absolute_error: 0.0092 - val_mean_squared_error: 1.2833e-04\n",
      "Epoch 75/300\n",
      "3600/3600 [==============================] - 0s 106us/sample - loss: 5.3980e-05 - mean_absolute_error: 0.0054 - mean_squared_error: 5.3980e-05 - val_loss: 4.9429e-05 - val_mean_absolute_error: 0.0056 - val_mean_squared_error: 4.9429e-05\n",
      "Epoch 76/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 5.4900e-05 - mean_absolute_error: 0.0055 - mean_squared_error: 5.4900e-05 - val_loss: 5.6967e-05 - val_mean_absolute_error: 0.0057 - val_mean_squared_error: 5.6967e-05\n",
      "Epoch 77/300\n",
      "3600/3600 [==============================] - 0s 128us/sample - loss: 6.0202e-05 - mean_absolute_error: 0.0058 - mean_squared_error: 6.0202e-05 - val_loss: 4.7205e-05 - val_mean_absolute_error: 0.0053 - val_mean_squared_error: 4.7205e-05\n",
      "Epoch 78/300\n",
      "3600/3600 [==============================] - 0s 115us/sample - loss: 6.5091e-05 - mean_absolute_error: 0.0060 - mean_squared_error: 6.5091e-05 - val_loss: 6.1013e-05 - val_mean_absolute_error: 0.0059 - val_mean_squared_error: 6.1013e-05\n",
      "Epoch 79/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 4.8020e-05 - mean_absolute_error: 0.0051 - mean_squared_error: 4.8020e-05 - val_loss: 5.3570e-05 - val_mean_absolute_error: 0.0054 - val_mean_squared_error: 5.3570e-05\n",
      "Epoch 80/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 5.3072e-05 - mean_absolute_error: 0.0054 - mean_squared_error: 5.3072e-05 - val_loss: 7.2483e-05 - val_mean_absolute_error: 0.0062 - val_mean_squared_error: 7.2483e-05\n",
      "Epoch 81/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 5.5761e-05 - mean_absolute_error: 0.0055 - mean_squared_error: 5.5761e-05 - val_loss: 5.8595e-05 - val_mean_absolute_error: 0.0058 - val_mean_squared_error: 5.8595e-05\n",
      "Epoch 82/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 6.8533e-05 - mean_absolute_error: 0.0062 - mean_squared_error: 6.8533e-05 - val_loss: 5.4795e-05 - val_mean_absolute_error: 0.0057 - val_mean_squared_error: 5.4795e-05\n",
      "Epoch 83/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 5.7042e-05 - mean_absolute_error: 0.0056 - mean_squared_error: 5.7042e-05 - val_loss: 4.5464e-05 - val_mean_absolute_error: 0.0050 - val_mean_squared_error: 4.5464e-05\n",
      "Epoch 84/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 5.1185e-05 - mean_absolute_error: 0.0053 - mean_squared_error: 5.1185e-05 - val_loss: 3.1735e-05 - val_mean_absolute_error: 0.0038 - val_mean_squared_error: 3.1735e-05\n",
      "Epoch 85/300\n",
      "3600/3600 [==============================] - 0s 107us/sample - loss: 5.4077e-05 - mean_absolute_error: 0.0055 - mean_squared_error: 5.4077e-05 - val_loss: 5.9680e-05 - val_mean_absolute_error: 0.0055 - val_mean_squared_error: 5.9680e-05\n",
      "Epoch 86/300\n",
      "3600/3600 [==============================] - 0s 107us/sample - loss: 5.2991e-05 - mean_absolute_error: 0.0054 - mean_squared_error: 5.2991e-05 - val_loss: 7.0240e-05 - val_mean_absolute_error: 0.0058 - val_mean_squared_error: 7.0240e-05\n",
      "Epoch 87/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 5.4922e-05 - mean_absolute_error: 0.0055 - mean_squared_error: 5.4922e-05 - val_loss: 4.1651e-05 - val_mean_absolute_error: 0.0046 - val_mean_squared_error: 4.1651e-05\n",
      "Epoch 88/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 5.6065e-05 - mean_absolute_error: 0.0056 - mean_squared_error: 5.6065e-05 - val_loss: 4.3679e-05 - val_mean_absolute_error: 0.0050 - val_mean_squared_error: 4.3679e-05\n",
      "Epoch 89/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 4.5008e-05 - mean_absolute_error: 0.0049 - mean_squared_error: 4.5008e-05 - val_loss: 4.1957e-05 - val_mean_absolute_error: 0.0048 - val_mean_squared_error: 4.1957e-05\n",
      "Epoch 90/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 4.4549e-05 - mean_absolute_error: 0.0049 - mean_squared_error: 4.4549e-05 - val_loss: 4.4967e-05 - val_mean_absolute_error: 0.0052 - val_mean_squared_error: 4.4967e-05\n",
      "Epoch 91/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 4.8102e-05 - mean_absolute_error: 0.0052 - mean_squared_error: 4.8102e-05 - val_loss: 4.1970e-05 - val_mean_absolute_error: 0.0048 - val_mean_squared_error: 4.1970e-05\n",
      "Epoch 92/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 5.6215e-05 - mean_absolute_error: 0.0056 - mean_squared_error: 5.6215e-05 - val_loss: 4.0418e-05 - val_mean_absolute_error: 0.0047 - val_mean_squared_error: 4.0418e-05\n",
      "Epoch 93/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 5.2149e-05 - mean_absolute_error: 0.0053 - mean_squared_error: 5.2149e-05 - val_loss: 3.5064e-05 - val_mean_absolute_error: 0.0042 - val_mean_squared_error: 3.5064e-05\n",
      "Epoch 94/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 5.4573e-05 - mean_absolute_error: 0.0054 - mean_squared_error: 5.4573e-05 - val_loss: 4.5003e-05 - val_mean_absolute_error: 0.0047 - val_mean_squared_error: 4.5003e-05\n",
      "Epoch 95/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 5.2983e-05 - mean_absolute_error: 0.0052 - mean_squared_error: 5.2983e-05 - val_loss: 1.1502e-04 - val_mean_absolute_error: 0.0083 - val_mean_squared_error: 1.1502e-04\n",
      "Epoch 96/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 5.2021e-05 - mean_absolute_error: 0.0053 - mean_squared_error: 5.2021e-05 - val_loss: 1.0359e-04 - val_mean_absolute_error: 0.0087 - val_mean_squared_error: 1.0359e-04\n",
      "Epoch 97/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 4.9943e-05 - mean_absolute_error: 0.0052 - mean_squared_error: 4.9943e-05 - val_loss: 3.7563e-05 - val_mean_absolute_error: 0.0041 - val_mean_squared_error: 3.7563e-05\n",
      "Epoch 98/300\n",
      "3600/3600 [==============================] - 0s 107us/sample - loss: 5.8185e-05 - mean_absolute_error: 0.0056 - mean_squared_error: 5.8185e-05 - val_loss: 5.8232e-05 - val_mean_absolute_error: 0.0053 - val_mean_squared_error: 5.8232e-05\n",
      "Epoch 99/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 4.5003e-05 - mean_absolute_error: 0.0049 - mean_squared_error: 4.5003e-05 - val_loss: 5.0452e-05 - val_mean_absolute_error: 0.0053 - val_mean_squared_error: 5.0452e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 100/300\n",
      "3600/3600 [==============================] - 0s 118us/sample - loss: 5.5036e-05 - mean_absolute_error: 0.0055 - mean_squared_error: 5.5036e-05 - val_loss: 3.7047e-05 - val_mean_absolute_error: 0.0044 - val_mean_squared_error: 3.7047e-05\n",
      "Epoch 101/300\n",
      "3600/3600 [==============================] - 0s 133us/sample - loss: 4.4231e-05 - mean_absolute_error: 0.0048 - mean_squared_error: 4.4231e-05 - val_loss: 4.7204e-05 - val_mean_absolute_error: 0.0054 - val_mean_squared_error: 4.7204e-05\n",
      "Epoch 102/300\n",
      "3600/3600 [==============================] - 0s 127us/sample - loss: 4.2772e-05 - mean_absolute_error: 0.0047 - mean_squared_error: 4.2772e-05 - val_loss: 5.2147e-05 - val_mean_absolute_error: 0.0054 - val_mean_squared_error: 5.2147e-05\n",
      "Epoch 103/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 6.7136e-05 - mean_absolute_error: 0.0059 - mean_squared_error: 6.7136e-05 - val_loss: 5.7398e-05 - val_mean_absolute_error: 0.0058 - val_mean_squared_error: 5.7398e-05\n",
      "Epoch 104/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 4.5007e-05 - mean_absolute_error: 0.0048 - mean_squared_error: 4.5007e-05 - val_loss: 4.9902e-05 - val_mean_absolute_error: 0.0049 - val_mean_squared_error: 4.9902e-05\n",
      "Epoch 105/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 4.3597e-05 - mean_absolute_error: 0.0048 - mean_squared_error: 4.3597e-05 - val_loss: 4.3945e-05 - val_mean_absolute_error: 0.0043 - val_mean_squared_error: 4.3945e-05\n",
      "Epoch 106/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 4.4339e-05 - mean_absolute_error: 0.0048 - mean_squared_error: 4.4339e-05 - val_loss: 5.9953e-05 - val_mean_absolute_error: 0.0053 - val_mean_squared_error: 5.9953e-05\n",
      "Epoch 107/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 4.0838e-05 - mean_absolute_error: 0.0046 - mean_squared_error: 4.0838e-05 - val_loss: 4.0735e-05 - val_mean_absolute_error: 0.0047 - val_mean_squared_error: 4.0735e-05\n",
      "Epoch 108/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 4.5825e-05 - mean_absolute_error: 0.0049 - mean_squared_error: 4.5825e-05 - val_loss: 2.9283e-05 - val_mean_absolute_error: 0.0036 - val_mean_squared_error: 2.9283e-05\n",
      "Epoch 109/300\n",
      "3600/3600 [==============================] - 0s 106us/sample - loss: 4.2230e-05 - mean_absolute_error: 0.0046 - mean_squared_error: 4.2230e-05 - val_loss: 3.6991e-05 - val_mean_absolute_error: 0.0040 - val_mean_squared_error: 3.6991e-05\n",
      "Epoch 110/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 4.6540e-05 - mean_absolute_error: 0.0049 - mean_squared_error: 4.6540e-05 - val_loss: 3.3347e-05 - val_mean_absolute_error: 0.0041 - val_mean_squared_error: 3.3347e-05\n",
      "Epoch 111/300\n",
      "3600/3600 [==============================] - 0s 105us/sample - loss: 4.6099e-05 - mean_absolute_error: 0.0049 - mean_squared_error: 4.6099e-05 - val_loss: 3.9212e-05 - val_mean_absolute_error: 0.0047 - val_mean_squared_error: 3.9212e-05\n",
      "Epoch 112/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 4.3431e-05 - mean_absolute_error: 0.0047 - mean_squared_error: 4.3431e-05 - val_loss: 3.5866e-05 - val_mean_absolute_error: 0.0043 - val_mean_squared_error: 3.5866e-05\n",
      "Epoch 113/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 6.0615e-05 - mean_absolute_error: 0.0056 - mean_squared_error: 6.0615e-05 - val_loss: 4.5530e-05 - val_mean_absolute_error: 0.0048 - val_mean_squared_error: 4.5530e-05\n",
      "Epoch 114/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 4.1966e-05 - mean_absolute_error: 0.0046 - mean_squared_error: 4.1966e-05 - val_loss: 5.0992e-05 - val_mean_absolute_error: 0.0056 - val_mean_squared_error: 5.0992e-05\n",
      "Epoch 115/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 4.8922e-05 - mean_absolute_error: 0.0050 - mean_squared_error: 4.8922e-05 - val_loss: 3.5820e-05 - val_mean_absolute_error: 0.0041 - val_mean_squared_error: 3.5820e-05\n",
      "Epoch 116/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 4.8747e-05 - mean_absolute_error: 0.0051 - mean_squared_error: 4.8747e-05 - val_loss: 3.7112e-05 - val_mean_absolute_error: 0.0044 - val_mean_squared_error: 3.7112e-05\n",
      "Epoch 117/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 3.9430e-05 - mean_absolute_error: 0.0045 - mean_squared_error: 3.9430e-05 - val_loss: 5.3445e-05 - val_mean_absolute_error: 0.0047 - val_mean_squared_error: 5.3445e-05\n",
      "Epoch 118/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 4.5985e-05 - mean_absolute_error: 0.0050 - mean_squared_error: 4.5985e-05 - val_loss: 3.6512e-05 - val_mean_absolute_error: 0.0040 - val_mean_squared_error: 3.6512e-05\n",
      "Epoch 119/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.5416e-05 - mean_absolute_error: 0.0042 - mean_squared_error: 3.5416e-05 - val_loss: 3.1586e-05 - val_mean_absolute_error: 0.0037 - val_mean_squared_error: 3.1586e-05\n",
      "Epoch 120/300\n",
      "3600/3600 [==============================] - 0s 107us/sample - loss: 3.9791e-05 - mean_absolute_error: 0.0045 - mean_squared_error: 3.9791e-05 - val_loss: 3.4378e-05 - val_mean_absolute_error: 0.0042 - val_mean_squared_error: 3.4378e-05\n",
      "Epoch 121/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 4.4626e-05 - mean_absolute_error: 0.0049 - mean_squared_error: 4.4626e-05 - val_loss: 4.6206e-05 - val_mean_absolute_error: 0.0048 - val_mean_squared_error: 4.6206e-05\n",
      "Epoch 122/300\n",
      "3600/3600 [==============================] - 0s 123us/sample - loss: 3.7990e-05 - mean_absolute_error: 0.0044 - mean_squared_error: 3.7990e-05 - val_loss: 6.6834e-05 - val_mean_absolute_error: 0.0059 - val_mean_squared_error: 6.6834e-05\n",
      "Epoch 123/300\n",
      "3600/3600 [==============================] - 1s 139us/sample - loss: 4.7247e-05 - mean_absolute_error: 0.0051 - mean_squared_error: 4.7247e-05 - val_loss: 4.2214e-05 - val_mean_absolute_error: 0.0049 - val_mean_squared_error: 4.2214e-05\n",
      "Epoch 124/300\n",
      "3600/3600 [==============================] - 0s 114us/sample - loss: 4.1036e-05 - mean_absolute_error: 0.0046 - mean_squared_error: 4.1036e-05 - val_loss: 4.4616e-05 - val_mean_absolute_error: 0.0048 - val_mean_squared_error: 4.4616e-05\n",
      "Epoch 125/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 4.6359e-05 - mean_absolute_error: 0.0049 - mean_squared_error: 4.6359e-05 - val_loss: 3.7029e-05 - val_mean_absolute_error: 0.0042 - val_mean_squared_error: 3.7029e-05\n",
      "Epoch 126/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 4.2527e-05 - mean_absolute_error: 0.0048 - mean_squared_error: 4.2527e-05 - val_loss: 3.0320e-05 - val_mean_absolute_error: 0.0037 - val_mean_squared_error: 3.0320e-05\n",
      "Epoch 127/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 4.8490e-05 - mean_absolute_error: 0.0051 - mean_squared_error: 4.8490e-05 - val_loss: 3.9339e-05 - val_mean_absolute_error: 0.0040 - val_mean_squared_error: 3.9339e-05\n",
      "Epoch 128/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 4.2344e-05 - mean_absolute_error: 0.0046 - mean_squared_error: 4.2344e-05 - val_loss: 4.0591e-05 - val_mean_absolute_error: 0.0047 - val_mean_squared_error: 4.0591e-05\n",
      "Epoch 129/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 3.9497e-05 - mean_absolute_error: 0.0045 - mean_squared_error: 3.9497e-05 - val_loss: 3.5510e-05 - val_mean_absolute_error: 0.0041 - val_mean_squared_error: 3.5510e-05\n",
      "Epoch 130/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 4.5545e-05 - mean_absolute_error: 0.0048 - mean_squared_error: 4.5545e-05 - val_loss: 5.1725e-05 - val_mean_absolute_error: 0.0052 - val_mean_squared_error: 5.1725e-05\n",
      "Epoch 131/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 4.6017e-05 - mean_absolute_error: 0.0050 - mean_squared_error: 4.6017e-05 - val_loss: 5.7381e-05 - val_mean_absolute_error: 0.0054 - val_mean_squared_error: 5.7381e-05\n",
      "Epoch 132/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 3.9078e-05 - mean_absolute_error: 0.0045 - mean_squared_error: 3.9078e-05 - val_loss: 3.2951e-05 - val_mean_absolute_error: 0.0041 - val_mean_squared_error: 3.2951e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 133/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 4.5920e-05 - mean_absolute_error: 0.0049 - mean_squared_error: 4.5920e-05 - val_loss: 3.2365e-05 - val_mean_absolute_error: 0.0036 - val_mean_squared_error: 3.2365e-05\n",
      "Epoch 134/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.4587e-05 - mean_absolute_error: 0.0041 - mean_squared_error: 3.4587e-05 - val_loss: 3.3199e-05 - val_mean_absolute_error: 0.0041 - val_mean_squared_error: 3.3199e-05\n",
      "Epoch 135/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 4.1067e-05 - mean_absolute_error: 0.0047 - mean_squared_error: 4.1067e-05 - val_loss: 4.7881e-05 - val_mean_absolute_error: 0.0048 - val_mean_squared_error: 4.7881e-05\n",
      "Epoch 136/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 4.5478e-05 - mean_absolute_error: 0.0049 - mean_squared_error: 4.5478e-05 - val_loss: 2.9492e-05 - val_mean_absolute_error: 0.0038 - val_mean_squared_error: 2.9492e-05\n",
      "Epoch 137/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 4.1527e-05 - mean_absolute_error: 0.0047 - mean_squared_error: 4.1527e-05 - val_loss: 3.4699e-05 - val_mean_absolute_error: 0.0044 - val_mean_squared_error: 3.4699e-05\n",
      "Epoch 138/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.7968e-05 - mean_absolute_error: 0.0043 - mean_squared_error: 3.7968e-05 - val_loss: 6.1477e-05 - val_mean_absolute_error: 0.0051 - val_mean_squared_error: 6.1477e-05\n",
      "Epoch 139/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 3.9997e-05 - mean_absolute_error: 0.0045 - mean_squared_error: 3.9997e-05 - val_loss: 3.1806e-05 - val_mean_absolute_error: 0.0041 - val_mean_squared_error: 3.1806e-05\n",
      "Epoch 140/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 4.0279e-05 - mean_absolute_error: 0.0045 - mean_squared_error: 4.0279e-05 - val_loss: 7.9131e-05 - val_mean_absolute_error: 0.0056 - val_mean_squared_error: 7.9131e-05\n",
      "Epoch 141/300\n",
      "3600/3600 [==============================] - 0s 107us/sample - loss: 4.1413e-05 - mean_absolute_error: 0.0045 - mean_squared_error: 4.1413e-05 - val_loss: 9.1156e-05 - val_mean_absolute_error: 0.0063 - val_mean_squared_error: 9.1156e-05\n",
      "Epoch 142/300\n",
      "3600/3600 [==============================] - 0s 106us/sample - loss: 4.7654e-05 - mean_absolute_error: 0.0050 - mean_squared_error: 4.7654e-05 - val_loss: 3.8683e-05 - val_mean_absolute_error: 0.0046 - val_mean_squared_error: 3.8683e-05\n",
      "Epoch 143/300\n",
      "3600/3600 [==============================] - 0s 106us/sample - loss: 4.3257e-05 - mean_absolute_error: 0.0047 - mean_squared_error: 4.3257e-05 - val_loss: 3.0641e-05 - val_mean_absolute_error: 0.0039 - val_mean_squared_error: 3.0641e-05\n",
      "Epoch 144/300\n",
      "3600/3600 [==============================] - 0s 107us/sample - loss: 3.6905e-05 - mean_absolute_error: 0.0043 - mean_squared_error: 3.6905e-05 - val_loss: 2.6142e-05 - val_mean_absolute_error: 0.0033 - val_mean_squared_error: 2.6142e-05\n",
      "Epoch 145/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.8003e-05 - mean_absolute_error: 0.0044 - mean_squared_error: 3.8003e-05 - val_loss: 3.0052e-05 - val_mean_absolute_error: 0.0038 - val_mean_squared_error: 3.0052e-05\n",
      "Epoch 146/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 3.9173e-05 - mean_absolute_error: 0.0045 - mean_squared_error: 3.9173e-05 - val_loss: 6.6496e-05 - val_mean_absolute_error: 0.0053 - val_mean_squared_error: 6.6496e-05\n",
      "Epoch 147/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 4.0205e-05 - mean_absolute_error: 0.0044 - mean_squared_error: 4.0205e-05 - val_loss: 2.9825e-05 - val_mean_absolute_error: 0.0036 - val_mean_squared_error: 2.9825e-05\n",
      "Epoch 148/300\n",
      "3600/3600 [==============================] - 0s 126us/sample - loss: 3.6786e-05 - mean_absolute_error: 0.0043 - mean_squared_error: 3.6786e-05 - val_loss: 3.7848e-05 - val_mean_absolute_error: 0.0040 - val_mean_squared_error: 3.7848e-05\n",
      "Epoch 149/300\n",
      "3600/3600 [==============================] - 0s 139us/sample - loss: 3.3513e-05 - mean_absolute_error: 0.0040 - mean_squared_error: 3.3513e-05 - val_loss: 5.0578e-05 - val_mean_absolute_error: 0.0046 - val_mean_squared_error: 5.0578e-05\n",
      "Epoch 150/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 3.3991e-05 - mean_absolute_error: 0.0041 - mean_squared_error: 3.3991e-05 - val_loss: 6.1328e-05 - val_mean_absolute_error: 0.0051 - val_mean_squared_error: 6.1328e-05\n",
      "Epoch 151/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 5.0368e-05 - mean_absolute_error: 0.0051 - mean_squared_error: 5.0368e-05 - val_loss: 6.6313e-05 - val_mean_absolute_error: 0.0061 - val_mean_squared_error: 6.6313e-05\n",
      "Epoch 152/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.8474e-05 - mean_absolute_error: 0.0043 - mean_squared_error: 3.8474e-05 - val_loss: 3.0635e-05 - val_mean_absolute_error: 0.0038 - val_mean_squared_error: 3.0635e-05\n",
      "Epoch 153/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 3.7740e-05 - mean_absolute_error: 0.0042 - mean_squared_error: 3.7740e-05 - val_loss: 6.6909e-05 - val_mean_absolute_error: 0.0069 - val_mean_squared_error: 6.6909e-05\n",
      "Epoch 154/300\n",
      "3600/3600 [==============================] - 0s 113us/sample - loss: 3.2489e-05 - mean_absolute_error: 0.0039 - mean_squared_error: 3.2489e-05 - val_loss: 5.7951e-05 - val_mean_absolute_error: 0.0057 - val_mean_squared_error: 5.7951e-05\n",
      "Epoch 155/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.7700e-05 - mean_absolute_error: 0.0043 - mean_squared_error: 3.7700e-05 - val_loss: 1.0254e-04 - val_mean_absolute_error: 0.0084 - val_mean_squared_error: 1.0254e-04\n",
      "Epoch 156/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 4.8670e-05 - mean_absolute_error: 0.0051 - mean_squared_error: 4.8670e-05 - val_loss: 4.0380e-05 - val_mean_absolute_error: 0.0047 - val_mean_squared_error: 4.0380e-05\n",
      "Epoch 157/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 3.5185e-05 - mean_absolute_error: 0.0042 - mean_squared_error: 3.5185e-05 - val_loss: 4.1194e-05 - val_mean_absolute_error: 0.0040 - val_mean_squared_error: 4.1194e-05\n",
      "Epoch 158/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 3.4208e-05 - mean_absolute_error: 0.0041 - mean_squared_error: 3.4208e-05 - val_loss: 2.8464e-05 - val_mean_absolute_error: 0.0034 - val_mean_squared_error: 2.8464e-05\n",
      "Epoch 159/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.6941e-05 - mean_absolute_error: 0.0043 - mean_squared_error: 3.6941e-05 - val_loss: 3.1153e-05 - val_mean_absolute_error: 0.0038 - val_mean_squared_error: 3.1153e-05\n",
      "Epoch 160/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 3.4856e-05 - mean_absolute_error: 0.0041 - mean_squared_error: 3.4856e-05 - val_loss: 3.6915e-05 - val_mean_absolute_error: 0.0038 - val_mean_squared_error: 3.6915e-05\n",
      "Epoch 161/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.9937e-05 - mean_absolute_error: 0.0044 - mean_squared_error: 3.9937e-05 - val_loss: 4.5743e-05 - val_mean_absolute_error: 0.0053 - val_mean_squared_error: 4.5743e-05\n",
      "Epoch 162/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.5923e-05 - mean_absolute_error: 0.0042 - mean_squared_error: 3.5923e-05 - val_loss: 4.0326e-05 - val_mean_absolute_error: 0.0043 - val_mean_squared_error: 4.0326e-05\n",
      "Epoch 163/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 9.1840e-05 - mean_absolute_error: 0.0067 - mean_squared_error: 9.1840e-05 - val_loss: 3.5177e-05 - val_mean_absolute_error: 0.0041 - val_mean_squared_error: 3.5177e-05\n",
      "Epoch 164/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.3899e-05 - mean_absolute_error: 0.0040 - mean_squared_error: 3.3899e-05 - val_loss: 3.1010e-05 - val_mean_absolute_error: 0.0036 - val_mean_squared_error: 3.1010e-05\n",
      "Epoch 165/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.6326e-05 - mean_absolute_error: 0.0042 - mean_squared_error: 3.6326e-05 - val_loss: 2.6676e-05 - val_mean_absolute_error: 0.0034 - val_mean_squared_error: 2.6676e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 166/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 3.2996e-05 - mean_absolute_error: 0.0039 - mean_squared_error: 3.2996e-05 - val_loss: 3.0407e-05 - val_mean_absolute_error: 0.0038 - val_mean_squared_error: 3.0407e-05\n",
      "Epoch 167/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.0521e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 3.0521e-05 - val_loss: 2.8986e-05 - val_mean_absolute_error: 0.0038 - val_mean_squared_error: 2.8986e-05\n",
      "Epoch 168/300\n",
      "3600/3600 [==============================] - 0s 106us/sample - loss: 3.1188e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 3.1188e-05 - val_loss: 2.8385e-05 - val_mean_absolute_error: 0.0034 - val_mean_squared_error: 2.8385e-05\n",
      "Epoch 169/300\n",
      "3600/3600 [==============================] - 0s 106us/sample - loss: 2.9721e-05 - mean_absolute_error: 0.0037 - mean_squared_error: 2.9721e-05 - val_loss: 3.5577e-05 - val_mean_absolute_error: 0.0040 - val_mean_squared_error: 3.5577e-05\n",
      "Epoch 170/300\n",
      "3600/3600 [==============================] - 0s 115us/sample - loss: 3.3766e-05 - mean_absolute_error: 0.0041 - mean_squared_error: 3.3766e-05 - val_loss: 3.4778e-05 - val_mean_absolute_error: 0.0044 - val_mean_squared_error: 3.4778e-05\n",
      "Epoch 171/300\n",
      "3600/3600 [==============================] - 0s 129us/sample - loss: 3.5251e-05 - mean_absolute_error: 0.0041 - mean_squared_error: 3.5251e-05 - val_loss: 5.6645e-05 - val_mean_absolute_error: 0.0050 - val_mean_squared_error: 5.6645e-05\n",
      "Epoch 172/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 3.6826e-05 - mean_absolute_error: 0.0043 - mean_squared_error: 3.6826e-05 - val_loss: 5.9836e-05 - val_mean_absolute_error: 0.0063 - val_mean_squared_error: 5.9836e-05\n",
      "Epoch 173/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.6975e-05 - mean_absolute_error: 0.0043 - mean_squared_error: 3.6975e-05 - val_loss: 3.6514e-05 - val_mean_absolute_error: 0.0040 - val_mean_squared_error: 3.6514e-05\n",
      "Epoch 174/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 4.0166e-05 - mean_absolute_error: 0.0044 - mean_squared_error: 4.0166e-05 - val_loss: 7.6561e-05 - val_mean_absolute_error: 0.0059 - val_mean_squared_error: 7.6561e-05\n",
      "Epoch 175/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 4.0330e-05 - mean_absolute_error: 0.0045 - mean_squared_error: 4.0330e-05 - val_loss: 4.0673e-05 - val_mean_absolute_error: 0.0045 - val_mean_squared_error: 4.0673e-05\n",
      "Epoch 176/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.9982e-05 - mean_absolute_error: 0.0046 - mean_squared_error: 3.9982e-05 - val_loss: 3.4739e-05 - val_mean_absolute_error: 0.0036 - val_mean_squared_error: 3.4739e-05\n",
      "Epoch 177/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 3.3983e-05 - mean_absolute_error: 0.0041 - mean_squared_error: 3.3983e-05 - val_loss: 3.3435e-05 - val_mean_absolute_error: 0.0042 - val_mean_squared_error: 3.3435e-05\n",
      "Epoch 178/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 4.2155e-05 - mean_absolute_error: 0.0047 - mean_squared_error: 4.2155e-05 - val_loss: 3.0506e-05 - val_mean_absolute_error: 0.0038 - val_mean_squared_error: 3.0506e-05\n",
      "Epoch 179/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.3061e-05 - mean_absolute_error: 0.0040 - mean_squared_error: 3.3061e-05 - val_loss: 3.2652e-05 - val_mean_absolute_error: 0.0040 - val_mean_squared_error: 3.2652e-05\n",
      "Epoch 180/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.7193e-05 - mean_absolute_error: 0.0044 - mean_squared_error: 3.7193e-05 - val_loss: 2.6395e-05 - val_mean_absolute_error: 0.0032 - val_mean_squared_error: 2.6395e-05\n",
      "Epoch 181/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.0433e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 3.0433e-05 - val_loss: 3.7305e-05 - val_mean_absolute_error: 0.0047 - val_mean_squared_error: 3.7305e-05\n",
      "Epoch 182/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.2998e-05 - mean_absolute_error: 0.0040 - mean_squared_error: 3.2998e-05 - val_loss: 2.3101e-05 - val_mean_absolute_error: 0.0030 - val_mean_squared_error: 2.3101e-05\n",
      "Epoch 183/300\n",
      "3600/3600 [==============================] - 0s 106us/sample - loss: 3.3291e-05 - mean_absolute_error: 0.0040 - mean_squared_error: 3.3291e-05 - val_loss: 2.9682e-05 - val_mean_absolute_error: 0.0040 - val_mean_squared_error: 2.9682e-05\n",
      "Epoch 184/300\n",
      "3600/3600 [==============================] - 0s 107us/sample - loss: 3.1014e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 3.1014e-05 - val_loss: 3.0656e-05 - val_mean_absolute_error: 0.0040 - val_mean_squared_error: 3.0656e-05\n",
      "Epoch 185/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 4.0793e-05 - mean_absolute_error: 0.0045 - mean_squared_error: 4.0793e-05 - val_loss: 5.9703e-05 - val_mean_absolute_error: 0.0064 - val_mean_squared_error: 5.9703e-05\n",
      "Epoch 186/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 3.4119e-05 - mean_absolute_error: 0.0040 - mean_squared_error: 3.4119e-05 - val_loss: 2.5052e-05 - val_mean_absolute_error: 0.0033 - val_mean_squared_error: 2.5052e-05\n",
      "Epoch 187/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.2586e-05 - mean_absolute_error: 0.0039 - mean_squared_error: 3.2586e-05 - val_loss: 3.0375e-05 - val_mean_absolute_error: 0.0037 - val_mean_squared_error: 3.0375e-05\n",
      "Epoch 188/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.9572e-05 - mean_absolute_error: 0.0044 - mean_squared_error: 3.9572e-05 - val_loss: 3.0653e-05 - val_mean_absolute_error: 0.0039 - val_mean_squared_error: 3.0653e-05\n",
      "Epoch 189/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.3101e-05 - mean_absolute_error: 0.0039 - mean_squared_error: 3.3101e-05 - val_loss: 8.5686e-05 - val_mean_absolute_error: 0.0072 - val_mean_squared_error: 8.5686e-05\n",
      "Epoch 190/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.5229e-05 - mean_absolute_error: 0.0041 - mean_squared_error: 3.5229e-05 - val_loss: 2.9184e-05 - val_mean_absolute_error: 0.0038 - val_mean_squared_error: 2.9184e-05\n",
      "Epoch 191/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 3.4394e-05 - mean_absolute_error: 0.0041 - mean_squared_error: 3.4394e-05 - val_loss: 2.8939e-05 - val_mean_absolute_error: 0.0037 - val_mean_squared_error: 2.8939e-05\n",
      "Epoch 192/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 2.9402e-05 - mean_absolute_error: 0.0037 - mean_squared_error: 2.9402e-05 - val_loss: 3.1019e-05 - val_mean_absolute_error: 0.0039 - val_mean_squared_error: 3.1019e-05\n",
      "Epoch 193/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 3.3597e-05 - mean_absolute_error: 0.0041 - mean_squared_error: 3.3597e-05 - val_loss: 4.1217e-05 - val_mean_absolute_error: 0.0048 - val_mean_squared_error: 4.1217e-05\n",
      "Epoch 194/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 3.4773e-05 - mean_absolute_error: 0.0041 - mean_squared_error: 3.4773e-05 - val_loss: 3.8600e-05 - val_mean_absolute_error: 0.0040 - val_mean_squared_error: 3.8600e-05\n",
      "Epoch 195/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.6387e-05 - mean_absolute_error: 0.0042 - mean_squared_error: 3.6387e-05 - val_loss: 3.3203e-05 - val_mean_absolute_error: 0.0039 - val_mean_squared_error: 3.3203e-05\n",
      "Epoch 196/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.2739e-05 - mean_absolute_error: 0.0040 - mean_squared_error: 3.2739e-05 - val_loss: 2.9984e-05 - val_mean_absolute_error: 0.0036 - val_mean_squared_error: 2.9984e-05\n",
      "Epoch 197/300\n",
      "3600/3600 [==============================] - 1s 140us/sample - loss: 4.0184e-05 - mean_absolute_error: 0.0045 - mean_squared_error: 4.0184e-05 - val_loss: 2.7532e-05 - val_mean_absolute_error: 0.0035 - val_mean_squared_error: 2.7532e-05\n",
      "Epoch 198/300\n",
      "3600/3600 [==============================] - 0s 137us/sample - loss: 3.3111e-05 - mean_absolute_error: 0.0040 - mean_squared_error: 3.3111e-05 - val_loss: 2.7649e-05 - val_mean_absolute_error: 0.0036 - val_mean_squared_error: 2.7649e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 199/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 3.1854e-05 - mean_absolute_error: 0.0039 - mean_squared_error: 3.1854e-05 - val_loss: 2.4946e-05 - val_mean_absolute_error: 0.0033 - val_mean_squared_error: 2.4946e-05\n",
      "Epoch 200/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.4912e-05 - mean_absolute_error: 0.0041 - mean_squared_error: 3.4912e-05 - val_loss: 2.3674e-05 - val_mean_absolute_error: 0.0031 - val_mean_squared_error: 2.3674e-05\n",
      "Epoch 201/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.7074e-05 - mean_absolute_error: 0.0043 - mean_squared_error: 3.7074e-05 - val_loss: 3.5604e-05 - val_mean_absolute_error: 0.0043 - val_mean_squared_error: 3.5604e-05\n",
      "Epoch 202/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.6501e-05 - mean_absolute_error: 0.0043 - mean_squared_error: 3.6501e-05 - val_loss: 3.1674e-05 - val_mean_absolute_error: 0.0038 - val_mean_squared_error: 3.1674e-05\n",
      "Epoch 203/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 4.3787e-05 - mean_absolute_error: 0.0046 - mean_squared_error: 4.3787e-05 - val_loss: 4.7227e-05 - val_mean_absolute_error: 0.0048 - val_mean_squared_error: 4.7227e-05\n",
      "Epoch 204/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 3.4210e-05 - mean_absolute_error: 0.0040 - mean_squared_error: 3.4210e-05 - val_loss: 3.2934e-05 - val_mean_absolute_error: 0.0042 - val_mean_squared_error: 3.2934e-05\n",
      "Epoch 205/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.1934e-05 - mean_absolute_error: 0.0039 - mean_squared_error: 3.1934e-05 - val_loss: 2.4590e-05 - val_mean_absolute_error: 0.0033 - val_mean_squared_error: 2.4590e-05\n",
      "Epoch 206/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 2.9393e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 2.9393e-05 - val_loss: 2.5605e-05 - val_mean_absolute_error: 0.0035 - val_mean_squared_error: 2.5605e-05\n",
      "Epoch 207/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 3.4255e-05 - mean_absolute_error: 0.0041 - mean_squared_error: 3.4255e-05 - val_loss: 2.4879e-05 - val_mean_absolute_error: 0.0032 - val_mean_squared_error: 2.4879e-05\n",
      "Epoch 208/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.1925e-05 - mean_absolute_error: 0.0039 - mean_squared_error: 3.1925e-05 - val_loss: 2.8657e-05 - val_mean_absolute_error: 0.0039 - val_mean_squared_error: 2.8657e-05\n",
      "Epoch 209/300\n",
      "3600/3600 [==============================] - 0s 107us/sample - loss: 2.8120e-05 - mean_absolute_error: 0.0036 - mean_squared_error: 2.8120e-05 - val_loss: 3.2857e-05 - val_mean_absolute_error: 0.0037 - val_mean_squared_error: 3.2857e-05\n",
      "Epoch 210/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 3.0812e-05 - mean_absolute_error: 0.0039 - mean_squared_error: 3.0812e-05 - val_loss: 3.4001e-05 - val_mean_absolute_error: 0.0044 - val_mean_squared_error: 3.4001e-05\n",
      "Epoch 211/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.5681e-05 - mean_absolute_error: 0.0043 - mean_squared_error: 3.5681e-05 - val_loss: 3.3449e-05 - val_mean_absolute_error: 0.0044 - val_mean_squared_error: 3.3449e-05\n",
      "Epoch 212/300\n",
      "3600/3600 [==============================] - 0s 105us/sample - loss: 3.7963e-05 - mean_absolute_error: 0.0044 - mean_squared_error: 3.7963e-05 - val_loss: 2.9219e-05 - val_mean_absolute_error: 0.0037 - val_mean_squared_error: 2.9219e-05\n",
      "Epoch 213/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 3.6972e-05 - mean_absolute_error: 0.0043 - mean_squared_error: 3.6972e-05 - val_loss: 2.4318e-05 - val_mean_absolute_error: 0.0032 - val_mean_squared_error: 2.4318e-05\n",
      "Epoch 214/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 2.7851e-05 - mean_absolute_error: 0.0036 - mean_squared_error: 2.7851e-05 - val_loss: 2.6497e-05 - val_mean_absolute_error: 0.0034 - val_mean_squared_error: 2.6497e-05\n",
      "Epoch 215/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.4359e-05 - mean_absolute_error: 0.0042 - mean_squared_error: 3.4359e-05 - val_loss: 3.8312e-05 - val_mean_absolute_error: 0.0046 - val_mean_squared_error: 3.8312e-05\n",
      "Epoch 216/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.0106e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 3.0106e-05 - val_loss: 3.4715e-05 - val_mean_absolute_error: 0.0044 - val_mean_squared_error: 3.4715e-05\n",
      "Epoch 217/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.0157e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 3.0157e-05 - val_loss: 2.2033e-05 - val_mean_absolute_error: 0.0029 - val_mean_squared_error: 2.2033e-05\n",
      "Epoch 218/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 3.6898e-05 - mean_absolute_error: 0.0044 - mean_squared_error: 3.6898e-05 - val_loss: 3.1270e-05 - val_mean_absolute_error: 0.0040 - val_mean_squared_error: 3.1270e-05\n",
      "Epoch 219/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.0529e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 3.0529e-05 - val_loss: 2.7071e-05 - val_mean_absolute_error: 0.0037 - val_mean_squared_error: 2.7071e-05\n",
      "Epoch 220/300\n",
      "3600/3600 [==============================] - 0s 106us/sample - loss: 3.5495e-05 - mean_absolute_error: 0.0042 - mean_squared_error: 3.5495e-05 - val_loss: 2.7460e-05 - val_mean_absolute_error: 0.0037 - val_mean_squared_error: 2.7460e-05\n",
      "Epoch 221/300\n",
      "3600/3600 [==============================] - 0s 107us/sample - loss: 3.4802e-05 - mean_absolute_error: 0.0041 - mean_squared_error: 3.4802e-05 - val_loss: 3.5965e-05 - val_mean_absolute_error: 0.0041 - val_mean_squared_error: 3.5965e-05\n",
      "Epoch 222/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 3.4085e-05 - mean_absolute_error: 0.0041 - mean_squared_error: 3.4085e-05 - val_loss: 2.7089e-05 - val_mean_absolute_error: 0.0036 - val_mean_squared_error: 2.7089e-05\n",
      "Epoch 223/300\n",
      "3600/3600 [==============================] - 0s 128us/sample - loss: 2.7783e-05 - mean_absolute_error: 0.0036 - mean_squared_error: 2.7783e-05 - val_loss: 3.2844e-05 - val_mean_absolute_error: 0.0042 - val_mean_squared_error: 3.2844e-05\n",
      "Epoch 224/300\n",
      "3600/3600 [==============================] - 0s 129us/sample - loss: 3.0522e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 3.0522e-05 - val_loss: 2.2431e-05 - val_mean_absolute_error: 0.0031 - val_mean_squared_error: 2.2431e-05\n",
      "Epoch 225/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 3.1756e-05 - mean_absolute_error: 0.0039 - mean_squared_error: 3.1756e-05 - val_loss: 2.9092e-05 - val_mean_absolute_error: 0.0034 - val_mean_squared_error: 2.9092e-05\n",
      "Epoch 226/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 2.8160e-05 - mean_absolute_error: 0.0036 - mean_squared_error: 2.8160e-05 - val_loss: 3.2898e-05 - val_mean_absolute_error: 0.0043 - val_mean_squared_error: 3.2898e-05\n",
      "Epoch 227/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 2.9984e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 2.9984e-05 - val_loss: 4.2114e-05 - val_mean_absolute_error: 0.0052 - val_mean_squared_error: 4.2114e-05\n",
      "Epoch 228/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 3.1364e-05 - mean_absolute_error: 0.0039 - mean_squared_error: 3.1364e-05 - val_loss: 2.8017e-05 - val_mean_absolute_error: 0.0037 - val_mean_squared_error: 2.8017e-05\n",
      "Epoch 229/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 2.9616e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 2.9616e-05 - val_loss: 3.1000e-05 - val_mean_absolute_error: 0.0042 - val_mean_squared_error: 3.1000e-05\n",
      "Epoch 230/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 3.1692e-05 - mean_absolute_error: 0.0039 - mean_squared_error: 3.1692e-05 - val_loss: 3.2052e-05 - val_mean_absolute_error: 0.0040 - val_mean_squared_error: 3.2052e-05\n",
      "Epoch 231/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.8705e-05 - mean_absolute_error: 0.0044 - mean_squared_error: 3.8705e-05 - val_loss: 2.3713e-05 - val_mean_absolute_error: 0.0030 - val_mean_squared_error: 2.3713e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 232/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.1580e-05 - mean_absolute_error: 0.0039 - mean_squared_error: 3.1580e-05 - val_loss: 2.8801e-05 - val_mean_absolute_error: 0.0038 - val_mean_squared_error: 2.8801e-05\n",
      "Epoch 233/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 2.9053e-05 - mean_absolute_error: 0.0037 - mean_squared_error: 2.9053e-05 - val_loss: 2.4233e-05 - val_mean_absolute_error: 0.0034 - val_mean_squared_error: 2.4233e-05\n",
      "Epoch 234/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.1580e-05 - mean_absolute_error: 0.0039 - mean_squared_error: 3.1580e-05 - val_loss: 2.7065e-05 - val_mean_absolute_error: 0.0037 - val_mean_squared_error: 2.7065e-05\n",
      "Epoch 235/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 2.8078e-05 - mean_absolute_error: 0.0036 - mean_squared_error: 2.8078e-05 - val_loss: 2.6629e-05 - val_mean_absolute_error: 0.0036 - val_mean_squared_error: 2.6629e-05\n",
      "Epoch 236/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 2.8090e-05 - mean_absolute_error: 0.0036 - mean_squared_error: 2.8090e-05 - val_loss: 3.2463e-05 - val_mean_absolute_error: 0.0035 - val_mean_squared_error: 3.2463e-05\n",
      "Epoch 237/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 2.7093e-05 - mean_absolute_error: 0.0035 - mean_squared_error: 2.7093e-05 - val_loss: 2.2649e-05 - val_mean_absolute_error: 0.0030 - val_mean_squared_error: 2.2649e-05\n",
      "Epoch 238/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.6089e-05 - mean_absolute_error: 0.0043 - mean_squared_error: 3.6089e-05 - val_loss: 4.6340e-05 - val_mean_absolute_error: 0.0045 - val_mean_squared_error: 4.6340e-05\n",
      "Epoch 239/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 2.8003e-05 - mean_absolute_error: 0.0036 - mean_squared_error: 2.8003e-05 - val_loss: 2.7666e-05 - val_mean_absolute_error: 0.0037 - val_mean_squared_error: 2.7666e-05\n",
      "Epoch 240/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.0273e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 3.0273e-05 - val_loss: 3.0666e-05 - val_mean_absolute_error: 0.0035 - val_mean_squared_error: 3.0666e-05\n",
      "Epoch 241/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 2.7735e-05 - mean_absolute_error: 0.0036 - mean_squared_error: 2.7735e-05 - val_loss: 2.4312e-05 - val_mean_absolute_error: 0.0033 - val_mean_squared_error: 2.4312e-05\n",
      "Epoch 242/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.0194e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 3.0194e-05 - val_loss: 4.3023e-05 - val_mean_absolute_error: 0.0051 - val_mean_squared_error: 4.3023e-05\n",
      "Epoch 243/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 3.0300e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 3.0300e-05 - val_loss: 5.9353e-05 - val_mean_absolute_error: 0.0059 - val_mean_squared_error: 5.9353e-05\n",
      "Epoch 244/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 3.4344e-05 - mean_absolute_error: 0.0041 - mean_squared_error: 3.4344e-05 - val_loss: 5.4634e-05 - val_mean_absolute_error: 0.0054 - val_mean_squared_error: 5.4634e-05\n",
      "Epoch 245/300\n",
      "3600/3600 [==============================] - 0s 106us/sample - loss: 3.2253e-05 - mean_absolute_error: 0.0040 - mean_squared_error: 3.2253e-05 - val_loss: 2.4563e-05 - val_mean_absolute_error: 0.0033 - val_mean_squared_error: 2.4563e-05\n",
      "Epoch 246/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 3.0668e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 3.0668e-05 - val_loss: 3.4742e-05 - val_mean_absolute_error: 0.0039 - val_mean_squared_error: 3.4742e-05\n",
      "Epoch 247/300\n",
      "3600/3600 [==============================] - 0s 124us/sample - loss: 2.9766e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 2.9766e-05 - val_loss: 7.0010e-05 - val_mean_absolute_error: 0.0061 - val_mean_squared_error: 7.0010e-05\n",
      "Epoch 248/300\n",
      "3600/3600 [==============================] - 0s 118us/sample - loss: 3.3381e-05 - mean_absolute_error: 0.0041 - mean_squared_error: 3.3381e-05 - val_loss: 3.1246e-05 - val_mean_absolute_error: 0.0040 - val_mean_squared_error: 3.1246e-05\n",
      "Epoch 249/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 2.9124e-05 - mean_absolute_error: 0.0037 - mean_squared_error: 2.9124e-05 - val_loss: 2.7559e-05 - val_mean_absolute_error: 0.0036 - val_mean_squared_error: 2.7559e-05\n",
      "Epoch 250/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 2.6231e-05 - mean_absolute_error: 0.0034 - mean_squared_error: 2.6231e-05 - val_loss: 6.2582e-05 - val_mean_absolute_error: 0.0048 - val_mean_squared_error: 6.2582e-05\n",
      "Epoch 251/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 3.1867e-05 - mean_absolute_error: 0.0039 - mean_squared_error: 3.1867e-05 - val_loss: 5.9642e-05 - val_mean_absolute_error: 0.0063 - val_mean_squared_error: 5.9642e-05\n",
      "Epoch 252/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.0993e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 3.0993e-05 - val_loss: 2.9847e-05 - val_mean_absolute_error: 0.0040 - val_mean_squared_error: 2.9847e-05\n",
      "Epoch 253/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 2.9551e-05 - mean_absolute_error: 0.0037 - mean_squared_error: 2.9551e-05 - val_loss: 2.2340e-05 - val_mean_absolute_error: 0.0031 - val_mean_squared_error: 2.2340e-05\n",
      "Epoch 254/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 3.3809e-05 - mean_absolute_error: 0.0040 - mean_squared_error: 3.3809e-05 - val_loss: 2.1999e-05 - val_mean_absolute_error: 0.0030 - val_mean_squared_error: 2.1999e-05\n",
      "Epoch 255/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 2.8188e-05 - mean_absolute_error: 0.0036 - mean_squared_error: 2.8188e-05 - val_loss: 4.0172e-05 - val_mean_absolute_error: 0.0047 - val_mean_squared_error: 4.0172e-05\n",
      "Epoch 256/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 3.0491e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 3.0491e-05 - val_loss: 3.2198e-05 - val_mean_absolute_error: 0.0043 - val_mean_squared_error: 3.2198e-05\n",
      "Epoch 257/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 2.9319e-05 - mean_absolute_error: 0.0037 - mean_squared_error: 2.9319e-05 - val_loss: 2.6162e-05 - val_mean_absolute_error: 0.0035 - val_mean_squared_error: 2.6162e-05\n",
      "Epoch 258/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 2.8727e-05 - mean_absolute_error: 0.0037 - mean_squared_error: 2.8727e-05 - val_loss: 2.4985e-05 - val_mean_absolute_error: 0.0033 - val_mean_squared_error: 2.4985e-05\n",
      "Epoch 259/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 2.8096e-05 - mean_absolute_error: 0.0037 - mean_squared_error: 2.8096e-05 - val_loss: 3.0301e-05 - val_mean_absolute_error: 0.0034 - val_mean_squared_error: 3.0301e-05\n",
      "Epoch 260/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 3.2348e-05 - mean_absolute_error: 0.0040 - mean_squared_error: 3.2348e-05 - val_loss: 3.1836e-05 - val_mean_absolute_error: 0.0040 - val_mean_squared_error: 3.1836e-05\n",
      "Epoch 261/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 3.0633e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 3.0633e-05 - val_loss: 5.3823e-05 - val_mean_absolute_error: 0.0051 - val_mean_squared_error: 5.3823e-05\n",
      "Epoch 262/300\n",
      "3600/3600 [==============================] - 0s 107us/sample - loss: 2.8820e-05 - mean_absolute_error: 0.0037 - mean_squared_error: 2.8820e-05 - val_loss: 2.5992e-05 - val_mean_absolute_error: 0.0036 - val_mean_squared_error: 2.5992e-05\n",
      "Epoch 263/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 2.7401e-05 - mean_absolute_error: 0.0035 - mean_squared_error: 2.7401e-05 - val_loss: 2.2788e-05 - val_mean_absolute_error: 0.0032 - val_mean_squared_error: 2.2788e-05\n",
      "Epoch 264/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.1669e-05 - mean_absolute_error: 0.0039 - mean_squared_error: 3.1669e-05 - val_loss: 4.9739e-05 - val_mean_absolute_error: 0.0051 - val_mean_squared_error: 4.9739e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 265/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 2.6357e-05 - mean_absolute_error: 0.0035 - mean_squared_error: 2.6357e-05 - val_loss: 2.1127e-05 - val_mean_absolute_error: 0.0029 - val_mean_squared_error: 2.1127e-05\n",
      "Epoch 266/300\n",
      "3600/3600 [==============================] - 0s 113us/sample - loss: 2.7107e-05 - mean_absolute_error: 0.0035 - mean_squared_error: 2.7107e-05 - val_loss: 3.0496e-05 - val_mean_absolute_error: 0.0039 - val_mean_squared_error: 3.0496e-05\n",
      "Epoch 267/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.4022e-05 - mean_absolute_error: 0.0041 - mean_squared_error: 3.4022e-05 - val_loss: 3.0062e-05 - val_mean_absolute_error: 0.0034 - val_mean_squared_error: 3.0062e-05\n",
      "Epoch 268/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 2.8231e-05 - mean_absolute_error: 0.0037 - mean_squared_error: 2.8231e-05 - val_loss: 2.3967e-05 - val_mean_absolute_error: 0.0030 - val_mean_squared_error: 2.3967e-05\n",
      "Epoch 269/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 2.5160e-05 - mean_absolute_error: 0.0033 - mean_squared_error: 2.5160e-05 - val_loss: 2.3260e-05 - val_mean_absolute_error: 0.0031 - val_mean_squared_error: 2.3260e-05\n",
      "Epoch 270/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.1971e-05 - mean_absolute_error: 0.0039 - mean_squared_error: 3.1971e-05 - val_loss: 2.0879e-05 - val_mean_absolute_error: 0.0029 - val_mean_squared_error: 2.0879e-05\n",
      "Epoch 271/300\n",
      "3600/3600 [==============================] - 0s 107us/sample - loss: 2.9361e-05 - mean_absolute_error: 0.0037 - mean_squared_error: 2.9361e-05 - val_loss: 2.4666e-05 - val_mean_absolute_error: 0.0034 - val_mean_squared_error: 2.4666e-05\n",
      "Epoch 272/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 3.0408e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 3.0408e-05 - val_loss: 2.7924e-05 - val_mean_absolute_error: 0.0035 - val_mean_squared_error: 2.7924e-05\n",
      "Epoch 273/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 2.6622e-05 - mean_absolute_error: 0.0035 - mean_squared_error: 2.6622e-05 - val_loss: 5.8085e-05 - val_mean_absolute_error: 0.0052 - val_mean_squared_error: 5.8085e-05\n",
      "Epoch 274/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 3.1465e-05 - mean_absolute_error: 0.0039 - mean_squared_error: 3.1465e-05 - val_loss: 2.9343e-05 - val_mean_absolute_error: 0.0040 - val_mean_squared_error: 2.9343e-05\n",
      "Epoch 275/300\n",
      "3600/3600 [==============================] - 0s 107us/sample - loss: 2.8400e-05 - mean_absolute_error: 0.0036 - mean_squared_error: 2.8400e-05 - val_loss: 2.5465e-05 - val_mean_absolute_error: 0.0033 - val_mean_squared_error: 2.5465e-05\n",
      "Epoch 276/300\n",
      "3600/3600 [==============================] - 0s 107us/sample - loss: 2.8429e-05 - mean_absolute_error: 0.0036 - mean_squared_error: 2.8429e-05 - val_loss: 3.9621e-05 - val_mean_absolute_error: 0.0048 - val_mean_squared_error: 3.9621e-05\n",
      "Epoch 277/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 2.7764e-05 - mean_absolute_error: 0.0036 - mean_squared_error: 2.7764e-05 - val_loss: 4.4339e-05 - val_mean_absolute_error: 0.0049 - val_mean_squared_error: 4.4339e-05\n",
      "Epoch 278/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 2.7184e-05 - mean_absolute_error: 0.0036 - mean_squared_error: 2.7184e-05 - val_loss: 2.1567e-05 - val_mean_absolute_error: 0.0030 - val_mean_squared_error: 2.1567e-05\n",
      "Epoch 279/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 2.9297e-05 - mean_absolute_error: 0.0037 - mean_squared_error: 2.9297e-05 - val_loss: 2.5145e-05 - val_mean_absolute_error: 0.0035 - val_mean_squared_error: 2.5145e-05\n",
      "Epoch 280/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 3.0025e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 3.0025e-05 - val_loss: 3.9531e-05 - val_mean_absolute_error: 0.0041 - val_mean_squared_error: 3.9531e-05\n",
      "Epoch 281/300\n",
      "3600/3600 [==============================] - 0s 119us/sample - loss: 3.1625e-05 - mean_absolute_error: 0.0039 - mean_squared_error: 3.1625e-05 - val_loss: 2.0712e-05 - val_mean_absolute_error: 0.0028 - val_mean_squared_error: 2.0712e-05\n",
      "Epoch 282/300\n",
      "3600/3600 [==============================] - 0s 127us/sample - loss: 2.5645e-05 - mean_absolute_error: 0.0034 - mean_squared_error: 2.5645e-05 - val_loss: 3.2495e-05 - val_mean_absolute_error: 0.0043 - val_mean_squared_error: 3.2495e-05\n",
      "Epoch 283/300\n",
      "3600/3600 [==============================] - 0s 113us/sample - loss: 2.4889e-05 - mean_absolute_error: 0.0033 - mean_squared_error: 2.4889e-05 - val_loss: 2.3989e-05 - val_mean_absolute_error: 0.0030 - val_mean_squared_error: 2.3989e-05\n",
      "Epoch 284/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 2.7919e-05 - mean_absolute_error: 0.0035 - mean_squared_error: 2.7919e-05 - val_loss: 2.7473e-05 - val_mean_absolute_error: 0.0036 - val_mean_squared_error: 2.7473e-05\n",
      "Epoch 285/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 3.1949e-05 - mean_absolute_error: 0.0039 - mean_squared_error: 3.1949e-05 - val_loss: 2.3180e-05 - val_mean_absolute_error: 0.0033 - val_mean_squared_error: 2.3180e-05\n",
      "Epoch 286/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 2.7255e-05 - mean_absolute_error: 0.0035 - mean_squared_error: 2.7255e-05 - val_loss: 4.9604e-05 - val_mean_absolute_error: 0.0050 - val_mean_squared_error: 4.9604e-05\n",
      "Epoch 287/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 3.1126e-05 - mean_absolute_error: 0.0038 - mean_squared_error: 3.1126e-05 - val_loss: 2.8061e-05 - val_mean_absolute_error: 0.0032 - val_mean_squared_error: 2.8061e-05\n",
      "Epoch 288/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 2.4654e-05 - mean_absolute_error: 0.0033 - mean_squared_error: 2.4654e-05 - val_loss: 2.3300e-05 - val_mean_absolute_error: 0.0029 - val_mean_squared_error: 2.3300e-05\n",
      "Epoch 289/300\n",
      "3600/3600 [==============================] - 0s 112us/sample - loss: 2.6670e-05 - mean_absolute_error: 0.0034 - mean_squared_error: 2.6670e-05 - val_loss: 2.1700e-05 - val_mean_absolute_error: 0.0029 - val_mean_squared_error: 2.1700e-05\n",
      "Epoch 290/300\n",
      "3600/3600 [==============================] - 0s 107us/sample - loss: 2.8237e-05 - mean_absolute_error: 0.0036 - mean_squared_error: 2.8237e-05 - val_loss: 2.0697e-05 - val_mean_absolute_error: 0.0029 - val_mean_squared_error: 2.0697e-05\n",
      "Epoch 291/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 2.8039e-05 - mean_absolute_error: 0.0036 - mean_squared_error: 2.8039e-05 - val_loss: 3.1453e-05 - val_mean_absolute_error: 0.0043 - val_mean_squared_error: 3.1453e-05\n",
      "Epoch 292/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 2.6870e-05 - mean_absolute_error: 0.0035 - mean_squared_error: 2.6870e-05 - val_loss: 4.7706e-05 - val_mean_absolute_error: 0.0043 - val_mean_squared_error: 4.7706e-05\n",
      "Epoch 293/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 2.7030e-05 - mean_absolute_error: 0.0035 - mean_squared_error: 2.7030e-05 - val_loss: 5.0801e-05 - val_mean_absolute_error: 0.0042 - val_mean_squared_error: 5.0801e-05\n",
      "Epoch 294/300\n",
      "3600/3600 [==============================] - 0s 110us/sample - loss: 2.6762e-05 - mean_absolute_error: 0.0035 - mean_squared_error: 2.6762e-05 - val_loss: 2.7678e-05 - val_mean_absolute_error: 0.0032 - val_mean_squared_error: 2.7678e-05\n",
      "Epoch 295/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 2.8109e-05 - mean_absolute_error: 0.0036 - mean_squared_error: 2.8109e-05 - val_loss: 4.1535e-05 - val_mean_absolute_error: 0.0042 - val_mean_squared_error: 4.1535e-05\n",
      "Epoch 296/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 2.8590e-05 - mean_absolute_error: 0.0036 - mean_squared_error: 2.8590e-05 - val_loss: 2.5012e-05 - val_mean_absolute_error: 0.0031 - val_mean_squared_error: 2.5012e-05\n",
      "Epoch 297/300\n",
      "3600/3600 [==============================] - 0s 111us/sample - loss: 2.5676e-05 - mean_absolute_error: 0.0034 - mean_squared_error: 2.5676e-05 - val_loss: 4.1193e-05 - val_mean_absolute_error: 0.0053 - val_mean_squared_error: 4.1193e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 298/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 2.8271e-05 - mean_absolute_error: 0.0036 - mean_squared_error: 2.8271e-05 - val_loss: 2.8076e-05 - val_mean_absolute_error: 0.0034 - val_mean_squared_error: 2.8076e-05\n",
      "Epoch 299/300\n",
      "3600/3600 [==============================] - 0s 109us/sample - loss: 2.7757e-05 - mean_absolute_error: 0.0035 - mean_squared_error: 2.7757e-05 - val_loss: 2.9865e-05 - val_mean_absolute_error: 0.0035 - val_mean_squared_error: 2.9865e-05\n",
      "Epoch 300/300\n",
      "3600/3600 [==============================] - 0s 108us/sample - loss: 2.6982e-05 - mean_absolute_error: 0.0035 - mean_squared_error: 2.6982e-05 - val_loss: 2.5358e-05 - val_mean_absolute_error: 0.0035 - val_mean_squared_error: 2.5358e-05\n"
     ]
    }
   ],
   "source": [
    "history = mod.fit(X_train, y_train, epochs=300,validation_split=0.2, verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 200,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = mod.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 201,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7fe84803b290>]"
      ]
     },
     "execution_count": 201,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(np.arange(y_data.shape[1]),y_test[10,:],'*')\n",
    "plt.plot(np.arange(y_data.shape[1]),y_pred[10])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 202,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<seaborn.axisgrid.JointGrid at 0x7fe82c7d52d0>"
      ]
     },
     "execution_count": 202,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.jointplot(y_test,y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}