train_model_DeepInsight_CNN.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "0a59ff85",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-17T09:22:24.325212Z",
     "start_time": "2021-10-17T09:22:22.607210Z"
    }
   },
   "outputs": [],
   "source": [
    "import joblib\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.preprocessing import LabelBinarizer, MinMaxScaler, RobustScaler, StandardScaler\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "521ff09a",
   "metadata": {},
   "source": [
    "## DeepInsight"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "1ecada4e",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-17T08:54:14.102076Z",
     "start_time": "2021-10-17T08:54:13.805076Z"
    }
   },
   "outputs": [],
   "source": [
    "from pyDeepInsight import ImageTransformer, LogScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "c3fc7c4f",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T09:21:31.800926Z",
     "start_time": "2021-10-16T09:21:31.138927Z"
    }
   },
   "outputs": [],
   "source": [
    "train = pd.read_csv('train_extra_2_no_scaling.csv')\n",
    "target = pd.read_csv('target_extra_for_gan.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "f122006d",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T09:21:31.830929Z",
     "start_time": "2021-10-16T09:21:31.803929Z"
    }
   },
   "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>CO2_working_capacity [mL/g]</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>105.284502</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>101.224774</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>118.987011</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>187.626004</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>79.210001</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68597</th>\n",
       "      <td>-12.943652</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68598</th>\n",
       "      <td>-12.985582</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68599</th>\n",
       "      <td>-13.187635</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68600</th>\n",
       "      <td>15.672698</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68601</th>\n",
       "      <td>3.144708</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>68602 rows × 1 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       CO2_working_capacity [mL/g]\n",
       "0                       105.284502\n",
       "1                       101.224774\n",
       "2                       118.987011\n",
       "3                       187.626004\n",
       "4                        79.210001\n",
       "...                            ...\n",
       "68597                   -12.943652\n",
       "68598                   -12.985582\n",
       "68599                   -13.187635\n",
       "68600                    15.672698\n",
       "68601                     3.144708\n",
       "\n",
       "[68602 rows x 1 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "target"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "86ce2b32",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T09:21:31.876926Z",
     "start_time": "2021-10-16T09:21:31.834930Z"
    }
   },
   "outputs": [],
   "source": [
    "target['is_good'] = 0\n",
    "target.loc[target['CO2_working_capacity [mL/g]']>200, 'is_good'] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "0410168f",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T09:21:31.892931Z",
     "start_time": "2021-10-16T09:21:31.877929Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    57143\n",
       "1    11459\n",
       "Name: is_good, dtype: int64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "target['is_good'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "56b71038",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T09:21:31.908927Z",
     "start_time": "2021-10-16T09:21:31.894929Z"
    }
   },
   "outputs": [],
   "source": [
    "is_good = target['is_good'].values\n",
    "target = target.drop(['is_good'], axis=1).values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "51edeaea",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T09:21:32.370931Z",
     "start_time": "2021-10-16T09:21:31.910930Z"
    }
   },
   "outputs": [],
   "source": [
    "ln = LogScaler()\n",
    "train_norm = ln.fit_transform(train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "48e9b3f9",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T09:21:32.432929Z",
     "start_time": "2021-10-16T09:21:32.372930Z"
    }
   },
   "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>volume</th>\n",
       "      <th>volume/g</th>\n",
       "      <th>density</th>\n",
       "      <th>weight</th>\n",
       "      <th>surface_area</th>\n",
       "      <th>void_fraction</th>\n",
       "      <th>void_volume</th>\n",
       "      <th>CO2/N2_selectivity</th>\n",
       "      <th>heat_adsorption</th>\n",
       "      <th>n_atoms</th>\n",
       "      <th>...</th>\n",
       "      <th>topology_1</th>\n",
       "      <th>topology_2</th>\n",
       "      <th>topology_3</th>\n",
       "      <th>topology_4</th>\n",
       "      <th>topology_5</th>\n",
       "      <th>topology_6</th>\n",
       "      <th>topology_7</th>\n",
       "      <th>topology_8</th>\n",
       "      <th>topology_9</th>\n",
       "      <th>topology_10</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.657034</td>\n",
       "      <td>0.327012</td>\n",
       "      <td>0.525629</td>\n",
       "      <td>0.717338</td>\n",
       "      <td>0.989935</td>\n",
       "      <td>0.068567</td>\n",
       "      <td>0.029998</td>\n",
       "      <td>0.357182</td>\n",
       "      <td>0.750837</td>\n",
       "      <td>0.626299</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.716299</td>\n",
       "      <td>0.323640</td>\n",
       "      <td>0.531518</td>\n",
       "      <td>0.787361</td>\n",
       "      <td>0.989546</td>\n",
       "      <td>0.116262</td>\n",
       "      <td>0.050286</td>\n",
       "      <td>0.399062</td>\n",
       "      <td>0.763476</td>\n",
       "      <td>0.731002</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.655650</td>\n",
       "      <td>0.345518</td>\n",
       "      <td>0.495302</td>\n",
       "      <td>0.709314</td>\n",
       "      <td>0.989673</td>\n",
       "      <td>0.124732</td>\n",
       "      <td>0.060382</td>\n",
       "      <td>0.338768</td>\n",
       "      <td>0.734841</td>\n",
       "      <td>0.635287</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.700179</td>\n",
       "      <td>0.382530</td>\n",
       "      <td>0.443224</td>\n",
       "      <td>0.745552</td>\n",
       "      <td>0.990121</td>\n",
       "      <td>0.177328</td>\n",
       "      <td>0.101780</td>\n",
       "      <td>0.369830</td>\n",
       "      <td>0.728883</td>\n",
       "      <td>0.668210</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.658108</td>\n",
       "      <td>0.325010</td>\n",
       "      <td>0.529111</td>\n",
       "      <td>0.719330</td>\n",
       "      <td>0.989774</td>\n",
       "      <td>0.067561</td>\n",
       "      <td>0.029233</td>\n",
       "      <td>0.386644</td>\n",
       "      <td>0.739671</td>\n",
       "      <td>0.649430</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68597</th>\n",
       "      <td>0.660623</td>\n",
       "      <td>0.314106</td>\n",
       "      <td>0.548841</td>\n",
       "      <td>0.726506</td>\n",
       "      <td>0.988915</td>\n",
       "      <td>0.456283</td>\n",
       "      <td>0.196649</td>\n",
       "      <td>0.363009</td>\n",
       "      <td>0.482876</td>\n",
       "      <td>0.674864</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68598</th>\n",
       "      <td>0.675011</td>\n",
       "      <td>0.287459</td>\n",
       "      <td>0.603240</td>\n",
       "      <td>0.755806</td>\n",
       "      <td>0.989763</td>\n",
       "      <td>0.010283</td>\n",
       "      <td>0.003519</td>\n",
       "      <td>0.220242</td>\n",
       "      <td>0.482876</td>\n",
       "      <td>0.681206</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68599</th>\n",
       "      <td>0.694694</td>\n",
       "      <td>0.292050</td>\n",
       "      <td>0.593168</td>\n",
       "      <td>0.777381</td>\n",
       "      <td>0.988804</td>\n",
       "      <td>0.505645</td>\n",
       "      <td>0.196649</td>\n",
       "      <td>0.268707</td>\n",
       "      <td>0.482876</td>\n",
       "      <td>0.736995</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68600</th>\n",
       "      <td>0.743410</td>\n",
       "      <td>0.300622</td>\n",
       "      <td>0.575180</td>\n",
       "      <td>0.830363</td>\n",
       "      <td>0.988873</td>\n",
       "      <td>0.485493</td>\n",
       "      <td>0.196649</td>\n",
       "      <td>0.206962</td>\n",
       "      <td>0.482876</td>\n",
       "      <td>0.808263</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68601</th>\n",
       "      <td>0.678459</td>\n",
       "      <td>0.248685</td>\n",
       "      <td>0.703209</td>\n",
       "      <td>0.781925</td>\n",
       "      <td>0.989585</td>\n",
       "      <td>0.014715</td>\n",
       "      <td>0.003877</td>\n",
       "      <td>0.144885</td>\n",
       "      <td>0.482876</td>\n",
       "      <td>0.672052</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>68602 rows × 144 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         volume  volume/g   density    weight  surface_area  void_fraction  \\\n",
       "0      0.657034  0.327012  0.525629  0.717338      0.989935       0.068567   \n",
       "1      0.716299  0.323640  0.531518  0.787361      0.989546       0.116262   \n",
       "2      0.655650  0.345518  0.495302  0.709314      0.989673       0.124732   \n",
       "3      0.700179  0.382530  0.443224  0.745552      0.990121       0.177328   \n",
       "4      0.658108  0.325010  0.529111  0.719330      0.989774       0.067561   \n",
       "...         ...       ...       ...       ...           ...            ...   \n",
       "68597  0.660623  0.314106  0.548841  0.726506      0.988915       0.456283   \n",
       "68598  0.675011  0.287459  0.603240  0.755806      0.989763       0.010283   \n",
       "68599  0.694694  0.292050  0.593168  0.777381      0.988804       0.505645   \n",
       "68600  0.743410  0.300622  0.575180  0.830363      0.988873       0.485493   \n",
       "68601  0.678459  0.248685  0.703209  0.781925      0.989585       0.014715   \n",
       "\n",
       "       void_volume  CO2/N2_selectivity  heat_adsorption   n_atoms  ...  \\\n",
       "0         0.029998            0.357182         0.750837  0.626299  ...   \n",
       "1         0.050286            0.399062         0.763476  0.731002  ...   \n",
       "2         0.060382            0.338768         0.734841  0.635287  ...   \n",
       "3         0.101780            0.369830         0.728883  0.668210  ...   \n",
       "4         0.029233            0.386644         0.739671  0.649430  ...   \n",
       "...            ...                 ...              ...       ...  ...   \n",
       "68597     0.196649            0.363009         0.482876  0.674864  ...   \n",
       "68598     0.003519            0.220242         0.482876  0.681206  ...   \n",
       "68599     0.196649            0.268707         0.482876  0.736995  ...   \n",
       "68600     0.196649            0.206962         0.482876  0.808263  ...   \n",
       "68601     0.003877            0.144885         0.482876  0.672052  ...   \n",
       "\n",
       "       topology_1  topology_2  topology_3  topology_4  topology_5  topology_6  \\\n",
       "0             0.0         0.0         0.0         0.0         1.0         0.0   \n",
       "1             0.0         1.0         0.0         0.0         0.0         0.0   \n",
       "2             0.0         0.0         0.0         0.0         1.0         0.0   \n",
       "3             0.0         0.0         0.0         0.0         0.0         0.0   \n",
       "4             0.0         0.0         0.0         0.0         1.0         0.0   \n",
       "...           ...         ...         ...         ...         ...         ...   \n",
       "68597         0.0         0.0         0.0         0.0         1.0         0.0   \n",
       "68598         0.0         1.0         0.0         0.0         0.0         0.0   \n",
       "68599         0.0         0.0         0.0         0.0         0.0         0.0   \n",
       "68600         0.0         0.0         0.0         0.0         0.0         0.0   \n",
       "68601         0.0         0.0         0.0         0.0         1.0         0.0   \n",
       "\n",
       "       topology_7  topology_8  topology_9  topology_10  \n",
       "0             0.0         0.0         0.0          0.0  \n",
       "1             0.0         0.0         0.0          0.0  \n",
       "2             0.0         0.0         0.0          0.0  \n",
       "3             0.0         1.0         0.0          0.0  \n",
       "4             0.0         0.0         0.0          0.0  \n",
       "...           ...         ...         ...          ...  \n",
       "68597         0.0         0.0         0.0          0.0  \n",
       "68598         0.0         0.0         0.0          0.0  \n",
       "68599         0.0         0.0         0.0          0.0  \n",
       "68600         0.0         0.0         0.0          0.0  \n",
       "68601         0.0         0.0         0.0          0.0  \n",
       "\n",
       "[68602 rows x 144 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_norm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "f2c7c645",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T09:21:32.462927Z",
     "start_time": "2021-10-16T09:21:32.437931Z"
    }
   },
   "outputs": [],
   "source": [
    "train_norm = train_norm.astype('float32')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "c5ce04ba",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T09:21:32.478930Z",
     "start_time": "2021-10-16T09:21:32.464929Z"
    }
   },
   "outputs": [],
   "source": [
    "it = ImageTransformer(feature_extractor='tsne', \n",
    "                      pixels=80, random_state=42, \n",
    "                      n_jobs=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "00f6fd45",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T09:21:33.918967Z",
     "start_time": "2021-10-16T09:21:32.480931Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\admin\\anaconda3\\envs\\chemistry-ml\\lib\\site-packages\\sklearn\\manifold\\_t_sne.py:699: FutureWarning: 'square_distances' has been introduced in 0.24 to help phase out legacy squaring behavior. The 'legacy' setting will be removed in 1.1 (renaming of 0.26), and the default setting will be changed to True. In 1.3, 'square_distances' will be removed altogether, and distances will be squared by default. Set 'square_distances'=True to silence this warning.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<pyDeepInsight.image_transformer.ImageTransformer at 0x1addfede2c8>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "plt.figure(figsize=(5, 5))\n",
    "it.fit(train_norm, plot=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "12be8794",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T09:21:33.934926Z",
     "start_time": "2021-10-16T09:21:33.920928Z"
    }
   },
   "outputs": [],
   "source": [
    "fdm = it.feature_density_matrix()\n",
    "fdm[fdm == 0] = np.nan # just for visualization purpose"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "1d5f6153",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T09:21:34.644931Z",
     "start_time": "2021-10-16T09:21:33.937930Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10, 7))\n",
    "\n",
    "ax = sns.heatmap(fdm, cmap=\"viridis\", linewidths=0.01, \n",
    "                 linecolor=\"lightgrey\", square=True)\n",
    "for _, spine in ax.spines.items():\n",
    "    spine.set_visible(True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "7a0e5d94",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T09:21:36.871929Z",
     "start_time": "2021-10-16T09:21:34.646928Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABZUAAAFhCAYAAAAIgUS2AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAAAsTAAALEwEAmpwYAACb+0lEQVR4nOzdebwkZXn28d8Fg7gAgrJkZDAQjUbmsOkENURFEEMU12gUI4pEJ1FURI2KvooaSTQaFaPRTASBV4JDWNTgBq9CCImiA0FgwD0uwyISBwU3Fq/3j6qDPc3pc3rvp7qv7+cznzn11FN33dWnp7rm6Xrukm0iIiIiIiIiIiIiIrqx2aQTiIiIiIiIiIiIiIjmyKByRERERERERERERHQtg8oRERERERERERER0bUMKkdERERERERERERE1zKoHBERERERERERERFdy6ByRERERERERERERHQtg8oxUpI+JOmNk84jIiIiImIa5Po6IqIZcr6OaZdB5eibpO9K+oWkmyXdJOm/JP2lpDvfV7b/0vZfD2l/fyXpynp//yPprzrkc0v959wm7zciIiIiZssErq/fLOm2luvYWyT9Tsv6vSVdIunn9d97D2m/J0m6tW2/m7esP1DS1+r9ni/pt4ex34iIYZnA+XrLepD6h5J+LOnfJO3csv4+ks6W9DNJ35P0nCHtV5LeJukaST+RdIGklW15nSjpp5Kul/TKYew3miGDyjGoJ9neGvht4O3Aa4ETRrQvAc8DtgMOBl4q6dkL5LNV/efxDd9vRERERMyecV5fA6xtuY7dyvZ3ACTdDfgE8FGq6+CTgU/U7cPwd237vaPe7/bAWcAbgfsA64C1Q9pnRMQwjfN8fRTwSGBP4H7ARuAfWtZ/ALgV2An4M+CDrYO/A3gmcATwKKpz8heB/9uy/s3A71K9Bo8FXiPp4CHsNxogg8oxFLZ/YvuTwLOA50uagzvvQnhb/fP+kjZIeo2kGyRdJ+mpkp4g6Rv1t22vX2Qff2f7Utu32/461UXufr3mKuluki6T9LJ6eXNJ/ynpTaPcb0REREREt8Zxfb2E/YFlwHtt/8r2+6hutjigvWN9h9wGSU+ql7eS9C1Jz+tjv08H1tv+V9u/pBqw2EvS7/V5HBERIzWm8/VuwOds/7A+N64FVtax7wX8CfBG27fYvgj4JHDYQoEkfVrS37csf0zSiYvs9yLb36m//PsosHvL+ucDf217o+2rgX8GDl/0BYupkUHlGCrbXwY2UH2LtZDfAu4O7Ay8ieqE81zgYfU2b5S021L7kaS6//q2VadK+pGkcyXt1SHHW+t9vlXSQ4DXAZsDx41yvxERERERvRrD9fWT6sGM9ZJe3NK+ErjctlvaLq/b23P8MdWdbP8saUfgPcBltk9ZZL8vqfd7iaQ/advvV1ti/wz49kL7jYgoyYjP1ycA+0m6n6R7Ut2N/Jl63YOA221/o6X/V+l83jwCOEzSAZL+DNiX6k7ohXwMeICkB0nagmoQ+bMAkrYDltNyzl5ivzFlMqgco3At1bSIhdwGHGf7NqqT0/bA8bZvtr0euAroZlD2zVTv34+0tP0ZsCvVtIvzgc9J2nahjW1fCbwN+DjwauCw+Sl3o9xvREREREQfRnV9fTrwEGAH4EXAmyQdWq/bCvhJW/+fAFsvFMj2ucC/Ap8HngD8xSLH8z6q6dI7UpW5OEnS/EzAnvYbEVGYUZ2vvwn8ALgG+CnVufut9bqt6rZWi52vrwdeTFXW6HjgebZv7rDf64CLgK8Dv6Aqh3F0y37n97XkfmP6ZFA5RmFn4Mcd1v1vy+DtL+q/f9iy/hf85sS0IEkvpapx/ETbv5pvt/2ftn9h++e2/xa4ic7fEEJ1Av1t4NO2v7nYPoe834iIiIiIXozk+tr2VbavtX2H7f+iGlx4Rr36FmCbtk22AToNPACsAeaAk2z/b6dOdWm5/63Ly30aOJWq7EW/+42IKMWoxkM+AGwJ3Be4F1Xt+fk7lfs5b/4b1Yztr9flMjp5E/D7wC5Ud1m/BfhCfbf0LS376na/MUUyqBxDJen3qU6ii52UBol/BFW5igNtb1iiu6nqvnXyj8A5wB9J+sMx7jciIiIioiujvr5u03odux7Ysy7/Nm9P7loGDqieU0I1qHwKVWmLBw6w3zvv1KtrhT6g034jIkox4vP13lRf2P24vsntH4B964ebfgNYJul3W/rvxeLnzeOAq4HlLTNUOu13re0N9ReBJ1E9vHV32xup7mRuvbt6qf3GFMmgcgyFpG0kHUI1heOjtq8YwT7+DPgb4KD5p1K3rLu/pP3qh/DdXdJfUU0l+c8OsQ6jqlt0OPBy4GRJC34jOMz9RkRERER0Y0zX10+RtJ0q+1JdF3+iXn0BcAfwcklb1rP2AL7QIdzrqQaHjwDeCZxSDzQvtN9n1A/z20zS46lqin6yXn02MCfpTyTdneouucttf22wo42IGI1xnK+BrwDPk3TvurbxS4Brbd9Y154/i+q5Ufeqywk9Bfi/HfJ9NPACqpnYzwf+QdLOi+z3mZJ2qs/ZhwFbAN+q158C/J/6s+T3qEopnTSMA47yZVA5BvVvkm6mqu3zBuDdVCenUXgb1VSPr0i6pf7zoXrd1sAHgY1UNYYOBv54oWl3ku4PvJeqbtAttv8FWEf1QJGR7TciIiIiogvjvL5+NtXAwM1UAwPvsH0y3Plw66dSDTrcRDVY/NS6fROSHga8kur6+g7gHVQDzK/rsN+jqK6db6IagH6R7Qvq/f4I+BOqu+g2Ag+v84yIKM04z9evBn5JVVv5R1S165/Wsv4lwD2AG4DTgBfXdZo3IWkbqvP9S21fY/s/qB4C+JG2mSnz3kH18L3LqM7ZRwN/Yvumev2xVA9T/R7w78A7bX92kAON5tCmD/ONiIiIiIiIiIiIiOgsdypHRERERERERERERNcyqBwRERERERERERERXcugckREREREREREDyQdLOnrkr4lqVP98IiIqZVB5YiIiIiIQmSQIiKifJI2Bz4A/DGwO3CopN0nm1VExHhlUDkiIiIiogAZpIiIaIx9gW/Z/o7tW4GPAU+ZcE4REWO1bJw722677Xy/+91vnLuMiBjYVVdddaPtHSadR0RETL07BykAJM0PUlzVaYNcX0dEE03B9fXOwA9aljcAD19sg5yvI6KJFjtfDzSoLOlg4Hhgc+DDtt++WP/73e9+rF+/vqvYV155JQBzc3ON7F9iTk3vX2JOpfY/es+3dNX/PZcf21P8QXJqcv899tjje111joiIGEzPgxQLXV93+qybtfYSc5rW9hJzGnZ7+/V1p+vovNbdtc/i9fX8+br9dRlkObGakUepsUrJYxyx5s/h7efuvNZL913sfN13+YtMz4uIiIiIiIiIGXQNsEvL8oq6bROSVktaJ2ndxo0bx5ZcRMQ4LBtg256n50VEREREREddD1IAqwGWL18+nswiIqLVV4DflbQb1Xn62cBz2jvZXgOsAVi5cqXn7wKE39wROIzlxGpGHqXGKiWPUcaav0N50nmV+vos1beTQR7Ut9D0vJ0HiBcRERERMcvuHKSQdDeqQYpPtneyvcb2Kturtttuu7EnGREx62zfDrwU+BxwNXC67e5qfUZETIlB7lSOiIiIiIghsX27pPlBis2BE7sZpOh0N0naJ7/vWWsvMadhtbff5TbpfCa572EeQ5PZ/jTw6V62aXpt1abFKiWPUmOVkscsxColj35jdSLbi3bouKH0SODNtv+oXj4GwPbftvVrnZ73sGuvvbar+O0H0rT+JebU9P4l5lRq/zyob7j999hjj0tsr+pqg4iIiDFauXKl165dO+k0IiJ6MovX1zlfR0QTLXa+XjZA3L5qCA2wv4iIiIiIaNP+RWmnL1Bnrb3EnKa1vcScprW9xJz6bY+IiGbru6ZyaghFRERERERExKyRdHdJX5b0VUnrJXU1VbR1gH1ubm7g5cRqRh6lxiolj1mIVUoe/cbqZJA7lfuqIRQRERERERER0WC/Ag6wfYukLYCLJH3G9pcmnVhExLgMNKgcERERERHDIWkX4BRgJ8DAGtvHTzariIho5+rhVLfUi1vUf5Ys99la+qO9DMggy4nVjDxKjVVKHrMQq5Q8eo3VSd8P6utHCtNHRBPN4oNEIiJi/CQtB5bbvlTS1sAlwFNtX9Vpm1xfR0QTTcP1taTNqc7TDwQ+YPu1i/XP+Toimmix83XfNZUjIiIiImJ4bF9n+9L655upnluy82SzioiIhdi+w/bewApgX0lLFiCdhtqqTYpVSh6lxiolj1mIVUoe/cbqZKDyF5JOBA4BbrC99N6gq6Rg8afdNqF/iTk1vX+JOc1a/xJzGlf/iIiIcZK0K7APcPFSfds/0zp91s1ae4k5TWt7iTlNa3uJOfXbPi1s3yTpfOBgYJODk7QaWA2wfPnyCWQXETE6Aw0qAycB76eq/RYREREREQOStBVwJvAK2z+ddD4REbEpSTsAt9UDyvcADgLe0d7P9hpgDVTlL6ahtmoTY5WSR6mxSsljFmKVkkevsToZaFDZ9oX1XRQRERERETEgSVtQDSifavusDn1y51tExGQtB06u6ypvBpxu+5wJ5xQRMVapqRwRERERUQBJAk4Arrb97k79bK+xvcr2qu222258CUZEBAC2L7e9j+09bc/Zfms3201DbdUmxSolj1JjlZLHLMQqJY9+Y3UyaPmLiIiIiIgYjv2Aw4ArJF1Wt73e9qcX26jTFMW0T37fs9ZeYk7T2l5iTv0cQ0RENJdsDxagKn9xTqcH9bVNz3vYtdde21Xc+Q+ebkbGS+xfYk5N719iTrPWv8ScxtF/jz32uMT2qq42iIiIGKOVK1d67dq1k04jIqIns3h9nfN1RDTRYufrkd+p3F6YftT7i4iIiIiYJe1flHb6AnXW2kvMaVrbS8xpWttLzKnf9oiIaLaBaipLOg34IvBgSRsk/flw0oqIiIiIiIiIKJOkbSWdIelrkq6W9MiltpmG2qpNilVKHqXGKiWPWYhVSh79xupkoDuVbR86yPYREREREREREQ10PPBZ28+QdDfgnpNOKCJinEZe/iIiIiIiIronaXNgHXCN7UMmnU9ERGxK0r2BRwOHA9i+Fbh1qe1aS3+0lwEZZDmxmpFHqbFKyWMWYpWSR6+xOhn4QX29SGH6iGiiWXyQSERETI6kVwKrgG2WGlTO9XVENFHTr68l7U317KirgL2AS4CjbP+s0zY5X0dEEy12vh6opnJERERERAyPpBXAE4EPTzqXiIjoaBnwUOCDtvcBfga8rr2TpNWS1klat3HjxqmordqkWKXkUWqsUvKYhVil5NFvrE7Gfqfy+vXru+q72NNum9C/xJya3r/EnGatf4k5jaN/0++kiIiI5pB0BvC3wNbAq7u5U7n9+rrTZ92stZeY07S2l5jTtLaXmFM/7U2/vpb0W8CXbO9aLz8KeJ3tJ3baJncqR0QTjeROZUm7SDpf0lWS1ks6qv8UIyIiIiJmm6RDgBtsXzLpXCIiojPb1wM/kPTguulAqlIYEREzY5DyF7cDr7K9O/AI4EhJuw8nrYiIiIiImbMf8GRJ3wU+Bhwg6aPtndqnU0dExES8DDhV0uXA3sDfLLXBNEyDb1KsUvIoNVYpecxCrFLy6DdWJ30PKtu+zval9c83A1cDO/cbLyIiIiJiltk+xvaKejr1s4Ev2H7uAv3W2F5le9V222039jwjIgJsX1afi/e0/VTb+ZYvImbKUGoqS9oVuBCYs/3TTv1SUzn9B+lfYk6z1r/EnFJTOSIippGk/emypnJqdEZE08zi9XXO1xHRRCOpqTxP0lbAmcArFhpQzvS8iIiIiIje2L5gqQHliIiIiIhJWTbIxpK2oBpQPtX2WQv1sb0GWAPVN3OD7C8iIiIiIjbVPvum06ycWWsvMadpbS8xp2ltLzGnftubTtJRwIsAAf9s+71LbTM3N3eX12WQ5cRqRh6lxiolj1mIVUoe/cbqpO87lSUJOAG42va7+40TEREREREREdEUkuaoBpT3BfYCDpH0wMlmFRExXssG2HY/4DDgCkmX1W2vt/3pgbOKiIiIiIiIiCjTQ4CLbf8cQNK/A08H/m6xjVrv+mu/A3CQ5cRqRh6lxiolj1mIVUoevcbqpO9BZdsXUU3ziIiIiIiIiIiYFVcCx0m6L/AL4AnAusmmFBExXrLHV+Y4TzuNiCaaxadTR0TEZEjaFvgwMAcYOML2Fzv1z/V1RDTRNFxfS/pz4CXAz4D1wK9sv6JT/5UrV3r9+vWNr63apFil5FFqrFLymIVYpeTRT6zFztd911SOiIiIiIihOx74rO3fo6rTefWE84mIiAXYPsH2w2w/GtgIfKO9j6TVktZJWrdx48bxJxkRMUJjv1N5/fr1XfVtHx1vWv8Sc2p6/xJzmrX+JeY0jv7TcCdFRESUT9K9gcuA33GXF+kLXV93+qybtfYSc5rW9hJzmtb2EnPqp30arq8l7Wj7Bkn3B84FHmH7pk79M7MkIpposfP1sn6DSro7cCGwZR3nDNvH9hsvIiIiImLG7Qb8CPiIpL2AS4CjbP9ssmlFRMQCzqxrKt8GHLnYgHJExDQapPzFr4ADbO8F7A0cLOkRQ8kqIiIiImL2LAMeCnzQ9j5UdTpf194p06kjIibP9qNs7257L9uf72ab1ru25+bmBl5OrGbkUWqsUvKYhVil5NFvrE76vlO5npJ3S724Rf1nfLU0IiIiIiKmywZgg+2L6+UzWGBQ2fYaYA1U06nHl15ERAxiviRI+8+DLidWM/IoNVYpecxCrFLy6DVWJ30PKgNI2pxqWt4DgQ+0XABHREREREQPbF8v6QeSHmz768CBwFVLbdfpwj/tk9/3rLWXmNO0tpeYUz/HEBERzTVI+Qts32F7b2AFsK+ku9wbnel5ERERERFdexlwqqTLqUrM/c1k04mIiGGZhmnwTYpVSh6lxiolj1mIVUoe/cbqZKA7lefZvknS+cDBwJVt6zI9LyIiIiKiC7YvA1b1sk37Rf/8XYGz3l5iTtPaXmJO09peYk79tjeBpBOBQ4AbbM/Vbe8EngTcCnwbeEEe0hcRs6jvQWVJOwC31QPK9wAOAt4xtMwiIiIiIiIiIibnJOD9wCktbecBx9i+XdI7gGOA13YTbBpqqzYxVil5lBqrlDxmIVYpefQaq5NByl8sB86vp+Z9BTjP9jkDxIuIiIiIiIiIKILtC4Eft7Wda/v2evFLVOVAIyJmTt+DyrYvt72P7T1tz9l+6zATi4iIiIiIiIgo2BHAZ7rtPA21VZsUq5Q8So1VSh6zEKuUPPqN1Unf5S/61Wv9pKb3H8c+Zq3/OPaR/pPfR2n9IyIixkHS0cALAQNXUNXq/OVi23T6TEv75Pc9a+0l5jSt7SXm1M8xNJmkNwC3A6cu0mc1sBpg+fLlY8osImI8ZI/v2XkrV6702rVrx7a/iIhh2GOPPS6xvWrSeURExHSTtDNwEbC77V9IOh34tO2TOm2T6+uIaKImXV9L2hU4Z/5BfXXb4cBfAAfa/nk3cXK+jogmWux8PfY7lbu5fRp+821mU/uXmFPT+5eY06z1LzGncfWPiIgYk2XAPSTdBtwTuHapDdo/0zp91s1ae4k5TWt7iTlNa3uJOfXb3lSSDgZeAzym2wHliIhpNMiD+gCQtLmk/5Z0zjASioiIiIiYRbavAd4FfB+4DviJ7XMnm1VExOySdBrwReDBkjZI+nPg/cDWwHmSLpP0oW7jTUNt1SbFKiWPUmOVkscsxColj35jdTKMO5WPAq4GthlCrIiIiIiImSRpO+ApwG7ATcC/Snqu7Y+29UuNzoiIMbB96ALNJ4w9kYiIAg00qCxpBfBE4DjglUPJKCIiIiJiNj0O+B/bPwKQdBbwB8Amg8q21wBroKrROe4kIyKiP62lP9rLgAyynFjNyKPUWKXkMQuxSsmj11idDHqn8nupagltPWCciIiIiIhZ933gEZLuCfwCOBBYt9RGnS780z75fc9ae4k5TWt7iTn1cwwREdFcfddUlnQIcIPtS5bot1rSOknrNm7c2O/uIiIiIiKmmu2LgTOAS4ErqK7V10w0qYiIGSbpREk3SLqype3Nkq6p6ylfJukJ3cabhtqqTYpVSh6lxiolj1mIVUoe/cbqZJA7lfcDnlyfQO8ObCPpo7af29op0/MiIiIiIrpj+1jg2F62ab/on78rcNbbS8xpWttLzGla20vMqd/2hjiJ6sF8p7S1v8f2u8afTkREOfoeVLZ9DHAMgKT9gVe3DyhHRERERERERDSR7Qsl7TqseNNQW7WJsUrJo9RYpeQxC7FKyaPXWJ30Xf4iIiIiIiIiImIGvVTS5XV5jO0mnUxExCQMZVDZ9gW2DxlGrIiIiIiIiIiIQn0QeACwN3Ad8PfdbjgNtVWbFKuUPEqNVUoesxCrlDz6jdWJ7PGVOV65cqXXrl07tv1FRAzDHnvscYntVZPOIyIipoOkE4FDqB56PVe33QdYC+wKfBf4U9tLPuU619cR0URNur6uy1+cM3++7nZdvX41sBpg+fLlDzv33HNHmGlExPAtdr5O+YuIiIhA0ockvXHSeUTMiJOAg9vaXgd83vbvAp+vlyMiojCSlrcsPg3oWHzU9hrbq2yv2m67VMmIiOnS94P6+tXN7dOw+NNum9C/xJya3r/EnGatf4k5jat/RJNJ+i6wE3A7cAdwFdVTzNfY/jWA7b8c4v7+Cng+8NvAjcA/2n7nAvncUTf9l+3HD2G/bwbeAPyqpXlP29+p1+8NnAA8BLga+HPblw2634hedXjw01OA/eufTwYuAF7bTbz2z7ROn3Wz1l5iTtPaXmJO09peYk79tjeBpNOozs3bS9oAHAvsX19TmGpmyV90G29ubu4ur8sgy4nVjDxKjVVKHrMQq5Q8+o3VybJF1y6h/k/hzVT/Kby9KdNXIiIiZtCTbP8/SfcGHgMcDzwceMEI9iXgecDlVDUHz5X0A9sfa89nBPtea/u5d0lIuhvwCeC9wD9S/QfwE5J+1/atI8gjolc72b6u/vl6qi9eIiJigmwfukDzCWNPJCKiQAMNKtcea/vGIcSJiIiIEbP9E+CTkq4HviTp721fKekkYIPt/yNpf+CjwPuAV1N9efxi4FaqQdntgXfZ/psO+/i7lsWvS/oEsB/wsYX6d1IPBH8ZOMH2P0jaHLgQ+Jztt/YSi+ouo2XAe109UOJ9kl4NHAB8tsdYESNl25I6PvikrUbn2PKKiIjBtN71134H4CDLidWMPEqNVUoesxCrlDx6jdVJaipHRETMINtfBjYAj+rQ5beAuwM7A28C/hl4LvCweps3Stptqf1IUt1/fduqUyX9SNK5kvbqkOOt9T7fKukhVDVmNweOW2SXT5L0Y0nrJb24pX0lcLk3fULx5XV7RAl+OF+ns/77hk4dU6MzIiIiIiZt0EFlU01pvaS+YyIiIiKa41rgPh3W3QYcZ/s2qjuMtweOt32z7fVUdZkXHAxu82aq642PtLT9GbArVc3l84HPSdp2oY1tXwm8Dfg41V3Th9m+Y6G+wOlU9ZJ3AF4EvEnS/LTVrYCftPX/CbB1F8cQMQ6fpKpFTv33JyaYS0REAJJOlHSDpCtb2vaW9CVJl0laJ2nfbuO11peem5sbeDmxmpFHqbFKyWMWYpWSR7+xOhm0/MUf2r5G0o7AeZK+ZvvCAWNGRETEeOwM/LjDuv9tGbz9Rf33D1vW/4JqoLYjSS+lqq38KNt3PjzP9n+2dPtbSc+nupv53zqEOpnq7uQzbX+z0/5sX9Wy+F+SjgeeAZwG3AJs07bJNlTPhogYqw4Pfno7cLqkPwe+B/xpt/E6TVFM++T3PWvtJeY0re0l5tTPMTTAScD7qR5wPO/vgLfY/oykJ9TL+48/tYiIydKms0AHCFQ9cf0W2+9qa2+t+fawa6+9tqt48x883YyMl9i/xJya3r/EnGatf4k5jaP/HnvscUkeRBpNVj9Y94WtD8aT9PvAxcBetq9YqKay7RV132VUdy7vZvu7ddtFwIdsf7TDPo8A3go82vZ3lsjvauC1tj/ZYf3pVHc77w881fZFXR73a4GH2366pMcDJwK7zJfAkPQ94C9sp6ZyNNbKlSu9du3aSacREdGTJl1fS9oVOMf2XL38OeBE22vrGVFPsv2cpeLkfB0RTbTY+brvO5Ul3QvYzPbN9c+Pp/rP4yZsrwHWQHUS7Xd/ERERMThJ2wCPBo6nGji+YgT7+DPgb6ge5vudtnX3B3YBvkI1UPwyqtIa/9kep+5/GFUd572AJwMnS9rL9i0L9H0K1YP8bgJ+H3g58Pp69QVUDxx8uaQPUZXHAPhCv8cZUYr2L0o7fYE6a+0l5jSt7SXmNK3tJebUb3uDvYKqdNe7qK5l/mCy6URETMYgNZV3Ai6S9FWqJ7N/Knf6REREFOvfJN0M/AB4A/Bu4AUj2tfbgPsCX5F0S/3nQ/W6rYEPAhuBa4CDgT+2/b/tQeoB6PcCz7N9i+1/AdYB7+mw32cD36IqaXEK8A7bJ8OdD/17KlU5jpuAI6juer514KONiIiIWfJi4GjbuwBHAyd0u+E01FZtUqxS8ig1Vil5zEKsUvLoN1Ynfd+pXN951M0DeiIiImKCbO/aRZ/DW36+AFjRsnw7oLb+f7hIrN0WWbce2HOpfOq+36canG5te9Yi/Q/ttK5e/99Udz1HRERE9Ov5wFH1z/8KfLhTx7ZyoKPPLCJijPoeVI6IiIiIiN5JOhE4BLihpUbnO4EnAbcC3wZeYPumiSUZERGdXAs8hqq01gHAYg8R3qQcaGvpj/YyIIMsJ1Yz8ig1Vil5zEKsUvLoNVYnQ3tQXzdSmD4imqhJDxKJiIjySXo0cAtwSsug8uOBL9i+XdI7AGy/dqlYub6OiCZqyvW1pNOoHhS8PfBD4Fjg61TPplgG/BJ4ie1LloqV83VENNFi5+tBaipHRERERESPbF8I/Lit7dy61AzAl2gpQRMREZNh+1Dby21vYXuF7RNsX2T7Ybb3sv3wbgaU501DbdUmxSolj1JjlZLHLMQqJY9+Y3UyUPkLSdtS1Q+aAwwcYfuLi23TTVKw+NNum9C/xJya3r/EnGatf4k5jat/00k6mOqOis2BD9t++4RTioiIzo4Aur6drf0zrdNn3ay1l5jTtLaXmNO0tpeYU7/tERHRbAMNKlMNUHzW9jMk3Q245xByioiIIZK0OfAB4CBgA/AVSZ+0fdVkM4uIiHaS3gDcDpw66VwiImK4pqG2ahNjlZJHqbFKyWMWYpWSR6+xOul7UFnSvYFHA4cD2L6V6sEiERFRln2Bb9n+DoCkjwFPAToOKm+33Xa+3/3uN6b0IiKG46qrrrrR9g6TzqNfkg6neoDfgV7kwSeSVgOrAZYvXz6e5CIiIiIiWgxyp/JuwI+Aj0jaC7gEOMr2z4aSWUREDMvOwA9aljcAD19sg/vd736sX79+k7ZJT5HMNNLpbS8xp3G1H73nWzZpf8/lx450v6M4hpLa99hjj+/RUHWZotcAj7H988X62l4DrIHqwU9jSC8iYiZJ2gU4BdiJquTnGtvHS3om8GbgIcC+ttd1E29ubu4un2ODLCdWM/IoNVYpecxCrFLy6DdWJ8sWXbu4ZcBDgZfZvljS8cDrgDcOEDMiIiIiYqpJOg3YH9he0gbgWOAYYEvgPEkAX7L9l93E63TBn/bJ73vW2kvMaVrbS8ypn2NogNuBV9m+VNLWwCWSzgOuBJ4O/FMvwVpfi/bXZZDlxGpGHqXGKiWPWYhVSh69xupkWVe9FrYB2GD74nr5DKpB5U1kel5ExMRdA+zSsryibttEztcREeNh+9AFmk8YeyIREbEo29cB19U/3yzpamBn2+cB1F8CRkTMpL4HlW1fL+kHkh5s++vAgSxQnzPT8yIiJu4rwCMl3Qj8kOqOi+e0d8r5OiKimUoqQVJSe4k5TWt7iTlNa3uJOfXb3jSSdgX2AS5eomtH0zANvkmxSsmj1Fil5DELsUrJo99YnSxbdO3SXgacKuluwHeAFwwYLyIihsz27ZLeBhwJPBB4q+31S2zW8QNkWttLzGnW2kvMadTt8zWUx73fcexjku+jiIiIYZK0FXAm8ArbP+1hu8wEjIipNdCgsu3LgFXDSSUiIkbF9t9JOh04x/Zxk84nIiIiIqIJJG1BNaB8qu2zetm2fSZg6xei7V+ODrKcWM3Io9RYpeQxC7FKyaPXWJ0MNKgcERHTq7QpkplGOr3tJeY0re0l5jTs9iaQdCJwCHCD7bm2da8C3gXsYPvGSeQXEREVVUWTTwCutv3uSecTEVGSsQ8q93rB3/T+49jHrPUfxz7Sf/L7KK1/RETEEJ0EvB84pbVR0i7A44Hv9xKs02da2ie/71lrLzGnaW0vMad+jqEB9gMOA66QdFnd9npgS+AfgB2AT0m6zPYfLRVsGmqrNilWKXmUGquUPGYhVil59Burk2WLro2IiJmRmm8REeNh+8L6gU/t3gO8BvjEeDOKiIiF2L4IUIfVZ48zl4iI0vQ9qCzpwcDalqbfAd5k+72LbbfQNMyFtI+ON63/IPs4es+3dNV//uE+z/vAeV31P+XIg/rKp9f+o8pnkJxG/Zr2Gr/X/tPwvm56/yar73w7BVgB7CLpKNvHt/drr/k23iwjImabpKcA19j+ajXbunullSAppb3EnKa1vcScprW9xJz6bZ9Frcfe/joMspxYzcij1Fil5DELsUrJo9dYnSzrqtcCbH8d2BtA0ubANeSbuoiIEt0O/ArYCtgceJekrW2/bbGNOn2QTGt7iTnNWnuJOU1re4k5DfPYmkbSPammUz9+0rlERERERHRjsyHFORD4tu3vDSleREQMie3rbB9se7ntLYBPAxdPOq+IiLjTA4DdgK9K+i7VzJJLJf3WQp0lrZa0TtK6jRs3jjHNiIjZImkXSedLukrSeklHta1/lSRL2r6beK13bc/NzQ28nFjNyKPUWKXkMQuxSsmj31id9H2ncptnA6cNKVZERIxIXcNzH7oYVG7/EJm/K3Da2kvMadbaS8xpWttLzGnY7U1k+wpgx/nlemB5le0bO/RPuaKIiPG4HXiV7UslbQ1cIuk821f1+3DViIhpMfCgsqS7AU8Gjhk8nYiIGBVJWwFnAq+w/dNJ5xMRMasknQbsD2wvaQNwrO0T+o1XWgmS0tpLzGla20vMaVrbS8xpGssV2b4OuK7++WZJVwM7A1fRx8NVW1+L9tdlkOXEakYepcYqJY9ZiFVKHr3G6mRZV70W98fApbZ/uNBKSauB1QDLly8fwu4iIqJXkragGlA+1fZZHfrkfB0RMQa2D11i/a5jSiUiIrrUOuNvkIerRkRMi2EMKh/KIqUvMj0vImKyJN0d2AAYWCHp3raPbe+X83VERDOVVoKklPZBYu2+/dM3ab/qxrMmcgz9th+951s2aX/P5ceOdL+jyLXXYxhWnNLbS8yp3/YmaZ3xR1USo6+Hq87Nzd3ldRlkeZZjzf/bbv833eRjGnesfrdd6LUfZNtOeXXTd5SvzzBjlZJHv7E6GWhQWdK9gIOAvxgkTkREjNQq4L7AFcCvgVdLusn2exbbqNMHyLS2l5jTrLWXmNO0tpeY0zCPLSIiYljaZ/xJ2oPfPFwVfvNw1X1tX9+2bWYCRsTUGmhQ2fbPqAYqIiKiULYvAgQg6Z7ARcB/TTSpiIiIiIjCqRo1PgG42va7obeHq7bPBGz9QrT9y9FBlmcx1vzdq5POYxpi9brtYq/9INu2L/fSd6nlpr7WpcTqZKBB5YiIaAZJmwOXAA8EPmD74qW2KW2KZKaRTm97iTlNa3uJOQ27vQkknQgcAtxge66l/WXAkcAdwKdsv2ZCKUZERGU/4DDgCkmX1W2vt/3pyaUUEVGGsQ8q93rB3/T+/WzT/m3MUk458qCe+o/6mEedTz/bjPo17TV+r/2n4X3d9P5NZ/sOYG9J2wJnS5qzPVsvQkREOU4C3g+cMt8g6bHAU4C9bP9K0o4dtr2LTp9pae9/m/kayqPKadTtna41S3ytO+Xa6zEMK05T2kvMqZ9jKF3rjL9F+uzabbxpqK1aaqxB6vyWckz91BAuIY+j93zLQHWPh3lMpcYqJY9+Y3WybNG1ERExVWzfJOl84GBgk0+I1HyLiBgP2xdK2rWt+cXA223/qu5zw9gTi4iIkWodoGkfrBlkObEGK8lQyjENUu5hknm09m/Kaz2pWKXk0WusTgYaVJZ0NPBCwFQPgHqB7V8uts1C0zAX0j46Xkr/533gvK76w2/udm1/4nAnnZ5EvFROpcUfVT6tOY36GLr9Pc//jpv+Oyvxd1Dav/0mk7QDcBtwM1UJjF2A57f3a6/5Ns4cIyKCBwGPknQc8Evg1ba/0s2GpZUg6bW9/Rqh0zVlp2uExa41ltqm1/bdt3/6Ju3zdzQ35TWdZGmdQX+foz6G0l7TQV7rprRHRESz9T2oLGln4OXA7rZ/Iel04NlU0/kiIqIcy4GTgZ2AewI/tH3OUht1uuCf1vYSc5q19hJzmtb2EnMa5rE11DLgPsAjgN8HTpf0O7bzJV9ExIRI2oWqVNFOVDfTrbF9vKS1wIPrbtsCN9nee6l40zANfpCyE+Ms79BakqGf/Q4rj1GWnSixJENTXutxxyolj35jdbLZomuXtgy4h6RlVAMV1w4YLyIihsz25cCTgKuBpwPfnGxGERGxgA3AWa58Gfg1sP1CHSWtlrRO0rqNGzeONcmIiBlzO/Aq27tTfel3pKTdbT/L9t71QPKZwMKF2CMiptiyfje0fY2kdwHfB34BnGv73KFlFhERw/Re4DXA1t1uUNoUyUwjnd72EnOa1vYScxp2e4N9HHgscL6kBwF3A25cqGPKFUVEjIft64Dr6p9vlnQ1sDNwFYAkAX8KHNBNvNbPqvbPrUGWJxmrl1rG46wZ3LqvQfY7aB69xCqlpvIgy015rScVq5Q8eo3VSd+DypK2o3pC9W7ATcC/Snqu7Y/2GzMiIoZP0iHADbYvkbT/hNOJiJh5kk4D9ge2l7QBOBY4EThR0pXArcDzuy190enCvynt7f8BXaq91zj95NSpfb6G8qjij/o1HfV+h5nTpI6hKflMct/DPIYmqR+wug9wcUvzo6hKy2UmYETMnEHKXzwO+B/bP7J9G9V0jz9o75TpeRERE7cf8GRJ3wU+Bhwg6S5fAOZ8HRExHrYPtb3c9ha2V9g+wfattp9re872Q21/YdJ5RkRERdJWVGUuXmH7py2rDgVO6zZO6yybubm5gZdLjNX6UMtJ51VKHqXGKiWPccQ6es+3DPzeLO2Yxhmrk77vVKYqe/EISfekKn9xILCuvVOm50VETJbtYyQdCtxMVRfu7rafu0C/nK8jIhqo/aJ/0iVISmkvMadht7f+Bxnu+mCkceVTYk6l5pP3dfPuXJa0BdWA8qm2z2ppX0b1vJKHLbLtamA1wPLly0ecaUTEePU9qGz7YklnAJdSDVL8N/VgREREFOmxwBzw6m46lzZFMtNIp7+9xJymtb3EnGZ1OnVERJSrrpl8AnC17Xe3rX4c8DXbGzpt337TRutnV/vn2CDLJcVqLRVTSl6l5FFqrFLyGGWsUmo9l/r6LNW3k2Vd9erA9rFUNeAiIqIBbF8AXDDhNCIiIiIimmA/4DDgCkmX1W2vt/1p4Nn0UPoiImLaDDSoHBERjWHgXEkG/qm+a2JRpU2RzDTS6W0vMadpbS8xp2G3N4GkE4FDqB6iOle37Q18CLg71SzAl9j+8sSSjIgIbF8EqMO6w3uNNzc3d5fPsUGWE6sZebQuH73nW+5ShmfSr898maBB8hrFa11qXk3Oo99YnSxbdO0I9HrBX1r/U448qKf+sPiTqBfSa06lxR91Pv1s02tOvf6em/47K/F3UNq//Snwh7avkbQjcJ6kr9m+cNJJRUTMqJOA9wOntLT9HfAW25+R9IR6ef9ugnX6TEv75Pc96vZO11eTfK1Ly6m0fPK+joiIaTH2QeWIiBg/29fUf98g6WxgX2CTQeU8SCQiYjxsXyhp1/ZmYJv653sD1441qYiIGLnWAfb2wfZBlhOrGXlAmTWnS6k33JS8piGPXmN1MtCgsqSjgBdRTQf5Z9vvXWqb9qfvdtLpqbydzB/w8z5wXlf95+9EHVX81n30esyj7t/razTq/t3+DuA3v4dRH3Ov74tZ/B2U9pr2Gr/X/k0m6V7AdsB7gT2AXYDXt/drf5DIGFOMiAh4BfA5Se8CNgP+oNsNSytBMqz29muBTp/5i107LLVN6cfQ6Xp/0sfVT2mdSR3btMXvZh+lv1+m4fo6IiIGGFSWNEc1oLwvcCvwWUnn2P7WsJKLiIih2AlYB/wM+Cnwdqqp14sa9ZTH0tpLzGnW2kvMaVrbS8xpmMfWUC8GjrZ9pqQ/BU4AHjfhnCIiZpqku1PN7tuSavzkDNvHStoN+BhwX+AS4DDbty4VbxpqqzYpVr/bLlTLd5Btm/b6LHUMg7w+pdRU7if2JF7rUt4/nWy26NrFPQS42PbPbd8O/Dvw9AHiRUTEaPwv8BPg/rZX2n6r7ZsmnFNERGzq+cBZ9c//SnXjxoIkrZa0TtK6jRs3jiW5iIgZ9SvgANt7AXsDB0t6BPAO4D22HwhsBP58cilGRExG33cqA1cCx0m6L/AL4AlUd8JFRERZdgN+BHxE0l5Ud1McZftni21U2hTJSU7ZTXt+B9PSXmJOw25vsGuBxwAXAAcA3+zUMeWKIiLGw7aBW+rFLeo/pjpPP6duPxl4M/DBpeK1fla1f24NspxYw912sVq+g2w7aF7jen2WOoZBXp9elkdZU3mQ2NOQR6+xOul7UNn21ZLeAZxLNaX6MuCOfuNFRMTILAMeCrzM9sWSjgdeB7xxsmlFRMwmSacB+wPbS9oAHEtVVu54ScuAX1I/OLUbnS78m97e/h+tfvsPM6dJHUOvcSbVvti6SR3btMafZE7DfL80gaTNqW7KeCDwAeDbwE31jG2ADcDOE0ovImJiBil/ge0TbD/M9qOppnx8o71PpudFREzcBmCD7Yvr5TOoBpk3kfN1RMR42D7U9nLbW9heUV9TX1RfV+9l++G2L5l0nhERAbbvsL03sIKqNNHvdbtt+/V16yybubm5gZe77Xv0nm/Z5AGOg8TqJvYgeQwrr1Je615en3EfYyl5THusXt/zJR5TJ33fqQwgaUfbN0i6P1U95Ue097nL9LzrB9ljRET0yvb1kn4g6cG2vw4cCFy1QL9Mp46IaKDW/6zAXR/2Mm/YpUYmtd9u2ye572G/RqW91u35dJNTU16L0vPsZ5tJv6ad2pvG9k2SzgceCWwraVl9t/IK4JoO22xyfT2pqeujLMnQSymEUkoOjPO1HlepiKa8PrMYq/U90NRj6mSgQWXgzLqm8m3AkXnwU0REeSQ9GNgR+G9Jojr3v2mp7UqbIjmOKZil5TRr7SXmNK3tJeY0zGOLiIgYBkk7ALfVA8r3AA6iekjf+cAzgI9RPWj1E5PLMiJiMgYtf/Eo27vX0/Q+P6ykIiJieGx/3faDbd8T2Ar4X+BfJpxWRERERETplgPnS7oc+Apwnu1zgNcCr5T0LeC+wAndBCtlOv6sxOp325JLVvQba1LHNMqSJ6WUdSn1dz7MWJ0MeqdyREQ0y4HAt21/b6mO7R8iw57yWEr7OPe9+/ZP36T9qhvPAsqb1txr/pP8HXR67XptH1b80n5nk3y/T6o9IiJiWGxfDuyzQPt3qOorR0TMLNnjK5u5cuVKr127dmz7i4gYhj322OMS26smnccwSDoRuNT2+xfrt3LlSq9fv36TtkkPFE3DIFsGlTOoPOn2EnMaZntTzteSdgFOAXYCDKyxfbyk+wBrgV2B7wJ/anvRJ6fm+joimqgp5+thyvk6IpposfP1QOUvIiKiOSTdDXgy8K8d1m/ydOqIiBiZ24FX2d6d6kHXR0raHXgd8Hnbvwt8vl6OiIiIiCjO2MtfPO8D53XV75QjD+qrfzc1P+A3d7kcsaG7sqInrngOcNe7tBbT6Q60Tubvchr1MfeaT6/9e82n2/6t24z6Ne31fVFa/FHl05pT09/XvcafEn9MdZfyDxda2f506nEmFhExS2xfB1xX/3yzpKuBnYGnAPvX3U4GLqCq27moYd05P+o7/Mc9I2AUOU3qzvxJ5d/t72aQfZT+PhrXa9pt/0H2XVp7E0i6O3AhsCXV+MkZto+VdAKwChDwDeBw27csFe/oPd9yl99f++vUy3KndfPvm1721W8eC+1rErGGmccgr083r30Jv7eS8mjf97D+jYzr39uk8ujn99TvMXWy5KByPVX6EOAG23N1W89T8yIiYnIkHQ28GbhZ0mnAC2z/crFtOn2ATGv7OPYx/2Vju/kLgXHnM6z8J/k76PTa9do+rPil/c4m+X6f5LE1iaRdqep1XgzsVA84A1xPVR4jIiIm51fAAbZvkbQFcJGkzwBH2/4pgKR3Ay8F3j7BPCMixm7JQWXgJOD9VHXf5s1PzXu7pNfVy0veRREREeMnaWfgKKrp1iuBfwaeTXV+j4iICZG0FXAm8ArbP5V05zrblrTgrBFJq4HVAMuXL+d+40g2ImIGuXoI1fwdyFvUf9wyoCzgHlT18ZfU+qV0+5ejgyy3r2v/8nuQWEstL7avccYaZh6DvD69vPaT/L2VmEf7vkuKNaz39TDzGOT31L68VN9OlhxUtn1hfQdFq76m5kVExMRsDvwu8DPgnsC1S21Q2hTJSU7ZHfU06ElNsy7lIXSD5Dpr06DH8VqXkmu/7U1R3/F2JnCq7flpAD+UtNz2dZKWAzcstO1dyhVdP5aUIyJmkqTNgUuABwIfsH1x3f4R4AnAVcCrJpdhRMRkdHOn8kIyNS8ioiFsXyPpXcD3gV8A59o+d8JpRUTMrPrOthOAq22/u2XVJ4HnU02hfj7wiW7iDascy6jLxkyqzMwwc5pUWZdJ5d/r76affTTlfTSpOON4rSfV3hS27wD2lrQtcLakOdtX2n5BPeD8D8CzgI8sFWtU9VAHjTWNdWwnVVN5UrF6fQ9Mqs5vU2pOD7tG97B+55P699XJZouu7UI9HaTjVA9JqyWtk7Ru48aUXY6IGDdJ21HNMNkNuB9wL0nPXaBfztcREeOxH3AYcICky+o/T6AaTD5I0jeBx5H6nBERxbB9E3A+cHBL2x3Ax4A/WWibXF9HxDTr907lrqbmwQLT8yIiYtweB/yP7R8BSDoL+APgo62dcr6OiBgP2xcB6rD6wF7jlVaapNd8dt/+6Zu0zz8UdND9drPv0soPlVZOpp/SOpMqVzTq992k8uznfd2Ucl5NIGkH4DbbN0m6B3AQ8HeSHmj7W/XMkycDX1to+/br69ZjH2Y91EFjTWMd22Ft25RYvb4HJlXntyk1p0dVo7up/7466XdQua+peRERMRHfB54gaX29/DPg/y61UacPkqa0D3PK7qinQU9qmvWopwr32r7YukyDHm57P9uUVg6gSYMSERHRWMuBk+syF5sBpwOfAv5D0jZUXxB+FXjx5FKMiJiMJctfSDoN+CLwYEkbJP05mZoXEdEkP6N6UN8yqnJFuwCfn2hGERERERGFs3257X1s72l7zvZbbf/a9n6296jb/sz2T7uJ13o399zc3CZ3cfezPKxYveR19J5v2aR/+3Iv2w5yjN3E6vf1mJVYw8pjkPfPuF6fft57/R7jMLct5ffWyZJ3Kts+tMOqnqfmRUTERDwE+Dfbfw4g6Y3AIVRPqu5oUlMkh7XfbuN3s49JTXnt1D6sKbvDijPoe2KQfYxq2vykp02P6t/HIMcw6td0WO0RERGlKrE0RK95tc9c6qVMwDDLFwxSRqGUkhWlxJrk+2dcr88kS8RMw7/7TpYcVB62U448aKT9e/0PxYkrntNT//n/cPVisanWCxn1MfeaT6/9e82nn/8Ejvo17fV9UVr8UecDzX9fz9jgw5XAcZLuC/wCeAKwbrIpRUTMLkm7AKcAO1HNIFlj+3hJ7wSeBNwKfBt4Qf1gqEWVVpqk13w6XV8Pa7/9bFPaa9SUdphcuaJRv+8mlWc/7+tJve96jR8REc22ZPmLiIhoNttXA+8AzgU+C1wG3NHeL0+njogYm9uBV9neHXgEcKSk3YHzgDnbewLfAI6ZYI4RETFkpU7tLzXWsPIYZkmGYb4+kyyFMOjvpduSJ8M8pm5Kr3R7jKWUYhlmHqN8/3Qy9juVF5p2vJBOUyQ76TTVcqn4o+rfuk2vx/C8D5zXVf/5Oz9Hfcy95j+q31lrTqN+jXqNP+r+R2z4l676z99xPKp8WrcZ9fuu12Me9bmi6WyfAJwAIOlvgA0L9Nnk6dRcP9YUIyJmhu3rgOvqn2+WdDWws+1zW7p9CXhGN/F6LREy6vJDve63U/9BS7QMM1apr+mofwfdxhkk1rhzHXb5oXGXyGrdZlT/dob9O+jU3gSS7g5cCGxJNX5yhu1jJQl4G/BMqps1Pmj7fZPLNCJi/JYcVJZ0IlXtzRtsz9VtzwTeTFWnc1/bmUYdETFhHc7X9wHWAg+gmkr9SuDpVHfGLWpSUySHtd9e4w8z1qinnQ5ryu6w4kxy2vSop81Patr0qP99wPBeu9Km9zeNpF2BfYCL21YdQXX+joiIyfkVcIDtWyRtAVwk6TNUYyG7AL9n+9eSduwmWKn1YkuNNaxatL3sZ6n+w3x9hvkeGOdr20sd7WEeUy/1vNvXD/I77zWvUW27VKxRvn866ab8xUnAwW1tV1INSlzY1V4iImIcTuKu5+vXAZ8HrgH2rH8+spsanRERMVqStgLOBF5h+6ct7W+gKpFxaoftUq4oImIMXLmlXtyi/mPgxcBbbf+67nfDhFKMiJiYJe9Utn1hfQdFa9vVANWMj4iIKMFC52vgKcD+tt8uaTlwge3PdxNvWFMnRx2n09TMTv3b27vJadRTXoc17bS0adb9TJse9TTiXt+Pw3rf9Xpco8qzm32XNr2/1/hNUd/xdiZwqu2zWtoPp5p1cqBtL7TtXcoVRUTEyEjaHLgEeCDwAdsXS3oA8CxJTwN+BLzc9jeXijU3N3eXz71Oy/Off+2fe639u43VzXKJsY7e8y2LHv+o8ljotS/h9Rkkr3EdUzfv26Xy7LRtr7F7ef+M+t/bJN4/wzymTpYcVI6IiEbbqa7dCXA9sNMkk4mICKhrcZ4AXG373S3tBwOvAR5j++fdxut0wd+pfdTlVXrdb6f+wyrRMsxYpb2mo/4d9Bqnn1iTynVS7cM63sXWTao0VD//NpvA9h3A3pK2Bc6WNEdVY/mXtldJejpwIvCo9m0lrQZWAyxfvnx8SUdEjMHIB5XbT6L3G/UOIyJiQbYtqeMdbXe96N1qXKlFRMya/YDDgCskXVa3vR54H9VAxXn1jMAv2f7LiWQYERGbsH2TpPOpys1tAOZH788GPtJhm01mlvRSw3SQOq2DLJcUa1w1lXt57Sf5+gyS17iOaZR1tHuN3cv7Z1z/3sb5/hnmMXUy8kHlu0zPu37Ue4yIiBY/lLTc9nV1+YuO9d7uOp36tnHlGBExU2xfBCxUR+7T/cQrpZRNaSV3puEYmvK7HMW+J1Wip/T9tm5Tyu+/3/hNIGkH4LZ6QPkewEHAO4CPA48F/gd4DPCNiSUZETEhIx9UjoiI8ZB0IvBkNr3F+GvAZfUF8QeAT3Qbb1hTJ0cdp9N/TDr179S+2L5HPeV1WNNOS5tm3Wuei60b1mvR6/txWO+7Xo9r1Hkutm5Y/w5G/X5p+nTqiIhohOXAyXVd5c2A022fI+ki4FRJRwO3AC/sJtg01OYddR3kYR1jKa91ak6Pt050t8fYT+3n0vMY5XugkyUHlSWdBuwPbC9pA3As8GPgH4AdgE9Jusz2Hy0VKyIiRur+VE+j3rLlfP23wHuArYF9gT+eXHoREREREc1h+3JgnwXabwKeOPaEIiIKsuSgsu1DO6w6e8i5RETEAGw/TtKuwDm2W+cfPlLSBcCrbf+423iTmjo57mmqo8h1WFNhmzJde9Dj7SbWsPa9+/ZP36R9/g7fUf+Oe81nVPl3k9Ow3ne9HtssTaeOiIjZ1HrXX/sdgIMs97rtKOvrDqsO8jDzKiWPYceahprTo6oTvVSsQWo/l5rHKN8DnSw5qDxsvV7wd3sg/cYfdX/o/RhOOfKgnvqP+hh6zX/UvzMY/WvUa/xR9z9xxXNGGr/X/jD6912vxzyO911ERMQwSNoFOAXYiWqGyRrbx7esfxXwLmAH2zcuFa/TZ+Co25tScmexdU05hqb8Loe570mV6GnKfhdbV9q/5Vx3R0RMp80mnUBERJRB0mpJ6ySt27hx46TTiYiYZrcDr7K9O/AI4EhJu8OdA86PB74/wfwiIgKQdHdJX5b0VUnrJb2lbj9A0qWSrpR0sqSubthrnX0zNzc38PI4Yh2951s2mR00qbyWymOh9aPIY5pijSKPft4v7dsPK69B8xjW+6mUYxrk330nY79TuX0KYyedpjwOu383LxIsPmV32Dn12v95Hzivq/7zd6NOw2va6z5GlVO/+fT6Oxv17/iIDf/SVX/4zR3Eo36flvY7ngW21wBrAFauXOkJpxMRMbVsXwdcV/98s6SrgZ2Bq6jq4L+GHh6s2v6ZtlTpkEm1j7o00EKf7cPOtbTXdFj5D3q8o8iplGMrbb+j2Pek3u8N8SvgANu3SNoCuEjS54CTgQNtf0PSW4HnAydMMtGIiHFbclBZ0onAIcAN8zU6Jb0TeBJwK/Bt4AV1ofqIiJiQ+nz9ZGCrlrb58/UK4J2Sntbt+XpSUycnNU11sW2GNeV1Uq/RpKZN9zOVd9TTjufr+o4q/rDyGXX+i60b1vuu12Ob1enUdS38fYCLJT0FuMb2VyVNNrGIiMC2gVvqxS3qP3cAt9r+Rt1+HnAMXQwqT6oe6iCxSqn5ulQeo6wTPe2xxvl7Wmp5mDW5B4k1qrrRkzymQf7dd9JN+YuTgIPb2s4D5mzvCXyD6gQaERGTdX+q2pxbStog6c+Bn1ENMm8B7AtcOsH8IiKihaStgDOBV1CVxHg98KYutku5ooiIMZG0uaTLgBuoxkK+DCyTtKru8gxglwmlFxExMUveqWz7wvoOita2c1sWv0R1Eo2IiAmy/bj6fH3O/MyS2psBJD2NHs7XpU3lHdU01RJy3X37p2/SPn+HZ1Nf01H8DkY9NbfX+L3+zoY17b9T/27zaY016tIFk5p+3RT1NOozgVNtnyVpD2A3YP4u5RXApZL2tX1967YpVxQRMT627wD2lrQtcDawEng28B5JWwLnUt29vKS5ubm7fI4NsjyJWPOf1+2f05POq+Q8uo3VzWvb3r/bvqPMY1SvTz95DCuvhfY9be+fXvLqZMlB5S4cAawdQpyIiBitnK8jIgqgatT4BOBq2+8GsH0FsGNLn+8Cq2zfuFS8Thf8pbWPujTQYv/xGVaupb2mw8p/WMc7jn1M6thK2+8w9z2p93vT2L5J0vnAwbbfBTwKQNLjgQcttI2k1cBqgOXLl2/yWrS/LoMsjzNWL+UNSjnGUvJYKlavpSOGVQph0DxG9foMWkpjkLxGVU6lpPdPL3l10k35i44kvYFqqt6pi/TJ9LyIiAnL+Toioij7AYcBB0i6rP7zhEknFRERm5K0Q32HMpLuARwEfE3SjnXblsBrgQ8ttL3tNbZX2V613XbbjSnriIjx6PtOZUmHUz3A78C6eP2CMj0vImKycr6OiCiL7YuARZ/EZ3vXbuOVViqnlPZR7GPUpXia2t66rrTXaFj5lBKnm3005TVtiOXAyZI2p7op73Tb50h6p6RD6rYP2v5CN8GaWJKhybFKyaPUWO3/pqetlMag5UGGmVfTX59O+hpUlnQw8BrgMbZ/3k+MiIgYLkknAk+mejDffNu/AE8Hvgl8XNLhtq/tJl6nD5CmtPc6lXMcOXVqn69/W0o+w3pNh/k7GPXU3F7j9/o7G9a0/079e81nsXXDOoZJTr+PiIgYBtuXA/ss0P5XwF+NP6OIiHIsOags6TRgf2B7SRuAY4FjgC2B8+oHiXzJ9l+OMM+IiFja/QEDW7acrx8O/Ijq4SG7Uz2xeuXEMoyIiIiImFGtX4i2fzk6yHJiNSOPUmONsnbxJOszD6vm9DDzaurr08mSg8q2D12g+YSuokdExNjYfpykXYFzbM/PM7zzfC3pGKqB566UMqV21O0l5jTp9t23f/om7fN3wTbpd9Bpau6op2WXPu27hJxG3R4REREREaO35KDysPV6wT/q/os9TXcY8fvZptf+pxx50Ejjl/ia9rqPUefUa/xef2ej/h2fuOI5PfXvZx+9HkNpv+Mmk3Qc8DzgJ8BjJ5xORMTMk7QLcAqwE9UMkzW2j6/XvQw4kmqGyadsv2apeJ0+09I+/FijLsXT9HYo7zUaVj6lxelnm9Je0yapayqvA66xfYikU4FVwG3Al4G/sH3bUnFKqqc7C7Hav2QfRc3gJscqJY/W5ZLqIHeKNck6yJOK1clmi66NiIjGs/0G27sApwIv7dRP0mpJ6ySt27hx4/gSjIiYPbcDr7K9O/AI4EhJu0t6LPAUYC/bK4F3TTLJiIi401HA1S3LpwK/B+wB3AN44SSSioiYpGXj3uFCU14X0j463rT+JeY0rv7P+8B5XfWfv3O1xNd0VMfQ/s3WUjpNUV4q/qjyGWdOpfWfEqcCn6aqtXwXttcAawBWrlzpMeYVETFTbF8HXFf/fLOkq4GdgRcBb7f9q3rdDd3EK60ESSnto9hH6SVuJpXPKPYxqWObVJmkbvc7yX0P+7VuCkkrgCcCxwGvBLD96Zb1XwZWdBOr9djbX4dBlhNr8eVR1gyehlil5AFl1kFuX55kHeRJxepkyTuVJZ0o6QZJV7a0/bWkyyVdJulcSffram8RETEykk6kmpb3wJa2363/fhXwPeA7k8kuIiIWUtfC3we4GHgQ8ChJF0v6d0m/P9HkIiIC4L3Aa4Bft6+QtAVwGPDZMecUETFxy7rocxLwfqq6b/PeafuNAJJeDrwJ+MuhZxcREb24P1Vtzi0lbaC6I/kJklZS3QH3C+D13Qbr9O3ktLaXmNOk2ucfzDfu/Q4zVqd6j6Ou9dmUWqKTzGkc76OmkLQVcCbwCts/lbQMuA9VSYzfB06X9Du2M3skImICJB0C3GD7Ekn7L9DlH4ELbf9HN/GmobZqU2JNa23eYeZVwu9p2L/z9tezlN/bsI5pUu/rTpa8U9n2hcCP29p+2rJ4L6pBjIiImCDbj6MahFhve4XtE2z/CXAl8IfADcD1k8wxIiIq9d1tZwKn2p7/JmcDcJYrX6a6K277BbZNDfyIiPHYD3iypO8CHwMOkPRRAEnHAjtQl8RYSM7XETHNurlTeUGSjgOeB/wEeOzQMoqIiKGR9BSqp1R/VVJP25ZSE3PU7SXmNGvtJeY0re0l5jTs9iZQdUI+Abja9rtbVn2c6rr6fEkPAu4G3Ni+fWrgR0SMh+1jgGMA6juVX237uZJeCPwRcKDtu5TFaNl+k/N162dV++fWIMuJddflaa/NO8y8Sn19BvmdjzKvcdZUbt/XpN7XnfQ9qGz7DcAbJB0DvJQOD36KiIjJkHRPqnIXj590LhERsYn9qGpwXiHpsrrt9cCJwIn1s0xuBZ7fTemLThf+aR9+rKaUuMlrXe5+h1WeaZL7Hub7pcE+RPW8ki/WN26cZfutk00pImK8lix/0YVTgT/ptDLTPSIiJuYBwG7AV+speyuASyX91kKdc76OiBgP2xfZlu09be9d//m07VttP9f2nO2H2v7CpHONiIiK7QtsH1L/vMz2A1rO4V0NKLfOspmbmxt4ObGakUenWEfv+ZY7a+ZOIq9+t10o71KOqdTfW+nHtFTfTvq6U1nS79r+Zr34FOBrnfpmel5ExHhIOhF4MrAVgO0rJP0j8CLgR8D9gNfYXrCucs7XERHN1H7RP+kSJKW0l5jTsNtb/1MLd31o0LjyaeIxlPLazfL7OiIimm3JQWVJpwH7A9tL2kBV5uIJkh5M9fCQ7wF/OcokIyKiK/enenDqli3na4D32H5Xfbfy/+s2WGnTl5s0ZTft/bWXmNO0tpeYU6ZTR0TEtGv97Gr/HBtkObGakUf7clNrKi+WdynHNMpY46ypXEqsTpYt1cH2oQs0n9BV9IiIGBvbj5O0K3CO7TkASW9uWb/rZDKLiIiIiIiIiGmy5KByREQ03kslPQ9YB7zKdlcFk0ubIplppNPbXmJO09peYk7Dbm8CSbsApwA7Uc0wWWP7eEl7Uz386e7A7cBLbH95YolGRAQAkjanupa+xvYhkk4CHgP8pO5yuO3LloozNzd3l8+xQZYTqxl5lBqrlDy6jTVftqi9XNGk82pSHv3G6kRdPFB6aFauXOm1a9eObX8REcOwxx57XGJ71aTz6MYCdyrvBNxINWjx18By20csFWflypVev379Jm2THijKINv0tpeY07S2l5jTMNubcr6WtJzqfHyppK2BS4CnAu+lKln0GUlPoKqDv/9isXJ9HRFN1JTz9TxJrwRWAdu0DCqfY/uMbmPkfB0RTbTY+XqzcScTERHjY/uHtu+w/Wvgn4F9O/WVtFrSOknrNm7s6mbmiIjog+3rbF9a/3wzcDWwM9UXgNvU3e4NXDuZDCMiYp6kFcATgQ9POpeIiJIsW6qDpBOBQ4Ab5u98a1n3KuBdwA62b+xmhwvdMbOQxe6waUL/EnNqev8Sc5q1/iXmNK7+TVCfr58MbNXSthx4BnAk1QDFTZ22t70GWAPVnRSjzDUiIir1DJN9gIuBVwCfk/Quqps//qCbGCXdLV5Se4k5TWt7iTlNa3uJOfXb3iDvBV4DbN3WfpykNwGfB15n+1dLBeplunk3y4nVjDxKjVVKHrMQq5Q8+o3VybJF11ZOAt5PVfftTnUtuMcD3+8iRkREjN79qe5y21LSBuBY4FlUgxL/A3wFeH23wTp9gExre4k5zVp7iTlNa3uJOQ3z2JpC0lbAmcArbP9U0tuAo22fKelPqR6O/biJJhkRMcMkHUJ1g90lkvZvWXUMcD1wN6qbMl4LvHWB7VcDqwGWL18+6nQjIsZqyUFl2xfWd1C0ew/Vt3WfGHZSERHRO9uPW6Cm8h8BT7X9/yaaXEREbELSFlQDyqfaPqtufj5wVP3zv9JhqnUGKSIixmY/4Ml1nfu7A9tI+qjt59brfyXpI8CrF9q4fSZg6xei7V+ODrKcWM3Io9RYpeQxC7FKyaPXWJ0sOai8EElPoXrq6Vcl9RMiIiLG40HAoyQdB/wSeLXtr3SzYWlTJDONdHrbS8xpWttLzGnY7U2g6gL6BOBq2+9uWXUt8BjgAuAA4JsLbZ9yRRER42H7GKq7kqnvVH617edKWm77uvp8/lSgOR9CERFD0vOgsqR7Uk2ffvzw04mIiCFbBtwHeATw+8Dpkn7HdgYhIiImZz/gMOAKSZfVba8HXgQcL2kZ1ReBq7sJ1mlAPe2T3/estZeY07S2l5hTP8fQYKdK2gEQcBnwl91sNA21VZsUq5Q8So1VSh6zEKuUPPqN1cmyRdcu7AHAbsD8XcorgEsl7Wv7+vbOmZ4XETFRG4Cz6kHkL0v6NbA98KP2jjlfR0SMh+2LqAYiFvKwceYSERHdsX0B1UwSbB8w0WQiIgrQ86Cy7SuAHeeXJX0XWGX7xg79Mz0vImIMJJ0IPBnYqqV5e+A0SdfXP+8I5HwdETFFSitBUkp7iTlNa3uJOU1re4k59ds+i1qPvf11GGQ5sZqRR6mxSsljFmKVkkevsTpZtlQHSacB+wPbS9oAHGv7hK6iR0TEON0fMLDl/Pka+APgRGBvYAuqB0J1NWDc6YNkWttLzGnW2kvMaVrbS8xpmMcWERERERGjtdlSHWwfanu57S1sr2gfULa9a6e7lCMiYnxsP46qbvL6+fO17Vvrp1PvAdwK/M1Ek4yIiIiIaBhJm0v6b0nn1MsHSrpU0mWSLpL0wG7itN61PTc3N/ByYjUjj1JjlZLHLMQqJY9+Y3Wy5J3KERExFR4F/ND2N7vdoP1DZP6uwGlrLzGnWWsvMadpbS8xp2G3N4GkuwMXAltSXY+fYftYSbsBHwPuC1wCHGb71sllGhERtaOAq4Ft6uUPAk+xfbWklwD/Bzh8QrlFREyEupwFPRQrV6702rVrx7a/iIhh2GOPPS6xvWrSeXRD0q7AObbn2to/CHzL9t93E2flypVev379Jm2THijKINv0tpeY07S2l5jTMNubcr5W9bTre9m+RdIWwEVUAxavpHq46sckfQj4qu0PLhYr19cR0URNOV8DSFoBnAwcB7zS9iGSvg48z/bFko4Btrb9+sXi5HwdEU202Pk6dypHREw5ScuApwMPW6LfamA1wPLly8eQWUTEbKpr299SL25R/zFwAPCcuv1k4M1Ud8NFRMTkvBd4DbB1S9sLgU9L+gXwU+ARE8grImKilhxUlnQicAhww/ydb5LeDLwI+FHd7fW2P93NDrupyQGL32HThP4l5tT0/iXmNGv9S8xpXP2boD5fPxnYqqVtb+A04F7AxyW9xPaXF9re9hpgDVR3Uow84YiIGSZpc6oSFw8EPgB8G7jJ9u11lw3Azt3EKulu8ZLaS8xpWttLzGla20vMacrLFc2PhVwiaf+WVUcDT6jvVP4r4N1UA82Lmpubu8vrMshyYjUjj1JjlZLHLMQqJY9+Y3WybNG1lZOA9wOntLW/x/a7utg+IiLG4/5Ud7ptKWkDcCzwLKqBieOB7wN/B+zfTbBOHyDT2l5iTrPWXmJO09peYk7DPLYmsH0HsLekbYGzgd+bbEYREbGA/YAnS3oCcHdgG0mfAn7P9sV1n7XAZxfaODMBI2KaLTmobPvCukZnREQUzPbj2msqS/pT4ATbayUdClw7yRwjImJTtm+SdD7wSGBbScvqu5VXANcstE0GKSIixsP2McAxAPWdyq8GngpcL+lBtr8BHET1EL+Ftt9kJmDrF6LtX44OspxYzcij1Fil5DELsUrJo9dYnSw5qLyIl0p6HrAOeJXtjQPEioiI0XgF8DlJ7wI2A/6g2w1LmyKZaaTT215iTtPaXmJOw25vAkk7ALfVA8r3oBqQeAdwPvAM4GPA84FPLLR9yhVFREyO7dslvQg4U9KvgY3AERNOKyJi7PodVP4g8NdU06z/Gvh7chKNiCjRi4GjbZ85f9cy8LgJ5xQRMeuWAyfXdZU3A063fY6kq4CPSXob8N9U5+wldRpQT/vk9z1r7SXmNK3tJebUzzE0ie0LgAvqn8+mKl3Uk2mordqkWKXkUWqsUvKYhVil5NFvrE6WLbq2A9s/nP9Z0j8D53Tqm+l5ERET9XzgqPrnfwU+3KljztcREeNh+3JgnwXavwPsO/6MIiJiHFoHaNoHawZZTqxm5FFqrFLymIVYpeTRa6xO+hpUlrTc9nX14tOAjnvL9LyIiPGQdCLwZGCrlub/Ba6sp+bdAny70/Y5X0dENFNpJUhKaS8xp2ltLzGnaW0vMad+2yMiotmWHFSWdBqwP7C9pA3AscD+kvamKn/xXeAvRpdiRER06f5U5+UtW87Xv6I61/8S2An4fLfBOl3wT2t7iTnNWnuJOU1re4k5DfPYIiIihkXSd4GbgTuA222vkvRM4M3AQ4B9ba/rJtY0TINvUqxS8ig1Vil5zEKsUvLoN1YnyxZdC9g+dIHmruq7RUTE+Nh+nKRdgXNszwFIejewrW1L2gX43CRzjIiIiIhooMfavrFl+Urg6cA/TSifiIiJW3JQOSIiGm098BTg48AzgV263bC0KZKZRjq97SXmNK3tJeY07PYmkHR34EJgS6rr8TNsHyvpVGAVcBvwZeAvbN82uUwjImIhtq8GkNTTdq2fVe2fW4MsJ1Yz8ig1Vil5zEKsUvLoNVYny7rqNUS9XvA3vf849jFr/cexj/Sf/D5K699gRwDvk/RG4JPArRPOJyIiqtJEB9i+RdIWwEWSPgOcCjy37vMvwAuBDy4VrNNnWtonv+9Zay8xp2ltLzGnfo6hIQycK8nAP9XPIYmImHmbTTqBiIgYHdtfs/142w8DTmORB/VJWi1pnaR1GzduHF+SEREzxpVb6sUt6j+2/el6nanuVF4xsSQjImLeH9p+KPDHwJGSHt1voNZZNnNzcwMvJ1Yz8ig1Vil5zEKsUvLoN1YnS96pLOlE4BDghvkanXX7y4AjqYrVf8r2a5bcW1vSi5n/NrOp/UvMqen9S8xp1vqXmNO4+jdBXTP5NOCBktYDa+rl24G1wMOBH0jazvZdRo3ruy7WAKxcudJjSzwiYgZJ2hy4BHgg8AHbF7es2wI4DDiqm1jtn2mdPutmrb3EnKa1vcScprW9xJz6bW8K29fUf98g6WxgX6oSRkuStBpYDbB8+fKR5RgRMQlLDioDJwHvB06Zb5D0WKoanXvZ/pWkHUeTXkRE9OB9VIMTmwPbAscAOwF/SfUF4D8CNwGvA167VLDSpkhmGun0t5eY07S2l5jTrE2ntn0HsLekbYGzJc3Znj+ofwQutP0fE0swIiKQdC9gM9s31z8/Hnhrt9u337TR+tnV/jk2yHJiNSOPUmOVkscsxColj15jdbJk+QvbFwI/bmt+MfB227+q+9zQ1d4iImJkbD/N9g62t7C9M3AxcD7wI2BP268DTgaeOsE0IyKihe2bqM7VBwNIOhbYAXhlp21SrigiYmx2oqp7/1WqskSfsv1ZSU+TtAF4JPApSZ+baJYREROwrM/tHgQ8StJxwC+BV9v+yvDSioiIQUjaFdiHamB5J9vX1auup7o4XlJpUyQzjXR620vMaVrbS8xp2O1NIGkH4DbbN0m6B3AQ8A5JLwT+CDjQ9q87bZ9yRRER42H7O8BeC7SfDZzda7y5ubm7fI4NspxYzcij1Fil5DELsUrJo99YnSxbdO3i290HeATw+8Dpkn6nfqhIRERMkKStgDOBV9j+qaQ719l2/eTqiIiYnOXAyXVd5c2A022fI+l24HvAF+tz91m2l5xm3emCP+2T3/estZeY07S2l5hTP8cQERHN1e+g8gaqi1wDX5b0a2B7qinWm0hh+oiI8akf7nQmcKrts+rmH0pabvs6ScuBBUsW5XwdETEeti+nmk3S3t7vtXlERDRA6wB7+2D7IMuJ1Yw8So1VSh6zEKuUPHqN1Um/F64fBx4LnC/pQcDdgBsX6pjpeRER4yFpF+ArwBbACkl32D4e+BpwWT3d+gPAJxbaPufriIhmKq0ESSntJeY0re0l5jSt7SXm1G97REQ025KDypJOA/YHtq8L0R8LnAicKOlK4Fbg+Sl9ERExcQ+lqpd8Rb38Dkm/BP4WeA+wNbAv8MfdBOt0wT+t7SXmNGvtJeY0re0l5jTMY4uIiBgWSd8FbgbuAG63vapl3auAdwE72F7wRrtW01BbtUmxSsmj1Fil5DELsUrJo99YnSxbdC1g+9AOq5671LYRETE+tj8B3FlAWdIngO/Y/hLwSEkXUD1Y9ccTSjEiIiIiooke2z5oXM8SfDzw/W6DtA7QtA/WDLKcWM3Io9RYpeQxC7FKyaPXWJ0sOagcERHNI2lXqnqdF/cbo7QpkplGOr3tJeY0re0l5jTs9iaQdHfgQmBLquvxM2wf27L+fcARtreaUIoREbG09wCvoUNpuYiIaTf2QeVeL/ib3n8c+5i1/uPYR/pPfh+l9W8SSVtRPazvFbZ/Oul8IiLiLn4FHGD7lvoBqxdJ+oztL0laBWzXS7BOn2lpn/y+Z629xJymtb3EnPo5hoYwcK4kA/9ke42kpwDX2P6qpCU2/41pmAbfpFil5FFqrFLymIVYpeTRb6xOli26NiIiGqUenDgTONX2WT1uuxpYDbB8+fIRZBcREQD1s0huqRe3qP9Y0ubAO4HnAE+bUHoREbGpP7R9jaQdgfMkfQ14PVXpi0Xl+joiptmSg8qSTgQOAW6wPVe3rQUeXHfZFrjJ9t7d7HChaZgLaR8db1r/EnNqev8Sc5q1/iXmNK7+TVDXdfsK1eDECkl32D5e0juBJwErgHdKeprtm9q3t70GWAOwcuXKPHw1ImKE6gHkS4AHAh+wfbGko4BP2r6u1zvfWnX6rJu19hJzmtb2EnOa1vYSc+q3vSlsX1P/fYOks4HHALsB83cprwAulbSv7evbtt3k+rr12Ntfh0GWE6sZeZQaq5Q8ZiFWKXn0GquTZV30OQl4P3DKfIPtZ83/LOnvgZ90tbeIiBilhwI7AVfUy++Q9EvgZ8BWVIPN+wKXAr+zVLBOHyTT2l5iTrPWXmJO09peYk7DPLYmsH0HsLekbYGzJT0aeCaw/yTzioiI35B0L2Az2zfXPz8eeKvtHVv6fBdY1f4gv4iIabfZUh1sXwj8eKF1qr6W+1PgtCHnFRERPbL9CduyvaftPYHPAd+x/WbbK2xvCTwX+OJkM42IiHn1zJHzgcdS3bX8rXqA4p6SvrXQNpJWS1onad3GjRvHlmtExAzaiaru/VeBLwOfsv3ZfoO13rU9Nzc38HJiNSOPUmOVkscsxColj35jddLNncqLeRTwQ9vfHDBOREQMkaRdgX2Ai9tWHQGs7SZG+4fI/F2B09ZeYk6z1l5iTtPaXmJOw25vAkk7ALfZvknSPYCDgHfY/q2WPrfYfuBC26dcUUTEeNj+DrDXEn12HU82ERFlGXRQ+VByl3JERFEkbUX1sL5X2P5pS/sbgNuBUyeVW0REALAcOLmuq7wZcLrtc/oNVloJktLaS8xpWttLzGla20vMaVrLFQ1T62vR/roMspxYzcij1Fil5DELsUrJo9dYnSzrqtcCJC0Dng48bIl+edppRMSYSNqCakD5VNtntbQfTvXQ1QNtL3hXW87XERHjYftyqtkki/XZakzpRERERET0rO9BZeBxwNdsb1isU6bnRUSMR13n/gTgatvvbmk/GHgN8BjbP++0fc7XERHNVFoJklLaS8xpWttLzGla20vMqd/2pqjr3N8M3AHcbnuVpLXAg+su2wI32d57qVhzc3N3eV0GWU6sZuRRaqxS8piFWKXk0W+sTpYtuhaQdBrVU6i3l7QBONb2CcCzSemLiIiSPA04DPiVpL8A/pfqzuOPAlsB35d0G/Bx289bKlinD5BpbS8xp1lrLzGnaW0vMadhHltERMSQPdb2jfMLtp81/7Okvwd+MpGsIiImaMlBZduHdmg/fOjZRETEIL4IPMz2pZK2Bi4BvgvsOl9bWdLLgd0nl2JERERExHSoZwr+KXBAN/1bvxBt/3J0kOXEakYepcYqJY9ZiFVKHr3G6mTJQeWIiGgG29cB19U/3yzpamBn21e1dLsX0FVpi9KmSGYa6fS2l5jTtLaXmNOw2yMiIobMwLmSDPxTXTJu3qOAH9r+5mRSi4iYnLEPKvd6wd/0/uPYx6z1H8c+0n/y+yitf9NI2pXqIVAX18vHAc+jmpr32MllFhERku4OXAhsSXU9fobtY+s73t4GPJOqducHbb9vqXidPtPSPvl9z1p7iTlNa3uJOfVzDA3xh7avkbQjcJ6kr9m+sF53KD2UBZ2G2qpNilVKHqXGKiWPWYhVSh79xupks0XXRkRE40jaCjgTeMV82Qvbb7C9C3Aq8NIO262WtE7Suo0bN44v4YiI2fMr4ADbewF7AwdLegRwOLAL8Hu2HwJ8bGIZRkQEALavqf++ATgb2BdA0jLg6cDaTtvm+joiptmyce9woWmYC2kfHW9a/xJzanr/EnOatf4l5jSu/k0haQuqAeVTbZ+1QJdTgU8Dx7avqKfyrQFYuXJlVyUyIiKid7YN3FIvblH/MfBi4Dm2f133u6GbeKWVICmlvcScprW9xJymtb3EnPptbwJJ9wI2q0vL3Qt4PPDWevXjgK/Z3tBp+/br69Zjb38dBllOrGbkUWqsUvKYhVil5NFrrE6WLdVB0onAIcANtufqtr2BDwF3B24HXmL7y13tMSIiRkLSLsBXqAYnVki6w/bxkn7X9jclvQp4F/Bv3cTr9EEyre0l5jRr7SXmNK3tJeY0zGNrAkmbUz1Q9YHAB2xfLOkBwLMkPQ34EfDy1OmMiJionYCzq+pELAP+xfZn63XPpofSFxER06ab8hcnAQe3tf0d8BbbewNvqpcjImKyHkp14XsNVS3Od0j6C+Dtkr4GvBn4BfD6iWUYEREA2L6jvpZeAewraY6qxvIvba8C/hk4caFtM506ImI8bH/H9l71n5W2j2tZd7jtD/USr/Wu7bm5uYGXE6sZeZQaq5Q8ZiFWKXn0G6uTJe9Utn1h/cCnTZqBbeqf7w1cu+SeIiJipGx/AtD8sqRPAN+x/SeSzgCeBXwCuL6beO0fIvN3BY66/eg937JJ+3suP3ak+x3FMaQ9v4NS20vMadjtTWP7JknnU93EsQGYL110NvCRDtukXFFERANNwzT4JsYqJY9SY5WSxyzEKiWPXmN1sqyrXnf1CuBzkt5FdbfzH/QZJyIiRqD+MnAf4GJJTwGusf3VeupeRERMkKQdgNvqAeV7AAcB7wA+DjwW+B/gMcA3uonX6cI/7ZPf96y1l5jTtLaXmFM/xxAREc3VTfmLhbwYONr2LsDRwAmdOmZ6XkTEeEnaiuphfa+gqnv/eqpSRUttl/N1RMR4LAfOl3Q5VS3882yfA7wd+BNJVwB/C7xwgjlGRMSQTcM0+CbFKiWPUmOVkscsxColj35jddLvncrPB46qf/5X4MOdOmZ6XkTE+EjagmpA+VTbZ0naA9gNmL9LeQVwqaR9bW9SBiPn64iI8bB9OdVskvb2m4An9hqv/aJ//q7AWW8vMadpbS8xp2ltLzGnftubQtK2VGMec1SlQI8Avg6sBXYFvgv8qe3clRERM6XfQeVrqabkXQAcAOSp1BERE6Zq1PgE4Grb7wawfQWwY0uf7wKrbN+4VLxJTYWcr6E87v2OYx9pX7y9xJymtb3EnDKdOiIiCnU88Fnbz5B0N+CeVDMBP2/77ZJeB7wOeO1SgVo/u9o/xwZZTqxm5FFqrFLymIVYpeTRa6xOli3VQdJpwP7A9pI2AMcCLwKOl7QM+CWwuqu9RUTEKO0HHAZcIemyuu31tj/da6Crrrrqxj322ON79eL2wJKD0C2a3n8c+0j/ye9j1vqPYx8l9P/tHuJFREQsStK9gUcDhwPYvhW4tX5myf51t5OpbrhbclA5ImKaLDmobPvQDqseNuRcIiJiALYvAhZ9Ep/tXbuMtcP8z5LW2V7VbR5N719iTrPWv8Scmt6/xJzGccwRERED2g34EfARSXsBl1CVAt3J9nV1n+uBnboJNjc3d+cdgPNlQQZZTqxm5FFqrFLymIVYpeTRb6xOlhxUHrZub6Gelv7j2Mes9R/HPtJ/8vsorX9ERMSwSLo7cCGwJdX1+Bm2j5V0IPBOqodp3wIcbvtbS8Xr9JmW9snve9baS8xpWttLzKmfY2iAZcBDgZfZvljS8VSlLu5k25IWfB6JpNXUM7uXL18+6lwjIsZK9viexbRy5UqvXbt2bPuLiBiGPfbY45JZvjuutLsPc0fn9PcvMaem9y8xp1m+U7mugX8v27fUD1i9iOrOt1OAp9i+WtJLgH1tH75YrFxfR0QTNeX6WtJvAV+an+0n6VFUg8oPBPa3fZ2k5cAFth+8WKycryOiiRY7X4/9TuWFnl67kPZbrpvWv8Scmt6/xJxmrX+JOY2r/4xbM2P9x7GP9J/8Pmat/zj2UVr/Yrm6q+OWenGL+o/rP9vU7femejj2kto/0zp91s1ae4k5TWt7iTlNa3uJOfXb3gS2r5f0A0kPtv114EDgqvrP84G3139/YoJpRkRMxGZLdZB0oqQbJF3Z0raXpC9KukLSv0naZrEYERHRXLZ7Gshpev9x7CP9J7+PWes/jn2U1r90kjavH6p6A3Ce7YuBFwKfrh+OfRjVYEVEREzWy4BTJV0O7A38DdX5+SBJ3wQeR5fn69YB9rm5uYGXE6sZeZQaq5Q8ZiFWKXn0G6uTJQeVgZOAg9vaPgy8zvYewNnAX3URJyIiIiIiANt32N4bWAHsK2kOOBp4gu0VwEeAdy+0raTVktZJWrdx48ax5RwRMYtsX2Z7le09bT/V9kbb/2v7QNu/a/txtn886TwjIsZtyfIXti+UtGtb84OoHi4CcB7wOeCNw00tIiIiImK62b5J0vnAHwN71XcsA6wFPtthmzXU5UBWrlw5vgekRETEQFpLf7SXARlkObGakUepsUrJYxZilZJHr7E6WXJQuYP1wFOAjwPPBHbpM05ERERExEyRtANwWz2gfA/gIOAdwL0lPcj2N+q2q7uJ1+nCP+2T3/estZeY07S2l5hTP8cQERHN1U35i4UcAbxE0iXA1sCtnTpmel5ERLNI+j1Jr5X0vvrPayU9ZIn+B0raqq29vXRSp+1PWWTdw+fr9ku6h6S31LX83yHp3gv0v5uk50l6XL38HEnvl3SkpC26ySeGT9KOPfa/76hyiSjEcuD8uj7nV6hqKp8DvAg4U9JXqWoqp8RcRMSESdpW0hmSvibpakmPlPTXki6XdJmkcyXdr5tY01BbtUmxSsmj1Fil5DELsUrJo99YnfR1p7LtrwGPB5D0IOCJi/TN9LyIiIaQ9FrgUOBjwJfr5hXAaZI+Zvvtbf1fDhxJdTfdCZKOsj3/9Ou/oW3qtqRPtu8SeKykbQFsP7lt/YnAXvXPxwM/p7qb70CqeqNPb+v/EarPtntKej6wFXBW3X9fqqdzT5ykHW3f0EP/+9r+3yHt+97AMcBTgR0BUz0o7BPA223f1NZ/m7r/CuAztv+lZd0/2n5JW//7tO8S+LKkfQC11xyU9HbgXbZvlLQKOB34df0lwPNs/3tb/1XAO4Fr6rxOpPrdfgNYbfu/FzjmZcCfA08D5v/Td019zCfYvq2t/+ZUD0xbAXzW9n+2rPs/tt/Wvo8F9vkN2w/qsO6lwMfqY35gfQx7Al8HXmj7irb+vwP8H+BaqgcBvQd4JNW/u7+y/d1ZPt4msn05sM8C7WdTPa+kJ+0X/fN3Bc56e4k5TWt7iTlNa3uJOfXb3iDHU30+PkPS3YB7AuttvxHuvB5+E/CXE8wxImLs+hpUnv/PsKTNqC76PzTctCIiYkL+HFi5wKDTu6lKH7U/2fpFwMNs31LX3z9D0q62j6caTGy3AriK6oGvrvusAv6+Qz6b2b69/nmV7YfWP18k6bIF+u9he896UO0a4H6275D0UeCrC+1gBgdZTwe+AOxv+/o6xm9RDbifTv2lcYuPAN8EzgSOkPQnwHNs/wp4BHd1I/C9tradgUupXtvfaVv3RNuvq39+J/As21+pv7T+F6r3R6t/BI4FtgX+Czja9kGSDqzXPXKBnP4vcBPwZmBD3baiPuaPAs9q6/9PVP9h/DLwPkn/bvuV9bqnA5sMskq6uT42+M37/p7z7ba3aYv/Ytvvr38+HniP7bMl7U91TbVfW/+TgNOAewNfovqdvJXqd3UicMCMH29ERMRI1NeJjwYOB7B9K3edqX0vfvO5uKhpqK3axFil5FFqrFLymIVYpeTRa6xOlix/Iek04IvAgyVtkPTnwKGSvgF8jeouko90tbeIiCjdr/nNnY2tltfr2m1m+xaA+u7B/YE/rgehFxpUXgVcArwB+IntC4Bf2P739sHS2pWSXlD//NV6AHV+lsxtC/TfrL6DZGuqQbL5EhlbAp3KX5wObKQaZL2P7fsCj63bTl+g/0fqYzsTeLakMyVtWa/rNMh6ScufdfxmkHXdAv2faPvG+uf5QdYHUtVXXWjw/R+BvwM+RTXI+k+27w28rl7Xblfb75gfUAawfb3tdwC/vUD/B9h+ne2P13eSXwp8YZESFX9FdQfqk23vZns3YEP9c/uAMsCy+ksAgHvY/kqd0zeofm/ttrD9GdunVd18Rt3/88DdO+T0MNsvtv0l2xvqP1+y/WIWuFsU2Nf2c2y/F3g4sJWks+rf80Lv649QPWfid21vbXtr4Pv1z+0DrLDpl/o71nenUv972HqB/lvb/mA9U2Ab239v+we2TwC2y/FGRESMzG7Aj4CPSPpvSR+WdC8AScdJ+gHwZ1R3KkdEzJQlB5VtH2p7ue0tbK+wfYLt420/qP7zOtspaxERMR1eAXxe0mckran/fBb4PHDUAv1/KGnv+YV6gPkQYHtgj/bOtn9t+z3AC4A3SHo/i8+aeSHwGEnfBnYHvijpO8A/1+vanUD1hedlVAPX/yrpn6lqln6swz5mbZD1e5JeI2mn+QZJO9WlT36wQP8t65lJ1HGPo3r9LwTucsy2/57qd/MmSe+WtDWL373zj8CnJR0AfFbS8ZIeI+ktVL/Hdr+U9HhJzwQs6an1MTwGuKPDPn4s6ZmtxyFpM0nPovryoN3dWo7ndturqe50/wJVSZX2Y3451R24p0l6eb2fxY75DEkn1WUezpb0Ckm/XX+B8v0F+v9a0oMk7Ut1R/D8lysPBDaf4uP9/S6PNyIiYlSWAQ8FPmh7H+BnVF/cY/sNtncBTgVe2k2waait2qRYpeRRaqxS8piFWKXk0W+sTjTO8eCVK1d67dq1Y9tfRMQw7LHHHpfYbp+CP7XqAaJ9qe6mhaqswlds32XATtIK4PbWAdmWdfu5pTZrh309EdjP9uuX6LcN1Z0iy6gGZH+4SN/7Adi+VlWt5sdR3UX55Q79zwX+H3DyfNx6wPVw4CDbj2vrfzVViZBft7QdTjV4vJXtuwxE16/Te6gGbY8FvtphQBlJLwOeRFVq5NFUd2aeRTXl/3dsH9bW/4t1zHsD7wKOsv3xepD179vfu5K2o/rP0FOAnagGA38IfBJ4h+9ajuPvgHNt/7+29oOBf7D9uwsdR93nycDrqQbuf2uRfvsDLwYeRPU7/gHVnbAf8V1LsexFdWf2r4Gj6+2eT/U+fZHt/1og/q5UtbgfS1UWAqryGecDr7P9P239Pwp81HZ7TfAXUv2ncsG73ut/Oy8Fnkn15UPHh/bU75kXAw+g+rJg/pjfYfsnbX3nS3v8mqrkzNFUNYnvXR/zJ9r6zx/vAVSDyKr7TvJ4X0BVa3LQ411t++Od9tM0qupZrwOusX2IpN2ovgC7L9XMhsNcTbXuKNfXEdFETbm+VlUi7Eu2d62XH0X1WfrElj73Bz5t+y4jMJJWA6sBli9f/rBzzz13LHlHRAzLYufrDCpHRCyhKRe90Z+2QdYd6+b5Qda3297Y1n+Sg6wn+jc1puf79jPI+ntUNXa/VN9dfucxtA8stvTfGbi4rf8f2/7MYv2p7h5+gO0r+4jfqf9DqMq0dNW/XvdwqgH0bwO/R1V7+Srbn+7Qf1+qO7+/Iml34GDga132fxTVAPa6LvuvrONfvUj/hwO/bun/x4vl37Ld/N3kx9t+7mJ927Y7xfbzuuy7HLjSVemYbuP/3/YvSJbofw7V3f4LleFpLEmvpCoLtE09qHw6cJbtj0n6ENUXUB9cLEauryOiiZp0fS3pP6geLPt1SW+mqqG8xvY36/UvAx5j+xmLxcn5OiKaqKhB5fXr13fVd7Gn3Tahf4k5Nb1/iTnNWv8ScxpH/yZd9MZwSXqB7a6fG9BNf0n34DeDrEOPv1R/VU8oPxK4Gtib6s7mT9TrLvVvHoY43/9lVHejdtu/1/j99H8JVZmTJfvX7cdSDcIuA86juhP/Aqo61Z+rS3os1v/hVHf5dtu/1/jD7v/J9teA6q7lLwC4KtuyWH9RDYp327/X+EPt31T1DIaTgeOAV1LNUPgR8Fu2b5f0SODNtv9osTgLXV93+qybtfYSc5rW9hJzmtb2EnPqp71J19eqSr19mKpc1Heoyrh9GHgw1Zf63wP+0vY1i8WZP1+3vy6DLCdWM/IoNVYpecxCrFLy6CfWYufrZQs1tpK0C3AKv5kiu8b28aqeZr8W2BX4LvCn7XdzRURENNxb6O1htEv2t/0L4MpRxe+i/4uoHuR2S10m4QxJu9o+noUfyra6x/69xu+n/6oe+gM8g2oAekvgemCF7Z9KehfV3dTHTVn/FcBVVP/hdf26/D4LP+gRYBdgfVv/VYv07zX+MPJZrH9TvRd4Db95WOF9gZtaZiNs4DdliCIiYkJsX0b1udjqTyaQSkREUZYcVAZuB15l+1JVD9u5RNJ5VLUmP2/77ZJeRzV1+LWjSzUiImL4JF3eaRXVF6pT1R/YbL5khO3v1qU2zpD02yw8KNv0/lDV/b4D+Lmkb9v+ab39LyQtVE6h6f1XUT1Y8w3AX9m+TNIvbP97h9fnYT327zX+qPNpHEmHADfYvqR+D/e6fWuNzuEmFxERIzN/F2D7z4MuJ1Yz8ig1Vil5zEKsUvLoNVYnSw4q274OuK7++WZVDyjamar25P51t5OppmFmUDkiIppmJ+CPqB5o1krAXeoRT0H/H0rau77rhvqO30OAE4E9prA/wK2S7mn751QDlgBIujfVtNWp6l/XHX6PpH+t//4hi1zzNb1/Q+0HPFnSE4C7A9sAxwPbSlpW3628gqo2+l3YXgOsgWo69XhSjoiIiIj4jZ4u0OtppvtQTbXcqR5whmoq5kJ3Q0VERJTuHGCr+UHKVpIumML+z6OahXSnegDreZL+aQr7Azza9q/qvq2DsFtQPdRw2vpT990APFPSE4Gfduo3Lf2bxPYxwDFw54M5X237z+qB9GcAH6P63X6im3id7iZJ++T3PWvtJeY0re0l5tTPMTSBpG2pSjLNUZVlOsL2F+vnThxJ9VDiT9l+zVKxml5btWmxSsmj1Fil5DELsUrJo99YnSxbdG0LSVsBZwKvqOv43bnOtiUteJdEpudFRETJbP/5IuueM4X9NyzS/z+nrX/d/qsO7TcCN05b/wX6fQr41FL9pqV/w70W+JiktwH/DZww4XwiIqKaSfJZ28+QdDfgnpIeSzV7ey/bv5K042RTjIgYP9lLz5iTtAXVnVCfs/3uuu3rwP62r5O0HLjA9oMXi7PQ06k7aR8db1r/EnNqev8Sc5q1/iXmNI7+TXo6dUREzJaFrq87fdbNWnuJOU1re4k5TWt7iTn1096U6+u61NRlwO+4ZfBE0unAGtv/r9tYK1eu9Nq1a4efZETECC12vt5sqY1V3ZJ8AnD1/IBy7ZP8Zspl19PzIiIiIiIiIiIaYDfgR8BHJP23pA9LuhfwIOBRki6W9O+Sfn+yaUZEjN+Sg8pUDxI5DDhA0mX1nycAbwcOkvRN4HH1ckRERERERETENFgGPBT4oO19gJ8Br6vb7wM8Avgr4HS11gjtoPWu7bm5uYGXE6sZeZQaq5Q8ZiFWKXn0G6uTJWsq276I6gnyCzlwyT1ERERERERERDTPBmCD7Yvr5TOoBpU3AGfVJTG+LOnXwPZUdzXfKc+Yiohp1vWD+iIiIiIiYngkbQ6sA66xfYikU4FVwG3Al4G/sH3bJHOMiJhltq+X9ANJD7b9daob664Cvg08Fjhf0oOAu7Hww3bXAGugqqk8X2caflNzehjLidWMPEqNVUoesxCrlDx6jdVJVw/qG5YUpo+IJmrKg0QiIqJZJL2SahB5m3pQ+QnAZ+rV/wJcaPuDi8XI9XVENFGTrq8l7Q18mGrg+DvAC6jKYJwI7A3cCrza9hcWi5PzdUQ00UAP6ouIiIiIiOGStAJ4ItVABQC2P+0a1Z3KKyaVX0REVGxfZnuV7T1tP9X2Rtu32n6u7TnbD11qQHneNNRWbVKsUvIoNVYpecxCrFLy6DdWJ0uWv5C0C3AKsBNgYI3t4yU9E3gz8BBgX9vrltxbW9KLmb/Vuqn9S8yp6f1LzGnW+peY07j6R0REDNl7gdcAW7evkLQF1YOyj+omUPtnWqfPullrLzGnaW0vMadpbS8xp37bIyKi2ZYcVAZuB15l+1JJWwOXSDoPuBJ4OvBPo0wwIiIiImKaSDoEuMH2JZL2X6DLP1KVvviPsSYWEREjNQ21VZsYq5Q8So1VSh6zEKuUPHqN1cmSg8q2rwOuq3++WdLVwM62zwOQ1NWOIiIiIiICgP2AJ9c1lO8ObCPpo7afK+lYYAfgLzptLGk1sBpg+fLl48g3IiIiImITPdVUlrQrsA9w8UiyiYiIiIiYcraPsb3C9q7As4Ev1APKLwT+CDjU9q8X2X5NXd9z1XbbbTemrCMiZpOkbSWdIelrkq6W9EhJe0n6oqQrJP2bpG26iTUNtVWbFKuUPEqNVUoesxCrlDz6jdWJqueALE3SVsC/A8fZPqul/QKqJ52uWyrGypUrvX79+q72t1gNqSb0LzGnpvcvMadZ619iTuPo36SnU0dERLPU5S9ebfsQSbcD3wNurlefZfuti22/cuVKr127drRJRkQMWZOuryWdDPyH7Q9LuhtwT+A8qnP3v0s6AtjN9hsXi5PzdUQ00WLn625qKs8/LORM4NTWAeUut830vIiIiIiIBdi+ALig/rmra/OIiBgPSfcGHg0cDmD7VuBWSQ8CLqy7nQd8Dlh0UDkiYtoseeGqqmjyCcDVtt/d6w5srwHWQPXNXM8ZRkRERERER+2zbzrNypm19hJzmtb2EnOa1vYSc+q3vSF2A34EfETSXsAlwFHAeuApwMeBZwK7dBNsbm7uLq/LIMuJ1Yw8So1VSh6zEKuUPPqN1Uk3NZX3Aw4DDpB0Wf3nCZKeJmkD8EjgU5I+10WsiIiIiIiIiIgmWAY8FPig7X2AnwGvA44AXiLpEmBr4NaFNpa0WtI6Ses2btw4rpwjIsZi2VIdbF8EqMPqs4ebTkREREREREREETYAG2xfXC+fAbyurp/8eIC6FMYTF9q4feZ2611/7XcADrKcWM3Io9RYpeQxC7FKyaPXWJ10c6dyREREREQMmaTNJf23pHPa2t8n6ZZJ5RURERXb1wM/kPTguulA4CpJOwJI2gz4P8CHJpRiRMTEyB5fmeM87TQimqhJT6eOiIjmkPRKYBWwje1D6rZVVPU6n2Z7q6Vi5Po6IpqoSdfXkvYGPgzcDfgO8ALgecCRdZezgGO8xODKypUrvX79+sbXVm1SrFLyKDVWKXnMQqxS8ugn1mLn62ULNUZERERExOhIWkE1Xfo44JV12+bAO4HnAE+bXHYRETHP9mVUXwC2Or7+ExExs5a8U1nSLsApwE6AgTW2j5f0TuBJVAXpvw28wPZNi8Wa/2auG+2j403rX2JOTe9fYk6z1r/EnMbRv0l3UkRERDNIOgP4W6oHPL3a9iGSjgI2s/0eSbd0e6dy+/V1p8+6WWsvMadpbS8xp2ltLzGnftpn8fo6M0siookWO193U1P5duBVtncHHgEcKWl34DxgzvaewDeAY4aVcERERETEtJJ0CHCD7Uta2u4HPBP4h4klFhERERHRpSUHlW1fZ/vS+uebgauBnW2fa/v2utuXgBWjSzMiIiIiYmrsBzxZ0neBjwEHAOuBBwLfqtvvKelbC20sabWkdZLWbdy4cUwpR0TMHkkPlnRZy5+fSnqFpHdK+pqkyyWdLWnbbuK13rU9Nzc38HJiNSOPUmOVkscsxColj35jddLNncp3krQrsA9wcduqI4DP9BIrIiIiImIW2T7G9grbuwLPBr5gezvbv2V717r957Yf2GH7NbZX2V613XbbjTHziIjZYvvrtve2vTfwMODnwNlk5nZExNI1le/sKG0F/DtwnO2zWtrfQFW0/undPu20GyXWVe2lf4k5Nb1/iTnNWv8Sc0pN5YiIaDJJ+1PXVG5r77qmcmp0RkTTNPH6WtLjgWNt79fW/jTgGbb/bLHtc76OiCZa7Hy9rJsAkrYAzgRObRtQPhw4BDiw04CypNXAaoDly5f3lnlERERExBSzfQFwwQLtSw4oR0TEWD0bOG2B9iOAjBZHxMxZclBZkoATgKttv7ul/WDgNcBjbP+80/a21wBroPpmbuCMIyIiIiLiTu2zbzrNypm19hJzmtb2EnOa1vYSc+q3vUkk3Q14Mm1lLuqZ27cDp3YTZ25u7i6vyyDLidWMPEqNVUoesxCrlDz6jdXJskXXVvYDDgOukHRZ3fZ64H3AlsB51bgzX7L9l13Ei4iIiIiIiIhoij8GLrX9w/mGzNyOiFm35KCy7YsALbDq08NPJyIiIiIiIiKiKIfSUvqi35nbrXf9td8BOMhyYjUjj1JjlZLHLMQqJY9eY3WyWVe9IiIiIiJiqCRtLum/JZ1TL0vScZK+IelqSS+fdI4REbNO0r2Ag4CzWprfD2xNNXP7MkkfmkhyERETpA6zNEYiTzuNiCZq4tOpIyKifJJeCawCtrF9iKQXAI8FDrf9a0k72r5hsRi5vo6IJprF6+uVK1d6/fr1ja+t2qRYpeRRaqxS8piFWKXk0U+sxc7XGVSOiFjCLF70RkTEaElaAZwMHAe8sh5U/jLwHNvf6jZOrq8joolm8fo65+uIaKLFztfLltpY0i7AKcBOgIE1to+X9NfAU4BfAzdQ3VFx7VLx5ke7l9I+Ot60/iXm1PT+JeY0a/1LzGlc/SMiIobsvVT1OLduaXsA8CxJTwN+BLzc9jeXCtT+mdbps27W2kvMaVrbS8xpWttLzKnf9oiIaLZuairfDrzK9u7AI4AjJe0OvNP2nrb3Bs4B3jS6NCMiIiIipoOkQ4AbbF/StmpL4Jf13SD/DJw49uQiIuJOkh5c10ye//NTSa+Q9GZJ17S0P6GbeK0D7HNzcwMvJ1Yz8ig1Vil5zEKsUvLoN1YnS96pbPs64Lr655slXQ3sbPuqlm73orqLOSIi4v+3d+ZRdxRlHn7ekAiGkBACQhSGKIgMIRAgAgoMGVEGgRPA5QieEYNKVGSLqKwzERVkGBbxqMwEIcpqREAQ2VRExZFAyAYYQFnCIiJCTNjODMg7f1R90HS67+26+e69fb/7e86pc7urf/V2VXdX9dt1q6uFEEI0ZhdgWuyEWAsYbWYXA4/x2oegrgLmFCU2sxnADIDx48e3P7dCCNGnuPt9wGQIH1cFHie0z4cAZ7v7Gd3LnRBCdJemncpZzGwCsB0wL66fAhwMrCB8VEQIIYQQQgjRAHc/HjgewMymAl9w9381s9MIPvVDwO7A/SXpZwOzIczR2YEsCyGEgD2AB9x9mZm1ZCA79Ud+GpDVWZet3shHXW3VJR/9YKsu+Ui1VUblTmUzGwVcARzt7isB3P1E4EQzOx44HJhV1Z4QQgghhBDidZwGXGJmM4HngE9VSVTm+Cu++/vut/g65mmoxtcxT62Uocc4ELgss364mR0MzCdMGbq8O9kSQojuUGVOZcxsBKFD+RJ3v7JAcgnwwZK0M8xsvpnNX75cbawQQgghhBADuPst7r5vXP6bu+/j7pPc/V3uvrjb+RNCCAFm9gZgGnB5jDqX8HHVyYTpQs+sYmcozK3aS7bqko+62qpLPvrBVl3y0aqtMpqOVLbwXsf5wFJ3PysT//bM16j3A+4tSq/X84QQQgghhGgfead/YFRgv8fXMU9DNb6OeRqq8XXMU6vxPcb7gQXu/iTAwC+AmZ0HXFuUSHPgCyGGMlWmv9gF+Bhwl5ktinEnAJ80s3cArwDLgM+0JYdCCCGEEEIIIYQQ3eMgMlNfmNl4d38irh4AFPaU5wfZDYW5VXvRVl3yUVdbdclHP9iqSz5SbZXRtFPZ3W8Fimahv67SHoQQQgghhBBCCCF6EDNbG3gf8OlM9OlmNhlw4OHcNiGE6AsqzakshBBCCCGEGFzMbA0zW2hm18b1PcxsgZktMrNbzWzzbudRCCH6HXd/3t3HufuKTNzH4vz327j7tMyo5YYMhblVe8lWXfJRV1t1yUc/2KpLPlq1VUaV6S8GldT5k3pd34l99Ju+E/uQvvv7qJteCCGEaANHAUuB0XH9XGA/d19qZocBJwHTmxkpu6cpvvv77rf4OuZpqMbXMU+tlEEIIUTvYu6d+3bexIkTfe7cuR3bnxBCDAaTJk26092ndDsfQgghhg5mtjHwfeAU4PPuvq+Z3Qcc7O7zzOx4YB13P6GRHfnXQohepB/9a7XXQohepFF73XSkspltAlwIbEiYL2i2u5+T2X4McAawgbv/tZm9KsOn4bV/M1P1M7c5uZL+7CWz2pqfVtJI3/19SN/9fdRVL4QQQgwy3wC+BKyTifsUcJ2ZvQisBHauYih/Tyu71/VbfB3zNFTj65inoRpfxzy1Gi+EEKK3qTKn8svAMe6+FcGx/ZyZbQWvdjjvCTzSviwKIYQQQggxdDCzfYG/uPuduU0zgb3dfWNgDnBWxzMnhBDiVczsHXGe+4Gw0syONrPJZnZbjJtvZjtWsTcU5lbtJVt1yUddbdUlH/1gqy75aNVWGU1HKscJ55+Iy8+a2VLgLcDvgbMJIyyubronIYQQQgghBMAuwDQz2xtYCxhtZj8FtnT3eVEzF7ihKLGZzQBmAIwfP74D2RVCiP7E3e8DJkP4uCrwOHAVcB5wsrtfH9vy04GpzexlR2nnR2yvzrps9UY+6mqrLvnoB1t1yUeqrTKqjFR+FTObAGwHzDOz/YDH3X1xig0hhBBCCCH6GXc/3t03dvcJwIHAzcB+wBgz2yLK3kf4iF9R+tnuPsXdp4wdO7YjeRZCCMEewAPuvowwNejAR1bHAH/qWq6EEKJLNB2pPICZjQKuAI4mTIlxAmHqCyGEEEIIIcRq4O4vm9mhwBVm9gqwHPhElbRlo0kU3/1991t8HfM0VOPrmKdWytBjHAhcFpePBm40szMIg/XeXcXA1ltv/erxGHi1fHXWZas38lFXW3XJRz/Yqks+WrVVRqWRymY2gtChfIm7XwlsBrwVWGxmDwMbAwvMbKOCtDPiHEPzly9fXmV3QgghhBBC9AXufou77xuXr3L3Se6+rbtPdfcHu50/IYQQYGZvAKYBl8eozwIz3X0Twnz455ekU3+IEGLIYu7eWGBmwPeBZ9z96BLNw8AUd/9rI1sTJ070e+65p1LG8r3jVfUztzm5kv7sJbNasl9V30oa6bu/D+m7v4866idNmnSnu0+plEAIIYToIEX+ddm9brDj8353mX/dqfwU3du7te9+i69jnoZqfB3z1Ep8r/nXcfrPz7n7nnF9BbCuu3vsM1nh7qMb2Zg4caLPnTu3A7kVQojBo1F7XWWk8i7Ax4D3ZL54uveg5lAIIYQQQgghhBCinhzEa1NfQJhDefe4/B7gDx3PkRBCdJmmncrufqu7m7tv4+6TY7gup5nQbJSyEEIIIYQQQgghRC9hZmsTPp56ZSb6UOBMM1sMnArMqGIrO2p76623Xu31qtqZ25z8urdLVsfWYOar3bbqko+62qpLPvrBVl3y0aqtMoY3VQghhBBCCCGEEEL0Ie7+PDAuF3crsEN3ciSEEPWg6ZzKg4nmEBJC9CK9NuebEEKI/kH+tRCiF+lH/1rttRCiF2nUXne0U9nMngKWFWxaH0iZPqNu+k7sQ/ru70P67u+jW/pN3X2DBDtCCCFER8j5171yX+2WvhP76Dd9J/Yhfff30Q593/nXmfY65Xi2S9tO23XJRzttKx+ds618dM52en+Iu3c9APN7WV/HPPWbvo556jd9HfPUiTIrKCgoKCjUJdTtPlk3fR3z1Ov6Ouap3/R1zJN86sE7Pu3S9kM++qGMdclHP5SxLvmoUxndvfmH+oQQQgghhBBCCCGEEEKIAdSpLIQQQgghhBBCCCGEEKIydelUnt3j+k7sQ/ru70P67u+jbnohhBCiTtTtPlk3fSf20W/6TuxD+u7vQz714JJyfNql7Yd8tNO28tE528pH52wnt90d/VCfEEIIIYQQQgghhBBCiN6mLiOVhRBCCCGEEEIIIYQQQvQA6lQWQgghhBBCCCGEEEIIURl1KgshhBBCCCGEEEIIIYSozPBu7NTMtgT2A94Sox4HrnH3pYNo/y3APHd/LhO/l7vfUKDfEXB3v8PMtgL2Au519+sq7u9Cdz+4onZXYEfgbne/qUSzE7DU3Vea2RuB44Dtgd8Dp7r7ipz+SOAqd3+0Yh7eABwI/Mndf25mHwXeDSwFZrv7SwVp3gZ8ANgE+DtwP3Cpu6+ssk8hegEze5O7/6Xb+RBCCCGa0Yo/neojF6Qv9Xlb8F+T/VHRflJ9ITMb5+5PtzNPQvQyqW111b6JFtrcvnue17Pd0KBXz2O/3B87PlLZzI4FfgAYcHsMBlxmZscl2jqkIO5I4GrgCOBuM9svs/nUAv0s4JvAuWb2deBbwNrAcWZ2YoH+mlz4CfCBgfUC/e2Z5UOj/XWAWQ3KewHwQlw+BxgD/EeMm1Og/yowz8x+Y2aHmdkGJXYHmAPsAxxlZhcBHwbmAe8EvltQhiOB/wLWipo1CTej28xsapN9DRnM7E1ttj+unfYHEzMbY2anmdm9ZvaMmT1tZktj3Lot2Lu+IG60mX3dzC6KD5rZbd8p0G9kZuea2bfNbJyZfdnM7jKzH5rZ+AL9erkwDrjdzMaa2XqpZRBCCCE6RSv+dAs+cpLPS7r/muSPdotU/28w/blUf6sF3ynJF4r7XT8uTzGzBwnPIMvMbPcC/RQz+6WZXWxmm5jZz8xshZndYWbbFeiHm9mnzewGM1sSw/Vm9hkzG1FyjNaIab5qZrvktp1UlKbAxv0Nth2eKfPmZvZrM/ubmc0zs0kF+reZ2QVm9jUzG2Vm55nZ3WZ2uZlNWN0y93p5+43UttrS+iYqt7k2BJ/nLff82EJ7tsDMTjKzzSrsq3I9NbORZvYlM/uima1lZtMt3DdPN7NRBbYr1zkzG2ZmnzCzn5rZ4liGHxSdw9S2wsy2ySyPiMfmGjM71cxG5rSDVu9X5zymnMOor3xPtfS+hcr3x3jMvmJm91i4Jz5lZreZ2fQCu6n39b1y5T0/Xq+XmtmGOW3SPboUd+9oIPwjNqIg/g3AHxJtPVIQdxcwKi5PAOYDR8X1hSX6NYCRwEpgdIx/I7CkQL8AuBiYCuwef5+Iy7sX6Bdmlu8ANojLawN3lZRraXZ/uW2LivZB+INgT+B84CngBuDjwDoF+iXxdzjwJLBGXLeSMt+V0YwEbonL/1ByTMcApwH3As8ATxNGnZwGrNvCNXN9Qdxo4OvARcBHc9u+U6DfCDgX+DYwDvhyLNcPgfEF+vVyYRzwMDAWWK9Av1eu/OcDS4BLgQ0L9KcB68flKcCDwB+BZSXX0QLgJGCzisdsCvDLeK1uAvwMWBGvwe1K0owCvgLcE7VPAbcB0wu0NwLHAhvljvGxwE0l9rcvCTsATxTor4jHaX/gmri+ZlG9iHE3EB6Uj4vH/thY9iOAqwv0rwAP5cJL8ffB1OtUQUFBQUGhU4EW/GnSfeRUnzfVf03yR+O2JB+TdH8x1f9L9eeS/DMS/a0WfKckX4jMs0ssxzvj8hbA/AL97cD7gYOAR4EPxfg9gN8V6C8j+Os7AxvHsHOMm1tyTXyX4G8fDdwJnFV2Hca4ZwnPfCvj8rOEUZvPAisL9Pdkln8KHBCXpwK/LdD/GvgswR+9GzgmnutPAjevbpl7vbz9Fkhsq0nomyChzSX9eT617az8LExCO0jC8yPp7dlDwBnAI4S2aibw5pLzWLmeEvoXzgS+A/yC8MfAbsB/AhcV2K5c5wh/FnwZ2BX4BuHZ/X3Az4EjctrUtmJBZvlM4HuE+/3ZwIU5bWo715bzmHIOo77yPZX0voXK90fCH/zT43X0eeDfgLcD3ye8ZZDVpt7Xs+fxu8DXgE3jsflxTpt0jy49rlWFgxUITuCmBfGbAvcVxC8pCXcB/9uoUsb1UfGCOIuSDtmi5bhepB8WT8jPgMkxrrQDClhMcETHFVxMC0vSXA4cEpfnAFMyF+QdjS6cuD4CmEZo/J4q0N9NuJGNJTgU68X4tcjcnLIVJHPhjs2WgzCNR8uVNbO9pzscSai8A8c0s1zFKU9tMJMbCNIat1XqarNtBAf25ljefHixQL8ot34i8FtCXSo6x9m6/EgjWzHumHhdTMoe57JyKSgoKCgo1CWQ6E/Hbak+cqrPm+q/JvmjcVu/dbIm+Vst+E5JvhChA394XL6t7Fhk4hZmlvO+2cIC/f0N9l24jUxHG+EPitnAlYSRmEX7+CZwIZmOriZlvi+zfEfZvjtV5l4vb78F0vs+FhYtx/VFufXKbS7pz/O16Mgi4fmR9PYsm+fdCJ3Af462Z+S0levpwHki/EH6Z8Ay60V1qHKdK1i/Lf6uSe6+SXpbkb32FhH/DCnKd2q9b9d5TDmH+WPdbFuTMi4qSF/5/ggszq3fEX+HEaa6aVTvm93XFzRIm19POo+lx66qcLACYU6gPwLXxwt7drxo/kjmH66M/klgMqFRyoYJhDnY8vqbiY5vJm444Wb69wL9PGDkwEnMxI8pOkmZ7RsTGvJv5U9ATvcwYdTCQ/F3fIwfVXQxZvb9PeCBmL+XYtpfAdsW6EtP+EDZcnEzo71lwJGEf9DOI9xsZhXojyJ0xJ5HuDEO3Lw2AH5doO+7DseUyhvjUp3y1AazUXkLrxfSGrebgC/xegd1Q8JD3c9L7N8NvL1k26Mlx2hYLm46YST1skb5B77W7JjG+IF6fBZhWhqNUFZQUFBQqH0g0Z+OaZJ85Iymqs+b6r8m+aMxTb91sib5W6m+U+78NvWFCIMxbgLeQxgtdw5hJNvJFI/A+x3hTcoPx/O8f4zfneJO99uiNvtMNgz4CGEe8KI83VsQNyue57JR+zvE+nBktN+ozKfE6/ptwAmEUX+bAocA1xbo7yR06u0I/JXXOvo2p7hDKanMNS7vO6uUt98C6X0flfsmSGhzSX+eX5Rb70pHFunPjyntWVE51ojnbE4uvnI9zZYXuCC3bXHBPivXuVjfNovL22fPHfD7nDaprYjXzgHAB1m1gzrfT5BU79t1HlPOYdxW+Z5K476FojJWvj8C/wPsGpenATdmtuU7t1P7RB4jDBA8Jp5TK8s3iffo0rpUVTiYIVbAneMF+8G4vEaJ9vyBA16w7dKSC3CjEv0uBXFrlmjXJ+NQNijLPuRGcVY8BiOBtzbRjAa2JTgCq0yhkNFt0cL+30wc6QqsC3wI2LGBfmLUbFnBdt91OKZU3hiX6pSnNpjJDQRpjdtYwpxd9wLLCa+gLo1xq7weGtN8CHhHybb9C+JOB95bEL8XxTfCrxBf683Fbw78qMk1O43gLPy5ah1SUFBQUFDoZiDBn476JB+5QFPJ56Wi/xq1qf5ov3WyJvlbqb5TTlPJFyK8kj2XMP3eXcB1wAyKX/HfljC6/Hpgy3h8/haP/7sL9BOi7b8Qpg24Py7PpeS5ifAKfVHn3KeAl5rUnyOB31AwUKngmplH6Dx5lvghNGBMgXYP4L54nnYljPD8QyzHfg3K/FQs74C2sMwdKu8hg1Te/Rvtp18CaX0fyX0TVO8zSHmer0VHFonPj5ltTdsz4AcJ53CgnjZtmwgjtYueSTcDbi2xP71KnSPcax6JdewhYKcYvwFwek6b1FYQOrbnZMKGMX4j4Bc5bVK9b9d5TDmHUV/5nkoLfQtUvD8C2xBG7C8HbiX258XzeGROm9onMisXBqbf3YhVpzFJukeXHteUk6CgUCXkKuszuco6tiRNT3c4plTeTJqyRmd4gTa1wUxuIFIatxi/JfDe/HGlZIRUJs0eVdM00L9/sO0T5irbulkZFBQUFBQUFLoTUn3MVH8xp1ndTtYif64V/yzJ31od36mKL9SCr/WPifqdCKN8xwG7AF8A9m5yDnbktalHtiJ0bpWmyel3A/49QT+R0HHWSL9TTt+0DFE7LoaLE+tF4bNGiXY88HSi/VX+IGmiv5Zch6RCb4XUtpM2dmSltDntbM9S2qaSNmkfMp3tDfQN2xjgXSntXS5t5bYiVZ9S7wkd0Z8H9qyg3Y3wbakq2oZ24zkcE5dHEvqIriX4EGMKtNk5zU8GflKkLdBXsZ3NR6ltwp+BmySch8r6VNtlYWB+FyE6gpkd4u5z2pmmHXozeyPhVZO765CfTuqL0lj4gvDnCA9ykwkf+rk6blvg7tsX2EhKY2ZHAIe3UZ9cBiGEEELUk6Ho/3XAd0q134r+MMKfAFX0swhzrQ4nzOW9I3AL4UNUN7r7KQXHLZ9mJ8L0cIVpUvfRAf01+TIRRiPeDODu05roDfjnBH2q/UHVi96nnW3h6jxztrO9TKnX7WyTErVtaytaaCdud/cd4/KhhPN0FWHE+k/c/bQG2sOAH1fUltqNmnsIU8O8bGazgecJo6z3iPEfaKB9AfhRkbadts1sRbT1AOGbaZe7+1MFxz9Zn2q7lNXtlVZQSAk0mItvsNJI3/5zQOIX5FtJUze9goKCgoKCQn1D3fynwdDXzRfqkH4Nwsitlbx+lFjh/LypaWqoX0B4TX0q4fX/qcATcXn3Av3CRH2q/bbmR6H3A21sC/PalDakxfYmxXalet3ONiNRm9pWVNa3YjuzfAevjWRfm1U/YtcWbYxfmlnOz1G+qFVtO23HYz2M0FF+PmGqpBuAjwPrlJzHSvpU22VhOEIMMma2pGwTYd671U4j/eDqW0gzzN2fA3D3h81sKvAjM9s06otITVM3vRBCCCG6SN38pw74Z3Xzhdqtf9nd/w68YGYPuPvKmPZFM3ulQN9KmrrppxA+onYi8EV3X2RmL7r7r0rKu0OiPtV+u/MjeoB2tm1tfOZsZ/uUUq/b2WakaFPrZoo+1fYwMxtL6MA0j6Nh3f15M3u5Q1qA7FtHi81sirvPN7MtCB+6bFXbTtvu7q8Qvt9wk5mNIIxWPwg4gzBVaav6VNuFqFNZtIMNgX8hzM2bxQgfgxuMNNIPrj41zZNmNtndFwG4+3Nmti9wATCpxH5qmrrphRBCCNFd6uY/tVtfN1+o3fr/M7OR7v4CodMCADMbA5R1KqemqZU+PtCfbWaXx98nafCM3ut60TO0s21r1zNnO9unlHrdzjajsraddbmFej8GuJNwjt3Mxrv7E2Y2ilU78NulhfCRwnPM7CTCRxF/Z2aPAo/Gba1q22n7deVw95eAa4BrzGxkQT5S9Km2C1GDL9rBtYRXSRblN5jZLYOURvrB1aemORh43b9/7v4ycLCZ/XeJ/dQ0ddMLIYQQorvUzX9qt75uvlC79f/k7v8bddkOkhGE13GLSE1TNz1R+xjwYTPbh/Bae0N6XS9qTzvbtnY9c7azfUqp1+1sM5Lbl3bW5apad59QsukV4IBOaKN+BTDdzEYDbyX0hz7m7k+ujrbNtj9SUkbinwuro0+1XYg+1CeEEEIIIYQQQgghhBCiMsO6nQEhhBBCCCGEEEIIIYQQvYM6lYUQQgghhBBCCCGEEEJURp3KQgghhBBCCCGEEEIIISqjTmUhhBBCCCGEEEIIIYQQlVGnshBCCCGEEEIIIYQQQojK/D86prIaALDlaAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 1800x504 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "px_sizes = [25, (25, 50), 50, 80]\n",
    "\n",
    "fig, ax = plt.subplots(1, len(px_sizes), figsize=(25, 7))\n",
    "for ix, px in enumerate(px_sizes):\n",
    "    it.pixels = px\n",
    "    fdm = it.feature_density_matrix()\n",
    "    fdm[fdm == 0] = np.nan\n",
    "    cax = sns.heatmap(fdm, cmap=\"viridis\", linewidth=0.01, \n",
    "                      linecolor=\"lightgrey\", square=True, \n",
    "                      ax=ax[ix], cbar=False)\n",
    "    cax.set_title('Dim {} x {}'.format(*it.pixels))\n",
    "    for _, spine in cax.spines.items():\n",
    "        spine.set_visible(True)\n",
    "\n",
    "it.pixels = 50"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "bcb02891",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T02:29:45.192344Z",
     "start_time": "2021-10-16T02:28:56.723054Z"
    }
   },
   "outputs": [],
   "source": [
    "mat_train = it.transform(train_norm)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "c27cf5b5",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T02:29:45.208309Z",
     "start_time": "2021-10-16T02:29:45.194321Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((68602, 50, 50, 3), numpy.ndarray)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mat_train.shape, type(mat_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "5720681f",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T02:29:49.523368Z",
     "start_time": "2021-10-16T02:29:45.219310Z"
    }
   },
   "outputs": [],
   "source": [
    "np.save('np_files/train_image_matrix_50_50.npy', mat_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "6f0d5646",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T03:35:09.588131Z",
     "start_time": "2021-10-15T03:34:43.251498Z"
    }
   },
   "outputs": [],
   "source": [
    "# test loading\n",
    "mat_train = np.load('np_files/train_image_matrix_50_50.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "597dca76",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T02:29:49.538310Z",
     "start_time": "2021-10-16T02:29:49.526311Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(68602, 50, 50, 3)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mat_train.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e624134d",
   "metadata": {},
   "source": [
    "## Train/val split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "b935ab22",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T02:29:50.653343Z",
     "start_time": "2021-10-16T02:29:49.542311Z"
    }
   },
   "outputs": [],
   "source": [
    "X_train, X_val, y_train, y_val = train_test_split(mat_train, target, \n",
    "                                                  test_size= 0.2, \n",
    "                                                  random_state=42,\n",
    "                                                  stratify=is_good)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "dd476a3c",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T02:29:50.669310Z",
     "start_time": "2021-10-16T02:29:50.658310Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((54881, 50, 50, 3), (54881, 1))"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape, y_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "642b1b5b",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T02:29:50.685353Z",
     "start_time": "2021-10-16T02:29:50.672310Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((13721, 50, 50, 3), (13721, 1))"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_val.shape, y_val.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "deab9ff4",
   "metadata": {},
   "source": [
    "### Tensorflow NN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "a728f2ed",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-17T09:22:42.184582Z",
     "start_time": "2021-10-17T09:22:38.878586Z"
    }
   },
   "outputs": [],
   "source": [
    "from tensorflow.keras import Sequential\n",
    "from tensorflow.keras.layers import InputLayer, Dense, Dropout, Conv1D, Conv2D, AveragePooling1D, AveragePooling2D, LeakyReLU\n",
    "from tensorflow.keras.layers import Input, Flatten, Reshape, Lambda, BatchNormalization, Conv2D, MaxPooling2D, GlobalAveragePooling2D\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "from tensorflow.keras.regularizers import L2\n",
    "from tensorflow.keras.callbacks import EarlyStopping\n",
    "from tensorflow.keras.applications.efficientnet import EfficientNetB1\n",
    "from tensorflow.keras.models import load_model\n",
    "import tensorflow as tf\n",
    "import tensorflow_addons as tfa\n",
    "import tensorflow.keras.backend as K\n",
    "\n",
    "\n",
    "from sklearn.model_selection import KFold\n",
    "from sklearn.metrics import mean_squared_error, mean_absolute_error, r2_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "186c5984",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T02:29:59.769267Z",
     "start_time": "2021-10-16T02:29:58.207281Z"
    },
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "physical_devices = tf.config.list_physical_devices('GPU')\n",
    "tf.config.experimental.set_memory_growth(physical_devices[0], True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "90dbbf25",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-17T09:22:42.199585Z",
     "start_time": "2021-10-17T09:22:42.186600Z"
    }
   },
   "outputs": [],
   "source": [
    "def r2_metric(y_true, y_pred):\n",
    "    numerator = K.sum((y_true - y_pred) ** 2, axis=0)\n",
    "    denominator = K.sum((y_true - K.mean(y_pred, axis=0)) ** 2, axis=0)\n",
    "    return 1-(numerator/denominator)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "9a9269db",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-17T09:22:42.215586Z",
     "start_time": "2021-10-17T09:22:42.202586Z"
    }
   },
   "outputs": [],
   "source": [
    "def LMAE(y_actual, y_pred): \n",
    "    lmae=tf.experimental.numpy.log10(K.mean(K.abs(y_actual - y_pred), axis=0))\n",
    "    return lmae"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "4116c84f",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T02:29:59.847267Z",
     "start_time": "2021-10-16T02:29:59.823266Z"
    }
   },
   "outputs": [],
   "source": [
    "tf.keras.backend.clear_session()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "778f25bf",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T02:29:59.877267Z",
     "start_time": "2021-10-16T02:29:59.854264Z"
    }
   },
   "outputs": [],
   "source": [
    "BATCH_SIZE = 256\n",
    "EPOCHS = 200\n",
    "INIT_LR = 0.001"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "877ae000",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T02:29:59.892267Z",
     "start_time": "2021-10-16T02:29:59.884266Z"
    }
   },
   "outputs": [],
   "source": [
    "# sklearn use GlorotUniform (from source code)\n",
    "xavier = tf.keras.initializers.GlorotUniform(seed=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "ca27a1f9",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T02:29:59.908264Z",
     "start_time": "2021-10-16T02:29:59.898265Z"
    }
   },
   "outputs": [],
   "source": [
    "optimizer = Adam(learning_rate=INIT_LR, beta_1=0.9, beta_2=0.999, epsilon=1e-8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 321,
   "id": "67639d51",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T19:43:08.378127Z",
     "start_time": "2021-10-15T19:43:08.363086Z"
    }
   },
   "outputs": [],
   "source": [
    "# , kernel_regularizer=L2(0.01)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 322,
   "id": "37aaa7f9",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T19:43:08.871779Z",
     "start_time": "2021-10-15T19:43:08.864779Z"
    }
   },
   "outputs": [],
   "source": [
    "# # 1st architect with 256 batch size\n",
    "# model = Sequential([\n",
    "#     InputLayer(input_shape=(80,80,3), dtype='float32'),\n",
    "#     Conv2D(32, 3, activation='relu', padding='same'),\n",
    "#     Conv2D(64, 3, activation='relu', padding='same'),\n",
    "#     Conv2D(64, 3, activation='relu', padding='same'),\n",
    "#     BatchNormalization(),\n",
    "#     MaxPooling2D((2,2)),\n",
    "#     Dropout(0.25),\n",
    "#     Conv2D(128, 3, activation='relu', padding='same'),\n",
    "#     Conv2D(128, 3, activation='relu', padding='same'),\n",
    "#     BatchNormalization(),\n",
    "#     MaxPooling2D((2,2)),\n",
    "#     Dropout(0.25),\n",
    "#     Conv2D(256, 3, activation='relu', padding='same'),\n",
    "#     Conv2D(256, 3, activation='relu', padding='same'),\n",
    "#     BatchNormalization(),\n",
    "#     MaxPooling2D((2,2)),\n",
    "#     Dropout(0.25),\n",
    "#     Flatten(),\n",
    "#     Dense(512, activation='relu'),\n",
    "#     Dense(256, activation='relu'),\n",
    "#     BatchNormalization(),\n",
    "#     Dropout(0.4),\n",
    "#     Dense(64, activation='relu'),\n",
    "#     BatchNormalization(),\n",
    "#     Dense(16, activation='relu'),\n",
    "#     Dense(1, activation='linear', kernel_initializer=xavier),\n",
    "# ])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "1548a363",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T02:30:37.701578Z",
     "start_time": "2021-10-16T02:30:37.479577Z"
    }
   },
   "outputs": [],
   "source": [
    "# 2nd architect\n",
    "model = Sequential([\n",
    "    InputLayer(input_shape=(50,50,3), dtype='float32'),\n",
    "    Conv2D(32, 3, activation='relu', padding='same'),\n",
    "    Conv2D(32, 3, activation='relu', padding='same'),\n",
    "    BatchNormalization(),\n",
    "    Dropout(0.2),\n",
    "    Conv2D(64, 3, activation='relu', padding='same'),\n",
    "    Conv2D(64, 3, activation='relu', padding='same'),\n",
    "    BatchNormalization(),\n",
    "    MaxPooling2D((2,2)),\n",
    "    Dropout(0.2),\n",
    "    Conv2D(128, 3, activation='relu', padding='same'),\n",
    "    Conv2D(128, 3, activation='relu', padding='same'),\n",
    "    BatchNormalization(),\n",
    "    MaxPooling2D((2,2)),\n",
    "    Dropout(0.2),\n",
    "    Conv2D(256, 3, activation='relu', padding='same'),\n",
    "    Conv2D(256, 3, activation='relu', padding='same'),\n",
    "    BatchNormalization(),\n",
    "    MaxPooling2D((2,2)),\n",
    "    Dropout(0.2),\n",
    "    GlobalAveragePooling2D(),\n",
    "    Dense(128, activation='relu'),\n",
    "    BatchNormalization(),\n",
    "    Dropout(0.5),\n",
    "    Dense(64, activation='relu'),\n",
    "    BatchNormalization(),\n",
    "    Dense(16, activation='relu'),\n",
    "    Dense(8, activation='relu'),\n",
    "    Dense(1, activation='linear'),\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "2250d681",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T02:30:37.731551Z",
     "start_time": "2021-10-16T02:30:37.704551Z"
    },
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_8 (Conv2D)            (None, 50, 50, 32)        896       \n",
      "_________________________________________________________________\n",
      "conv2d_9 (Conv2D)            (None, 50, 50, 32)        9248      \n",
      "_________________________________________________________________\n",
      "batch_normalization_6 (Batch (None, 50, 50, 32)        128       \n",
      "_________________________________________________________________\n",
      "dropout_5 (Dropout)          (None, 50, 50, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_10 (Conv2D)           (None, 50, 50, 64)        18496     \n",
      "_________________________________________________________________\n",
      "conv2d_11 (Conv2D)           (None, 50, 50, 64)        36928     \n",
      "_________________________________________________________________\n",
      "batch_normalization_7 (Batch (None, 50, 50, 64)        256       \n",
      "_________________________________________________________________\n",
      "max_pooling2d_3 (MaxPooling2 (None, 25, 25, 64)        0         \n",
      "_________________________________________________________________\n",
      "dropout_6 (Dropout)          (None, 25, 25, 64)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_12 (Conv2D)           (None, 25, 25, 128)       73856     \n",
      "_________________________________________________________________\n",
      "conv2d_13 (Conv2D)           (None, 25, 25, 128)       147584    \n",
      "_________________________________________________________________\n",
      "batch_normalization_8 (Batch (None, 25, 25, 128)       512       \n",
      "_________________________________________________________________\n",
      "max_pooling2d_4 (MaxPooling2 (None, 12, 12, 128)       0         \n",
      "_________________________________________________________________\n",
      "dropout_7 (Dropout)          (None, 12, 12, 128)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_14 (Conv2D)           (None, 12, 12, 256)       295168    \n",
      "_________________________________________________________________\n",
      "conv2d_15 (Conv2D)           (None, 12, 12, 256)       590080    \n",
      "_________________________________________________________________\n",
      "batch_normalization_9 (Batch (None, 12, 12, 256)       1024      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_5 (MaxPooling2 (None, 6, 6, 256)         0         \n",
      "_________________________________________________________________\n",
      "dropout_8 (Dropout)          (None, 6, 6, 256)         0         \n",
      "_________________________________________________________________\n",
      "global_average_pooling2d_1 ( (None, 256)               0         \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 128)               32896     \n",
      "_________________________________________________________________\n",
      "batch_normalization_10 (Batc (None, 128)               512       \n",
      "_________________________________________________________________\n",
      "dropout_9 (Dropout)          (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense_6 (Dense)              (None, 64)                8256      \n",
      "_________________________________________________________________\n",
      "batch_normalization_11 (Batc (None, 64)                256       \n",
      "_________________________________________________________________\n",
      "dense_7 (Dense)              (None, 16)                1040      \n",
      "_________________________________________________________________\n",
      "dense_8 (Dense)              (None, 8)                 136       \n",
      "_________________________________________________________________\n",
      "dense_9 (Dense)              (None, 1)                 9         \n",
      "=================================================================\n",
      "Total params: 1,217,281\n",
      "Trainable params: 1,215,937\n",
      "Non-trainable params: 1,344\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "83a71891",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T02:30:43.222416Z",
     "start_time": "2021-10-16T02:30:43.209385Z"
    }
   },
   "outputs": [],
   "source": [
    "# callbacks\n",
    "early_stopping = EarlyStopping(patience=10, restore_best_weights=True, monitor='val_loss', mode='min')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "930eee81",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T02:30:43.704846Z",
     "start_time": "2021-10-16T02:30:43.681810Z"
    }
   },
   "outputs": [],
   "source": [
    "model.compile(optimizer=optimizer, loss=LMAE, metrics=[])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "9bd318b9",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T03:50:56.863538Z",
     "start_time": "2021-10-16T02:30:44.378486Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "215/215 [==============================] - 72s 325ms/step - loss: 1.8891 - val_loss: 2.5073\n",
      "Epoch 2/200\n",
      "215/215 [==============================] - 86s 399ms/step - loss: 1.5038 - val_loss: 1.8548\n",
      "Epoch 3/200\n",
      "215/215 [==============================] - 99s 463ms/step - loss: 1.4496 - val_loss: 1.9176\n",
      "Epoch 4/200\n",
      "215/215 [==============================] - 118s 552ms/step - loss: 1.4209 - val_loss: 1.8079\n",
      "Epoch 5/200\n",
      "215/215 [==============================] - 128s 597ms/step - loss: 1.4063 - val_loss: 1.6469\n",
      "Epoch 6/200\n",
      "215/215 [==============================] - 134s 623ms/step - loss: 1.3937 - val_loss: 1.8381\n",
      "Epoch 7/200\n",
      "215/215 [==============================] - 144s 670ms/step - loss: 1.3871 - val_loss: 1.8556\n",
      "Epoch 8/200\n",
      "215/215 [==============================] - 146s 682ms/step - loss: 1.3804 - val_loss: 2.1213\n",
      "Epoch 9/200\n",
      "215/215 [==============================] - 151s 704ms/step - loss: 1.3749 - val_loss: 1.5904\n",
      "Epoch 10/200\n",
      "215/215 [==============================] - 154s 717ms/step - loss: 1.3728 - val_loss: 1.5461\n",
      "Epoch 11/200\n",
      "215/215 [==============================] - 165s 770ms/step - loss: 1.3652 - val_loss: 1.6484\n",
      "Epoch 12/200\n",
      "215/215 [==============================] - 157s 731ms/step - loss: 1.3650 - val_loss: 1.9548\n",
      "Epoch 13/200\n",
      "215/215 [==============================] - 163s 756ms/step - loss: 1.3638 - val_loss: 1.7090\n",
      "Epoch 14/200\n",
      "215/215 [==============================] - 158s 738ms/step - loss: 1.3585 - val_loss: 1.7558\n",
      "Epoch 15/200\n",
      "215/215 [==============================] - 162s 754ms/step - loss: 1.3545 - val_loss: 1.6672\n",
      "Epoch 16/200\n",
      "215/215 [==============================] - 165s 768ms/step - loss: 1.3528 - val_loss: 1.8481\n",
      "Epoch 17/200\n",
      "215/215 [==============================] - 167s 778ms/step - loss: 1.3522 - val_loss: 1.4111\n",
      "Epoch 18/200\n",
      "215/215 [==============================] - 150s 699ms/step - loss: 1.3493 - val_loss: 1.5298\n",
      "Epoch 19/200\n",
      "215/215 [==============================] - 102s 472ms/step - loss: 1.3466 - val_loss: 1.5438\n",
      "Epoch 20/200\n",
      "215/215 [==============================] - 88s 411ms/step - loss: 1.3451 - val_loss: 1.5951\n",
      "Epoch 21/200\n",
      "215/215 [==============================] - 82s 380ms/step - loss: 1.3432 - val_loss: 1.5032\n",
      "Epoch 22/200\n",
      "215/215 [==============================] - 86s 401ms/step - loss: 1.3448 - val_loss: 1.5432\n",
      "Epoch 23/200\n",
      "215/215 [==============================] - 83s 386ms/step - loss: 1.3406 - val_loss: 1.3650\n",
      "Epoch 24/200\n",
      "215/215 [==============================] - 83s 385ms/step - loss: 1.3405 - val_loss: 1.6688\n",
      "Epoch 25/200\n",
      "215/215 [==============================] - 80s 371ms/step - loss: 1.3370 - val_loss: 1.7240\n",
      "Epoch 26/200\n",
      "215/215 [==============================] - 85s 396ms/step - loss: 1.3355 - val_loss: 1.4734\n",
      "Epoch 27/200\n",
      "215/215 [==============================] - 81s 375ms/step - loss: 1.3349 - val_loss: 1.5513\n",
      "Epoch 28/200\n",
      "215/215 [==============================] - 82s 383ms/step - loss: 1.3348 - val_loss: 1.7251\n",
      "Epoch 29/200\n",
      "215/215 [==============================] - 79s 366ms/step - loss: 1.3323 - val_loss: 1.4915\n",
      "Epoch 30/200\n",
      "215/215 [==============================] - 81s 379ms/step - loss: 1.3311 - val_loss: 1.3519\n",
      "Epoch 31/200\n",
      "215/215 [==============================] - 92s 429ms/step - loss: 1.3310 - val_loss: 1.6364\n",
      "Epoch 32/200\n",
      "215/215 [==============================] - 97s 452ms/step - loss: 1.3316 - val_loss: 1.3511\n",
      "Epoch 33/200\n",
      "215/215 [==============================] - 111s 515ms/step - loss: 1.3277 - val_loss: 1.5791\n",
      "Epoch 34/200\n",
      "215/215 [==============================] - 106s 493ms/step - loss: 1.3276 - val_loss: 1.7843\n",
      "Epoch 35/200\n",
      "215/215 [==============================] - 109s 509ms/step - loss: 1.3257 - val_loss: 1.6244\n",
      "Epoch 36/200\n",
      "215/215 [==============================] - 108s 502ms/step - loss: 1.3263 - val_loss: 1.6297\n",
      "Epoch 37/200\n",
      "215/215 [==============================] - 106s 492ms/step - loss: 1.3272 - val_loss: 1.4666\n",
      "Epoch 38/200\n",
      "215/215 [==============================] - 109s 509ms/step - loss: 1.3252 - val_loss: 1.3830\n",
      "Epoch 39/200\n",
      "215/215 [==============================] - 111s 514ms/step - loss: 1.3245 - val_loss: 1.3855\n",
      "Epoch 40/200\n",
      "215/215 [==============================] - 110s 514ms/step - loss: 1.3231 - val_loss: 1.5983\n",
      "Epoch 41/200\n",
      "215/215 [==============================] - 110s 513ms/step - loss: 1.3192 - val_loss: 1.6872\n",
      "Epoch 42/200\n",
      "215/215 [==============================] - 110s 513ms/step - loss: 1.3192 - val_loss: 1.8177\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x2bf2cfd51c8>"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(X_train, y_train,\n",
    "          validation_data=(X_val, y_val),\n",
    "          batch_size=BATCH_SIZE,\n",
    "          epochs=EPOCHS,\n",
    "          verbose=1,\n",
    "          callbacks=[early_stopping]\n",
    "         )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "d025d3da",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T03:54:37.916545Z",
     "start_time": "2021-10-16T03:54:13.515728Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MAE (more accurate):  22.213853073506375\n",
      "r2 square: 0.8792058807593872\n"
     ]
    }
   ],
   "source": [
    "y_hat = model.predict(mat_train).flatten()\n",
    "print('MAE (more accurate): ',mean_absolute_error(target, y_hat))\n",
    "## look at MAE is more accurate\n",
    "print('r2 square:', r2_score(target, y_hat))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 331,
   "id": "5d32d982",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T02:10:32.569235Z",
     "start_time": "2021-10-16T02:10:32.271181Z"
    }
   },
   "outputs": [],
   "source": [
    "model.save('models/2nd_deepinsight_cnn_80_80.h5', save_format='h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6b383721",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-13T15:48:52.896747Z",
     "start_time": "2021-10-13T15:48:52.896747Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# TF model with KFold cross-validation\n",
    "kfold = KFold(n_splits=5, shuffle=True)\n",
    "cv_loss_per_fold = []\n",
    "train_loss_per_fold = []\n",
    "fold_no = 1\n",
    "inputs = np.array(clean_X)\n",
    "targets = np.array(y)\n",
    "for train, test in kfold.split(inputs, targets):\n",
    "    model = Sequential([\n",
    "    InputLayer(input_shape=(clean_X.shape[1],), dtype='float32'),\n",
    "    Dense(128, activation='relu'),\n",
    "    Dense(256, activation='relu'),\n",
    "    Dropout(0.5),\n",
    "    Dense(256, activation='relu'),\n",
    "    Dropout(0.5),\n",
    "    Dense(64, activation='relu'),\n",
    "    Dense(16, activation='relu'),\n",
    "    Dense(1, activation='linear')\n",
    "    ])\n",
    "    \n",
    "    model.compile(optimizer=optimizer, loss='mean_absolute_error', metrics=[])\n",
    "    \n",
    "    print('------------------------------------------------------------------------')\n",
    "    print(f'Training for fold {fold_no} ...')\n",
    "    \n",
    "    H = model.fit(inputs[train], targets[train],\n",
    "          batch_size=BATCH_SIZE,\n",
    "          epochs=EPOCHS,\n",
    "          verbose=1,\n",
    "#           validation_split=0.2,\n",
    "#           callbacks=[early_stopping]\n",
    "          )\n",
    "    \n",
    "    # Generate generalization metrics\n",
    "    scores = model.evaluate(inputs[test], targets[test], verbose=0)\n",
    "    print(f'Training loss for fold {fold_no}: {H.history[\"loss\"][-1]}')\n",
    "    print(f'CV Score for fold {fold_no}: {model.metrics_names[0]} of {scores}')\n",
    "    cv_loss_per_fold.append(scores)\n",
    "    train_loss_per_fold.append(H.history[\"loss\"][-1])\n",
    "    fold_no += 1\n",
    "    \n",
    "# == Provide average scores ==\n",
    "print('------------------------------------------------------------------------')\n",
    "print('CV Score per fold')\n",
    "for i in range(0, len(cv_loss_per_fold)):\n",
    "    print('------------------------------------------------------------------------')\n",
    "    print(f'> Fold {i+1} - Loss: {cv_loss_per_fold[i]}')\n",
    "print('------------------------------------------------------------------------')\n",
    "print('Average scores for all folds:')\n",
    "print(f'> Train Average Loss: {np.mean(train_loss_per_fold)}')\n",
    "print(f'> CV Average Loss: {np.mean(cv_loss_per_fold)}')\n",
    "print('------------------------------------------------------------------------')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f459a6ca",
   "metadata": {},
   "source": [
    "## Test and Submission"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "6ba7520f",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T09:21:52.628909Z",
     "start_time": "2021-10-16T09:21:52.407912Z"
    }
   },
   "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>volume</th>\n",
       "      <th>volume/g</th>\n",
       "      <th>density</th>\n",
       "      <th>weight</th>\n",
       "      <th>surface_area</th>\n",
       "      <th>void_fraction</th>\n",
       "      <th>void_volume</th>\n",
       "      <th>CO2/N2_selectivity</th>\n",
       "      <th>heat_adsorption</th>\n",
       "      <th>n_atoms</th>\n",
       "      <th>...</th>\n",
       "      <th>topology_1</th>\n",
       "      <th>topology_2</th>\n",
       "      <th>topology_3</th>\n",
       "      <th>topology_4</th>\n",
       "      <th>topology_5</th>\n",
       "      <th>topology_6</th>\n",
       "      <th>topology_7</th>\n",
       "      <th>topology_8</th>\n",
       "      <th>topology_9</th>\n",
       "      <th>topology_10</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1208.301332</td>\n",
       "      <td>0.912181</td>\n",
       "      <td>1.096273</td>\n",
       "      <td>797.70936</td>\n",
       "      <td>586.54</td>\n",
       "      <td>0.11392</td>\n",
       "      <td>0.1039</td>\n",
       "      <td>36.639791</td>\n",
       "      <td>7.005640</td>\n",
       "      <td>84</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4126.414623</td>\n",
       "      <td>0.665563</td>\n",
       "      <td>1.502488</td>\n",
       "      <td>3733.65779</td>\n",
       "      <td>852.49</td>\n",
       "      <td>0.21367</td>\n",
       "      <td>0.1422</td>\n",
       "      <td>18.390691</td>\n",
       "      <td>5.119399</td>\n",
       "      <td>158</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1602.148373</td>\n",
       "      <td>1.291250</td>\n",
       "      <td>0.774443</td>\n",
       "      <td>747.21048</td>\n",
       "      <td>3155.73</td>\n",
       "      <td>0.33883</td>\n",
       "      <td>0.4375</td>\n",
       "      <td>13.062850</td>\n",
       "      <td>5.045400</td>\n",
       "      <td>60</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2436.629312</td>\n",
       "      <td>1.473557</td>\n",
       "      <td>0.678630</td>\n",
       "      <td>995.80232</td>\n",
       "      <td>3521.09</td>\n",
       "      <td>0.40464</td>\n",
       "      <td>0.5963</td>\n",
       "      <td>9.601198</td>\n",
       "      <td>5.106238</td>\n",
       "      <td>94</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3123.418006</td>\n",
       "      <td>1.406289</td>\n",
       "      <td>0.711091</td>\n",
       "      <td>1337.53800</td>\n",
       "      <td>2678.46</td>\n",
       "      <td>0.38959</td>\n",
       "      <td>0.5479</td>\n",
       "      <td>12.974954</td>\n",
       "      <td>5.287639</td>\n",
       "      <td>118</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16995</th>\n",
       "      <td>32660.944605</td>\n",
       "      <td>4.163884</td>\n",
       "      <td>0.240160</td>\n",
       "      <td>4723.68288</td>\n",
       "      <td>5720.14</td>\n",
       "      <td>0.77614</td>\n",
       "      <td>3.2318</td>\n",
       "      <td>4.536626</td>\n",
       "      <td>3.146698</td>\n",
       "      <td>416</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16996</th>\n",
       "      <td>5070.998617</td>\n",
       "      <td>2.037061</td>\n",
       "      <td>0.490903</td>\n",
       "      <td>1499.13262</td>\n",
       "      <td>4017.28</td>\n",
       "      <td>0.59192</td>\n",
       "      <td>1.2058</td>\n",
       "      <td>6.745508</td>\n",
       "      <td>3.658871</td>\n",
       "      <td>124</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16997</th>\n",
       "      <td>4669.804446</td>\n",
       "      <td>2.127168</td>\n",
       "      <td>0.470109</td>\n",
       "      <td>1322.04892</td>\n",
       "      <td>4288.76</td>\n",
       "      <td>0.54950</td>\n",
       "      <td>1.1689</td>\n",
       "      <td>4.666206</td>\n",
       "      <td>3.593052</td>\n",
       "      <td>112</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16998</th>\n",
       "      <td>4682.120862</td>\n",
       "      <td>2.323536</td>\n",
       "      <td>0.430379</td>\n",
       "      <td>1213.51148</td>\n",
       "      <td>4331.86</td>\n",
       "      <td>0.60643</td>\n",
       "      <td>1.4091</td>\n",
       "      <td>4.823305</td>\n",
       "      <td>3.454497</td>\n",
       "      <td>104</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16999</th>\n",
       "      <td>25702.628698</td>\n",
       "      <td>3.229246</td>\n",
       "      <td>0.309670</td>\n",
       "      <td>4793.21712</td>\n",
       "      <td>5006.39</td>\n",
       "      <td>0.71730</td>\n",
       "      <td>2.3163</td>\n",
       "      <td>4.278184</td>\n",
       "      <td>3.169874</td>\n",
       "      <td>420</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>17000 rows × 144 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             volume  volume/g   density      weight  surface_area  \\\n",
       "0       1208.301332  0.912181  1.096273   797.70936        586.54   \n",
       "1       4126.414623  0.665563  1.502488  3733.65779        852.49   \n",
       "2       1602.148373  1.291250  0.774443   747.21048       3155.73   \n",
       "3       2436.629312  1.473557  0.678630   995.80232       3521.09   \n",
       "4       3123.418006  1.406289  0.711091  1337.53800       2678.46   \n",
       "...             ...       ...       ...         ...           ...   \n",
       "16995  32660.944605  4.163884  0.240160  4723.68288       5720.14   \n",
       "16996   5070.998617  2.037061  0.490903  1499.13262       4017.28   \n",
       "16997   4669.804446  2.127168  0.470109  1322.04892       4288.76   \n",
       "16998   4682.120862  2.323536  0.430379  1213.51148       4331.86   \n",
       "16999  25702.628698  3.229246  0.309670  4793.21712       5006.39   \n",
       "\n",
       "       void_fraction  void_volume  CO2/N2_selectivity  heat_adsorption  \\\n",
       "0            0.11392       0.1039           36.639791         7.005640   \n",
       "1            0.21367       0.1422           18.390691         5.119399   \n",
       "2            0.33883       0.4375           13.062850         5.045400   \n",
       "3            0.40464       0.5963            9.601198         5.106238   \n",
       "4            0.38959       0.5479           12.974954         5.287639   \n",
       "...              ...          ...                 ...              ...   \n",
       "16995        0.77614       3.2318            4.536626         3.146698   \n",
       "16996        0.59192       1.2058            6.745508         3.658871   \n",
       "16997        0.54950       1.1689            4.666206         3.593052   \n",
       "16998        0.60643       1.4091            4.823305         3.454497   \n",
       "16999        0.71730       2.3163            4.278184         3.169874   \n",
       "\n",
       "       n_atoms  ...  topology_1  topology_2  topology_3  topology_4  \\\n",
       "0           84  ...           0           0           0           0   \n",
       "1          158  ...           0           0           0           0   \n",
       "2           60  ...           0           0           0           0   \n",
       "3           94  ...           0           0           0           0   \n",
       "4          118  ...           0           0           0           0   \n",
       "...        ...  ...         ...         ...         ...         ...   \n",
       "16995      416  ...           0           0           0           1   \n",
       "16996      124  ...           0           0           0           1   \n",
       "16997      112  ...           0           0           0           0   \n",
       "16998      104  ...           0           0           0           0   \n",
       "16999      420  ...           0           0           0           0   \n",
       "\n",
       "       topology_5  topology_6  topology_7  topology_8  topology_9  topology_10  \n",
       "0               1           0           0           0           0            0  \n",
       "1               0           0           0           0           0            0  \n",
       "2               1           0           0           0           0            0  \n",
       "3               1           0           0           0           0            0  \n",
       "4               1           0           0           0           0            0  \n",
       "...           ...         ...         ...         ...         ...          ...  \n",
       "16995           0           0           0           0           0            0  \n",
       "16996           0           0           0           0           0            0  \n",
       "16997           1           0           0           0           0            0  \n",
       "16998           1           0           0           0           0            0  \n",
       "16999           1           0           0           0           0            0  \n",
       "\n",
       "[17000 rows x 144 columns]"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test = pd.read_csv('test_extra_2_no_scaling.csv')\n",
    "test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "53276659",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T09:21:54.044050Z",
     "start_time": "2021-10-16T09:21:53.835054Z"
    }
   },
   "outputs": [],
   "source": [
    "test_norm = ln.transform(test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "4badcf64",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T09:22:03.962129Z",
     "start_time": "2021-10-16T09:21:56.361130Z"
    }
   },
   "outputs": [],
   "source": [
    "mat_test = it.transform(test_norm)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "d90096a0",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-16T09:23:37.633634Z",
     "start_time": "2021-10-16T09:23:36.543666Z"
    }
   },
   "outputs": [],
   "source": [
    "np.save('np_files/test_image_matrix_50_50.npy', mat_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 176,
   "id": "b0146822",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T15:44:30.680350Z",
     "start_time": "2021-10-15T15:44:20.447391Z"
    }
   },
   "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>id</th>\n",
       "      <th>CO2_working_capacity [mL/g]</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>68614</td>\n",
       "      <td>177.120590</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>68615</td>\n",
       "      <td>60.852951</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>68616</td>\n",
       "      <td>62.032269</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>68617</td>\n",
       "      <td>59.481808</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>68618</td>\n",
       "      <td>61.565498</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16995</th>\n",
       "      <td>85609</td>\n",
       "      <td>-7.604535</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16996</th>\n",
       "      <td>85610</td>\n",
       "      <td>0.004563</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16997</th>\n",
       "      <td>85611</td>\n",
       "      <td>0.004563</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16998</th>\n",
       "      <td>85612</td>\n",
       "      <td>-1.422112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16999</th>\n",
       "      <td>85613</td>\n",
       "      <td>-5.238155</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>17000 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          id  CO2_working_capacity [mL/g]\n",
       "0      68614                   177.120590\n",
       "1      68615                    60.852951\n",
       "2      68616                    62.032269\n",
       "3      68617                    59.481808\n",
       "4      68618                    61.565498\n",
       "...      ...                          ...\n",
       "16995  85609                    -7.604535\n",
       "16996  85610                     0.004563\n",
       "16997  85611                     0.004563\n",
       "16998  85612                    -1.422112\n",
       "16999  85613                    -5.238155\n",
       "\n",
       "[17000 rows x 2 columns]"
      ]
     },
     "execution_count": 176,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predict = model.predict(mat_test).flatten()\n",
    "\n",
    "predicted_df = pd.DataFrame({\"id\":range(68614,68614+len(test)), \"CO2_working_capacity [mL/g]\":predict})\n",
    "predicted_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "id": "6d3dbf76",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T15:45:09.683848Z",
     "start_time": "2021-10-15T15:45:09.604580Z"
    }
   },
   "outputs": [],
   "source": [
    "predicted_df.to_csv(\"output/phase2/submission.csv\",index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 206,
   "id": "3a1fab72",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T15:58:32.883586Z",
     "start_time": "2021-10-15T15:58:32.824572Z"
    }
   },
   "outputs": [],
   "source": [
    "best = pd.read_csv('output/phase2/best.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 207,
   "id": "fbc9c991",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T15:58:33.285372Z",
     "start_time": "2021-10-15T15:58:33.272374Z"
    }
   },
   "outputs": [],
   "source": [
    "fun_ensemble = best[['id']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 208,
   "id": "9e68347f",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T15:58:35.488766Z",
     "start_time": "2021-10-15T15:58:35.466766Z"
    }
   },
   "outputs": [],
   "source": [
    "fun_ensemble['CO2_working_capacity [mL/g]'] = (0.8*best['CO2_working_capacity [mL/g]']) + (0.2 * predicted_df['CO2_working_capacity [mL/g]'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 209,
   "id": "aa5c14da",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T15:58:36.326039Z",
     "start_time": "2021-10-15T15:58:36.247031Z"
    }
   },
   "outputs": [],
   "source": [
    "fun_ensemble.to_csv('output/phase2/submission.csv', index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "983d32b9",
   "metadata": {},
   "source": [
    "## Turn model to feature extractor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 180,
   "id": "a11ec796",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T15:47:31.228148Z",
     "start_time": "2021-10-15T15:47:30.276011Z"
    }
   },
   "outputs": [],
   "source": [
    "loaded_model = load_model('models/deepinsight_cnn_80_80.h5', custom_objects={'LMAE':LMAE})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "id": "d9c29187",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T15:47:41.180607Z",
     "start_time": "2021-10-15T15:47:41.158609Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_2\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_14 (Conv2D)           (None, 80, 80, 32)        896       \n",
      "_________________________________________________________________\n",
      "conv2d_15 (Conv2D)           (None, 80, 80, 64)        18496     \n",
      "_________________________________________________________________\n",
      "conv2d_16 (Conv2D)           (None, 80, 80, 64)        36928     \n",
      "_________________________________________________________________\n",
      "batch_normalization_10 (Batc (None, 80, 80, 64)        256       \n",
      "_________________________________________________________________\n",
      "max_pooling2d_3 (MaxPooling2 (None, 40, 40, 64)        0         \n",
      "_________________________________________________________________\n",
      "dropout_8 (Dropout)          (None, 40, 40, 64)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_17 (Conv2D)           (None, 40, 40, 128)       73856     \n",
      "_________________________________________________________________\n",
      "conv2d_18 (Conv2D)           (None, 40, 40, 128)       147584    \n",
      "_________________________________________________________________\n",
      "batch_normalization_11 (Batc (None, 40, 40, 128)       512       \n",
      "_________________________________________________________________\n",
      "max_pooling2d_4 (MaxPooling2 (None, 20, 20, 128)       0         \n",
      "_________________________________________________________________\n",
      "dropout_9 (Dropout)          (None, 20, 20, 128)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_19 (Conv2D)           (None, 20, 20, 256)       295168    \n",
      "_________________________________________________________________\n",
      "conv2d_20 (Conv2D)           (None, 20, 20, 256)       590080    \n",
      "_________________________________________________________________\n",
      "batch_normalization_12 (Batc (None, 20, 20, 256)       1024      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_5 (MaxPooling2 (None, 10, 10, 256)       0         \n",
      "_________________________________________________________________\n",
      "dropout_10 (Dropout)         (None, 10, 10, 256)       0         \n",
      "_________________________________________________________________\n",
      "flatten_2 (Flatten)          (None, 25600)             0         \n",
      "_________________________________________________________________\n",
      "dense_10 (Dense)             (None, 512)               13107712  \n",
      "_________________________________________________________________\n",
      "dense_11 (Dense)             (None, 256)               131328    \n",
      "_________________________________________________________________\n",
      "batch_normalization_13 (Batc (None, 256)               1024      \n",
      "_________________________________________________________________\n",
      "dropout_11 (Dropout)         (None, 256)               0         \n",
      "_________________________________________________________________\n",
      "dense_12 (Dense)             (None, 64)                16448     \n",
      "_________________________________________________________________\n",
      "batch_normalization_14 (Batc (None, 64)                256       \n",
      "_________________________________________________________________\n",
      "dense_13 (Dense)             (None, 16)                1040      \n",
      "_________________________________________________________________\n",
      "dense_14 (Dense)             (None, 1)                 17        \n",
      "=================================================================\n",
      "Total params: 14,422,625\n",
      "Trainable params: 14,421,089\n",
      "Non-trainable params: 1,536\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "loaded_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "id": "2e43c2f4",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T15:52:07.679029Z",
     "start_time": "2021-10-15T15:52:07.660031Z"
    }
   },
   "outputs": [],
   "source": [
    "feats_ext = tf.keras.Model(inputs = loaded_model.input, outputs = loaded_model.layers[-2].output)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "id": "b3675981",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T15:52:11.700767Z",
     "start_time": "2021-10-15T15:52:11.684699Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_3 (InputLayer)         [(None, 80, 80, 3)]       0         \n",
      "_________________________________________________________________\n",
      "conv2d_14 (Conv2D)           (None, 80, 80, 32)        896       \n",
      "_________________________________________________________________\n",
      "conv2d_15 (Conv2D)           (None, 80, 80, 64)        18496     \n",
      "_________________________________________________________________\n",
      "conv2d_16 (Conv2D)           (None, 80, 80, 64)        36928     \n",
      "_________________________________________________________________\n",
      "batch_normalization_10 (Batc (None, 80, 80, 64)        256       \n",
      "_________________________________________________________________\n",
      "max_pooling2d_3 (MaxPooling2 (None, 40, 40, 64)        0         \n",
      "_________________________________________________________________\n",
      "dropout_8 (Dropout)          (None, 40, 40, 64)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_17 (Conv2D)           (None, 40, 40, 128)       73856     \n",
      "_________________________________________________________________\n",
      "conv2d_18 (Conv2D)           (None, 40, 40, 128)       147584    \n",
      "_________________________________________________________________\n",
      "batch_normalization_11 (Batc (None, 40, 40, 128)       512       \n",
      "_________________________________________________________________\n",
      "max_pooling2d_4 (MaxPooling2 (None, 20, 20, 128)       0         \n",
      "_________________________________________________________________\n",
      "dropout_9 (Dropout)          (None, 20, 20, 128)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_19 (Conv2D)           (None, 20, 20, 256)       295168    \n",
      "_________________________________________________________________\n",
      "conv2d_20 (Conv2D)           (None, 20, 20, 256)       590080    \n",
      "_________________________________________________________________\n",
      "batch_normalization_12 (Batc (None, 20, 20, 256)       1024      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_5 (MaxPooling2 (None, 10, 10, 256)       0         \n",
      "_________________________________________________________________\n",
      "dropout_10 (Dropout)         (None, 10, 10, 256)       0         \n",
      "_________________________________________________________________\n",
      "flatten_2 (Flatten)          (None, 25600)             0         \n",
      "_________________________________________________________________\n",
      "dense_10 (Dense)             (None, 512)               13107712  \n",
      "_________________________________________________________________\n",
      "dense_11 (Dense)             (None, 256)               131328    \n",
      "_________________________________________________________________\n",
      "batch_normalization_13 (Batc (None, 256)               1024      \n",
      "_________________________________________________________________\n",
      "dropout_11 (Dropout)         (None, 256)               0         \n",
      "_________________________________________________________________\n",
      "dense_12 (Dense)             (None, 64)                16448     \n",
      "_________________________________________________________________\n",
      "batch_normalization_14 (Batc (None, 64)                256       \n",
      "_________________________________________________________________\n",
      "dense_13 (Dense)             (None, 16)                1040      \n",
      "=================================================================\n",
      "Total params: 14,422,608\n",
      "Trainable params: 14,421,072\n",
      "Non-trainable params: 1,536\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "feats_ext.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "id": "16d6f339",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T16:01:57.835951Z",
     "start_time": "2021-10-15T16:01:57.103916Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:Compiled the loaded model, but the compiled metrics have yet to be built. `model.compile_metrics` will be empty until you train or evaluate the model.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:Compiled the loaded model, but the compiled metrics have yet to be built. `model.compile_metrics` will be empty until you train or evaluate the model.\n"
     ]
    }
   ],
   "source": [
    "feats_ext.save('models/di_cnn_feats_ext.h5', save_format='h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c62c09d9",
   "metadata": {},
   "source": [
    "## Train new feats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "0e3d0411",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-17T09:22:53.248058Z",
     "start_time": "2021-10-17T09:22:51.910062Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:No training configuration found in the save file, so the model was *not* compiled. Compile it manually.\n"
     ]
    }
   ],
   "source": [
    "feats_ext = load_model('models/di_cnn_feats_ext.h5', custom_objects={'LMAE':LMAE})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "f7b89b48",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-17T08:56:27.702348Z",
     "start_time": "2021-10-17T08:56:23.734412Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting pydot\n",
      "  Downloading pydot-1.4.2-py2.py3-none-any.whl (21 kB)\n",
      "Requirement already satisfied: graphviz in c:\\users\\admin\\appdata\\roaming\\python\\python37\\site-packages (0.17)\n",
      "Requirement already satisfied: pyparsing>=2.1.4 in c:\\users\\admin\\anaconda3\\envs\\chemistry-ml\\lib\\site-packages (from pydot) (2.4.7)\n",
      "Installing collected packages: pydot\n",
      "Successfully installed pydot-1.4.2\n"
     ]
    }
   ],
   "source": [
    "!pip install pydot graphviz"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "68221041",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-17T09:23:48.826669Z",
     "start_time": "2021-10-17T09:23:48.619670Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tf.keras.utils.plot_model(feats_ext, to_file=\"my_model.png\", show_shapes=True, show_layer_names=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "id": "821e189a",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T15:53:54.197251Z",
     "start_time": "2021-10-15T15:52:23.335803Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(68602, 16)"
      ]
     },
     "execution_count": 199,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_feats = feats_ext.predict(mat_train)\n",
    "new_feats.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 228,
   "id": "4f3d4000",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T17:20:38.468125Z",
     "start_time": "2021-10-15T17:20:38.434087Z"
    }
   },
   "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>di_cnn_0</th>\n",
       "      <th>di_cnn_1</th>\n",
       "      <th>di_cnn_2</th>\n",
       "      <th>di_cnn_3</th>\n",
       "      <th>di_cnn_4</th>\n",
       "      <th>di_cnn_5</th>\n",
       "      <th>di_cnn_6</th>\n",
       "      <th>di_cnn_7</th>\n",
       "      <th>di_cnn_8</th>\n",
       "      <th>di_cnn_9</th>\n",
       "      <th>di_cnn_10</th>\n",
       "      <th>di_cnn_11</th>\n",
       "      <th>di_cnn_12</th>\n",
       "      <th>di_cnn_13</th>\n",
       "      <th>di_cnn_14</th>\n",
       "      <th>di_cnn_15</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>21.167830</td>\n",
       "      <td>30.934912</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>21.876890</td>\n",
       "      <td>30.938911</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>24.025457</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>23.965372</td>\n",
       "      <td>35.462387</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>25.400829</td>\n",
       "      <td>35.742039</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>26.279602</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>25.874134</td>\n",
       "      <td>39.247131</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>27.537067</td>\n",
       "      <td>39.353321</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.338663</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>26.913395</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>28.681208</td>\n",
       "      <td>44.552822</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>31.782713</td>\n",
       "      <td>44.529415</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>20.764702</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>29.767157</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>21.825914</td>\n",
       "      <td>31.975189</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>22.386007</td>\n",
       "      <td>31.988407</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>24.455626</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68597</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.536157</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68598</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68599</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68600</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68601</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>68602 rows × 16 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       di_cnn_0   di_cnn_1   di_cnn_2  di_cnn_3  di_cnn_4   di_cnn_5  \\\n",
       "0           0.0  21.167830  30.934912       0.0       0.0  21.876890   \n",
       "1           0.0  23.965372  35.462387       0.0       0.0  25.400829   \n",
       "2           0.0  25.874134  39.247131       0.0       0.0  27.537067   \n",
       "3           0.0  28.681208  44.552822       0.0       0.0  31.782713   \n",
       "4           0.0  21.825914  31.975189       0.0       0.0  22.386007   \n",
       "...         ...        ...        ...       ...       ...        ...   \n",
       "68597       0.0   0.000000   0.000000       0.0       0.0   0.000000   \n",
       "68598       0.0   0.000000   0.000000       0.0       0.0   0.000000   \n",
       "68599       0.0   0.000000   0.000000       0.0       0.0   0.000000   \n",
       "68600       0.0   0.000000   0.000000       0.0       0.0   0.000000   \n",
       "68601       0.0   0.000000   0.000000       0.0       0.0   0.000000   \n",
       "\n",
       "        di_cnn_6  di_cnn_7  di_cnn_8  di_cnn_9  di_cnn_10  di_cnn_11  \\\n",
       "0      30.938911       0.0  0.000000       0.0   0.000000        0.0   \n",
       "1      35.742039       0.0  0.000000       0.0   0.000000        0.0   \n",
       "2      39.353321       0.0  0.000000       0.0   5.338663        0.0   \n",
       "3      44.529415       0.0  0.000000       0.0  20.764702        0.0   \n",
       "4      31.988407       0.0  0.000000       0.0   0.000000        0.0   \n",
       "...          ...       ...       ...       ...        ...        ...   \n",
       "68597   0.000000       0.0  2.536157       0.0   0.000000        0.0   \n",
       "68598   0.000000       0.0  0.000000       0.0   0.000000        0.0   \n",
       "68599   0.000000       0.0  0.000000       0.0   0.000000        0.0   \n",
       "68600   0.000000       0.0  0.000000       0.0   0.000000        0.0   \n",
       "68601   0.000000       0.0  0.000000       0.0   0.000000        0.0   \n",
       "\n",
       "       di_cnn_12  di_cnn_13  di_cnn_14  di_cnn_15  \n",
       "0            0.0        0.0  24.025457        0.0  \n",
       "1            0.0        0.0  26.279602        0.0  \n",
       "2            0.0        0.0  26.913395        0.0  \n",
       "3            0.0        0.0  29.767157        0.0  \n",
       "4            0.0        0.0  24.455626        0.0  \n",
       "...          ...        ...        ...        ...  \n",
       "68597        0.0        0.0   0.000000        0.0  \n",
       "68598        0.0        0.0   0.000000        0.0  \n",
       "68599        0.0        0.0   0.000000        0.0  \n",
       "68600        0.0        0.0   0.000000        0.0  \n",
       "68601        0.0        0.0   0.000000        0.0  \n",
       "\n",
       "[68602 rows x 16 columns]"
      ]
     },
     "execution_count": 228,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_feats_cols = [f'di_cnn_{i}' for i in range(new_feats.shape[1])]\n",
    "new_feats_df = pd.DataFrame(new_feats, columns=new_feats_cols)\n",
    "new_feats_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 229,
   "id": "129ff692",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T17:20:40.157226Z",
     "start_time": "2021-10-15T17:20:40.022226Z"
    }
   },
   "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>di_cnn_0</th>\n",
       "      <th>di_cnn_1</th>\n",
       "      <th>di_cnn_2</th>\n",
       "      <th>di_cnn_3</th>\n",
       "      <th>di_cnn_4</th>\n",
       "      <th>di_cnn_5</th>\n",
       "      <th>di_cnn_6</th>\n",
       "      <th>di_cnn_7</th>\n",
       "      <th>di_cnn_8</th>\n",
       "      <th>di_cnn_9</th>\n",
       "      <th>di_cnn_10</th>\n",
       "      <th>di_cnn_11</th>\n",
       "      <th>di_cnn_12</th>\n",
       "      <th>di_cnn_13</th>\n",
       "      <th>di_cnn_14</th>\n",
       "      <th>di_cnn_15</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>68602.0</td>\n",
       "      <td>68602.000000</td>\n",
       "      <td>68602.000000</td>\n",
       "      <td>68602.0</td>\n",
       "      <td>68602.000000</td>\n",
       "      <td>68602.000000</td>\n",
       "      <td>68602.000000</td>\n",
       "      <td>68602.0</td>\n",
       "      <td>68602.000000</td>\n",
       "      <td>68602.0</td>\n",
       "      <td>68602.000000</td>\n",
       "      <td>68602.0</td>\n",
       "      <td>68602.000000</td>\n",
       "      <td>68602.000000</td>\n",
       "      <td>68602.000000</td>\n",
       "      <td>68602.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.0</td>\n",
       "      <td>20.601782</td>\n",
       "      <td>32.062370</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.574373</td>\n",
       "      <td>23.895941</td>\n",
       "      <td>31.967796</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.746188</td>\n",
       "      <td>0.0</td>\n",
       "      <td>16.154501</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.756109</td>\n",
       "      <td>0.003251</td>\n",
       "      <td>23.036751</td>\n",
       "      <td>0.900197</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.0</td>\n",
       "      <td>9.974702</td>\n",
       "      <td>16.408279</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.164420</td>\n",
       "      <td>12.789223</td>\n",
       "      <td>16.666117</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.101822</td>\n",
       "      <td>0.0</td>\n",
       "      <td>27.574968</td>\n",
       "      <td>0.0</td>\n",
       "      <td>10.320706</td>\n",
       "      <td>0.061450</td>\n",
       "      <td>11.238010</td>\n",
       "      <td>2.457077</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.0</td>\n",
       "      <td>14.836861</td>\n",
       "      <td>21.866644</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>17.017932</td>\n",
       "      <td>21.248284</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>16.419960</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.0</td>\n",
       "      <td>23.272974</td>\n",
       "      <td>34.557562</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>24.386808</td>\n",
       "      <td>34.433174</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>24.338228</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.0</td>\n",
       "      <td>29.121634</td>\n",
       "      <td>45.967256</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.252267</td>\n",
       "      <td>32.766131</td>\n",
       "      <td>46.018899</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>22.609877</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>31.317194</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>0.0</td>\n",
       "      <td>43.842312</td>\n",
       "      <td>57.914997</td>\n",
       "      <td>0.0</td>\n",
       "      <td>10.258031</td>\n",
       "      <td>88.345787</td>\n",
       "      <td>64.400414</td>\n",
       "      <td>0.0</td>\n",
       "      <td>43.081436</td>\n",
       "      <td>0.0</td>\n",
       "      <td>182.934509</td>\n",
       "      <td>0.0</td>\n",
       "      <td>119.085663</td>\n",
       "      <td>2.493887</td>\n",
       "      <td>61.993038</td>\n",
       "      <td>15.900460</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       di_cnn_0      di_cnn_1      di_cnn_2  di_cnn_3      di_cnn_4  \\\n",
       "count   68602.0  68602.000000  68602.000000   68602.0  68602.000000   \n",
       "mean        0.0     20.601782     32.062370       0.0      0.574373   \n",
       "std         0.0      9.974702     16.408279       0.0      1.164420   \n",
       "min         0.0      0.000000      0.000000       0.0      0.000000   \n",
       "25%         0.0     14.836861     21.866644       0.0      0.000000   \n",
       "50%         0.0     23.272974     34.557562       0.0      0.000000   \n",
       "75%         0.0     29.121634     45.967256       0.0      0.252267   \n",
       "max         0.0     43.842312     57.914997       0.0     10.258031   \n",
       "\n",
       "           di_cnn_5      di_cnn_6  di_cnn_7      di_cnn_8  di_cnn_9  \\\n",
       "count  68602.000000  68602.000000   68602.0  68602.000000   68602.0   \n",
       "mean      23.895941     31.967796       0.0      0.746188       0.0   \n",
       "std       12.789223     16.666117       0.0      4.101822       0.0   \n",
       "min        0.000000      0.000000       0.0      0.000000       0.0   \n",
       "25%       17.017932     21.248284       0.0      0.000000       0.0   \n",
       "50%       24.386808     34.433174       0.0      0.000000       0.0   \n",
       "75%       32.766131     46.018899       0.0      0.000000       0.0   \n",
       "max       88.345787     64.400414       0.0     43.081436       0.0   \n",
       "\n",
       "          di_cnn_10  di_cnn_11     di_cnn_12     di_cnn_13     di_cnn_14  \\\n",
       "count  68602.000000    68602.0  68602.000000  68602.000000  68602.000000   \n",
       "mean      16.154501        0.0      2.756109      0.003251     23.036751   \n",
       "std       27.574968        0.0     10.320706      0.061450     11.238010   \n",
       "min        0.000000        0.0      0.000000      0.000000      0.000000   \n",
       "25%        0.000000        0.0      0.000000      0.000000     16.419960   \n",
       "50%        0.000000        0.0      0.000000      0.000000     24.338228   \n",
       "75%       22.609877        0.0      0.000000      0.000000     31.317194   \n",
       "max      182.934509        0.0    119.085663      2.493887     61.993038   \n",
       "\n",
       "          di_cnn_15  \n",
       "count  68602.000000  \n",
       "mean       0.900197  \n",
       "std        2.457077  \n",
       "min        0.000000  \n",
       "25%        0.000000  \n",
       "50%        0.000000  \n",
       "75%        0.000000  \n",
       "max       15.900460  "
      ]
     },
     "execution_count": 229,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_feats_df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 230,
   "id": "e4d52570",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T17:20:40.576875Z",
     "start_time": "2021-10-15T17:20:40.548917Z"
    }
   },
   "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>di_cnn_1</th>\n",
       "      <th>di_cnn_2</th>\n",
       "      <th>di_cnn_4</th>\n",
       "      <th>di_cnn_5</th>\n",
       "      <th>di_cnn_6</th>\n",
       "      <th>di_cnn_8</th>\n",
       "      <th>di_cnn_10</th>\n",
       "      <th>di_cnn_12</th>\n",
       "      <th>di_cnn_13</th>\n",
       "      <th>di_cnn_14</th>\n",
       "      <th>di_cnn_15</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>21.167830</td>\n",
       "      <td>30.934912</td>\n",
       "      <td>0.0</td>\n",
       "      <td>21.876890</td>\n",
       "      <td>30.938911</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>24.025457</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>23.965372</td>\n",
       "      <td>35.462387</td>\n",
       "      <td>0.0</td>\n",
       "      <td>25.400829</td>\n",
       "      <td>35.742039</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>26.279602</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>25.874134</td>\n",
       "      <td>39.247131</td>\n",
       "      <td>0.0</td>\n",
       "      <td>27.537067</td>\n",
       "      <td>39.353321</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>5.338663</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>26.913395</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>28.681208</td>\n",
       "      <td>44.552822</td>\n",
       "      <td>0.0</td>\n",
       "      <td>31.782713</td>\n",
       "      <td>44.529415</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>20.764702</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>29.767157</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>21.825914</td>\n",
       "      <td>31.975189</td>\n",
       "      <td>0.0</td>\n",
       "      <td>22.386007</td>\n",
       "      <td>31.988407</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>24.455626</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68597</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.536157</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68598</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68599</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68600</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68601</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>68602 rows × 11 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        di_cnn_1   di_cnn_2  di_cnn_4   di_cnn_5   di_cnn_6  di_cnn_8  \\\n",
       "0      21.167830  30.934912       0.0  21.876890  30.938911  0.000000   \n",
       "1      23.965372  35.462387       0.0  25.400829  35.742039  0.000000   \n",
       "2      25.874134  39.247131       0.0  27.537067  39.353321  0.000000   \n",
       "3      28.681208  44.552822       0.0  31.782713  44.529415  0.000000   \n",
       "4      21.825914  31.975189       0.0  22.386007  31.988407  0.000000   \n",
       "...          ...        ...       ...        ...        ...       ...   \n",
       "68597   0.000000   0.000000       0.0   0.000000   0.000000  2.536157   \n",
       "68598   0.000000   0.000000       0.0   0.000000   0.000000  0.000000   \n",
       "68599   0.000000   0.000000       0.0   0.000000   0.000000  0.000000   \n",
       "68600   0.000000   0.000000       0.0   0.000000   0.000000  0.000000   \n",
       "68601   0.000000   0.000000       0.0   0.000000   0.000000  0.000000   \n",
       "\n",
       "       di_cnn_10  di_cnn_12  di_cnn_13  di_cnn_14  di_cnn_15  \n",
       "0       0.000000        0.0        0.0  24.025457        0.0  \n",
       "1       0.000000        0.0        0.0  26.279602        0.0  \n",
       "2       5.338663        0.0        0.0  26.913395        0.0  \n",
       "3      20.764702        0.0        0.0  29.767157        0.0  \n",
       "4       0.000000        0.0        0.0  24.455626        0.0  \n",
       "...          ...        ...        ...        ...        ...  \n",
       "68597   0.000000        0.0        0.0   0.000000        0.0  \n",
       "68598   0.000000        0.0        0.0   0.000000        0.0  \n",
       "68599   0.000000        0.0        0.0   0.000000        0.0  \n",
       "68600   0.000000        0.0        0.0   0.000000        0.0  \n",
       "68601   0.000000        0.0        0.0   0.000000        0.0  \n",
       "\n",
       "[68602 rows x 11 columns]"
      ]
     },
     "execution_count": 230,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# drop those feats containing only 0 values (maybe the effect from dropout)\n",
    "new_feats_df = new_feats_df.drop(['di_cnn_0', 'di_cnn_3', 'di_cnn_7', 'di_cnn_9', 'di_cnn_11'], axis=1)\n",
    "new_feats_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 231,
   "id": "61d87f80",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T17:20:45.970681Z",
     "start_time": "2021-10-15T17:20:45.149732Z"
    }
   },
   "outputs": [],
   "source": [
    "new_feats_df.to_csv('train_feats_from_di_cnn.csv', index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5cfded4f",
   "metadata": {},
   "source": [
    "## Test new feats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 232,
   "id": "ab164f73",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T17:21:12.489530Z",
     "start_time": "2021-10-15T17:21:01.411008Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(17000, 16)"
      ]
     },
     "execution_count": 232,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_new_feats = feats_ext.predict(mat_test)\n",
    "test_new_feats.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 233,
   "id": "617a31b2",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T17:21:12.535568Z",
     "start_time": "2021-10-15T17:21:12.493535Z"
    }
   },
   "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>di_cnn_0</th>\n",
       "      <th>di_cnn_1</th>\n",
       "      <th>di_cnn_2</th>\n",
       "      <th>di_cnn_3</th>\n",
       "      <th>di_cnn_4</th>\n",
       "      <th>di_cnn_5</th>\n",
       "      <th>di_cnn_6</th>\n",
       "      <th>di_cnn_7</th>\n",
       "      <th>di_cnn_8</th>\n",
       "      <th>di_cnn_9</th>\n",
       "      <th>di_cnn_10</th>\n",
       "      <th>di_cnn_11</th>\n",
       "      <th>di_cnn_12</th>\n",
       "      <th>di_cnn_13</th>\n",
       "      <th>di_cnn_14</th>\n",
       "      <th>di_cnn_15</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>30.440174</td>\n",
       "      <td>48.356892</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>34.556770</td>\n",
       "      <td>48.609718</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>28.531464</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>33.908691</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>9.247261</td>\n",
       "      <td>14.186121</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>11.476697</td>\n",
       "      <td>13.499871</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>19.669582</td>\n",
       "      <td>10.738886</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>14.510868</td>\n",
       "      <td>21.447094</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.171774</td>\n",
       "      <td>16.769449</td>\n",
       "      <td>20.630054</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>14.299532</td>\n",
       "      <td>0.456302</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>13.006152</td>\n",
       "      <td>19.137091</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.882986</td>\n",
       "      <td>15.593858</td>\n",
       "      <td>18.270979</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>13.343773</td>\n",
       "      <td>3.108078</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>11.602341</td>\n",
       "      <td>16.963276</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.470842</td>\n",
       "      <td>14.365107</td>\n",
       "      <td>16.177500</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>14.241302</td>\n",
       "      <td>8.076712</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16995</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.084170</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>26.112291</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16996</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16997</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16998</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.955123</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16999</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>18.208992</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>17000 rows × 16 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       di_cnn_0   di_cnn_1   di_cnn_2  di_cnn_3  di_cnn_4   di_cnn_5  \\\n",
       "0           0.0  30.440174  48.356892       0.0  0.000000  34.556770   \n",
       "1           0.0   9.247261  14.186121       0.0  0.000000  11.476697   \n",
       "2           0.0  14.510868  21.447094       0.0  4.171774  16.769449   \n",
       "3           0.0  13.006152  19.137091       0.0  3.882986  15.593858   \n",
       "4           0.0  11.602341  16.963276       0.0  1.470842  14.365107   \n",
       "...         ...        ...        ...       ...       ...        ...   \n",
       "16995       0.0   0.000000   0.000000       0.0  1.084170   0.000000   \n",
       "16996       0.0   0.000000   0.000000       0.0  0.000000   0.000000   \n",
       "16997       0.0   0.000000   0.000000       0.0  0.000000   0.000000   \n",
       "16998       0.0   0.000000   0.000000       0.0  0.000000   0.000000   \n",
       "16999       0.0   0.000000   0.000000       0.0  0.000000   0.000000   \n",
       "\n",
       "        di_cnn_6  di_cnn_7   di_cnn_8  di_cnn_9  di_cnn_10  di_cnn_11  \\\n",
       "0      48.609718       0.0   0.000000       0.0  28.531464        0.0   \n",
       "1      13.499871       0.0   0.000000       0.0   0.000000        0.0   \n",
       "2      20.630054       0.0   0.000000       0.0   0.000000        0.0   \n",
       "3      18.270979       0.0   0.000000       0.0   0.000000        0.0   \n",
       "4      16.177500       0.0   0.000000       0.0   0.000000        0.0   \n",
       "...          ...       ...        ...       ...        ...        ...   \n",
       "16995   0.000000       0.0  26.112291       0.0   0.000000        0.0   \n",
       "16996   0.000000       0.0   0.000000       0.0   0.000000        0.0   \n",
       "16997   0.000000       0.0   0.000000       0.0   0.000000        0.0   \n",
       "16998   0.000000       0.0   4.955123       0.0   0.000000        0.0   \n",
       "16999   0.000000       0.0  18.208992       0.0   0.000000        0.0   \n",
       "\n",
       "       di_cnn_12  di_cnn_13  di_cnn_14  di_cnn_15  \n",
       "0            0.0        0.0  33.908691   0.000000  \n",
       "1            0.0        0.0  19.669582  10.738886  \n",
       "2            0.0        0.0  14.299532   0.456302  \n",
       "3            0.0        0.0  13.343773   3.108078  \n",
       "4            0.0        0.0  14.241302   8.076712  \n",
       "...          ...        ...        ...        ...  \n",
       "16995        0.0        0.0   0.000000   0.000000  \n",
       "16996        0.0        0.0   0.000000   0.000000  \n",
       "16997        0.0        0.0   0.000000   0.000000  \n",
       "16998        0.0        0.0   0.000000   0.000000  \n",
       "16999        0.0        0.0   0.000000   0.000000  \n",
       "\n",
       "[17000 rows x 16 columns]"
      ]
     },
     "execution_count": 233,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_new_feats_df = pd.DataFrame(test_new_feats, columns=new_feats_cols)\n",
    "test_new_feats_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 234,
   "id": "94f24b85",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T17:21:15.599784Z",
     "start_time": "2021-10-15T17:21:15.518757Z"
    }
   },
   "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>di_cnn_0</th>\n",
       "      <th>di_cnn_1</th>\n",
       "      <th>di_cnn_2</th>\n",
       "      <th>di_cnn_3</th>\n",
       "      <th>di_cnn_4</th>\n",
       "      <th>di_cnn_5</th>\n",
       "      <th>di_cnn_6</th>\n",
       "      <th>di_cnn_7</th>\n",
       "      <th>di_cnn_8</th>\n",
       "      <th>di_cnn_9</th>\n",
       "      <th>di_cnn_10</th>\n",
       "      <th>di_cnn_11</th>\n",
       "      <th>di_cnn_12</th>\n",
       "      <th>di_cnn_13</th>\n",
       "      <th>di_cnn_14</th>\n",
       "      <th>di_cnn_15</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>17000.0</td>\n",
       "      <td>17000.000000</td>\n",
       "      <td>17000.000000</td>\n",
       "      <td>17000.0</td>\n",
       "      <td>17000.000000</td>\n",
       "      <td>17000.000000</td>\n",
       "      <td>17000.000000</td>\n",
       "      <td>17000.0</td>\n",
       "      <td>17000.000000</td>\n",
       "      <td>17000.0</td>\n",
       "      <td>17000.000000</td>\n",
       "      <td>17000.0</td>\n",
       "      <td>17000.000000</td>\n",
       "      <td>17000.000000</td>\n",
       "      <td>17000.000000</td>\n",
       "      <td>17000.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.0</td>\n",
       "      <td>19.246334</td>\n",
       "      <td>29.680969</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.704428</td>\n",
       "      <td>22.128553</td>\n",
       "      <td>29.515322</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.029979</td>\n",
       "      <td>0.0</td>\n",
       "      <td>12.721861</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.026930</td>\n",
       "      <td>0.000407</td>\n",
       "      <td>21.350115</td>\n",
       "      <td>1.106024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.0</td>\n",
       "      <td>9.865600</td>\n",
       "      <td>16.034487</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.255562</td>\n",
       "      <td>12.330595</td>\n",
       "      <td>16.267271</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.702759</td>\n",
       "      <td>0.0</td>\n",
       "      <td>24.747637</td>\n",
       "      <td>0.0</td>\n",
       "      <td>8.989811</td>\n",
       "      <td>0.018484</td>\n",
       "      <td>10.709275</td>\n",
       "      <td>2.687207</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.0</td>\n",
       "      <td>13.518953</td>\n",
       "      <td>19.884085</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>15.989784</td>\n",
       "      <td>19.227555</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>15.226677</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.0</td>\n",
       "      <td>20.854296</td>\n",
       "      <td>30.630587</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>21.998467</td>\n",
       "      <td>30.351612</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>21.668471</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.0</td>\n",
       "      <td>27.885930</td>\n",
       "      <td>42.871400</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.031594</td>\n",
       "      <td>30.410144</td>\n",
       "      <td>42.926318</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>13.867979</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>29.007347</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>0.0</td>\n",
       "      <td>43.491451</td>\n",
       "      <td>57.931873</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9.288282</td>\n",
       "      <td>87.367409</td>\n",
       "      <td>64.169914</td>\n",
       "      <td>0.0</td>\n",
       "      <td>42.195904</td>\n",
       "      <td>0.0</td>\n",
       "      <td>180.508804</td>\n",
       "      <td>0.0</td>\n",
       "      <td>116.977852</td>\n",
       "      <td>1.817728</td>\n",
       "      <td>61.627983</td>\n",
       "      <td>14.934413</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       di_cnn_0      di_cnn_1      di_cnn_2  di_cnn_3      di_cnn_4  \\\n",
       "count   17000.0  17000.000000  17000.000000   17000.0  17000.000000   \n",
       "mean        0.0     19.246334     29.680969       0.0      0.704428   \n",
       "std         0.0      9.865600     16.034487       0.0      1.255562   \n",
       "min         0.0      0.000000      0.000000       0.0      0.000000   \n",
       "25%         0.0     13.518953     19.884085       0.0      0.000000   \n",
       "50%         0.0     20.854296     30.630587       0.0      0.000000   \n",
       "75%         0.0     27.885930     42.871400       0.0      1.031594   \n",
       "max         0.0     43.491451     57.931873       0.0      9.288282   \n",
       "\n",
       "           di_cnn_5      di_cnn_6  di_cnn_7      di_cnn_8  di_cnn_9  \\\n",
       "count  17000.000000  17000.000000   17000.0  17000.000000   17000.0   \n",
       "mean      22.128553     29.515322       0.0      1.029979       0.0   \n",
       "std       12.330595     16.267271       0.0      4.702759       0.0   \n",
       "min        0.000000      0.000000       0.0      0.000000       0.0   \n",
       "25%       15.989784     19.227555       0.0      0.000000       0.0   \n",
       "50%       21.998467     30.351612       0.0      0.000000       0.0   \n",
       "75%       30.410144     42.926318       0.0      0.000000       0.0   \n",
       "max       87.367409     64.169914       0.0     42.195904       0.0   \n",
       "\n",
       "          di_cnn_10  di_cnn_11     di_cnn_12     di_cnn_13     di_cnn_14  \\\n",
       "count  17000.000000    17000.0  17000.000000  17000.000000  17000.000000   \n",
       "mean      12.721861        0.0      2.026930      0.000407     21.350115   \n",
       "std       24.747637        0.0      8.989811      0.018484     10.709275   \n",
       "min        0.000000        0.0      0.000000      0.000000      0.000000   \n",
       "25%        0.000000        0.0      0.000000      0.000000     15.226677   \n",
       "50%        0.000000        0.0      0.000000      0.000000     21.668471   \n",
       "75%       13.867979        0.0      0.000000      0.000000     29.007347   \n",
       "max      180.508804        0.0    116.977852      1.817728     61.627983   \n",
       "\n",
       "          di_cnn_15  \n",
       "count  17000.000000  \n",
       "mean       1.106024  \n",
       "std        2.687207  \n",
       "min        0.000000  \n",
       "25%        0.000000  \n",
       "50%        0.000000  \n",
       "75%        0.000000  \n",
       "max       14.934413  "
      ]
     },
     "execution_count": 234,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_new_feats_df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 235,
   "id": "cbb0864e",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T17:21:16.064933Z",
     "start_time": "2021-10-15T17:21:16.031960Z"
    }
   },
   "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>di_cnn_1</th>\n",
       "      <th>di_cnn_2</th>\n",
       "      <th>di_cnn_4</th>\n",
       "      <th>di_cnn_5</th>\n",
       "      <th>di_cnn_6</th>\n",
       "      <th>di_cnn_8</th>\n",
       "      <th>di_cnn_10</th>\n",
       "      <th>di_cnn_12</th>\n",
       "      <th>di_cnn_13</th>\n",
       "      <th>di_cnn_14</th>\n",
       "      <th>di_cnn_15</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>30.440174</td>\n",
       "      <td>48.356892</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>34.556770</td>\n",
       "      <td>48.609718</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>28.531464</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>33.908691</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>9.247261</td>\n",
       "      <td>14.186121</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>11.476697</td>\n",
       "      <td>13.499871</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>19.669582</td>\n",
       "      <td>10.738886</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>14.510868</td>\n",
       "      <td>21.447094</td>\n",
       "      <td>4.171774</td>\n",
       "      <td>16.769449</td>\n",
       "      <td>20.630054</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>14.299532</td>\n",
       "      <td>0.456302</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>13.006152</td>\n",
       "      <td>19.137091</td>\n",
       "      <td>3.882986</td>\n",
       "      <td>15.593858</td>\n",
       "      <td>18.270979</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>13.343773</td>\n",
       "      <td>3.108078</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>11.602341</td>\n",
       "      <td>16.963276</td>\n",
       "      <td>1.470842</td>\n",
       "      <td>14.365107</td>\n",
       "      <td>16.177500</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>14.241302</td>\n",
       "      <td>8.076712</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16995</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.084170</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>26.112291</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16996</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16997</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16998</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.955123</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16999</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>18.208992</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>17000 rows × 11 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        di_cnn_1   di_cnn_2  di_cnn_4   di_cnn_5   di_cnn_6   di_cnn_8  \\\n",
       "0      30.440174  48.356892  0.000000  34.556770  48.609718   0.000000   \n",
       "1       9.247261  14.186121  0.000000  11.476697  13.499871   0.000000   \n",
       "2      14.510868  21.447094  4.171774  16.769449  20.630054   0.000000   \n",
       "3      13.006152  19.137091  3.882986  15.593858  18.270979   0.000000   \n",
       "4      11.602341  16.963276  1.470842  14.365107  16.177500   0.000000   \n",
       "...          ...        ...       ...        ...        ...        ...   \n",
       "16995   0.000000   0.000000  1.084170   0.000000   0.000000  26.112291   \n",
       "16996   0.000000   0.000000  0.000000   0.000000   0.000000   0.000000   \n",
       "16997   0.000000   0.000000  0.000000   0.000000   0.000000   0.000000   \n",
       "16998   0.000000   0.000000  0.000000   0.000000   0.000000   4.955123   \n",
       "16999   0.000000   0.000000  0.000000   0.000000   0.000000  18.208992   \n",
       "\n",
       "       di_cnn_10  di_cnn_12  di_cnn_13  di_cnn_14  di_cnn_15  \n",
       "0      28.531464        0.0        0.0  33.908691   0.000000  \n",
       "1       0.000000        0.0        0.0  19.669582  10.738886  \n",
       "2       0.000000        0.0        0.0  14.299532   0.456302  \n",
       "3       0.000000        0.0        0.0  13.343773   3.108078  \n",
       "4       0.000000        0.0        0.0  14.241302   8.076712  \n",
       "...          ...        ...        ...        ...        ...  \n",
       "16995   0.000000        0.0        0.0   0.000000   0.000000  \n",
       "16996   0.000000        0.0        0.0   0.000000   0.000000  \n",
       "16997   0.000000        0.0        0.0   0.000000   0.000000  \n",
       "16998   0.000000        0.0        0.0   0.000000   0.000000  \n",
       "16999   0.000000        0.0        0.0   0.000000   0.000000  \n",
       "\n",
       "[17000 rows x 11 columns]"
      ]
     },
     "execution_count": 235,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# drop those feats containing only 0 values (maybe the effect from dropout)\n",
    "test_new_feats_df = test_new_feats_df.drop(['di_cnn_0', 'di_cnn_3', 'di_cnn_7', 'di_cnn_9', 'di_cnn_11'], axis=1)\n",
    "test_new_feats_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 236,
   "id": "926f67a5",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-10-15T17:21:18.058049Z",
     "start_time": "2021-10-15T17:21:17.846077Z"
    }
   },
   "outputs": [],
   "source": [
    "test_new_feats_df.to_csv('test_feats_from_di_cnn.csv', index=False)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "chemistry-ml",
   "language": "python",
   "name": "chemistry-ml"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}