001_2_PrepDataForBatchGenerator.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Callin Switzer\n",
    "## Prep data for batch-generating with Keras\n",
    "## 27 Aug 2018"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\calli\\Anaconda3\\envs\\deeplearning\\lib\\site-packages\\h5py\\__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  from ._conv import register_converters as _register_converters\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TensorFlow successfully installed.\n",
      "The installed version of TensorFlow includes GPU support.\n",
      "3.6.4 |Anaconda custom (64-bit)| (default, Jan 16 2018, 10:22:32) [MSC v.1900 64 bit (AMD64)] \n",
      "\n",
      "last run on 2018-08-27 14:47:22.995128\n"
     ]
    }
   ],
   "source": [
    "# import custom Utils functions\n",
    "from DeepLearningUtils import *\n",
    "\n",
    "\n",
    "# define directories\n",
    "baseDir = os.getcwd()\n",
    "dataDir = r'D:\\MothSimulations\\11c-AggressiveManeuver\\Qstore\\hws_am_con'\n",
    "figDir = r'D:\\Dropbox\\mothMachineLearning_dataAndFigs\\Figs'\n",
    "dataOutput = r'D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput'\n",
    "if not os.path.exists(dataOutput):\n",
    "    os.mkdir(dataOutput)\n",
    "    \n",
    "    \n",
    "# open file explorer\n",
    "# aa = r'explorer /select,\"'\n",
    "# subprocess.Popen(str(aa + figDir + '\\\"'))\n",
    "# subprocess.Popen(str(aa + dataDir + '\\\"'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_10_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_11_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_12_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_13_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_14_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_15_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_16_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_17_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_18_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_19_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_1_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_20_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_21_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_22_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_23_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_24_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_25_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_26_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_27_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_28_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_29_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_2_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_30_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_31_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_32_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_33_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_34_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_35_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_36_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_3_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_4_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_5_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_6_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_7_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_8_hws_am_con.hdf5',\n",
       " 'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_9_hws_am_con.hdf5']"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataDir2 = r\"D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\"\n",
    "dataFiles = listdir_fullpath(dataDir2)\n",
    "dataFiles"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'train': ['D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_24_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_26_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_14_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_18_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_27_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_8_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_25_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_2_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_32_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_10_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_35_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_36_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_30_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_34_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_4_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_7_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_29_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_19_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_16_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_23_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_12_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_9_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_13_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_21_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_33_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_5_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_15_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_11_hws_am_con.hdf5'],\n",
       " 'validation': ['D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_17_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_1_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_20_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_22_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_28_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_31_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_3_hws_am_con.hdf5',\n",
       "  'D:\\\\Dropbox\\\\mothMachineLearning_dataAndFigs\\\\DataOutput\\\\Qstore_6_hws_am_con.hdf5']}"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# define train and validation dataset\n",
    "np.random.seed(36513654)\n",
    "\n",
    "partition = {\"train\": np.random.choice(dataFiles, size = int(0.8*len(dataFiles)), replace = False).tolist()}\n",
    "partition[\"validation\"] = np.setdiff1d(dataFiles, partition[\"train\"]).tolist()\n",
    "partition"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_10_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_11_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_12_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_13_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_14_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_15_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_16_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_17_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_18_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_19_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_1_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_20_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_21_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_22_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_23_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_24_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_25_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_26_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_27_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_28_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_29_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_2_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_30_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_31_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_32_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_33_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_34_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_35_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_36_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_3_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_4_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_5_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_6_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_7_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_8_hws_am_con.hdf5\n",
      "D:\\Dropbox\\mothMachineLearning_dataAndFigs\\DataOutput\\Qstore_9_hws_am_con.hdf5\n"
     ]
    }
   ],
   "source": [
    "dfList = []\n",
    "\n",
    "for dataFile in dataFiles:\n",
    "    bf2 = pd.read_hdf(dataFile, \"data\")\n",
    "    bf2[['simNum', 'timestep']] = bf2[['simNum', 'timestep']].astype(int)\n",
    "    bf2['count'] = bf2.groupby([\"simNum\", \"timestep\"]).cumcount()\n",
    "    sdf = bf2.loc[(bf2[\"count\"] == 0) | (bf2[\"count\"] == 99), :]\n",
    "    ss = sdf.pivot_table(index=['simNum', 'timestep'], columns = \"count\",aggfunc='first')\n",
    "    ss.columns = [f'{i}_{j}' if j != '' else f'{i}' for i,j in ss.columns]\n",
    "    ss.reset_index(inplace=True) \n",
    "    dfList.append(ss)\n",
    "    print(dataFile)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "trainDF = pd.concat(dfList)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "trainDF = trainDF.drop(columns=['F_99', 'alpha_99', \n",
    "                               'tau_99'])\n",
    "trainDF=trainDF.rename(columns = {'F_0':'F', \n",
    "                                 \"alpha_0\": \"alpha\", \n",
    "                                 \"tau_0\": \"tau\"})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(9000000, 21)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>simNum</th>\n",
       "      <th>timestep</th>\n",
       "      <th>F</th>\n",
       "      <th>alpha</th>\n",
       "      <th>phi_0</th>\n",
       "      <th>phi_99</th>\n",
       "      <th>phi_dot_0</th>\n",
       "      <th>phi_dot_99</th>\n",
       "      <th>tau</th>\n",
       "      <th>theta_0</th>\n",
       "      <th>...</th>\n",
       "      <th>theta_dot_0</th>\n",
       "      <th>theta_dot_99</th>\n",
       "      <th>x_0</th>\n",
       "      <th>x_99</th>\n",
       "      <th>x_dot_0</th>\n",
       "      <th>x_dot_99</th>\n",
       "      <th>y_0</th>\n",
       "      <th>y_99</th>\n",
       "      <th>y_dot_0</th>\n",
       "      <th>y_dot_99</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>38304.433962</td>\n",
       "      <td>5.340270</td>\n",
       "      <td>3.926991</td>\n",
       "      <td>3.869604</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-5.674295</td>\n",
       "      <td>10963.268558</td>\n",
       "      <td>0.785398</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-4.770986</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>6.853314</td>\n",
       "      <td>0.000100</td>\n",
       "      <td>683.734561</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-1.298536</td>\n",
       "      <td>0.000100</td>\n",
       "      <td>-133.752363</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>34264.536249</td>\n",
       "      <td>2.465501</td>\n",
       "      <td>3.948428</td>\n",
       "      <td>4.019544</td>\n",
       "      <td>2.009397</td>\n",
       "      <td>5.425513</td>\n",
       "      <td>72580.767201</td>\n",
       "      <td>0.768478</td>\n",
       "      <td>...</td>\n",
       "      <td>0.087406</td>\n",
       "      <td>10.479995</td>\n",
       "      <td>-0.019322</td>\n",
       "      <td>-6.079672</td>\n",
       "      <td>-0.145396</td>\n",
       "      <td>-606.129200</td>\n",
       "      <td>2.520087</td>\n",
       "      <td>6.366770</td>\n",
       "      <td>250.908304</td>\n",
       "      <td>120.273149</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>12045.791872</td>\n",
       "      <td>4.573819</td>\n",
       "      <td>3.966608</td>\n",
       "      <td>3.933017</td>\n",
       "      <td>-0.124272</td>\n",
       "      <td>-2.775517</td>\n",
       "      <td>93435.855346</td>\n",
       "      <td>0.768515</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.029996</td>\n",
       "      <td>3.897680</td>\n",
       "      <td>0.509507</td>\n",
       "      <td>2.986736</td>\n",
       "      <td>52.490438</td>\n",
       "      <td>193.914740</td>\n",
       "      <td>4.834518</td>\n",
       "      <td>2.470847</td>\n",
       "      <td>-18.848146</td>\n",
       "      <td>-211.050011</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>35562.854879</td>\n",
       "      <td>0.767089</td>\n",
       "      <td>4.000007</td>\n",
       "      <td>4.111777</td>\n",
       "      <td>3.434103</td>\n",
       "      <td>7.246049</td>\n",
       "      <td>-96088.338473</td>\n",
       "      <td>0.773977</td>\n",
       "      <td>...</td>\n",
       "      <td>2.037211</td>\n",
       "      <td>0.318982</td>\n",
       "      <td>-0.836633</td>\n",
       "      <td>-4.334782</td>\n",
       "      <td>-187.046694</td>\n",
       "      <td>-157.064454</td>\n",
       "      <td>6.838940</td>\n",
       "      <td>17.508792</td>\n",
       "      <td>218.755312</td>\n",
       "      <td>844.589134</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>20856.636628</td>\n",
       "      <td>2.424378</td>\n",
       "      <td>4.009324</td>\n",
       "      <td>3.995061</td>\n",
       "      <td>-2.309667</td>\n",
       "      <td>0.686804</td>\n",
       "      <td>-27142.805375</td>\n",
       "      <td>0.815904</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.673328</td>\n",
       "      <td>-1.304712</td>\n",
       "      <td>0.429770</td>\n",
       "      <td>2.895126</td>\n",
       "      <td>313.730247</td>\n",
       "      <td>-65.350386</td>\n",
       "      <td>8.392823</td>\n",
       "      <td>6.771320</td>\n",
       "      <td>-58.222983</td>\n",
       "      <td>-102.458917</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   simNum  timestep             F     alpha     phi_0    phi_99  phi_dot_0  \\\n",
       "0       0         0  38304.433962  5.340270  3.926991  3.869604   0.000000   \n",
       "1       0         1  34264.536249  2.465501  3.948428  4.019544   2.009397   \n",
       "2       0         2  12045.791872  4.573819  3.966608  3.933017  -0.124272   \n",
       "3       0         3  35562.854879  0.767089  4.000007  4.111777   3.434103   \n",
       "4       0         4  20856.636628  2.424378  4.009324  3.995061  -2.309667   \n",
       "\n",
       "   phi_dot_99           tau   theta_0     ...      theta_dot_0  theta_dot_99  \\\n",
       "0   -5.674295  10963.268558  0.785398     ...         0.000000     -4.770986   \n",
       "1    5.425513  72580.767201  0.768478     ...         0.087406     10.479995   \n",
       "2   -2.775517  93435.855346  0.768515     ...        -1.029996      3.897680   \n",
       "3    7.246049 -96088.338473  0.773977     ...         2.037211      0.318982   \n",
       "4    0.686804 -27142.805375  0.815904     ...        -0.673328     -1.304712   \n",
       "\n",
       "        x_0      x_99     x_dot_0    x_dot_99       y_0       y_99  \\\n",
       "0  0.000000  6.853314    0.000100  683.734561  0.000000  -1.298536   \n",
       "1 -0.019322 -6.079672   -0.145396 -606.129200  2.520087   6.366770   \n",
       "2  0.509507  2.986736   52.490438  193.914740  4.834518   2.470847   \n",
       "3 -0.836633 -4.334782 -187.046694 -157.064454  6.838940  17.508792   \n",
       "4  0.429770  2.895126  313.730247  -65.350386  8.392823   6.771320   \n",
       "\n",
       "      y_dot_0    y_dot_99  \n",
       "0    0.000100 -133.752363  \n",
       "1  250.908304  120.273149  \n",
       "2  -18.848146 -211.050011  \n",
       "3  218.755312  844.589134  \n",
       "4  -58.222983 -102.458917  \n",
       "\n",
       "[5 rows x 21 columns]"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trainDF.reset_index(inplace = True, drop = True)\n",
    "print(trainDF.shape)\n",
    "trainDF.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>simNum</th>\n",
       "      <th>timestep</th>\n",
       "      <th>F</th>\n",
       "      <th>alpha</th>\n",
       "      <th>phi_0</th>\n",
       "      <th>phi_99</th>\n",
       "      <th>phi_dot_0</th>\n",
       "      <th>phi_dot_99</th>\n",
       "      <th>tau</th>\n",
       "      <th>theta_0</th>\n",
       "      <th>...</th>\n",
       "      <th>theta_dot_0</th>\n",
       "      <th>theta_dot_99</th>\n",
       "      <th>x_0</th>\n",
       "      <th>x_99</th>\n",
       "      <th>x_dot_0</th>\n",
       "      <th>x_dot_99</th>\n",
       "      <th>y_0</th>\n",
       "      <th>y_99</th>\n",
       "      <th>y_dot_0</th>\n",
       "      <th>y_dot_99</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>38304.433962</td>\n",
       "      <td>5.340270</td>\n",
       "      <td>3.926991</td>\n",
       "      <td>3.869604</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-5.674295</td>\n",
       "      <td>10963.268558</td>\n",
       "      <td>0.785398</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-4.770986</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.853314</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>683.734561</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.298536</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>-133.752363</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>250000</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>6143.354049</td>\n",
       "      <td>5.961418</td>\n",
       "      <td>3.926991</td>\n",
       "      <td>3.916364</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.645207</td>\n",
       "      <td>87574.132880</td>\n",
       "      <td>0.785398</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.559808</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.048256</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>100.376445</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.278338</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>33.038324</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>500000</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>39225.428283</td>\n",
       "      <td>2.578644</td>\n",
       "      <td>3.926991</td>\n",
       "      <td>3.965986</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.896474</td>\n",
       "      <td>-4419.593159</td>\n",
       "      <td>0.785398</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.493059</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-6.930938</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>-691.619443</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.807255</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>-184.465021</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>750000</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>19970.007383</td>\n",
       "      <td>3.039763</td>\n",
       "      <td>3.926991</td>\n",
       "      <td>3.937839</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.881673</td>\n",
       "      <td>-63747.079038</td>\n",
       "      <td>0.785398</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-3.654178</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-2.919590</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>-292.316692</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-2.360690</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>-234.655195</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1000000</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>10481.140921</td>\n",
       "      <td>3.949336</td>\n",
       "      <td>3.926991</td>\n",
       "      <td>3.905200</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.850964</td>\n",
       "      <td>82273.215370</td>\n",
       "      <td>0.785398</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.021020</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.167448</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>16.533703</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-2.151740</td>\n",
       "      <td>0.0001</td>\n",
       "      <td>-211.757325</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         simNum  timestep             F     alpha     phi_0    phi_99  \\\n",
       "0             0         0  38304.433962  5.340270  3.926991  3.869604   \n",
       "250000        0         0   6143.354049  5.961418  3.926991  3.916364   \n",
       "500000        0         0  39225.428283  2.578644  3.926991  3.965986   \n",
       "750000        0         0  19970.007383  3.039763  3.926991  3.937839   \n",
       "1000000       0         0  10481.140921  3.949336  3.926991  3.905200   \n",
       "\n",
       "         phi_dot_0  phi_dot_99           tau   theta_0     ...      \\\n",
       "0              0.0   -5.674295  10963.268558  0.785398     ...       \n",
       "250000         0.0   -0.645207  87574.132880  0.785398     ...       \n",
       "500000         0.0    3.896474  -4419.593159  0.785398     ...       \n",
       "750000         0.0    0.881673 -63747.079038  0.785398     ...       \n",
       "1000000        0.0   -1.850964  82273.215370  0.785398     ...       \n",
       "\n",
       "         theta_dot_0  theta_dot_99  x_0      x_99  x_dot_0    x_dot_99  y_0  \\\n",
       "0                0.0     -4.770986  0.0  6.853314   0.0001  683.734561  0.0   \n",
       "250000           0.0      5.559808  0.0  1.048256   0.0001  100.376445  0.0   \n",
       "500000           0.0      3.493059  0.0 -6.930938   0.0001 -691.619443  0.0   \n",
       "750000           0.0     -3.654178  0.0 -2.919590   0.0001 -292.316692  0.0   \n",
       "1000000          0.0      4.021020  0.0  0.167448   0.0001   16.533703  0.0   \n",
       "\n",
       "             y_99  y_dot_0    y_dot_99  \n",
       "0       -1.298536   0.0001 -133.752363  \n",
       "250000   0.278338   0.0001   33.038324  \n",
       "500000  -1.807255   0.0001 -184.465021  \n",
       "750000  -2.360690   0.0001 -234.655195  \n",
       "1000000 -2.151740   0.0001 -211.757325  \n",
       "\n",
       "[5 rows x 21 columns]"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trainDF.sort_values([\"timestep\", \"simNum\"], ascending = [True, True]).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['simNum', 'timestep', 'F', 'alpha', 'phi_0', 'phi_99', 'phi_dot_0',\n",
       "       'phi_dot_99', 'tau', 'theta_0', 'theta_99', 'theta_dot_0',\n",
       "       'theta_dot_99', 'x_0', 'x_99', 'x_dot_0', 'x_dot_99', 'y_0', 'y_99',\n",
       "       'y_dot_0', 'y_dot_99'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trainDF.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "# convert to fx and fy\n",
    "trainDF[\"Fx\"] = trainDF.F * np.cos(trainDF.alpha)\n",
    "trainDF[\"Fy\"] = trainDF.F * np.sin(trainDF.alpha)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Keras models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "# make dataset\n",
    "X = trainDF.loc[:, [\"x_0\", \"y_0\", \"phi_0\", \"theta_0\", \n",
    "                    \"x_99\", \"y_99\", \"phi_99\", \"theta_99\"]]\n",
    "Y = trainDF.loc[:, [\"Fx\", \"Fy\", \"tau\"] ]\n",
    "\n",
    "# # # subtract x_0 and y_0 from everything\n",
    "# X.x_99 = X.x_99 - X.x_0\n",
    "# X.y_99 = X.y_99 - X.y_0\n",
    "\n",
    "# X.x_0 = X.y_0 = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "Xtrain, Xtest, Ytrain, Ytest = train_test_split(X, Y, test_size=0.3, random_state = 123)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler  \n",
    "scalerX = StandardScaler()  \n",
    "scalerY = StandardScaler()  \n",
    "\n",
    "# Don't cheat - fit only on training data\n",
    "scalerX.fit(Xtrain)  \n",
    "scalerY.fit(Ytrain) \n",
    "\n",
    "Xtrain_scaled = scalerX.transform(Xtrain)  \n",
    "Ytrain_scaled = scalerY.transform(Ytrain)  \n",
    "\n",
    "# apply same transformation to test data\n",
    "Xtest_scaled = scalerX.transform(Xtest)\n",
    "Ytest_scaled = scalerY.transform(Ytest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "# from sklearn.preprocessing import MinMaxScaler  \n",
    "# scalerX = MinMaxScaler()  \n",
    "# scalerY = MinMaxScaler()  \n",
    "\n",
    "# # Don't cheat - fit only on training data\n",
    "# scalerX.fit(Xtrain)  \n",
    "# scalerY.fit(Ytrain) \n",
    "\n",
    "# Xtrain_scaled = scalerX.transform(Xtrain)  \n",
    "# Ytrain_scaled = scalerY.transform(Ytrain)  \n",
    "\n",
    "# # apply same transformation to test data\n",
    "# Xtest_scaled = scalerX.transform(Xtest)\n",
    "# Ytest_scaled = scalerY.transform(Ytest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "# build Keras model with functional API, instead of Sequential\n",
    "from keras.layers import Input, Dense\n",
    "from keras.models import Model\n",
    "\n",
    "# callcacks\n",
    "from keras.callbacks import EarlyStopping\n",
    "from keras.callbacks import ReduceLROnPlateau"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "inputs = Input(shape=(Xtrain_scaled.shape[1],))\n",
    "\n",
    "x = Dense(400, activation='tanh')(inputs)\n",
    "x = Dense(400, activation='tanh')(x)\n",
    "x = Dense(400, activation='tanh')(x)\n",
    "x = Dense(16, activation='tanh')(x)\n",
    "predictions = Dense(Ytrain_scaled.shape[1], activation='linear')(x)\n",
    "\n",
    "# This creates a model that includes\n",
    "# the Input layer and three Dense layers\n",
    "model = Model(inputs=inputs, outputs=predictions)\n",
    "model.compile(loss=\"mse\", optimizer=\"adam\", metrics = ['mse'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 4410000 samples, validate on 1890000 samples\n",
      "Epoch 1/1000\n",
      " - 9s - loss: 0.3027 - mean_squared_error: 0.3027 - val_loss: 0.2815 - val_mean_squared_error: 0.2815\n",
      "Epoch 2/1000\n",
      " - 6s - loss: 0.2757 - mean_squared_error: 0.2757 - val_loss: 0.2709 - val_mean_squared_error: 0.2709\n",
      "Epoch 3/1000\n",
      " - 6s - loss: 0.2628 - mean_squared_error: 0.2628 - val_loss: 0.2546 - val_mean_squared_error: 0.2546\n",
      "Epoch 4/1000\n",
      " - 6s - loss: 0.2417 - mean_squared_error: 0.2417 - val_loss: 0.2295 - val_mean_squared_error: 0.2295\n",
      "Epoch 5/1000\n",
      " - 6s - loss: 0.2121 - mean_squared_error: 0.2121 - val_loss: 0.1946 - val_mean_squared_error: 0.1946\n",
      "Epoch 6/1000\n",
      " - 6s - loss: 0.1734 - mean_squared_error: 0.1734 - val_loss: 0.1518 - val_mean_squared_error: 0.1518\n",
      "Epoch 7/1000\n",
      " - 6s - loss: 0.1325 - mean_squared_error: 0.1325 - val_loss: 0.1168 - val_mean_squared_error: 0.1168\n",
      "Epoch 8/1000\n",
      " - 6s - loss: 0.0967 - mean_squared_error: 0.0967 - val_loss: 0.0871 - val_mean_squared_error: 0.0871\n",
      "Epoch 9/1000\n",
      " - 6s - loss: 0.0702 - mean_squared_error: 0.0702 - val_loss: 0.0620 - val_mean_squared_error: 0.0620\n",
      "Epoch 10/1000\n",
      " - 6s - loss: 0.0518 - mean_squared_error: 0.0518 - val_loss: 0.0512 - val_mean_squared_error: 0.0512\n",
      "Epoch 11/1000\n",
      " - 6s - loss: 0.0388 - mean_squared_error: 0.0388 - val_loss: 0.0323 - val_mean_squared_error: 0.0323\n",
      "Epoch 12/1000\n",
      " - 6s - loss: 0.0299 - mean_squared_error: 0.0299 - val_loss: 0.0252 - val_mean_squared_error: 0.0252\n",
      "Epoch 13/1000\n",
      " - 6s - loss: 0.0237 - mean_squared_error: 0.0237 - val_loss: 0.0221 - val_mean_squared_error: 0.0221\n",
      "Epoch 14/1000\n",
      " - 6s - loss: 0.0193 - mean_squared_error: 0.0193 - val_loss: 0.0169 - val_mean_squared_error: 0.0169\n",
      "Epoch 15/1000\n",
      " - 6s - loss: 0.0162 - mean_squared_error: 0.0162 - val_loss: 0.0159 - val_mean_squared_error: 0.0159\n",
      "Epoch 16/1000\n",
      " - 6s - loss: 0.0140 - mean_squared_error: 0.0140 - val_loss: 0.0121 - val_mean_squared_error: 0.0121\n",
      "Epoch 17/1000\n",
      " - 6s - loss: 0.0125 - mean_squared_error: 0.0125 - val_loss: 0.0251 - val_mean_squared_error: 0.0251\n",
      "Epoch 18/1000\n",
      " - 6s - loss: 0.0112 - mean_squared_error: 0.0112 - val_loss: 0.0110 - val_mean_squared_error: 0.0110\n",
      "Epoch 19/1000\n",
      " - 6s - loss: 0.0101 - mean_squared_error: 0.0101 - val_loss: 0.0085 - val_mean_squared_error: 0.0085\n",
      "Epoch 20/1000\n",
      " - 6s - loss: 0.0092 - mean_squared_error: 0.0092 - val_loss: 0.0066 - val_mean_squared_error: 0.0066\n",
      "Epoch 21/1000\n",
      " - 6s - loss: 0.0085 - mean_squared_error: 0.0085 - val_loss: 0.0069 - val_mean_squared_error: 0.0069\n",
      "Epoch 22/1000\n",
      " - 6s - loss: 0.0080 - mean_squared_error: 0.0080 - val_loss: 0.0084 - val_mean_squared_error: 0.0084\n",
      "Epoch 23/1000\n",
      " - 6s - loss: 0.0076 - mean_squared_error: 0.0076 - val_loss: 0.0053 - val_mean_squared_error: 0.0053\n",
      "Epoch 24/1000\n",
      " - 6s - loss: 0.0072 - mean_squared_error: 0.0072 - val_loss: 0.0042 - val_mean_squared_error: 0.0042\n",
      "Epoch 25/1000\n",
      " - 6s - loss: 0.0067 - mean_squared_error: 0.0067 - val_loss: 0.0091 - val_mean_squared_error: 0.0091\n",
      "Epoch 26/1000\n",
      " - 6s - loss: 0.0064 - mean_squared_error: 0.0064 - val_loss: 0.0089 - val_mean_squared_error: 0.0089\n",
      "Epoch 27/1000\n",
      " - 6s - loss: 0.0064 - mean_squared_error: 0.0064 - val_loss: 0.0042 - val_mean_squared_error: 0.0042\n",
      "Epoch 28/1000\n",
      " - 6s - loss: 0.0057 - mean_squared_error: 0.0057 - val_loss: 0.0091 - val_mean_squared_error: 0.0091\n",
      "Epoch 29/1000\n",
      " - 6s - loss: 0.0058 - mean_squared_error: 0.0058 - val_loss: 0.0035 - val_mean_squared_error: 0.0035\n",
      "Epoch 30/1000\n",
      " - 6s - loss: 0.0054 - mean_squared_error: 0.0054 - val_loss: 0.0047 - val_mean_squared_error: 0.0047\n",
      "Epoch 31/1000\n",
      " - 6s - loss: 0.0053 - mean_squared_error: 0.0053 - val_loss: 0.0038 - val_mean_squared_error: 0.0038\n",
      "Epoch 32/1000\n",
      " - 6s - loss: 0.0051 - mean_squared_error: 0.0051 - val_loss: 0.0033 - val_mean_squared_error: 0.0033\n",
      "Epoch 33/1000\n",
      " - 6s - loss: 0.0050 - mean_squared_error: 0.0050 - val_loss: 0.0062 - val_mean_squared_error: 0.0062\n",
      "Epoch 34/1000\n",
      " - 6s - loss: 0.0048 - mean_squared_error: 0.0048 - val_loss: 0.0078 - val_mean_squared_error: 0.0078\n",
      "Epoch 35/1000\n",
      " - 6s - loss: 0.0046 - mean_squared_error: 0.0046 - val_loss: 0.0045 - val_mean_squared_error: 0.0045\n",
      "Epoch 36/1000\n",
      " - 6s - loss: 0.0046 - mean_squared_error: 0.0046 - val_loss: 0.0039 - val_mean_squared_error: 0.0039\n",
      "Epoch 37/1000\n",
      " - 6s - loss: 0.0043 - mean_squared_error: 0.0043 - val_loss: 0.0027 - val_mean_squared_error: 0.0027\n",
      "Epoch 38/1000\n",
      " - 6s - loss: 0.0045 - mean_squared_error: 0.0045 - val_loss: 0.0048 - val_mean_squared_error: 0.0048\n",
      "Epoch 39/1000\n",
      " - 6s - loss: 0.0042 - mean_squared_error: 0.0042 - val_loss: 0.0036 - val_mean_squared_error: 0.0036\n",
      "Epoch 40/1000\n",
      " - 6s - loss: 0.0041 - mean_squared_error: 0.0041 - val_loss: 0.0057 - val_mean_squared_error: 0.0057\n",
      "Epoch 41/1000\n",
      " - 6s - loss: 0.0040 - mean_squared_error: 0.0040 - val_loss: 0.0053 - val_mean_squared_error: 0.0053\n",
      "Epoch 42/1000\n",
      " - 6s - loss: 0.0039 - mean_squared_error: 0.0039 - val_loss: 0.0029 - val_mean_squared_error: 0.0029\n",
      "Epoch 43/1000\n",
      " - 6s - loss: 0.0038 - mean_squared_error: 0.0038 - val_loss: 0.0029 - val_mean_squared_error: 0.0029\n",
      "Epoch 44/1000\n",
      " - 6s - loss: 0.0039 - mean_squared_error: 0.0039 - val_loss: 0.0019 - val_mean_squared_error: 0.0019\n",
      "Epoch 45/1000\n",
      " - 6s - loss: 0.0037 - mean_squared_error: 0.0037 - val_loss: 0.0052 - val_mean_squared_error: 0.0052\n",
      "Epoch 46/1000\n",
      " - 6s - loss: 0.0036 - mean_squared_error: 0.0036 - val_loss: 0.0021 - val_mean_squared_error: 0.0021\n",
      "Epoch 47/1000\n",
      " - 6s - loss: 0.0036 - mean_squared_error: 0.0036 - val_loss: 0.0019 - val_mean_squared_error: 0.0019\n",
      "Epoch 48/1000\n",
      " - 6s - loss: 0.0034 - mean_squared_error: 0.0034 - val_loss: 0.0027 - val_mean_squared_error: 0.0027\n",
      "Epoch 49/1000\n",
      " - 6s - loss: 0.0035 - mean_squared_error: 0.0035 - val_loss: 0.0020 - val_mean_squared_error: 0.0020\n",
      "Epoch 50/1000\n",
      " - 6s - loss: 0.0034 - mean_squared_error: 0.0034 - val_loss: 0.0028 - val_mean_squared_error: 0.0028\n",
      "Epoch 51/1000\n",
      " - 6s - loss: 0.0034 - mean_squared_error: 0.0034 - val_loss: 0.0024 - val_mean_squared_error: 0.0024\n",
      "Epoch 52/1000\n",
      " - 6s - loss: 0.0032 - mean_squared_error: 0.0032 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 53/1000\n",
      " - 6s - loss: 0.0032 - mean_squared_error: 0.0032 - val_loss: 0.0030 - val_mean_squared_error: 0.0030\n",
      "Epoch 54/1000\n",
      " - 6s - loss: 0.0032 - mean_squared_error: 0.0032 - val_loss: 0.0043 - val_mean_squared_error: 0.0043\n",
      "Epoch 55/1000\n",
      " - 6s - loss: 0.0031 - mean_squared_error: 0.0031 - val_loss: 0.0021 - val_mean_squared_error: 0.0021\n",
      "Epoch 56/1000\n",
      " - 6s - loss: 0.0030 - mean_squared_error: 0.0030 - val_loss: 0.0024 - val_mean_squared_error: 0.0024\n",
      "Epoch 57/1000\n",
      " - 6s - loss: 0.0030 - mean_squared_error: 0.0030 - val_loss: 0.0023 - val_mean_squared_error: 0.0023\n",
      "Epoch 58/1000\n",
      " - 6s - loss: 0.0030 - mean_squared_error: 0.0030 - val_loss: 0.0038 - val_mean_squared_error: 0.0038\n",
      "Epoch 59/1000\n",
      " - 6s - loss: 0.0029 - mean_squared_error: 0.0029 - val_loss: 0.0030 - val_mean_squared_error: 0.0030\n",
      "Epoch 60/1000\n",
      " - 6s - loss: 0.0029 - mean_squared_error: 0.0029 - val_loss: 0.0022 - val_mean_squared_error: 0.0022\n",
      "Epoch 61/1000\n",
      " - 6s - loss: 0.0028 - mean_squared_error: 0.0028 - val_loss: 0.0018 - val_mean_squared_error: 0.0018\n",
      "Epoch 62/1000\n",
      " - 6s - loss: 0.0028 - mean_squared_error: 0.0028 - val_loss: 0.0020 - val_mean_squared_error: 0.0020\n",
      "Epoch 63/1000\n",
      " - 6s - loss: 0.0027 - mean_squared_error: 0.0027 - val_loss: 0.0027 - val_mean_squared_error: 0.0027\n",
      "Epoch 64/1000\n",
      " - 6s - loss: 0.0027 - mean_squared_error: 0.0027 - val_loss: 0.0020 - val_mean_squared_error: 0.0020\n",
      "Epoch 65/1000\n",
      " - 6s - loss: 0.0027 - mean_squared_error: 0.0027 - val_loss: 0.0025 - val_mean_squared_error: 0.0025\n",
      "Epoch 66/1000\n",
      " - 6s - loss: 0.0027 - mean_squared_error: 0.0027 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 67/1000\n",
      " - 6s - loss: 0.0026 - mean_squared_error: 0.0026 - val_loss: 0.0023 - val_mean_squared_error: 0.0023\n",
      "Epoch 68/1000\n",
      " - 6s - loss: 0.0026 - mean_squared_error: 0.0026 - val_loss: 0.0037 - val_mean_squared_error: 0.0037\n",
      "Epoch 69/1000\n",
      " - 6s - loss: 0.0027 - mean_squared_error: 0.0027 - val_loss: 0.0025 - val_mean_squared_error: 0.0025\n",
      "Epoch 70/1000\n",
      " - 6s - loss: 0.0025 - mean_squared_error: 0.0025 - val_loss: 0.0024 - val_mean_squared_error: 0.0024\n",
      "Epoch 71/1000\n",
      " - 6s - loss: 0.0026 - mean_squared_error: 0.0026 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 72/1000\n",
      " - 6s - loss: 0.0025 - mean_squared_error: 0.0025 - val_loss: 0.0023 - val_mean_squared_error: 0.0023\n",
      "Epoch 73/1000\n",
      " - 6s - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0072 - val_mean_squared_error: 0.0072\n",
      "Epoch 74/1000\n",
      " - 6s - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0018 - val_mean_squared_error: 0.0018\n",
      "Epoch 75/1000\n",
      " - 6s - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0026 - val_mean_squared_error: 0.0026\n",
      "Epoch 76/1000\n",
      " - 6s - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0022 - val_mean_squared_error: 0.0022\n",
      "Epoch 77/1000\n",
      " - 6s - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0020 - val_mean_squared_error: 0.0020\n",
      "Epoch 78/1000\n",
      " - 6s - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0025 - val_mean_squared_error: 0.0025\n",
      "Epoch 79/1000\n",
      " - 6s - loss: 0.0023 - mean_squared_error: 0.0023 - val_loss: 0.0013 - val_mean_squared_error: 0.0013\n",
      "Epoch 80/1000\n",
      " - 6s - loss: 0.0024 - mean_squared_error: 0.0024 - val_loss: 0.0031 - val_mean_squared_error: 0.0031\n",
      "Epoch 81/1000\n",
      " - 6s - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0022 - val_mean_squared_error: 0.0022\n",
      "Epoch 82/1000\n",
      " - 6s - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 83/1000\n",
      " - 6s - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 84/1000\n",
      " - 6s - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0019 - val_mean_squared_error: 0.0019\n",
      "Epoch 85/1000\n",
      " - 6s - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0021 - val_mean_squared_error: 0.0021\n",
      "Epoch 86/1000\n",
      " - 6s - loss: 0.0022 - mean_squared_error: 0.0022 - val_loss: 0.0025 - val_mean_squared_error: 0.0025\n",
      "Epoch 87/1000\n",
      " - 6s - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0041 - val_mean_squared_error: 0.0041\n",
      "Epoch 88/1000\n",
      " - 6s - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0025 - val_mean_squared_error: 0.0025\n",
      "Epoch 89/1000\n",
      " - 6s - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0022 - val_mean_squared_error: 0.0022\n",
      "Epoch 90/1000\n",
      " - 6s - loss: 0.0020 - mean_squared_error: 0.0020 - val_loss: 0.0035 - val_mean_squared_error: 0.0035\n",
      "Epoch 91/1000\n",
      " - 6s - loss: 0.0020 - mean_squared_error: 0.0020 - val_loss: 0.0028 - val_mean_squared_error: 0.0028\n",
      "Epoch 92/1000\n",
      " - 6s - loss: 0.0021 - mean_squared_error: 0.0021 - val_loss: 0.0026 - val_mean_squared_error: 0.0026\n",
      "Epoch 93/1000\n",
      " - 6s - loss: 0.0020 - mean_squared_error: 0.0020 - val_loss: 0.0018 - val_mean_squared_error: 0.0018\n",
      "Epoch 94/1000\n",
      " - 6s - loss: 0.0020 - mean_squared_error: 0.0020 - val_loss: 0.0019 - val_mean_squared_error: 0.0019\n",
      "Epoch 95/1000\n",
      " - 6s - loss: 0.0020 - mean_squared_error: 0.0020 - val_loss: 0.0047 - val_mean_squared_error: 0.0047\n",
      "Epoch 96/1000\n",
      " - 6s - loss: 0.0019 - mean_squared_error: 0.0019 - val_loss: 0.0019 - val_mean_squared_error: 0.0019\n",
      "Epoch 97/1000\n",
      " - 6s - loss: 0.0020 - mean_squared_error: 0.0020 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 98/1000\n",
      " - 6s - loss: 0.0019 - mean_squared_error: 0.0019 - val_loss: 0.0018 - val_mean_squared_error: 0.0018\n",
      "Epoch 99/1000\n",
      " - 6s - loss: 0.0019 - mean_squared_error: 0.0019 - val_loss: 0.0023 - val_mean_squared_error: 0.0023\n",
      "Epoch 100/1000\n",
      " - 6s - loss: 0.0019 - mean_squared_error: 0.0019 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 101/1000\n",
      " - 6s - loss: 0.0019 - mean_squared_error: 0.0019 - val_loss: 0.0028 - val_mean_squared_error: 0.0028\n",
      "Epoch 102/1000\n",
      " - 6s - loss: 0.0019 - mean_squared_error: 0.0019 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 103/1000\n",
      " - 6s - loss: 0.0018 - mean_squared_error: 0.0018 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 104/1000\n",
      " - 6s - loss: 0.0018 - mean_squared_error: 0.0018 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 105/1000\n",
      " - 6s - loss: 0.0018 - mean_squared_error: 0.0018 - val_loss: 9.4052e-04 - val_mean_squared_error: 9.4052e-04\n",
      "Epoch 106/1000\n",
      " - 6s - loss: 0.0018 - mean_squared_error: 0.0018 - val_loss: 0.0012 - val_mean_squared_error: 0.0012\n",
      "Epoch 107/1000\n",
      " - 6s - loss: 0.0018 - mean_squared_error: 0.0018 - val_loss: 0.0013 - val_mean_squared_error: 0.0013\n",
      "Epoch 108/1000\n",
      " - 6s - loss: 0.0018 - mean_squared_error: 0.0018 - val_loss: 0.0013 - val_mean_squared_error: 0.0013\n",
      "Epoch 109/1000\n",
      " - 6s - loss: 0.0017 - mean_squared_error: 0.0017 - val_loss: 0.0013 - val_mean_squared_error: 0.0013\n",
      "Epoch 110/1000\n",
      " - 6s - loss: 0.0017 - mean_squared_error: 0.0017 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 111/1000\n",
      " - 6s - loss: 0.0017 - mean_squared_error: 0.0017 - val_loss: 0.0019 - val_mean_squared_error: 0.0019\n",
      "Epoch 112/1000\n",
      " - 6s - loss: 0.0017 - mean_squared_error: 0.0017 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 113/1000\n",
      " - 6s - loss: 0.0017 - mean_squared_error: 0.0017 - val_loss: 0.0023 - val_mean_squared_error: 0.0023\n",
      "Epoch 114/1000\n",
      " - 6s - loss: 0.0017 - mean_squared_error: 0.0017 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 115/1000\n",
      " - 6s - loss: 0.0016 - mean_squared_error: 0.0016 - val_loss: 0.0043 - val_mean_squared_error: 0.0043\n",
      "Epoch 116/1000\n",
      " - 6s - loss: 0.0017 - mean_squared_error: 0.0017 - val_loss: 0.0013 - val_mean_squared_error: 0.0013\n",
      "Epoch 117/1000\n",
      " - 6s - loss: 0.0017 - mean_squared_error: 0.0017 - val_loss: 0.0013 - val_mean_squared_error: 0.0013\n",
      "Epoch 118/1000\n",
      " - 6s - loss: 0.0016 - mean_squared_error: 0.0016 - val_loss: 0.0020 - val_mean_squared_error: 0.0020\n",
      "Epoch 119/1000\n",
      " - 6s - loss: 0.0016 - mean_squared_error: 0.0016 - val_loss: 0.0035 - val_mean_squared_error: 0.0035\n",
      "Epoch 120/1000\n",
      " - 6s - loss: 0.0016 - mean_squared_error: 0.0016 - val_loss: 0.0021 - val_mean_squared_error: 0.0021\n",
      "Epoch 121/1000\n",
      " - 6s - loss: 0.0016 - mean_squared_error: 0.0016 - val_loss: 0.0019 - val_mean_squared_error: 0.0019\n",
      "Epoch 122/1000\n",
      " - 6s - loss: 0.0016 - mean_squared_error: 0.0016 - val_loss: 0.0023 - val_mean_squared_error: 0.0023\n",
      "Epoch 123/1000\n",
      " - 6s - loss: 0.0016 - mean_squared_error: 0.0016 - val_loss: 0.0020 - val_mean_squared_error: 0.0020\n",
      "Epoch 124/1000\n",
      " - 6s - loss: 0.0016 - mean_squared_error: 0.0016 - val_loss: 0.0019 - val_mean_squared_error: 0.0019\n",
      "Epoch 125/1000\n",
      " - 6s - loss: 0.0015 - mean_squared_error: 0.0015 - val_loss: 0.0010 - val_mean_squared_error: 0.0010\n",
      "Epoch 126/1000\n",
      " - 6s - loss: 0.0016 - mean_squared_error: 0.0016 - val_loss: 9.5449e-04 - val_mean_squared_error: 9.5449e-04\n",
      "Epoch 127/1000\n",
      " - 6s - loss: 0.0015 - mean_squared_error: 0.0015 - val_loss: 0.0013 - val_mean_squared_error: 0.0013\n",
      "Epoch 128/1000\n",
      " - 6s - loss: 0.0015 - mean_squared_error: 0.0015 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 129/1000\n",
      " - 6s - loss: 0.0015 - mean_squared_error: 0.0015 - val_loss: 0.0013 - val_mean_squared_error: 0.0013\n",
      "Epoch 130/1000\n",
      " - 6s - loss: 0.0015 - mean_squared_error: 0.0015 - val_loss: 0.0018 - val_mean_squared_error: 0.0018\n",
      "Epoch 131/1000\n",
      " - 6s - loss: 0.0015 - mean_squared_error: 0.0015 - val_loss: 9.4679e-04 - val_mean_squared_error: 9.4679e-04\n",
      "Epoch 132/1000\n",
      " - 6s - loss: 0.0014 - mean_squared_error: 0.0014 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 133/1000\n",
      " - 6s - loss: 0.0015 - mean_squared_error: 0.0015 - val_loss: 0.0018 - val_mean_squared_error: 0.0018\n",
      "Epoch 134/1000\n",
      " - 6s - loss: 0.0015 - mean_squared_error: 0.0015 - val_loss: 8.4127e-04 - val_mean_squared_error: 8.4127e-04\n",
      "Epoch 135/1000\n",
      " - 6s - loss: 0.0014 - mean_squared_error: 0.0014 - val_loss: 0.0010 - val_mean_squared_error: 0.0010\n",
      "Epoch 136/1000\n",
      " - 6s - loss: 0.0015 - mean_squared_error: 0.0015 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 137/1000\n",
      " - 6s - loss: 0.0014 - mean_squared_error: 0.0014 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 138/1000\n",
      " - 6s - loss: 0.0014 - mean_squared_error: 0.0014 - val_loss: 0.0029 - val_mean_squared_error: 0.0029\n",
      "Epoch 139/1000\n",
      " - 6s - loss: 0.0014 - mean_squared_error: 0.0014 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 140/1000\n",
      " - 6s - loss: 0.0014 - mean_squared_error: 0.0014 - val_loss: 7.6953e-04 - val_mean_squared_error: 7.6953e-04\n",
      "Epoch 141/1000\n",
      " - 6s - loss: 0.0014 - mean_squared_error: 0.0014 - val_loss: 0.0012 - val_mean_squared_error: 0.0012\n",
      "Epoch 142/1000\n",
      " - 6s - loss: 0.0014 - mean_squared_error: 0.0014 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 143/1000\n",
      " - 6s - loss: 0.0014 - mean_squared_error: 0.0014 - val_loss: 0.0029 - val_mean_squared_error: 0.0029\n",
      "Epoch 144/1000\n",
      " - 6s - loss: 0.0014 - mean_squared_error: 0.0014 - val_loss: 0.0020 - val_mean_squared_error: 0.0020\n",
      "Epoch 145/1000\n",
      " - 6s - loss: 0.0014 - mean_squared_error: 0.0014 - val_loss: 0.0029 - val_mean_squared_error: 0.0029\n",
      "Epoch 146/1000\n",
      " - 6s - loss: 0.0014 - mean_squared_error: 0.0014 - val_loss: 0.0012 - val_mean_squared_error: 0.0012\n",
      "Epoch 147/1000\n",
      " - 6s - loss: 0.0014 - mean_squared_error: 0.0014 - val_loss: 0.0010 - val_mean_squared_error: 0.0010\n",
      "Epoch 148/1000\n",
      " - 6s - loss: 0.0013 - mean_squared_error: 0.0013 - val_loss: 9.4367e-04 - val_mean_squared_error: 9.4367e-04\n",
      "Epoch 149/1000\n",
      " - 6s - loss: 0.0013 - mean_squared_error: 0.0013 - val_loss: 0.0012 - val_mean_squared_error: 0.0012\n",
      "Epoch 150/1000\n",
      " - 6s - loss: 0.0014 - mean_squared_error: 0.0014 - val_loss: 9.6281e-04 - val_mean_squared_error: 9.6281e-04\n",
      "Epoch 151/1000\n",
      " - 6s - loss: 0.0013 - mean_squared_error: 0.0013 - val_loss: 0.0012 - val_mean_squared_error: 0.0012\n",
      "Epoch 152/1000\n",
      " - 6s - loss: 0.0013 - mean_squared_error: 0.0013 - val_loss: 0.0021 - val_mean_squared_error: 0.0021\n",
      "Epoch 153/1000\n",
      " - 6s - loss: 0.0013 - mean_squared_error: 0.0013 - val_loss: 0.0010 - val_mean_squared_error: 0.0010\n",
      "Epoch 154/1000\n",
      " - 6s - loss: 0.0013 - mean_squared_error: 0.0013 - val_loss: 0.0013 - val_mean_squared_error: 0.0013\n",
      "Epoch 155/1000\n",
      " - 6s - loss: 0.0013 - mean_squared_error: 0.0013 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 156/1000\n",
      " - 6s - loss: 0.0012 - mean_squared_error: 0.0012 - val_loss: 0.0012 - val_mean_squared_error: 0.0012\n",
      "Epoch 157/1000\n",
      " - 6s - loss: 0.0013 - mean_squared_error: 0.0013 - val_loss: 0.0012 - val_mean_squared_error: 0.0012\n",
      "Epoch 158/1000\n",
      " - 6s - loss: 0.0013 - mean_squared_error: 0.0013 - val_loss: 0.0011 - val_mean_squared_error: 0.0011\n",
      "Epoch 159/1000\n",
      " - 6s - loss: 0.0013 - mean_squared_error: 0.0013 - val_loss: 7.6809e-04 - val_mean_squared_error: 7.6809e-04\n",
      "Epoch 160/1000\n",
      " - 6s - loss: 0.0012 - mean_squared_error: 0.0012 - val_loss: 0.0029 - val_mean_squared_error: 0.0029\n",
      "Epoch 161/1000\n",
      " - 6s - loss: 0.0013 - mean_squared_error: 0.0013 - val_loss: 0.0021 - val_mean_squared_error: 0.0021\n",
      "Epoch 162/1000\n",
      " - 6s - loss: 0.0013 - mean_squared_error: 0.0013 - val_loss: 8.2955e-04 - val_mean_squared_error: 8.2955e-04\n",
      "Epoch 163/1000\n",
      " - 6s - loss: 0.0012 - mean_squared_error: 0.0012 - val_loss: 0.0011 - val_mean_squared_error: 0.0011\n",
      "Epoch 164/1000\n",
      " - 6s - loss: 0.0013 - mean_squared_error: 0.0013 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 165/1000\n",
      " - 6s - loss: 0.0012 - mean_squared_error: 0.0012 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 166/1000\n",
      " - 6s - loss: 0.0012 - mean_squared_error: 0.0012 - val_loss: 0.0012 - val_mean_squared_error: 0.0012\n",
      "Epoch 167/1000\n",
      " - 6s - loss: 0.0012 - mean_squared_error: 0.0012 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 168/1000\n",
      " - 6s - loss: 0.0012 - mean_squared_error: 0.0012 - val_loss: 0.0020 - val_mean_squared_error: 0.0020\n",
      "Epoch 169/1000\n",
      " - 6s - loss: 0.0012 - mean_squared_error: 0.0012 - val_loss: 0.0032 - val_mean_squared_error: 0.0032\n",
      "Epoch 170/1000\n",
      " - 6s - loss: 0.0012 - mean_squared_error: 0.0012 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 171/1000\n",
      " - 6s - loss: 0.0012 - mean_squared_error: 0.0012 - val_loss: 8.2537e-04 - val_mean_squared_error: 8.2537e-04\n",
      "Epoch 172/1000\n",
      " - 6s - loss: 0.0012 - mean_squared_error: 0.0012 - val_loss: 0.0021 - val_mean_squared_error: 0.0021\n",
      "Epoch 173/1000\n",
      " - 6s - loss: 0.0012 - mean_squared_error: 0.0012 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 174/1000\n",
      " - 6s - loss: 0.0011 - mean_squared_error: 0.0011 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 175/1000\n",
      " - 6s - loss: 0.0012 - mean_squared_error: 0.0012 - val_loss: 7.2333e-04 - val_mean_squared_error: 7.2333e-04\n",
      "Epoch 176/1000\n",
      " - 6s - loss: 0.0012 - mean_squared_error: 0.0012 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 177/1000\n",
      " - 6s - loss: 0.0011 - mean_squared_error: 0.0011 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 178/1000\n",
      " - 6s - loss: 0.0012 - mean_squared_error: 0.0012 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 179/1000\n",
      " - 6s - loss: 0.0011 - mean_squared_error: 0.0011 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 180/1000\n",
      " - 6s - loss: 0.0011 - mean_squared_error: 0.0011 - val_loss: 6.3729e-04 - val_mean_squared_error: 6.3729e-04\n",
      "Epoch 181/1000\n",
      " - 6s - loss: 0.0012 - mean_squared_error: 0.0012 - val_loss: 0.0011 - val_mean_squared_error: 0.0011\n",
      "Epoch 182/1000\n",
      " - 6s - loss: 0.0011 - mean_squared_error: 0.0011 - val_loss: 0.0021 - val_mean_squared_error: 0.0021\n",
      "Epoch 183/1000\n",
      " - 6s - loss: 0.0011 - mean_squared_error: 0.0011 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 184/1000\n",
      " - 6s - loss: 0.0011 - mean_squared_error: 0.0011 - val_loss: 9.9884e-04 - val_mean_squared_error: 9.9884e-04\n",
      "Epoch 185/1000\n",
      " - 6s - loss: 0.0011 - mean_squared_error: 0.0011 - val_loss: 0.0013 - val_mean_squared_error: 0.0013\n",
      "Epoch 186/1000\n",
      " - 6s - loss: 0.0011 - mean_squared_error: 0.0011 - val_loss: 5.8603e-04 - val_mean_squared_error: 5.8603e-04\n",
      "Epoch 187/1000\n",
      " - 6s - loss: 0.0011 - mean_squared_error: 0.0011 - val_loss: 6.7535e-04 - val_mean_squared_error: 6.7535e-04\n",
      "Epoch 188/1000\n",
      " - 6s - loss: 0.0011 - mean_squared_error: 0.0011 - val_loss: 9.5708e-04 - val_mean_squared_error: 9.5708e-04\n",
      "Epoch 189/1000\n",
      " - 6s - loss: 0.0011 - mean_squared_error: 0.0011 - val_loss: 7.2285e-04 - val_mean_squared_error: 7.2285e-04\n",
      "Epoch 190/1000\n",
      " - 6s - loss: 0.0011 - mean_squared_error: 0.0011 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 191/1000\n",
      " - 6s - loss: 0.0011 - mean_squared_error: 0.0011 - val_loss: 9.8937e-04 - val_mean_squared_error: 9.8937e-04\n",
      "Epoch 192/1000\n",
      " - 6s - loss: 0.0011 - mean_squared_error: 0.0011 - val_loss: 6.9010e-04 - val_mean_squared_error: 6.9010e-04\n",
      "Epoch 193/1000\n",
      " - 6s - loss: 0.0010 - mean_squared_error: 0.0010 - val_loss: 7.7321e-04 - val_mean_squared_error: 7.7321e-04\n",
      "Epoch 194/1000\n",
      " - 6s - loss: 0.0011 - mean_squared_error: 0.0011 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 195/1000\n",
      " - 6s - loss: 0.0010 - mean_squared_error: 0.0010 - val_loss: 0.0027 - val_mean_squared_error: 0.0027\n",
      "Epoch 196/1000\n",
      " - 6s - loss: 0.0011 - mean_squared_error: 0.0011 - val_loss: 8.7746e-04 - val_mean_squared_error: 8.7746e-04\n",
      "Epoch 197/1000\n",
      " - 6s - loss: 0.0011 - mean_squared_error: 0.0011 - val_loss: 8.2613e-04 - val_mean_squared_error: 8.2613e-04\n",
      "Epoch 198/1000\n",
      " - 6s - loss: 0.0010 - mean_squared_error: 0.0010 - val_loss: 0.0011 - val_mean_squared_error: 0.0011\n",
      "Epoch 199/1000\n",
      " - 6s - loss: 0.0011 - mean_squared_error: 0.0011 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 200/1000\n",
      " - 6s - loss: 0.0010 - mean_squared_error: 0.0010 - val_loss: 0.0013 - val_mean_squared_error: 0.0013\n",
      "Epoch 201/1000\n",
      " - 6s - loss: 0.0011 - mean_squared_error: 0.0011 - val_loss: 8.7141e-04 - val_mean_squared_error: 8.7141e-04\n",
      "Epoch 202/1000\n",
      " - 6s - loss: 9.9329e-04 - mean_squared_error: 9.9329e-04 - val_loss: 0.0011 - val_mean_squared_error: 0.0011\n",
      "Epoch 203/1000\n",
      " - 6s - loss: 0.0010 - mean_squared_error: 0.0010 - val_loss: 0.0013 - val_mean_squared_error: 0.0013\n",
      "Epoch 204/1000\n",
      " - 6s - loss: 0.0010 - mean_squared_error: 0.0010 - val_loss: 9.9648e-04 - val_mean_squared_error: 9.9648e-04\n",
      "Epoch 205/1000\n",
      " - 6s - loss: 0.0010 - mean_squared_error: 0.0010 - val_loss: 5.5541e-04 - val_mean_squared_error: 5.5541e-04\n",
      "Epoch 206/1000\n",
      " - 6s - loss: 0.0010 - mean_squared_error: 0.0010 - val_loss: 6.0907e-04 - val_mean_squared_error: 6.0907e-04\n",
      "Epoch 207/1000\n",
      " - 6s - loss: 9.7166e-04 - mean_squared_error: 9.7166e-04 - val_loss: 9.8694e-04 - val_mean_squared_error: 9.8694e-04\n",
      "Epoch 208/1000\n",
      " - 6s - loss: 0.0010 - mean_squared_error: 0.0010 - val_loss: 7.6815e-04 - val_mean_squared_error: 7.6815e-04\n",
      "Epoch 209/1000\n",
      " - 6s - loss: 9.9381e-04 - mean_squared_error: 9.9381e-04 - val_loss: 8.7050e-04 - val_mean_squared_error: 8.7050e-04\n",
      "Epoch 210/1000\n",
      " - 6s - loss: 0.0010 - mean_squared_error: 0.0010 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 211/1000\n",
      " - 6s - loss: 9.8465e-04 - mean_squared_error: 9.8465e-04 - val_loss: 4.9753e-04 - val_mean_squared_error: 4.9753e-04\n",
      "Epoch 212/1000\n",
      " - 6s - loss: 0.0010 - mean_squared_error: 0.0010 - val_loss: 5.4798e-04 - val_mean_squared_error: 5.4798e-04\n",
      "Epoch 213/1000\n",
      " - 6s - loss: 9.2705e-04 - mean_squared_error: 9.2705e-04 - val_loss: 0.0010 - val_mean_squared_error: 0.0010\n",
      "Epoch 214/1000\n",
      " - 6s - loss: 9.8152e-04 - mean_squared_error: 9.8152e-04 - val_loss: 0.0011 - val_mean_squared_error: 0.0011\n",
      "Epoch 215/1000\n",
      " - 6s - loss: 9.8723e-04 - mean_squared_error: 9.8723e-04 - val_loss: 8.6951e-04 - val_mean_squared_error: 8.6951e-04\n",
      "Epoch 216/1000\n",
      " - 6s - loss: 9.7243e-04 - mean_squared_error: 9.7243e-04 - val_loss: 8.3244e-04 - val_mean_squared_error: 8.3244e-04\n",
      "Epoch 217/1000\n",
      " - 6s - loss: 9.4875e-04 - mean_squared_error: 9.4875e-04 - val_loss: 8.9410e-04 - val_mean_squared_error: 8.9410e-04\n",
      "Epoch 218/1000\n",
      " - 6s - loss: 9.6123e-04 - mean_squared_error: 9.6123e-04 - val_loss: 9.6661e-04 - val_mean_squared_error: 9.6661e-04\n",
      "Epoch 219/1000\n",
      " - 6s - loss: 9.7310e-04 - mean_squared_error: 9.7310e-04 - val_loss: 7.3947e-04 - val_mean_squared_error: 7.3947e-04\n",
      "Epoch 220/1000\n",
      " - 6s - loss: 9.5141e-04 - mean_squared_error: 9.5141e-04 - val_loss: 7.3623e-04 - val_mean_squared_error: 7.3623e-04\n",
      "Epoch 221/1000\n",
      " - 6s - loss: 9.4385e-04 - mean_squared_error: 9.4385e-04 - val_loss: 7.0041e-04 - val_mean_squared_error: 7.0041e-04\n",
      "Epoch 222/1000\n",
      " - 6s - loss: 9.2691e-04 - mean_squared_error: 9.2691e-04 - val_loss: 0.0026 - val_mean_squared_error: 0.0026\n",
      "Epoch 223/1000\n",
      " - 6s - loss: 9.6417e-04 - mean_squared_error: 9.6417e-04 - val_loss: 8.4144e-04 - val_mean_squared_error: 8.4144e-04\n",
      "Epoch 224/1000\n",
      " - 6s - loss: 9.1642e-04 - mean_squared_error: 9.1642e-04 - val_loss: 6.9098e-04 - val_mean_squared_error: 6.9098e-04\n",
      "Epoch 225/1000\n",
      " - 6s - loss: 9.5456e-04 - mean_squared_error: 9.5456e-04 - val_loss: 5.5632e-04 - val_mean_squared_error: 5.5632e-04\n",
      "Epoch 226/1000\n",
      " - 6s - loss: 9.4830e-04 - mean_squared_error: 9.4830e-04 - val_loss: 4.8736e-04 - val_mean_squared_error: 4.8736e-04\n",
      "Epoch 227/1000\n",
      " - 6s - loss: 8.9082e-04 - mean_squared_error: 8.9082e-04 - val_loss: 7.3945e-04 - val_mean_squared_error: 7.3945e-04\n",
      "Epoch 228/1000\n",
      " - 6s - loss: 9.3499e-04 - mean_squared_error: 9.3499e-04 - val_loss: 9.3091e-04 - val_mean_squared_error: 9.3091e-04\n",
      "Epoch 229/1000\n",
      " - 6s - loss: 9.3379e-04 - mean_squared_error: 9.3379e-04 - val_loss: 8.4384e-04 - val_mean_squared_error: 8.4384e-04\n",
      "Epoch 230/1000\n",
      " - 6s - loss: 9.1217e-04 - mean_squared_error: 9.1217e-04 - val_loss: 0.0017 - val_mean_squared_error: 0.0017\n",
      "Epoch 231/1000\n",
      " - 6s - loss: 9.3835e-04 - mean_squared_error: 9.3835e-04 - val_loss: 0.0013 - val_mean_squared_error: 0.0013\n",
      "Epoch 232/1000\n",
      " - 6s - loss: 8.9230e-04 - mean_squared_error: 8.9230e-04 - val_loss: 6.2428e-04 - val_mean_squared_error: 6.2428e-04\n",
      "Epoch 233/1000\n",
      " - 6s - loss: 9.2161e-04 - mean_squared_error: 9.2161e-04 - val_loss: 7.0693e-04 - val_mean_squared_error: 7.0693e-04\n",
      "Epoch 234/1000\n",
      " - 6s - loss: 9.0095e-04 - mean_squared_error: 9.0095e-04 - val_loss: 9.3522e-04 - val_mean_squared_error: 9.3522e-04\n",
      "Epoch 235/1000\n",
      " - 6s - loss: 8.9638e-04 - mean_squared_error: 8.9638e-04 - val_loss: 6.3054e-04 - val_mean_squared_error: 6.3054e-04\n",
      "Epoch 236/1000\n",
      " - 6s - loss: 8.9698e-04 - mean_squared_error: 8.9698e-04 - val_loss: 8.0044e-04 - val_mean_squared_error: 8.0044e-04\n",
      "Epoch 237/1000\n",
      " - 6s - loss: 9.0353e-04 - mean_squared_error: 9.0353e-04 - val_loss: 7.6730e-04 - val_mean_squared_error: 7.6730e-04\n",
      "Epoch 238/1000\n",
      " - 6s - loss: 9.1347e-04 - mean_squared_error: 9.1347e-04 - val_loss: 9.3901e-04 - val_mean_squared_error: 9.3901e-04\n",
      "Epoch 239/1000\n",
      " - 6s - loss: 8.7165e-04 - mean_squared_error: 8.7165e-04 - val_loss: 6.3140e-04 - val_mean_squared_error: 6.3140e-04\n",
      "Epoch 240/1000\n",
      " - 6s - loss: 8.8279e-04 - mean_squared_error: 8.8279e-04 - val_loss: 7.6665e-04 - val_mean_squared_error: 7.6665e-04\n",
      "Epoch 241/1000\n",
      " - 6s - loss: 8.9412e-04 - mean_squared_error: 8.9412e-04 - val_loss: 0.0012 - val_mean_squared_error: 0.0012\n",
      "Epoch 242/1000\n",
      " - 6s - loss: 8.8975e-04 - mean_squared_error: 8.8975e-04 - val_loss: 5.8948e-04 - val_mean_squared_error: 5.8948e-04\n",
      "Epoch 243/1000\n",
      " - 6s - loss: 8.6410e-04 - mean_squared_error: 8.6410e-04 - val_loss: 7.9189e-04 - val_mean_squared_error: 7.9189e-04\n",
      "Epoch 244/1000\n",
      " - 6s - loss: 8.7710e-04 - mean_squared_error: 8.7710e-04 - val_loss: 5.2834e-04 - val_mean_squared_error: 5.2834e-04\n",
      "Epoch 245/1000\n",
      " - 6s - loss: 8.8264e-04 - mean_squared_error: 8.8264e-04 - val_loss: 7.7641e-04 - val_mean_squared_error: 7.7641e-04\n",
      "Epoch 246/1000\n",
      " - 6s - loss: 8.7918e-04 - mean_squared_error: 8.7918e-04 - val_loss: 0.0012 - val_mean_squared_error: 0.0012\n",
      "Epoch 247/1000\n",
      " - 6s - loss: 8.6240e-04 - mean_squared_error: 8.6240e-04 - val_loss: 6.0138e-04 - val_mean_squared_error: 6.0138e-04\n",
      "Epoch 248/1000\n",
      " - 6s - loss: 8.5742e-04 - mean_squared_error: 8.5742e-04 - val_loss: 6.7214e-04 - val_mean_squared_error: 6.7214e-04\n",
      "Epoch 249/1000\n",
      " - 6s - loss: 8.4358e-04 - mean_squared_error: 8.4358e-04 - val_loss: 6.2775e-04 - val_mean_squared_error: 6.2775e-04\n",
      "Epoch 250/1000\n",
      " - 6s - loss: 8.6349e-04 - mean_squared_error: 8.6349e-04 - val_loss: 0.0011 - val_mean_squared_error: 0.0011\n",
      "Epoch 251/1000\n",
      " - 6s - loss: 8.6879e-04 - mean_squared_error: 8.6879e-04 - val_loss: 8.4442e-04 - val_mean_squared_error: 8.4442e-04\n",
      "Epoch 252/1000\n",
      " - 6s - loss: 8.4809e-04 - mean_squared_error: 8.4809e-04 - val_loss: 0.0012 - val_mean_squared_error: 0.0012\n",
      "Epoch 253/1000\n",
      " - 6s - loss: 8.4197e-04 - mean_squared_error: 8.4197e-04 - val_loss: 0.0010 - val_mean_squared_error: 0.0010\n",
      "Epoch 254/1000\n",
      " - 6s - loss: 8.4473e-04 - mean_squared_error: 8.4473e-04 - val_loss: 4.6048e-04 - val_mean_squared_error: 4.6048e-04\n",
      "Epoch 255/1000\n",
      " - 6s - loss: 8.4523e-04 - mean_squared_error: 8.4523e-04 - val_loss: 4.7838e-04 - val_mean_squared_error: 4.7838e-04\n",
      "Epoch 256/1000\n",
      " - 6s - loss: 8.4681e-04 - mean_squared_error: 8.4681e-04 - val_loss: 6.2050e-04 - val_mean_squared_error: 6.2050e-04\n",
      "Epoch 257/1000\n",
      " - 6s - loss: 8.2572e-04 - mean_squared_error: 8.2572e-04 - val_loss: 3.6578e-04 - val_mean_squared_error: 3.6578e-04\n",
      "Epoch 258/1000\n",
      " - 6s - loss: 8.5090e-04 - mean_squared_error: 8.5090e-04 - val_loss: 0.0013 - val_mean_squared_error: 0.0013\n",
      "Epoch 259/1000\n",
      " - 6s - loss: 8.2405e-04 - mean_squared_error: 8.2405e-04 - val_loss: 5.8770e-04 - val_mean_squared_error: 5.8770e-04\n",
      "Epoch 260/1000\n",
      " - 6s - loss: 8.3019e-04 - mean_squared_error: 8.3019e-04 - val_loss: 9.6353e-04 - val_mean_squared_error: 9.6353e-04\n",
      "Epoch 261/1000\n",
      " - 6s - loss: 8.2503e-04 - mean_squared_error: 8.2503e-04 - val_loss: 4.5386e-04 - val_mean_squared_error: 4.5386e-04\n",
      "Epoch 262/1000\n",
      " - 6s - loss: 8.2927e-04 - mean_squared_error: 8.2927e-04 - val_loss: 5.5901e-04 - val_mean_squared_error: 5.5901e-04\n",
      "Epoch 263/1000\n",
      " - 6s - loss: 8.1528e-04 - mean_squared_error: 8.1528e-04 - val_loss: 0.0011 - val_mean_squared_error: 0.0011\n",
      "Epoch 264/1000\n",
      " - 6s - loss: 8.2358e-04 - mean_squared_error: 8.2358e-04 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 265/1000\n",
      " - 6s - loss: 8.4043e-04 - mean_squared_error: 8.4043e-04 - val_loss: 5.9475e-04 - val_mean_squared_error: 5.9475e-04\n",
      "Epoch 266/1000\n",
      " - 6s - loss: 8.0026e-04 - mean_squared_error: 8.0026e-04 - val_loss: 4.1623e-04 - val_mean_squared_error: 4.1623e-04\n",
      "Epoch 267/1000\n",
      " - 6s - loss: 8.1044e-04 - mean_squared_error: 8.1044e-04 - val_loss: 7.3996e-04 - val_mean_squared_error: 7.3996e-04\n",
      "Epoch 268/1000\n",
      " - 6s - loss: 8.1877e-04 - mean_squared_error: 8.1877e-04 - val_loss: 8.5281e-04 - val_mean_squared_error: 8.5281e-04\n",
      "Epoch 269/1000\n",
      " - 6s - loss: 7.8731e-04 - mean_squared_error: 7.8731e-04 - val_loss: 0.0013 - val_mean_squared_error: 0.0013\n",
      "Epoch 270/1000\n",
      " - 6s - loss: 8.1236e-04 - mean_squared_error: 8.1236e-04 - val_loss: 5.5710e-04 - val_mean_squared_error: 5.5710e-04\n",
      "Epoch 271/1000\n",
      " - 6s - loss: 8.1344e-04 - mean_squared_error: 8.1344e-04 - val_loss: 6.6800e-04 - val_mean_squared_error: 6.6800e-04\n",
      "Epoch 272/1000\n",
      " - 6s - loss: 8.0014e-04 - mean_squared_error: 8.0014e-04 - val_loss: 6.8248e-04 - val_mean_squared_error: 6.8248e-04\n",
      "Epoch 273/1000\n",
      " - 6s - loss: 7.9148e-04 - mean_squared_error: 7.9148e-04 - val_loss: 5.0099e-04 - val_mean_squared_error: 5.0099e-04\n",
      "Epoch 274/1000\n",
      " - 6s - loss: 7.8672e-04 - mean_squared_error: 7.8672e-04 - val_loss: 0.0011 - val_mean_squared_error: 0.0011\n",
      "Epoch 275/1000\n",
      " - 6s - loss: 7.8179e-04 - mean_squared_error: 7.8179e-04 - val_loss: 0.0015 - val_mean_squared_error: 0.0015\n",
      "Epoch 276/1000\n",
      " - 6s - loss: 7.8294e-04 - mean_squared_error: 7.8294e-04 - val_loss: 5.8408e-04 - val_mean_squared_error: 5.8408e-04\n",
      "Epoch 277/1000\n",
      " - 6s - loss: 7.9800e-04 - mean_squared_error: 7.9800e-04 - val_loss: 9.6703e-04 - val_mean_squared_error: 9.6703e-04\n",
      "Epoch 278/1000\n",
      " - 6s - loss: 7.8533e-04 - mean_squared_error: 7.8533e-04 - val_loss: 5.2439e-04 - val_mean_squared_error: 5.2439e-04\n",
      "Epoch 279/1000\n",
      " - 6s - loss: 7.6540e-04 - mean_squared_error: 7.6540e-04 - val_loss: 7.7138e-04 - val_mean_squared_error: 7.7138e-04\n",
      "Epoch 280/1000\n",
      " - 6s - loss: 7.6294e-04 - mean_squared_error: 7.6294e-04 - val_loss: 5.4937e-04 - val_mean_squared_error: 5.4937e-04\n",
      "Epoch 281/1000\n",
      " - 6s - loss: 7.7561e-04 - mean_squared_error: 7.7561e-04 - val_loss: 8.0387e-04 - val_mean_squared_error: 8.0387e-04\n",
      "Epoch 282/1000\n",
      " - 6s - loss: 7.7244e-04 - mean_squared_error: 7.7244e-04 - val_loss: 5.3930e-04 - val_mean_squared_error: 5.3930e-04\n",
      "Epoch 283/1000\n",
      " - 6s - loss: 7.6062e-04 - mean_squared_error: 7.6062e-04 - val_loss: 0.0012 - val_mean_squared_error: 0.0012\n",
      "Epoch 284/1000\n",
      " - 6s - loss: 7.6356e-04 - mean_squared_error: 7.6356e-04 - val_loss: 6.0934e-04 - val_mean_squared_error: 6.0934e-04\n",
      "Epoch 285/1000\n",
      " - 6s - loss: 7.4658e-04 - mean_squared_error: 7.4658e-04 - val_loss: 0.0010 - val_mean_squared_error: 0.0010\n",
      "Epoch 286/1000\n",
      " - 6s - loss: 7.4106e-04 - mean_squared_error: 7.4106e-04 - val_loss: 5.9914e-04 - val_mean_squared_error: 5.9914e-04\n",
      "Epoch 287/1000\n",
      " - 6s - loss: 7.6032e-04 - mean_squared_error: 7.6032e-04 - val_loss: 5.6169e-04 - val_mean_squared_error: 5.6169e-04\n",
      "Epoch 288/1000\n",
      " - 6s - loss: 7.4814e-04 - mean_squared_error: 7.4814e-04 - val_loss: 4.8700e-04 - val_mean_squared_error: 4.8700e-04\n",
      "Epoch 289/1000\n",
      " - 6s - loss: 7.7295e-04 - mean_squared_error: 7.7295e-04 - val_loss: 3.5899e-04 - val_mean_squared_error: 3.5899e-04\n",
      "Epoch 290/1000\n",
      " - 6s - loss: 7.5245e-04 - mean_squared_error: 7.5245e-04 - val_loss: 6.5140e-04 - val_mean_squared_error: 6.5140e-04\n",
      "Epoch 291/1000\n",
      " - 6s - loss: 7.1617e-04 - mean_squared_error: 7.1617e-04 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 292/1000\n",
      " - 6s - loss: 7.5167e-04 - mean_squared_error: 7.5167e-04 - val_loss: 4.5636e-04 - val_mean_squared_error: 4.5636e-04\n",
      "Epoch 293/1000\n",
      " - 6s - loss: 7.1993e-04 - mean_squared_error: 7.1993e-04 - val_loss: 7.7104e-04 - val_mean_squared_error: 7.7104e-04\n",
      "Epoch 294/1000\n",
      " - 6s - loss: 7.4459e-04 - mean_squared_error: 7.4459e-04 - val_loss: 6.4613e-04 - val_mean_squared_error: 6.4613e-04\n",
      "Epoch 295/1000\n",
      " - 6s - loss: 7.3516e-04 - mean_squared_error: 7.3516e-04 - val_loss: 4.2627e-04 - val_mean_squared_error: 4.2627e-04\n",
      "Epoch 296/1000\n",
      " - 6s - loss: 7.1398e-04 - mean_squared_error: 7.1398e-04 - val_loss: 0.0010 - val_mean_squared_error: 0.0010\n",
      "Epoch 297/1000\n",
      " - 6s - loss: 7.3326e-04 - mean_squared_error: 7.3326e-04 - val_loss: 4.5212e-04 - val_mean_squared_error: 4.5212e-04\n",
      "Epoch 298/1000\n",
      " - 6s - loss: 7.5073e-04 - mean_squared_error: 7.5073e-04 - val_loss: 0.0014 - val_mean_squared_error: 0.0014\n",
      "Epoch 299/1000\n",
      " - 6s - loss: 7.0957e-04 - mean_squared_error: 7.0957e-04 - val_loss: 8.9403e-04 - val_mean_squared_error: 8.9403e-04\n",
      "Epoch 300/1000\n",
      " - 6s - loss: 7.1487e-04 - mean_squared_error: 7.1487e-04 - val_loss: 8.6945e-04 - val_mean_squared_error: 8.6945e-04\n",
      "Epoch 301/1000\n",
      " - 6s - loss: 7.2388e-04 - mean_squared_error: 7.2388e-04 - val_loss: 5.3813e-04 - val_mean_squared_error: 5.3813e-04\n",
      "Epoch 302/1000\n",
      " - 6s - loss: 7.2532e-04 - mean_squared_error: 7.2532e-04 - val_loss: 0.0016 - val_mean_squared_error: 0.0016\n",
      "Epoch 303/1000\n",
      " - 6s - loss: 7.1387e-04 - mean_squared_error: 7.1387e-04 - val_loss: 0.0011 - val_mean_squared_error: 0.0011\n",
      "Epoch 304/1000\n",
      " - 6s - loss: 7.1670e-04 - mean_squared_error: 7.1670e-04 - val_loss: 6.6803e-04 - val_mean_squared_error: 6.6803e-04\n",
      "Epoch 305/1000\n",
      " - 6s - loss: 7.0594e-04 - mean_squared_error: 7.0594e-04 - val_loss: 5.6532e-04 - val_mean_squared_error: 5.6532e-04\n",
      "Epoch 306/1000\n",
      " - 6s - loss: 6.9598e-04 - mean_squared_error: 6.9598e-04 - val_loss: 0.0010 - val_mean_squared_error: 0.0010\n",
      "Epoch 307/1000\n",
      " - 6s - loss: 7.2644e-04 - mean_squared_error: 7.2644e-04 - val_loss: 0.0013 - val_mean_squared_error: 0.0013\n",
      "Epoch 00307: early stopping\n"
     ]
    }
   ],
   "source": [
    "earlystop = EarlyStopping(monitor='val_mean_squared_error', patience=50, \n",
    "                          verbose=1, mode='auto', min_delta = 0.0001)\n",
    "\n",
    "history = model.fit(Xtrain_scaled, Ytrain_scaled, validation_split = 0.3, \n",
    "                    verbose = 2, batch_size=2**13, epochs = 1000, \n",
    "                    callbacks = [earlystop])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_model_history(model_history, saveFig = False):\n",
    "    fig, axs = plt.subplots(1,1,figsize=(10,5))\n",
    "    # summarize history for accuracy\n",
    "    axs.plot(range(1,len(model_history.history['mean_squared_error'])+1),\n",
    "             model_history.history['mean_squared_error'])\n",
    "    axs.plot(range(1,len(model_history.history['val_mean_squared_error'])+1),\n",
    "             model_history.history['val_mean_squared_error'])\n",
    "    axs.set_title('Model MSE')\n",
    "    axs.set_ylabel('mean_squared_error')\n",
    "    axs.set_xlabel('Epoch')\n",
    "    axs.set_xticks(np.arange(1,len(model_history.history['val_mean_squared_error'])+1),\n",
    "                   len(model_history.history['val_mean_squared_error'])/10)\n",
    "    axs.legend(['train', 'val'], loc='best')\n",
    "    if saveFig:\n",
    "        fig.savefig(os.path.join(figDir, \"ModelTraining.png\"), dpi = 120, bbox_inches='tight')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1fff3b62dd8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_model_history(history, saveFig = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "nnpreds = model.predict(Xtest_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ffe27e9ba8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.close(\"all\")\n",
    "fig, axs = plt.subplots(1,3, figsize=(15, 5), facecolor='w', edgecolor='k')\n",
    "fig.subplots_adjust(hspace = 0.1, wspace=0.3)\n",
    "fig.suptitle('Predicted vs. acutal ', fontsize=14, fontweight='bold')\n",
    "\n",
    "axs = axs.ravel()\n",
    "\n",
    "nms = [\"Fx\", \"Fy\", \"tau\"]\n",
    "\n",
    "for ii in range(3):\n",
    "    # read in image\n",
    "    \n",
    "    try:\n",
    "        axs[ii].scatter(y = Ytest_scaled[:,ii],x = nnpreds[:,ii], s = 1, alpha = 0.1 )\n",
    "        axs[ii].set_xlabel(\"Predicted Value (scaled)\")\n",
    "        if(ii == 0):\n",
    "            axs[ii].set_ylabel(\"Actual Value\\n(scaled)\")\n",
    "        axs[ii].set_title(nms[ii])\n",
    "        axs[ii].plot(Ytest_scaled[:,ii], Ytest_scaled[:,ii], 'b')\n",
    "    except:\n",
    "        pass\n",
    "\n",
    "    \n",
    "fig.savefig(os.path.join(figDir, \"NNPreds_newData_control_bigbatch_FxFy.png\"), dpi = 120, bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# refref: idea - predict whole paths, rather than just the endpoint"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1fff3b582b0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.close(\"all\")\n",
    "fig, axs = plt.subplots(1,3, figsize=(15, 5), facecolor='w', edgecolor='k')\n",
    "fig.subplots_adjust(hspace = 0.1, wspace=0.3)\n",
    "fig.suptitle('Predicted vs. acutal ', fontsize=14, fontweight='bold')\n",
    "\n",
    "axs = axs.ravel()\n",
    "\n",
    "nms = [\"Fx\", \"Fy\", \"tau\"]\n",
    "\n",
    "\n",
    "# replace lightest colors with white\n",
    "import matplotlib.colors\n",
    "\n",
    "cmap = plt.cm.magma_r\n",
    "cmaplist = np.array([cmap(i) for i in range(cmap.N)])\n",
    "#cmaplist[0:2] = np.repeat((0.0,0.0,0.0,0.0), 2).reshape(-1,4)\n",
    "cmaplist[:,0:3] = np.divide(cmaplist[:, 0:3], 1.1)\n",
    "#cmaplist[:,0:3] = np.transpose(np.vstack([np.linspace(1, 0, 256), np.linspace(1, 0, 256), np.linspace(1, 0, 256)]))\n",
    "cmaplist[0] = (1,1,1,0.5)\n",
    "cmap = matplotlib.colors.LinearSegmentedColormap.from_list('mcm',cmaplist, cmap.N)\n",
    "\n",
    "for ii in range(3):\n",
    "    # read in image\n",
    "    \n",
    "    try:\n",
    "        axs[ii].hexbin(y = Ytest_scaled[:,ii],x = nnpreds[:,ii], gridsize = 150, cmap = cmap)\n",
    "        axs[ii].set_xlabel(\"Predicted Value (scaled)\")\n",
    "        if(ii == 0):\n",
    "            axs[ii].set_ylabel(\"Actual Value\\n(scaled)\")\n",
    "        axs[ii].set_title(nms[ii])\n",
    "        axs[ii].plot(Ytest_scaled[:,ii], Ytest_scaled[:,ii], 'grey', linewidth = 1, linestyle  = \"--\")\n",
    "    except:\n",
    "        pass\n",
    "\n",
    "fig.savefig(os.path.join(figDir, \"NNPreds_newData_control_bigbatch_FxFy_hexbin.png\"), dpi = 120, bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "bf2[['simNum', 'timestep']] = bf2[['simNum', 'timestep']].astype(int)\n",
    "\n",
    "bf2['count'] = bf2.groupby([\"simNum\", \"timestep\"]).cumcount()\n",
    "\n",
    "bf2.head()\n",
    "\n",
    "bf2.tail()\n",
    "\n",
    "sdf = bf2.loc[(bf2[\"count\"] == 0) | (bf2[\"count\"] == 99), :]\n",
    "sdf.head()\n",
    "\n",
    "ss = sdf.pivot_table(index='simNum', columns = \"count\",aggfunc='first')\n",
    "ss.columns = [f'{i}_{j}' if j != '' else f'{i}' for i,j in ss.columns]\n",
    "ss[\"simNum\"] = ss.index\n",
    "ss.index.name = None\n",
    "ss.head()\n",
    "\n",
    "np.array(ss).shape\n",
    "\n",
    "np.vstack([np.array(ss), np.array(ss)]).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "deeplearning",
   "language": "python",
   "name": "deeplearning"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}