NBA.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Benchmarking"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The NBA data is included in a compressed format because it's pretty big. The first time you run this notebook you'll need to uncomment the line below to unzip the data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# !tar -xzvf data/nba.tar.gz -C data/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SEASON_YEAR</th>\n",
       "      <th>PLAYER_ID</th>\n",
       "      <th>PLAYER_NAME</th>\n",
       "      <th>TEAM_ID</th>\n",
       "      <th>TEAM_ABBREVIATION</th>\n",
       "      <th>TEAM_NAME</th>\n",
       "      <th>GAME_ID</th>\n",
       "      <th>GAME_DATE</th>\n",
       "      <th>MATCHUP</th>\n",
       "      <th>WL</th>\n",
       "      <th>...</th>\n",
       "      <th>BLKA</th>\n",
       "      <th>PF</th>\n",
       "      <th>PFD</th>\n",
       "      <th>PTS</th>\n",
       "      <th>PLUS_MINUS</th>\n",
       "      <th>NBA_FANTASY_PTS</th>\n",
       "      <th>DD2</th>\n",
       "      <th>TD3</th>\n",
       "      <th>Season</th>\n",
       "      <th>date</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>559702</th>\n",
       "      <td>1996-97</td>\n",
       "      <td>239</td>\n",
       "      <td>Darrick Martin</td>\n",
       "      <td>1610612746</td>\n",
       "      <td>LAC</td>\n",
       "      <td>Los Angeles Clippers</td>\n",
       "      <td>29600013</td>\n",
       "      <td>1996-11-01</td>\n",
       "      <td>LAC @ GSW</td>\n",
       "      <td>W</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>13</td>\n",
       "      <td>7</td>\n",
       "      <td>19.7</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1996-97</td>\n",
       "      <td>1996-11-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>559506</th>\n",
       "      <td>1996-97</td>\n",
       "      <td>63</td>\n",
       "      <td>Michael Smith</td>\n",
       "      <td>1610612758</td>\n",
       "      <td>SAC</td>\n",
       "      <td>Sacramento Kings</td>\n",
       "      <td>29600009</td>\n",
       "      <td>1996-11-01</td>\n",
       "      <td>SAC @ HOU</td>\n",
       "      <td>L</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>-10</td>\n",
       "      <td>23.5</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1996-97</td>\n",
       "      <td>1996-11-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>559507</th>\n",
       "      <td>1996-97</td>\n",
       "      <td>251</td>\n",
       "      <td>Sean Elliott</td>\n",
       "      <td>1610612759</td>\n",
       "      <td>SAS</td>\n",
       "      <td>San Antonio Spurs</td>\n",
       "      <td>29600008</td>\n",
       "      <td>1996-11-01</td>\n",
       "      <td>SAS @ MIN</td>\n",
       "      <td>L</td>\n",
       "      <td>...</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>14</td>\n",
       "      <td>-11</td>\n",
       "      <td>29.1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1996-97</td>\n",
       "      <td>1996-11-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>559508</th>\n",
       "      <td>1996-97</td>\n",
       "      <td>112</td>\n",
       "      <td>Dontonio Wingfield</td>\n",
       "      <td>1610612757</td>\n",
       "      <td>POR</td>\n",
       "      <td>Portland Trail Blazers</td>\n",
       "      <td>29600014</td>\n",
       "      <td>1996-11-01</td>\n",
       "      <td>POR @ VAN</td>\n",
       "      <td>W</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>16</td>\n",
       "      <td>29.8</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1996-97</td>\n",
       "      <td>1996-11-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>559509</th>\n",
       "      <td>1996-97</td>\n",
       "      <td>210</td>\n",
       "      <td>Terrell Brandon</td>\n",
       "      <td>1610612739</td>\n",
       "      <td>CLE</td>\n",
       "      <td>Cleveland Cavaliers</td>\n",
       "      <td>29600002</td>\n",
       "      <td>1996-11-01</td>\n",
       "      <td>CLE @ NJN</td>\n",
       "      <td>W</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>15</td>\n",
       "      <td>45.7</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1996-97</td>\n",
       "      <td>1996-11-01</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 37 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       SEASON_YEAR  PLAYER_ID         PLAYER_NAME     TEAM_ID  \\\n",
       "559702     1996-97        239      Darrick Martin  1610612746   \n",
       "559506     1996-97         63       Michael Smith  1610612758   \n",
       "559507     1996-97        251        Sean Elliott  1610612759   \n",
       "559508     1996-97        112  Dontonio Wingfield  1610612757   \n",
       "559509     1996-97        210     Terrell Brandon  1610612739   \n",
       "\n",
       "       TEAM_ABBREVIATION               TEAM_NAME   GAME_ID   GAME_DATE  \\\n",
       "559702               LAC    Los Angeles Clippers  29600013  1996-11-01   \n",
       "559506               SAC        Sacramento Kings  29600009  1996-11-01   \n",
       "559507               SAS       San Antonio Spurs  29600008  1996-11-01   \n",
       "559508               POR  Portland Trail Blazers  29600014  1996-11-01   \n",
       "559509               CLE     Cleveland Cavaliers  29600002  1996-11-01   \n",
       "\n",
       "          MATCHUP WL    ...      BLKA  PF  PFD  PTS  PLUS_MINUS  \\\n",
       "559702  LAC @ GSW  W    ...         1   3    0   13           7   \n",
       "559506  SAC @ HOU  L    ...         0   4    0    8         -10   \n",
       "559507  SAS @ MIN  L    ...         4   3    0   14         -11   \n",
       "559508  POR @ VAN  W    ...         1   6    0    4          16   \n",
       "559509  CLE @ NJN  W    ...         1   4    0   25          15   \n",
       "\n",
       "        NBA_FANTASY_PTS  DD2  TD3   Season       date  \n",
       "559702             19.7    0    0  1996-97 1996-11-01  \n",
       "559506             23.5    0    0  1996-97 1996-11-01  \n",
       "559507             29.1    0    0  1996-97 1996-11-01  \n",
       "559508             29.8    0    0  1996-97 1996-11-01  \n",
       "559509             45.7    0    0  1996-97 1996-11-01  \n",
       "\n",
       "[5 rows x 37 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv('data/nba.csv')\n",
    "df['date'] = pd.to_datetime(df['GAME_DATE'])\n",
    "df.sort_values(by=['date'], inplace=True)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(559703, 37)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "559702    13\n",
       "559506     8\n",
       "559507    14\n",
       "559508     4\n",
       "559509    25\n",
       "559510     0\n",
       "559511    12\n",
       "559505     2\n",
       "559512     6\n",
       "559514    14\n",
       "Name: PTS, dtype: int64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['PTS'][:10]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I created some basic statistical features that can be used as forecasts. I also threw in home/away since it seemed like it might be important."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_features(player_df, focus='PTS'):\n",
    "    player_df = player_df.assign(\n",
    "        target=player_df[focus].shift(-1),\n",
    "        sample_mean=player_df[focus].expanding().mean(),\n",
    "        ewma=player_df[focus].ewm(alpha=0.1).mean(),\n",
    "        last_day=player_df[focus],\n",
    "        moving_window=player_df[focus].rolling(10).mean(),\n",
    "        home=player_df['MATCHUP'].shift(-1).apply(lambda x: x if type(x)==np.float else int(x.split()[1]=='vs.'))\n",
    "    )\n",
    "    \n",
    "    features = ['date', 'last_day', 'sample_mean', 'moving_window', 'ewma', 'home', 'target']    \n",
    "    features = player_df[features].copy()\n",
    "    features = features[10:-1] # drop rows with missing data\n",
    "    return features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Split the df by player, generate the features, then recombine into a single df\n",
    "combined = None\n",
    "\n",
    "for player in df['PLAYER_ID'].unique():\n",
    "    player_df = df[df['PLAYER_ID']==player].reset_index(drop=True)\n",
    "    features = make_features(player_df)\n",
    "    if combined is None: combined = features\n",
    "    else: combined = pd.concat([combined, features])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date</th>\n",
       "      <th>last_day</th>\n",
       "      <th>sample_mean</th>\n",
       "      <th>moving_window</th>\n",
       "      <th>ewma</th>\n",
       "      <th>home</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>1996-11-21</td>\n",
       "      <td>7</td>\n",
       "      <td>5.090909</td>\n",
       "      <td>4.3</td>\n",
       "      <td>5.084948</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>1996-11-23</td>\n",
       "      <td>4</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>4.5</td>\n",
       "      <td>4.933751</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>1996-11-25</td>\n",
       "      <td>5</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>5.0</td>\n",
       "      <td>4.942634</td>\n",
       "      <td>1.0</td>\n",
       "      <td>18.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>1996-11-27</td>\n",
       "      <td>18</td>\n",
       "      <td>5.928571</td>\n",
       "      <td>6.8</td>\n",
       "      <td>6.635687</td>\n",
       "      <td>0.0</td>\n",
       "      <td>18.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>1996-11-29</td>\n",
       "      <td>18</td>\n",
       "      <td>6.733333</td>\n",
       "      <td>7.9</td>\n",
       "      <td>8.066764</td>\n",
       "      <td>1.0</td>\n",
       "      <td>12.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         date  last_day  sample_mean  moving_window      ewma  home  target\n",
       "10 1996-11-21         7     5.090909            4.3  5.084948   0.0     4.0\n",
       "11 1996-11-23         4     5.000000            4.5  4.933751   1.0     5.0\n",
       "12 1996-11-25         5     5.000000            5.0  4.942634   1.0    18.0\n",
       "13 1996-11-27        18     5.928571            6.8  6.635687   0.0    18.0\n",
       "14 1996-11-29        18     6.733333            7.9  8.066764   1.0    12.0"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "combined.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(537795, 7)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "combined.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "def mae(targ, pred): return (pred-targ).abs().mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5.551758569715226"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mae(combined['target'], combined['last_day'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4.9634851880663335"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mae(combined['target'], combined['sample_mean'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4.319840755362905"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mae(combined['target'], combined['ewma'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4.389793508679368"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mae(combined['target'], combined['moving_window'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The averaging techniques perform better than looking at the last day. This may suggest that the player has an underlying level of skill that is predictive of points, and short-term blips are not very telling. The exponentially weighted average performs best, perhaps because it's able to pick up on trends such as a player improving or getting worse."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "import xgboost as xgb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "combined = combined.sort_values(by='date').reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": false
   },
   "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>date</th>\n",
       "      <th>last_day</th>\n",
       "      <th>sample_mean</th>\n",
       "      <th>moving_window</th>\n",
       "      <th>ewma</th>\n",
       "      <th>home</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1996-11-17</td>\n",
       "      <td>11</td>\n",
       "      <td>16.818182</td>\n",
       "      <td>16.5</td>\n",
       "      <td>16.445124</td>\n",
       "      <td>1.0</td>\n",
       "      <td>20.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1996-11-17</td>\n",
       "      <td>10</td>\n",
       "      <td>7.090909</td>\n",
       "      <td>7.0</td>\n",
       "      <td>6.909297</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1996-11-17</td>\n",
       "      <td>6</td>\n",
       "      <td>8.909091</td>\n",
       "      <td>8.3</td>\n",
       "      <td>8.480440</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1996-11-17</td>\n",
       "      <td>4</td>\n",
       "      <td>4.272727</td>\n",
       "      <td>3.9</td>\n",
       "      <td>4.283124</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1996-11-17</td>\n",
       "      <td>18</td>\n",
       "      <td>9.727273</td>\n",
       "      <td>9.8</td>\n",
       "      <td>9.540422</td>\n",
       "      <td>1.0</td>\n",
       "      <td>14.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        date  last_day  sample_mean  moving_window       ewma  home  target\n",
       "0 1996-11-17        11    16.818182           16.5  16.445124   1.0    20.0\n",
       "1 1996-11-17        10     7.090909            7.0   6.909297   1.0     3.0\n",
       "2 1996-11-17         6     8.909091            8.3   8.480440   1.0     6.0\n",
       "3 1996-11-17         4     4.272727            3.9   4.283124   1.0     6.0\n",
       "4 1996-11-17        18     9.727273            9.8   9.540422   1.0    14.0"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "combined.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "combined.drop(columns='date', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "403346"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "split = int(len(combined) * .75); split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "combined_train = combined.iloc[:split]\n",
    "combined_test = combined.iloc[split:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((403346, 6), (134449, 6))"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "combined_train.shape, combined_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = combined_train.drop(columns=['target'])\n",
    "y_train = combined_train['target']\n",
    "\n",
    "X_test = combined_test.drop(columns=['target'])\n",
    "y_test = combined_test['target']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/envs/fastai/lib/python3.6/site-packages/xgboost/core.py:587: FutureWarning: Series.base is deprecated and will be removed in a future version\n",
      "  if getattr(data, 'base', None) is not None and \\\n",
      "/opt/conda/envs/fastai/lib/python3.6/site-packages/xgboost/core.py:588: FutureWarning: Series.base is deprecated and will be removed in a future version\n",
      "  data.base is not None and isinstance(data, np.ndarray) \\\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[13:45:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "XGBRegressor(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "             colsample_bynode=1, colsample_bytree=1, gamma=0,\n",
       "             importance_type='gain', learning_rate=0.1, max_delta_step=0,\n",
       "             max_depth=3, min_child_weight=1, missing=None, n_estimators=1000,\n",
       "             n_jobs=1, nthread=None, objective='reg:linear', random_state=0,\n",
       "             reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
       "             silent=None, subsample=1, verbosity=1)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# training model; may take several minutes\n",
    "reg = xgb.XGBRegressor(n_estimators=1000)\n",
    "reg.fit(X_train, y_train,\n",
    "       verbose=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fa968f12e80>"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "xgb.plot_importance(reg)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "preds = reg.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4.3907351797624115"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mae(y_test, preds)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Of course there are many other features that could improve performance, such as playoff/regular season, days since last game, etc."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## fastai time series"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('data/nba.csv')\n",
    "df['date'] = pd.to_datetime(df['GAME_DATE'])\n",
    "df.sort_values(by=['date'], inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SEASON_YEAR</th>\n",
       "      <th>PLAYER_ID</th>\n",
       "      <th>PLAYER_NAME</th>\n",
       "      <th>TEAM_ID</th>\n",
       "      <th>TEAM_ABBREVIATION</th>\n",
       "      <th>TEAM_NAME</th>\n",
       "      <th>GAME_ID</th>\n",
       "      <th>GAME_DATE</th>\n",
       "      <th>MATCHUP</th>\n",
       "      <th>WL</th>\n",
       "      <th>...</th>\n",
       "      <th>BLKA</th>\n",
       "      <th>PF</th>\n",
       "      <th>PFD</th>\n",
       "      <th>PTS</th>\n",
       "      <th>PLUS_MINUS</th>\n",
       "      <th>NBA_FANTASY_PTS</th>\n",
       "      <th>DD2</th>\n",
       "      <th>TD3</th>\n",
       "      <th>Season</th>\n",
       "      <th>date</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>559702</th>\n",
       "      <td>1996-97</td>\n",
       "      <td>239</td>\n",
       "      <td>Darrick Martin</td>\n",
       "      <td>1610612746</td>\n",
       "      <td>LAC</td>\n",
       "      <td>Los Angeles Clippers</td>\n",
       "      <td>29600013</td>\n",
       "      <td>1996-11-01</td>\n",
       "      <td>LAC @ GSW</td>\n",
       "      <td>W</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>13</td>\n",
       "      <td>7</td>\n",
       "      <td>19.7</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1996-97</td>\n",
       "      <td>1996-11-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>559506</th>\n",
       "      <td>1996-97</td>\n",
       "      <td>63</td>\n",
       "      <td>Michael Smith</td>\n",
       "      <td>1610612758</td>\n",
       "      <td>SAC</td>\n",
       "      <td>Sacramento Kings</td>\n",
       "      <td>29600009</td>\n",
       "      <td>1996-11-01</td>\n",
       "      <td>SAC @ HOU</td>\n",
       "      <td>L</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>-10</td>\n",
       "      <td>23.5</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1996-97</td>\n",
       "      <td>1996-11-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>559507</th>\n",
       "      <td>1996-97</td>\n",
       "      <td>251</td>\n",
       "      <td>Sean Elliott</td>\n",
       "      <td>1610612759</td>\n",
       "      <td>SAS</td>\n",
       "      <td>San Antonio Spurs</td>\n",
       "      <td>29600008</td>\n",
       "      <td>1996-11-01</td>\n",
       "      <td>SAS @ MIN</td>\n",
       "      <td>L</td>\n",
       "      <td>...</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>14</td>\n",
       "      <td>-11</td>\n",
       "      <td>29.1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1996-97</td>\n",
       "      <td>1996-11-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>559508</th>\n",
       "      <td>1996-97</td>\n",
       "      <td>112</td>\n",
       "      <td>Dontonio Wingfield</td>\n",
       "      <td>1610612757</td>\n",
       "      <td>POR</td>\n",
       "      <td>Portland Trail Blazers</td>\n",
       "      <td>29600014</td>\n",
       "      <td>1996-11-01</td>\n",
       "      <td>POR @ VAN</td>\n",
       "      <td>W</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>16</td>\n",
       "      <td>29.8</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1996-97</td>\n",
       "      <td>1996-11-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>559509</th>\n",
       "      <td>1996-97</td>\n",
       "      <td>210</td>\n",
       "      <td>Terrell Brandon</td>\n",
       "      <td>1610612739</td>\n",
       "      <td>CLE</td>\n",
       "      <td>Cleveland Cavaliers</td>\n",
       "      <td>29600002</td>\n",
       "      <td>1996-11-01</td>\n",
       "      <td>CLE @ NJN</td>\n",
       "      <td>W</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>15</td>\n",
       "      <td>45.7</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1996-97</td>\n",
       "      <td>1996-11-01</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 37 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       SEASON_YEAR  PLAYER_ID         PLAYER_NAME     TEAM_ID  \\\n",
       "559702     1996-97        239      Darrick Martin  1610612746   \n",
       "559506     1996-97         63       Michael Smith  1610612758   \n",
       "559507     1996-97        251        Sean Elliott  1610612759   \n",
       "559508     1996-97        112  Dontonio Wingfield  1610612757   \n",
       "559509     1996-97        210     Terrell Brandon  1610612739   \n",
       "\n",
       "       TEAM_ABBREVIATION               TEAM_NAME   GAME_ID   GAME_DATE  \\\n",
       "559702               LAC    Los Angeles Clippers  29600013  1996-11-01   \n",
       "559506               SAC        Sacramento Kings  29600009  1996-11-01   \n",
       "559507               SAS       San Antonio Spurs  29600008  1996-11-01   \n",
       "559508               POR  Portland Trail Blazers  29600014  1996-11-01   \n",
       "559509               CLE     Cleveland Cavaliers  29600002  1996-11-01   \n",
       "\n",
       "          MATCHUP WL    ...      BLKA  PF  PFD  PTS  PLUS_MINUS  \\\n",
       "559702  LAC @ GSW  W    ...         1   3    0   13           7   \n",
       "559506  SAC @ HOU  L    ...         0   4    0    8         -10   \n",
       "559507  SAS @ MIN  L    ...         4   3    0   14         -11   \n",
       "559508  POR @ VAN  W    ...         1   6    0    4          16   \n",
       "559509  CLE @ NJN  W    ...         1   4    0   25          15   \n",
       "\n",
       "        NBA_FANTASY_PTS  DD2  TD3   Season       date  \n",
       "559702             19.7    0    0  1996-97 1996-11-01  \n",
       "559506             23.5    0    0  1996-97 1996-11-01  \n",
       "559507             29.1    0    0  1996-97 1996-11-01  \n",
       "559508             29.8    0    0  1996-97 1996-11-01  \n",
       "559509             45.7    0    0  1996-97 1996-11-01  \n",
       "\n",
       "[5 rows x 37 columns]"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Make data\n",
    "First I split the df by player (each player is really their own time series). Then I generate the training data as follows: X is 10 contiguous games, y is the game immediately after that (11th game). Do this as many times as the player series allows, stepping forward one day each time. This does mean there's a lot of overlap between the samples - I don't know if that's a problem. I split train/valid/test by season, partly for convenience, partly because I'm worried that the overlap in the samples could allow some data leakage. Because the test set is a season the model hasn't seen at all, the score on the test set should at least be a fair measure of how well it's working."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_arrays(player_df, window=10):\n",
    "    pts=player_df['PTS']\n",
    "    xs,ys = [],[]\n",
    "    for start in range(len(pts) - window):\n",
    "        end = start + window\n",
    "        x = pts[start:end]\n",
    "        y = pts[end]\n",
    "        xs.append(x)\n",
    "        ys.append(y)\n",
    "    return xs,ys"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_data(df):\n",
    "    all_xs, all_ys = [],[]\n",
    "    for player in df['PLAYER_ID'].unique():\n",
    "        player_df = df[df['PLAYER_ID']==player].reset_index(drop=True)\n",
    "        player_xs, player_ys = make_arrays(player_df)\n",
    "        all_xs += player_xs\n",
    "        all_ys += player_ys\n",
    "\n",
    "    X = np.array(all_xs).astype('float32')\n",
    "    X = X[:,np.newaxis,:] # Data must be in the shape samples, features, sequence length. In this case we only have one feature, points.\n",
    "    y = np.array(all_ys).astype('float32')\n",
    "    return X, y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_df = df[df['SEASON_YEAR'].isin(['2016-17', '2015-16', '2014-15'])].reset_index(drop=True)\n",
    "valid_df = df[df['SEASON_YEAR']=='2017-18'].reset_index(drop=True)\n",
    "test_df = df[df['SEASON_YEAR']=='2018-19'].reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "X_train, y_train = make_data(train_df)\n",
    "X_valid, y_valid = make_data(valid_df)\n",
    "X_test, y_test = make_data(test_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(71833, 1, 10), (71833,), (21144, 1, 10), (21144,), (21168, 1, 10), (21168,)]"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list(map(lambda x: x.shape, (X_train, y_train, X_valid, y_valid, X_test, y_test)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>.container { width:100% !important; }</style>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/storage/timeseriesAI\n",
      "fastai : 1.0.58\n",
      "torch  : 1.0.0\n",
      "device : cuda\n"
     ]
    }
   ],
   "source": [
    "import fastai, os\n",
    "from fastai_timeseries import *\n",
    "from torchtimeseries.models import *\n",
    "path = Path(os.getcwd())\n",
    "print(path)\n",
    "print('fastai :', fastai.__version__)\n",
    "print('torch  :', torch.__version__)\n",
    "print('device :', device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "bs = 64                            # ✳️\n",
    "seed = 1234                        # ✳️\n",
    "scale_type = 'standardize'         # ✳️ 'standardize', 'normalize'\n",
    "scale_subtype = 'per_channel'      # ✳️ 'all_samples', 'per_channel', 'per_sample'\n",
    "scale_range = (-1, 1)              # ✳️ for normalization only: usually left to (-1, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TSDataBunch;\n",
       "\n",
       "Train: LabelList (71833 items)\n",
       "x: TimeSeriesList\n",
       "TimeSeries(ch=1, seq_len=10),TimeSeries(ch=1, seq_len=10),TimeSeries(ch=1, seq_len=10),TimeSeries(ch=1, seq_len=10),TimeSeries(ch=1, seq_len=10)\n",
       "y: FloatList\n",
       "0.0,9.0,2.0,9.0,8.0\n",
       "Path: .;\n",
       "\n",
       "Valid: LabelList (21144 items)\n",
       "x: TimeSeriesList\n",
       "TimeSeries(ch=1, seq_len=10),TimeSeries(ch=1, seq_len=10),TimeSeries(ch=1, seq_len=10),TimeSeries(ch=1, seq_len=10),TimeSeries(ch=1, seq_len=10)\n",
       "y: FloatList\n",
       "3.0,11.0,11.0,9.0,0.0\n",
       "Path: .;\n",
       "\n",
       "Test: None"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "db = (ItemLists('.', TimeSeriesList(X_train), TimeSeriesList(X_valid))\n",
    "      .label_from_lists(y_train, y_valid)\n",
    "      .databunch(bs=min(bs, len(X_train)), val_bs=min(len(X_valid), bs * 2), num_workers=cpus, device=device)\n",
    "      .scale(scale_type=scale_type, scale_subtype=scale_subtype, scale_range=scale_range)\n",
    "     )\n",
    "db"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 25 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "db.show_batch()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torchtimeseries.models import *\n",
    "# Select one arch from these state-of-the-art time series/ 1D models:\n",
    "# ResCNN, FCN, InceptionTime, ResNet\n",
    "arch = InceptionTime                     # ✳️   \n",
    "arch_kwargs = dict()                     # ✳️ \n",
    "opt_func=Ranger                          # ✳️ a state-of-the-art optimizer\n",
    "loss_func = FlattenedLoss(nn.L1Loss)     # ✳️ this is MAE loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "InceptionTime(\n",
      "  (block): InceptionBlock(\n",
      "    (inc_mods): ModuleList(\n",
      "      (0): Inception(\n",
      "        (conv_layers): ModuleList(\n",
      "          (0): Conv1d(1, 32, kernel_size=(39,), stride=(1,), padding=(19,))\n",
      "          (1): Conv1d(1, 32, kernel_size=(19,), stride=(1,), padding=(9,))\n",
      "          (2): Conv1d(1, 32, kernel_size=(9,), stride=(1,), padding=(4,))\n",
      "        )\n",
      "        (maxpool): MaxPool1d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=False)\n",
      "        (conv): Conv1d(1, 32, kernel_size=(1,), stride=(1,))\n",
      "        (bn): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (act): ReLU()\n",
      "      )\n",
      "      (1): Inception(\n",
      "        (bottleneck): Conv1d(128, 32, kernel_size=(1,), stride=(1,))\n",
      "        (conv_layers): ModuleList(\n",
      "          (0): Conv1d(32, 32, kernel_size=(39,), stride=(1,), padding=(19,))\n",
      "          (1): Conv1d(32, 32, kernel_size=(19,), stride=(1,), padding=(9,))\n",
      "          (2): Conv1d(32, 32, kernel_size=(9,), stride=(1,), padding=(4,))\n",
      "        )\n",
      "        (maxpool): MaxPool1d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=False)\n",
      "        (conv): Conv1d(128, 32, kernel_size=(1,), stride=(1,))\n",
      "        (bn): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (act): ReLU()\n",
      "      )\n",
      "      (2): Inception(\n",
      "        (bottleneck): Conv1d(128, 32, kernel_size=(1,), stride=(1,))\n",
      "        (conv_layers): ModuleList(\n",
      "          (0): Conv1d(32, 32, kernel_size=(39,), stride=(1,), padding=(19,))\n",
      "          (1): Conv1d(32, 32, kernel_size=(19,), stride=(1,), padding=(9,))\n",
      "          (2): Conv1d(32, 32, kernel_size=(9,), stride=(1,), padding=(4,))\n",
      "        )\n",
      "        (maxpool): MaxPool1d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=False)\n",
      "        (conv): Conv1d(128, 32, kernel_size=(1,), stride=(1,))\n",
      "        (bn): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (act): ReLU()\n",
      "      )\n",
      "      (3): Inception(\n",
      "        (bottleneck): Conv1d(128, 32, kernel_size=(1,), stride=(1,))\n",
      "        (conv_layers): ModuleList(\n",
      "          (0): Conv1d(32, 32, kernel_size=(39,), stride=(1,), padding=(19,))\n",
      "          (1): Conv1d(32, 32, kernel_size=(19,), stride=(1,), padding=(9,))\n",
      "          (2): Conv1d(32, 32, kernel_size=(9,), stride=(1,), padding=(4,))\n",
      "        )\n",
      "        (maxpool): MaxPool1d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=False)\n",
      "        (conv): Conv1d(128, 32, kernel_size=(1,), stride=(1,))\n",
      "        (bn): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (act): ReLU()\n",
      "      )\n",
      "      (4): Inception(\n",
      "        (bottleneck): Conv1d(128, 32, kernel_size=(1,), stride=(1,))\n",
      "        (conv_layers): ModuleList(\n",
      "          (0): Conv1d(32, 32, kernel_size=(39,), stride=(1,), padding=(19,))\n",
      "          (1): Conv1d(32, 32, kernel_size=(19,), stride=(1,), padding=(9,))\n",
      "          (2): Conv1d(32, 32, kernel_size=(9,), stride=(1,), padding=(4,))\n",
      "        )\n",
      "        (maxpool): MaxPool1d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=False)\n",
      "        (conv): Conv1d(128, 32, kernel_size=(1,), stride=(1,))\n",
      "        (bn): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (act): ReLU()\n",
      "      )\n",
      "      (5): Inception(\n",
      "        (bottleneck): Conv1d(128, 32, kernel_size=(1,), stride=(1,))\n",
      "        (conv_layers): ModuleList(\n",
      "          (0): Conv1d(32, 32, kernel_size=(39,), stride=(1,), padding=(19,))\n",
      "          (1): Conv1d(32, 32, kernel_size=(19,), stride=(1,), padding=(9,))\n",
      "          (2): Conv1d(32, 32, kernel_size=(9,), stride=(1,), padding=(4,))\n",
      "        )\n",
      "        (maxpool): MaxPool1d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=False)\n",
      "        (conv): Conv1d(128, 32, kernel_size=(1,), stride=(1,))\n",
      "        (bn): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (act): ReLU()\n",
      "      )\n",
      "    )\n",
      "    (res_layers): ModuleList(\n",
      "      (0): None\n",
      "      (1): None\n",
      "      (2): Sequential(\n",
      "        (0): Conv1d(1, 128, kernel_size=(1,), stride=(1,))\n",
      "        (1): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "      (3): None\n",
      "      (4): None\n",
      "      (5): Sequential(\n",
      "        (0): Conv1d(128, 128, kernel_size=(1,), stride=(1,))\n",
      "        (1): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "    )\n",
      "    (act): ReLU()\n",
      "  )\n",
      "  (gap): AdaptiveAvgPool1d(output_size=1)\n",
      "  (fc): Linear(in_features=128, out_features=1, bias=True)\n",
      ")\n",
      "InceptionTime\n",
      "======================================================================\n",
      "Layer (type)         Output Shape         Param #    Trainable \n",
      "======================================================================\n",
      "Conv1d               [32, 10]             1,280      True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             640        True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             320        True      \n",
      "______________________________________________________________________\n",
      "MaxPool1d            [1, 10]              0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             64         True      \n",
      "______________________________________________________________________\n",
      "BatchNorm1d          [128, 10]            256        True      \n",
      "______________________________________________________________________\n",
      "ReLU                 [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             4,128      True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             39,968     True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             19,488     True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             9,248      True      \n",
      "______________________________________________________________________\n",
      "MaxPool1d            [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             4,128      True      \n",
      "______________________________________________________________________\n",
      "BatchNorm1d          [128, 10]            256        True      \n",
      "______________________________________________________________________\n",
      "ReLU                 [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             4,128      True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             39,968     True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             19,488     True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             9,248      True      \n",
      "______________________________________________________________________\n",
      "MaxPool1d            [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             4,128      True      \n",
      "______________________________________________________________________\n",
      "BatchNorm1d          [128, 10]            256        True      \n",
      "______________________________________________________________________\n",
      "ReLU                 [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             4,128      True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             39,968     True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             19,488     True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             9,248      True      \n",
      "______________________________________________________________________\n",
      "MaxPool1d            [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             4,128      True      \n",
      "______________________________________________________________________\n",
      "BatchNorm1d          [128, 10]            256        True      \n",
      "______________________________________________________________________\n",
      "ReLU                 [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             4,128      True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             39,968     True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             19,488     True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             9,248      True      \n",
      "______________________________________________________________________\n",
      "MaxPool1d            [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             4,128      True      \n",
      "______________________________________________________________________\n",
      "BatchNorm1d          [128, 10]            256        True      \n",
      "______________________________________________________________________\n",
      "ReLU                 [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             4,128      True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             39,968     True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             19,488     True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             9,248      True      \n",
      "______________________________________________________________________\n",
      "MaxPool1d            [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             4,128      True      \n",
      "______________________________________________________________________\n",
      "BatchNorm1d          [128, 10]            256        True      \n",
      "______________________________________________________________________\n",
      "ReLU                 [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [128, 10]            256        True      \n",
      "______________________________________________________________________\n",
      "BatchNorm1d          [128, 10]            256        True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [128, 10]            16,512     True      \n",
      "______________________________________________________________________\n",
      "BatchNorm1d          [128, 10]            256        True      \n",
      "______________________________________________________________________\n",
      "ReLU                 [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "AdaptiveAvgPool1d    [128, 1]             0          False     \n",
      "______________________________________________________________________\n",
      "Linear               [1]                  129        True      \n",
      "______________________________________________________________________\n",
      "\n",
      "Total params: 406,049\n",
      "Total trainable params: 406,049\n",
      "Total non-trainable params: 0\n",
      "Optimized with 92695a7b8\n",
      "Using true weight decay as discussed in https://www.fast.ai/2018/07/02/adam-weight-decay/ \n",
      "Loss function : FlattenedLoss\n",
      "======================================================================\n",
      "Callbacks functions applied \n",
      "\n"
     ]
    }
   ],
   "source": [
    "model = arch(db.features, 1, **arch_kwargs).to(device) # c_out must be 1 for regression\n",
    "learn = Learner(db, model, opt_func=opt_func, loss_func=loss_func)\n",
    "learn.save('stage_0')\n",
    "print(learn.model)\n",
    "print(learn.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LR Finder is complete, type {learner_name}.recorder.plot() to see the graph.\n",
      "Min numerical gradient: 1.91E-06\n",
      "Min loss divided by 10: 3.31E-01\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "learn.load('stage_0')\n",
    "learn.lr_find()\n",
    "learn.recorder.plot(suggestion=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "epochs = 25         # ✳️ \n",
    "max_lr = 1e-1        # ✳️ \n",
    "warmup = False       # ✳️\n",
    "pct_start = .7       # ✳️\n",
    "metrics = [mean_absolute_error] # ✳️\n",
    "wd = 1e-2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>mean_absolute_error</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>6.122210</td>\n",
       "      <td>6.332006</td>\n",
       "      <td>6.333297</td>\n",
       "      <td>00:32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>6.099464</td>\n",
       "      <td>6.480971</td>\n",
       "      <td>6.485870</td>\n",
       "      <td>00:32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>6.099688</td>\n",
       "      <td>6.294734</td>\n",
       "      <td>6.286690</td>\n",
       "      <td>00:32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>6.223199</td>\n",
       "      <td>6.322987</td>\n",
       "      <td>6.291809</td>\n",
       "      <td>00:32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>6.081867</td>\n",
       "      <td>6.277673</td>\n",
       "      <td>6.255686</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>6.196229</td>\n",
       "      <td>6.294348</td>\n",
       "      <td>6.298598</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>6.193324</td>\n",
       "      <td>6.327833</td>\n",
       "      <td>6.332403</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>6.192964</td>\n",
       "      <td>6.303276</td>\n",
       "      <td>6.297862</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>6.150563</td>\n",
       "      <td>6.273915</td>\n",
       "      <td>6.203588</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>6.220777</td>\n",
       "      <td>6.305011</td>\n",
       "      <td>6.304469</td>\n",
       "      <td>00:32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>10</td>\n",
       "      <td>6.135925</td>\n",
       "      <td>6.276954</td>\n",
       "      <td>6.277526</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>11</td>\n",
       "      <td>6.050456</td>\n",
       "      <td>6.274265</td>\n",
       "      <td>6.267895</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>12</td>\n",
       "      <td>6.105549</td>\n",
       "      <td>6.342024</td>\n",
       "      <td>6.431219</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>13</td>\n",
       "      <td>6.133523</td>\n",
       "      <td>6.332522</td>\n",
       "      <td>6.376991</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>14</td>\n",
       "      <td>6.074009</td>\n",
       "      <td>6.305774</td>\n",
       "      <td>6.304619</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15</td>\n",
       "      <td>6.133453</td>\n",
       "      <td>6.429334</td>\n",
       "      <td>6.459049</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>16</td>\n",
       "      <td>6.184906</td>\n",
       "      <td>6.251750</td>\n",
       "      <td>6.197723</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>17</td>\n",
       "      <td>6.119113</td>\n",
       "      <td>6.320592</td>\n",
       "      <td>6.384859</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>18</td>\n",
       "      <td>6.244925</td>\n",
       "      <td>6.265233</td>\n",
       "      <td>6.227043</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>19</td>\n",
       "      <td>6.113158</td>\n",
       "      <td>6.275912</td>\n",
       "      <td>6.272141</td>\n",
       "      <td>00:32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>20</td>\n",
       "      <td>6.124684</td>\n",
       "      <td>6.238652</td>\n",
       "      <td>6.159026</td>\n",
       "      <td>00:32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>21</td>\n",
       "      <td>6.251988</td>\n",
       "      <td>6.303951</td>\n",
       "      <td>6.336017</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>22</td>\n",
       "      <td>6.166596</td>\n",
       "      <td>6.312424</td>\n",
       "      <td>6.337815</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>23</td>\n",
       "      <td>6.117054</td>\n",
       "      <td>6.298711</td>\n",
       "      <td>6.321068</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>24</td>\n",
       "      <td>6.132538</td>\n",
       "      <td>6.301471</td>\n",
       "      <td>6.310572</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "learn.metrics = metrics\n",
    "learn.load('stage_0')\n",
    "learn.fit_one_cycle(epochs, max_lr=max_lr, pct_start=pct_start, moms=(.95, .85) if warmup else (.95, .95),\n",
    "                    div_factor=25.0 if warmup else 1., wd=wd)\n",
    "learn.save('stage_1')\n",
    "learn.recorder.plot_lr()\n",
    "learn.recorder.plot_losses()\n",
    "learn.recorder.plot_metrics()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_data = TimeSeriesList(X_test)\n",
    "db.add_test(test_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [],
   "source": [
    "preds, _ = learn.get_preds(DatasetType.Test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor(6.4602)"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mae(preds,torch.tensor(y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The model seems to train as intended, but doesn't learn anything useful. In fact, it performs much worse than a simple average over all the points in a sequence. I could see why it wouldn't be able to BEAT an average - maybe there's just not much predictive value in the sequence - but it's weird that it does so much WORSE."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[9.1175],\n",
       "        [8.8943],\n",
       "        [8.9365],\n",
       "        [9.3123],\n",
       "        [9.3003],\n",
       "        [8.8859],\n",
       "        [8.7192],\n",
       "        [8.8442],\n",
       "        [8.7674],\n",
       "        [8.8648]])"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "preds[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 4., 21., 27., 14., 21., 10., 20., 15., 14., 21.], dtype=float32)"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test[:10]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So the model is predicting nearly the same thing every time, while the actual targets have a lot of variance."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Trying to overfit with a very small subset of the data, to see if the model is capable of just memorizing the patterns."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_small = X_train[:10,:,:]\n",
    "y_small = y_train[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TSDataBunch;\n",
       "\n",
       "Train: LabelList (10 items)\n",
       "x: TimeSeriesList\n",
       "TimeSeries(ch=1, seq_len=10),TimeSeries(ch=1, seq_len=10),TimeSeries(ch=1, seq_len=10),TimeSeries(ch=1, seq_len=10),TimeSeries(ch=1, seq_len=10)\n",
       "y: FloatList\n",
       "0.0,9.0,2.0,9.0,8.0\n",
       "Path: .;\n",
       "\n",
       "Valid: LabelList (10 items)\n",
       "x: TimeSeriesList\n",
       "TimeSeries(ch=1, seq_len=10),TimeSeries(ch=1, seq_len=10),TimeSeries(ch=1, seq_len=10),TimeSeries(ch=1, seq_len=10),TimeSeries(ch=1, seq_len=10)\n",
       "y: FloatList\n",
       "0.0,9.0,2.0,9.0,8.0\n",
       "Path: .;\n",
       "\n",
       "Test: None"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "db = (ItemLists('.', TimeSeriesList(X_small), TimeSeriesList(X_small))\n",
    "      .label_from_lists(y_small, y_small)\n",
    "      .databunch(bs=min(bs, len(X_small)), val_bs=min(len(X_small), bs * 2), num_workers=cpus, device=device)\n",
    "      .scale(scale_type=scale_type, scale_subtype=scale_subtype, scale_range=scale_range)\n",
    "     )\n",
    "db"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsgAAALICAYAAABiqwZ2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzs3Xl4W+d1Lvp3YyZIECBAkAQBUhRJkRoIyIonTRZpO45tkUnT46ZN2rRJkzSxfTqc256e2+Y0t7e3PZ1up9zm2M48N2naZmhIx5NkaqAky7IdgZpFkJJAEiSIgQPmad8/wE1BMiUSIPYAYP2eJ88TUSDwWRLAtb/9fmsxLMuCEEIIIYQQkiUTewGEEEIIIYRICRXIhBBCCCGE5KACmRBCCCGEkBxUIBNCCCGEEJKDCmRCCCGEEEJyUIFMCCGEEEJIDiqQCSGEEEIIyUEFcgljGEbNMMxXGYa5zjDMEsMw7zAM8+RdHv9/MAwzwzDMAsMwX2MYRi3kegkhAMMw2xiGObz8PhxjGOYX7/JYes8SIiCGYX6bYZgzDMPEGYb5xm2/9yjDMJcYhokwDPM6wzCb7vI8bcuPiSx/z3t5XzwpKiqQS5sCgBtALwA9gM8B+AHDMG23P5BhmMcB/BGARwG0AWgH8GcCrZMQAoBhGAWAnwAYBGAE8GkA32EYpmuVx9J7lhDhTQP4CwBfy/0iwzD1AH6I7M9ZI4AzAP71Ls/zPQDvADAB+J8A/p1hGDMfCyb8YGiSXnlhGMYJ4M9Ylv2P277+LwCusSz72eVfPwrguyzLNomwTEIqEsMwPQBOAdCxyx++DMO8AuANlmU/d9tj6T1LiEgYhvkLADaWZT++/OtPA/g4y7J7l39dDcAHYBfLspdu+94uAKMA6lmWXVr+2jFk378vCPdfQTaCdpDLCMMwjQC6AJxf5bd3ADib8+uzABoZhjEJsTZCCACAucPXelb5Or1nCZGOW96PLMuGAbiWv77aY8e54njZ2Ts8lkgUFchlgmEYJYDvAvjm7Vezy2oALOT8mvv/Or7XRghZcQmAF8AfMgyjZBjmfchGpLSrPJbes4RIx+3vRyz/erX3Yz6PJRJFBXIZYBhGBuDbABIAfvsODwsBqM35Nff/l1Z5LCGEByzLJgF8EEA/gBkAfwDgBwAmV3k4vWcJkY7b349Y/vVq78d8HkskigrkEscwDAPgqwAaATy1/AN4NecB7Mz59U4AsyzL+nleIiEkB8uyTpZle1mWNbEs+ziyh+9Or/JQes8SIh23vB+XM8gdWD3SeB5AO8MwuTvGO+/wWCJRVCCXvucBbAPwfpZlo3d53LcAfJJhmO0Mw9QB+BMA3xBgfYSQHAzDOBiG0TAMo2UY5r8DsGD19yK9ZwkRGMMwCoZhNADkAOTL71UFgB8B6GEY5qnl3/+/ADhXizSyLHsFwM8B/Ony9/8iAAeA/7j9sUS6qEAuYcs9GD8D4B4AMwzDhJb/92sMw7Qu//9WAGBZ9iUAfwvgdQDXl//3p2KtnZAK9usAPMhmkR8F8BjLsnF6zxIiCX8CIIpsi8WPLv//P2FZdg7AUwD+F4AggAcBfJj7JoZhXmAYJrdDxYcB3Lf82L8G8EvLz0FKBLV5I4QQQgghJAftIBNCCCGEEJKDCmRCCCGEEEJyUIFMCCGEEEJIDiqQCSGEEEIIyaHI58H19fVsW1sbT0shhADAW2+95WNZ1rzR56H3KyHCKMZ7lt6vhAhjve/XvArktrY2nDlzpvBVEULWxDDM9WI8D71fCRFGMd6z9H4lRBjrfb9SxIIQQgghhJAcVCATQgghhBCSgwpkQgghhBBCclCBTAghhBBCSA4qkAkhhBBCCMlBBTIhhBBCCCE5qEAmRACpdAaLsaTYy3gXlmURiqfEXgYhhBAiKVQgEyKAF8/NYN9fHcbV2SWxl3KLT37zDD79Leq9ypdT43488vfDWIhI7+KIEEKE9pcvXsTvfO8dsZexLlQgE8IzlmXx/LAL5lo1Osw1Yi/nFlsaavDmtQDtIvPkjfEAxufCGL7iFXsphBAiqqn5KL52fAJHLpfG5yEVyITwbPjKHC56FvF0bwdkMkbs5dyit8uMZJrFSZdf7KWUJXcwAgA4cmVO5JUQQoi4vnJsHKkMi8VYqiTuqlGBTAjPnn/dBYtegw/eYxV7Ke9yX5sRWpUcR2iHkxfuQLZAPnrFh0yGFXk1hBAijkA4ge+fdqOxVg0AuLH82ShlVCATwqMz1wI4fS2A33qoHSqF9N5uKoUMezvqMXx5DixLBVyxTQajqFbJ4QvFccGzKPZyCCFEFN8YmUA0mcZnD24DcPPumpRJ7yc2IWXk+WEX6rRKfPiBFrGXcke93WZMBqMY94XFXkpZSaYz8CxE8Qu7sncOKGZBCKlEoXgK3zx5He/b3oiHtzYAoB1kQirapZlFHLrkxcf3boZWpRB7OXfU12UGAAxfpgKumKbno8iwwD0tBvRYazFcIgdTCCGkmL73xg0sRJN49uFO1GqUMGiVK/EzKaMCmRCePD/sglYlx8f2bhJ7KXfVYtSi3VxNO5xF5g5EAQAtdVr0dTXg7RvzWIhK/2AKIYQUSzyVxleOj2Nvhwn3tBgAAK1GLe0gE1Kpbvgj+OnZafzqA60waFViL2dNfV0NeGPcj1gyLfZSygaXsWsxVqG324x0hsWJMZ/IqyKEEOH86O0pzC7G8Uxfx8rXWuq0mAxGRVzV+lCBTAgPvnTMBbmMwaceahd7KevS221GPJXByXFq91Ys7kAEChkDi74Ku1oM0GkUFGMhhFSMdIbFF4+Ow27VY39n/crXW4xaTAYjSEu8sw8VyIQUmXcphh+cmcRT77GhSa8Reznr8uBmI9QKGY5QAVc0k8Eomg1VkMsYKOQyPLSlHkeuULcQQkhl+Nk5DyZ8YTzb1wGGuTkDoMVYhWSaxexiTMTVrY0KZEKK7Osj15BMZ/DpA6WxewwAGqUcezpMlEMuIncwAltd1cqve7vMmFmM4bLExo0TQkixsSyL5153ob2+Gu/b0XTL77UatQCk38mCCmRCimgxlsR3Tl7HwR4L2iU2VnotvV1mTPjCuO6ndm/F4A5E0VKnXfl1b1e2vRHt0hNCyt3Rqz5cWJ4gK79tgiz3uSj1ThZUIBNSRN8+eR1L8dQtBxJKRV/3cgFHu8gbFk2k4QvF0WK8uYPcpNdga5OO/nwJIWXvudfH0FSrwQd3vXuCbLOhCjKGCmRCKkYsmcbXRyZwoMuMHqte7OXkrc2kRatRSzucRTC50sFCe8vXe7vMePNaAKF4SoxlEUII7966HsQbEwF86qHNq06QVSlksOir4JZ4JwsqkAkpkn8744YvlMAzvaW3ewwADMOgr9uMEy4/4ilq97YRXIs3W91tBXK3Gck0i5Mu6hZCCClPzw+7YNAq8ZEHWu/4mBZjFWWQCakEqXQGXzw6jl2tBuxuN4q9nIL1dpkRTabx5kRQ7KWUtJUhITkRCwC4b5MR1So5TdUjhJSlyzNLeO3iLD6+tw3V6jtPkG2p01LEgpBKMOj0YDIYxbN9nbe0syk1ezpMUMllOHKFCriNcAci0ChlMNeob/m6SiHD3k5q90YIKU8vHFmeILun7a6PazVq4V2KS3o4FRXIhGxQJsPi+WEXtjTU4NGtDWIvZ0O0KgUe2GykgRYblG3xpl31Yqm3y4zJYBTjPuoWQggpH+5ABP95dhofeaAVddV3nyDLnc/gzmtIERXIhGzQ4UteXJ5dwjN9HZDJSnf3mNPbZcZVbwhT89I+QCFl2RZvVav+Xm+XGQDoIoQQUla+fGwcMgb41EOb13xsSwn0QqYCmZANYFkWzw2PwWqowvt3Nou9nKLo684WcEepHVnB3MHIuzpYcFqMWnSYq6ndGyGkbMwtxfGvb7rxX3bZYNGvvjmQizufwZ3XkCIqkAnZgNMTAbx9Yx6fPtAOpbw83k6dDTVo1mvoIFmBFiJJLMVStwwJuV1vVwNOjfsRTUg3f0cIIev1jRMTSKQz+HTv+ibImmvU0ChltINMSLl6btgFU7UKv3xfi9hLKRqGYdDbbcbImB/JdEbs5ZScmy3e7ryL0tdtRiKVwakJavdGCCltS7EkvnXyOp7saULHOifIMgwj+U4WVCATUqDz0ws4cmUOn9i/GVUqudjLKarergaE4im8dZ3aveXrTkNCcj2w2QiNUkZDWQghJe87p25gKZbCM72deX1fq1FLO8iElKPnh12oUSvw0d2bxF5K0e3rNEEhYygnW4CVHsh3iVholHLsbjfRny8hpKTFkml89fgEHtpSD7stvwmyLUYtJoNRyba8pAKZkAJc84Xx4qgHv7a7FfoqpdjLKTqdRol7N9XRDmcB3MEIdBoF9Nq7/7vo6zJjwhfGdT+1eyOElKZ/f2sSvlAcz/TlP0G2xahFKJ5CMJLkYWUbRwUyIQX44tFxKOQyfHLf2u1sSlVvtxkXPIvwLsbEXkpJcQcid9095vR2Z3tm0y4yIaQUZSfIunBPiwF72k15fz/XClOqOWQqkAnJ0+xiDP/x1iR+6V4bGmo1Yi+HN1y/Xirg8uMORt81Yno1m+urscmkpV16QkhJGhr1wB2I4pm+joImyLaapN0LmQpkQvL01eMTSGUy+MyB9bWzKVXbLbUw69QYpgJ53ViWxWRwfTvIQPYi5ITLj3iK2r0RQkoHy96cIPvYtsaCnoP7nHRLdJoeFciE5GEhksR3T13HgKMZm0zVYi+HVwzDoLfLjONXfUhRu7d1mQvFEUtm7trBIldvlxnRZBpvTlC3EEJI6Xj9sheXZpbwdG/hE2Sr1QqYqlUUsSCkHHzr5DWEE2k83Zv/gYRS1NdtxkI0ibOTC2IvpSSsdLBYR8QCAPZ0mKCSy3DkCg1lIYSUjuded8FqqMIH7tnYBNkWCbd6owKZkHWKJtL4+olreLjbjO3NtWIvRxD7O+shY4AjNFVvXVZ6IK8zYqFVKfDAZiOGKYdMCCkRb14L4Mz1IH7roc0bniDbYtRKdtw0FciErNO/vnkDgXACzz6cXzP0UmbQqnBPi4EO6q0Td6vQts4CGcju0l/1hjA1L80fEoQQkuu518dgrFbhV+5v3fBztRqrMDUflWSMjwpkQtYhmc7gy8cmcN+mOtzfZhR7OYLq626Ac2oB/lBc7KVInjsQRX2NOq/JiivdQmgXmRAicRemF/H65Tn85t62okyQbanTIp1h4VmQXjtRKpAJWYef/HwaU/NRPPtwZWSPc/V2mcGywLGrPrGXInmT8xHY6taXP+Z0NtSgWa+hHDIhRPJeOOJCtUqO39jTVpTna10+0CzFg3pUIBOyhkyGxQtHXNjapMPDy8MdKondqoexWkUxi3VwB6Lr7mDBYRgGvd0NGBnzIynB24yEEAIA1/1hDDqn8dHdm9acFLpe3OelFFu9UYFMyBpevTiLMW+o4GbopU4mY3BgSz2OXplDJsOKvRzJSmdYTM9HV6ZD5aO3y4xQPIW3rlO7N0KINH3x6DgUMhk+sb94E2Qteg3kMkaSnSyoQCbkLliWxXPDLrQatei3W8Rejmj6uhvgDydwbpravd2JZyGKVIbNewcZAPZ1mqCQMbRLTwiRJO9iDP9+ZhJP3WtDYxEnyCrkMjQbNJLsZEEFMiF3cXLcj7PueXz6QDsUG2xnU8oe2lIPhgG1I7uLlR7IeXSw4Og0Sty7qY7+fAkhkvTVkewE2ad7iz9BtlWivZAr9yc+Ievw/LAL9TVq/NK9NrGXIipTjRp2q552OO+Cy9Ctd0jI7Xq7zbjoWcTsovROcxNCKtdCNInvnrqBfp4myLbUaVd6yEsJFciE3MHo5AKOXfXhk/s3Q6PceDubUtfXZcY7N4JYiCTFXookTQYikDFAs6GwArmvK3sA9ChdhBBCJOQ7p64jFE/xsnsMZA/q+UIJhOMpXp6/UFQgE3IHzx8Zg06jwEd3b7wZejno7TYjwwLHxqiAW407GIVFX1XwZKltFh0adGoMU4FMCJGIaCKNrx2fQF+3GTua9by8BnduYzIorRwyFciErMI1F8LPzs3gN/Zsgk5TnHY2pW6nzQB9lZIGWtyBO5B/D+RcDMOgt8uM41d9kpwqRQipPD8444Y/nMCzffxNkOV6IUsth0wFMiGr+OIRF1RyGX5zX/Ha2ZQ6hVyG/VvqceTKHFiW2r3dzh2MFNTBIldvtxkL0STOTs4XaVWEEFKYZDqDLx0dx72b6nB/Wx1vr8O1xpTasBAqkAm5jWchih+9M4Vfub8F9TVqsZcjKb1dZniX4rjoWRJ7KZISS6YxuxgvqINFrv2d9ZAxNHaaECK+n55dniDL8wwAY7UK1So57SATInVfOTaBDAv81kP8HEgoZX1dZgDAMI1FvsXUfDY7t5GIBQAYtCrsaq2jbiGEEFFlMiyeH3ahu1GHR7byO0GWYRi0GKXXyYIKZEJyBMMJfO/0DXxgZ/OGb5eXo4ZaDbZZammH8zbc4ZJi/Jvp7TLDObUAfyi+4eeqBJFEiiI/hBTZaxdncVXACbItEuyFTAUyITm+9+YNRBJpPN3bIfZSJKuv24y3rgexFKN2bxwuO1doD+Rcfd1msCxw7Kpvw89V7pZiSTz4vw7hh29Pib0UQsrK10YmYKurwoBDmAmyLXVauANRSV3sUoFMSI6fnvXgPa0GdDfpxF6KZPV2mZHKsBgZ84u9FMlwByNQyWVo1G18BGtPsx6mahWGL1OMZS1XvSEsxVM44aJ/i4QUy+xiDG9MBPChe1sEmyDbaqxCNJmGL5QQ5PXWgwpkQpa55kK46FlEv6NZ7KVI2r2b6lCjVlBONsdkIAprXRVkso3fipTJGBzoMuPoVR8yGenspkiRyxsCAIxOUdcPQorlZ6MesCzQ72gS7DW5eJpbQjlkKpAJWfai0wMAOGgX7kOhFCnlMuzrNOHIZa+kboeJyR3cWA/k2/V2mREIJ3BueqFoz1mOxuayBfKYNyS5KVyElKqhUQ+2NunQ2SDcnVSuF7KUWr1RgUzIsqFRD+5vq4NFX7xCp1z1djVgeiGGseUdvErnDmy8B3Kuh7bUg2GAYToMeVcubxgAkGGBC55FkVdDSOnzLETx5rWgYNljjq2OCmRCJGnMu4RLM0votwv7oVCqeruz7d4oZgGE4ikEI8kN90DOZapRw2HV05/vGsbnQrh3U3aAwegk7bYTslEvjs4AAA4K/LOwSiWHWaeWVCcLKpAJATDo9IBhgCepQF4Xq6EKWxpqaIcTxe1gkau3y4x3bgQxH5HOoRUpSaQyuB6IYG+HCY21aoxOUYFMyEYNOaex3VKLdnON4K/dUlcFdyAq+OveCRXIhAAYcnpwf5sRjbUb70JQKXq7zDg9EUAkUdnZz5UCuYg7yADQ292ADAscH6N2b6u57g8jnWHRYa6B3aqHk8ZzE7IhU/NRvH1jHv0Cxys4rRLrhUwFMql4V2aXcNUbEjxzVer6uhuQSGdwssJbbLmLOCQk106bHvoqJQ1luQPX8gG9bIFswLgvjBAd1COkYD8bzR5UF+tnYYtRC89CFMl0RpTXvx0VyKTiDTo9kDHAEz3UvSIf92+uQ5VSXvE5WXcgAq1KjjqtsqjPq5DLsH9LPY5cmaNuIatwzWUP6LWbq+Gw6cGywHmKWRBSsJ86PbBb9dhkqhbl9VuMWmRYYHpeGjELKpBJRWNZFoPOaTy42YSGIgx5qCRqhRx7O0wVXyBPBqNoqdPyMo61r8sM71IcFz1LRX/uUufyhtCs16BarUCPVQ8AlEMmpEDuQARn3eLFK4Cbrd6kErOgAplUtEszSxifC4v6oVDKervNuO6PYMIXFnspopkMRop+QI/T25XtFjJ8habq3c41F0JHQ/YgkVmnhkWvgZM6WRBSkBeX4xVidnJaGRYikYN6VCCTijZE8YoN6etqAAAcqdCxyCzLwh2IrPTwLLaGWg22W2oph3wblmXhmgujI+ekvd2qxznaQSakIEOjHuy06Yt+liIfTbUaKOUM7SATIjaWZTE06sGeDhPqa9RiL6cktZq02FxfjeEKjVkEI0mEE2lef6j0dpvx1vUglmJJ3l6j1MwuxhGKp9BhvpmVdNj0GPeFsUh/ToTk5YY/Aufkguh3UuUyBlZDlWTGTVOBTCrWBc8iJnxhDDiaxV5KSevtMuPUuB+xZFrspQjuZos3/qYv9naZkcqwGBmr7G4huXI7WHC4HDLtIhOSn8HRaQDCDwdZTYtRK5lpelQgk4o16PRALmPw+A6KV2xEb7cZsWQGpycCvL+Wc3Ier5yf4f111ovb6eBzB/neTXWoUSsq/jBkLq5A7my4NWIBUIFMSL6GnB7sajXwFhXLBxXIhIiMZVkMOT3Y22GCsVol9nJK2u7NJqgUMl6n6qXSGfzjq1fwi8+dwDPffVsyu9XcYRI+C2SlXIZ9nSYcueyldm/Lxrwh6NQKmHU3o1GmGjWshio6qEdIHiZ8YZyfXhT1cF6uVqMWwUhSEpEyKpBJRTo3tYgbgQgNBymCKpUcu9tNOMJTp4Xr/jA+9MWT+Pyhq9hm0SGdYXHBs8jLa+XLHYygTqtEjVrB6+v0dTdgeiGGMW+I19cpFa65ENobat7VWs9h01OrN0LywHWvkEK8Arg5kVQKnSyoQCYVaXB0GgoZg/dtp3hFMfR2meGaCxf11hjLsvjBGTcOfv4YXN4Q/vkju/Dl37gPADAqkV1CdyAiyKnvA8vt3ihmkeXyhm85oMfpsepx3R/BQkT83SdCSsGg04N7N9Wh2cDfOYp8SKkXMhXIpOJw8Yp9nfWoo3hFUfQWuYALhhN49rtv43/8uxN2mx4v/bcDeP/OZjTValBfo5bMbXRuSAjfrIYqbGmo4TXGUipC8RRmFmO35I85DttyDnlaGv8+CJEy11wIFz3SiVcAWOkpPymBThZUIJOKc3ZyAZPBKMUriqjDXA1bXVVRCuTjV3144vNH8drFWfzRk1vx3U/tXtndYBhm+Tb6/IZfZ6MyGRZTwShsPA0JuV1ftxmnJwKIJFKCvJ5Uubzv7mDB4Q7qSeUCihApG3J6wDDSiVcAgL5KCZ1GQTvIhIhhyDkNpZziFcXEMAx6u8w4MeZDIpUp6DniqTT+YvACPvrVN1CjVuBHz+7D070dkMtuzZn2WPUY84ZELxRnl2JIpDOCnfzu7WpAIp3BSVdlt3tbrcUbx6BVocVYJYkLKEKkbsjpwf2bjGjSa8ReygqGYdBSJ41OFlQgk4rCxSse2mKGXqsUezllpa+7AeFEGmeu59/u7crsEn7hCyP4yvEJ/PruTRj8nYdW+trezmHVI8MCF6bFPag3GVzuYMFjD+Rc92+uQ5VSXvE5ZNdcCAoZg02m1S9MHFYD7SATsoars0u4PLsk+nCQ1bQatbSDTIjQ3nHPY3ohJqnMVbnY02GCUs7kNRaZZVl8fWQCA/98HL5QHF/7+H348w/2oEolv+P32G3SuI2+MiREoNGsaoUceztMVCB7w2g1aaGUr/7jy27TYzIYRTCcEHhlhJSOodFsvOLJHundSW0xVmEyGEUmI25bSyqQSUUZcnqgksvw2I5GsZdSdmrUCty3ybjuAs67GMPHvv4m/uynF7C/sx4/+70DeGTr2n8vjbUaNOjUog+E4NoQWQU8/d3bbcZ1fwQTvrBgryk1rrkQOleJV3Acy3ceqN0bIatjWRaDTg8eaDOioVY68QpOq1GLeCqDuVBc1HVQgUwqRiaTjVcc6DKjVkPxCj70dZtxaWYJMwuxuz7ulfMzeOLzx3B6wo8//2APvvqx+24Z+rAWh00Pp9gFcjCCxlo1NMo773YXW19XAwDgyGV+ek5LXSqdwTV/GB2rdLDg7KACmZC7ujIbwpg3hIGdzWIvZVU2I9cLWdyYBRXIpGK8fSOImcUYda/gUW831+5t9QIukkjhj3/oxKe//RYseg0Gf2c/fn33pncNfFiL3WqAay6EUFy8g3ruQESQFm+5Wk1abK6vxnCFxixuBCJIptlVD+hx9FVKtJm0kumVTYjUDDmnIWOAJ3ZIL14BSKcXctEL5LmlOJLpwk6xE8KnQacHKoUMj25rEHspZau7UYemWs2qMYuz7nn0/3/H8f033Xi6twM/enYfOht0Bb2Ow6YHywLnRdwlnAxGBcsf5+rtMuPUuF8y47aF5JrLRktWGxKSy24z0A4yIatgWRaDox7sbjfldddOSFxsTexpekUtkC/PLGH/3xzGf/58uphPS8iGZTIsXhz1oK/LDB3FK3jDtXs7dtWH1PKFcjrD4guHr+Kp508glkzjXz61G3/05FaoFIV//PSIfBs9mc7AsxAVrINFrt5uM2LJDE5P5N8tpNSttHi7S8QCAOzWWkzNR+ETOcNIiNRc9CxhfC4sye4VHI1SjqZaTXntIG9pqEGbqRovHHGJfvqQkFxnrgfhXYpL+kOhXPR2m7EUS+Ed9zzcgQg+/KWT+LtXruCJnia89HsHsKfDtOHXMOvUsOg1ohXI0/NRZNibWTkh7d5sgkohq8ipemPeEBp06jXPENitBgCUQybkdkOj05DLGMnGKzgtxiq4RZ6mV9QCWSZj8ExfB656Q3jt4mwxn5qQDRlyTkOtkOG926h7Bd/2ddZDLmPw969cxsHPH8NFzxL+4Zd34p8/squovaftVr1oOVPu1p/QGWQAqFLJsbvddMecdzlzzYXumj/m9FhrAQDnKIdMyApuDsDeDhNMNdKMV3BajOIPC1EU+wkHHBb83SuX8dywC49tb8z78A0hxZbOsHjx3Awe2dqAanXR/8mT2+irlHhPqwGnxgO4b1Md/vFX7uElq2u36vHKhVksxZKCx2a4nY0WgcZM3663y4w/H7yA9/z5qyj2J6xGKce3P/kA2tdRiAqJZVm4vCF84J61T97rNEq0m6tF73RCiJScn17ENX8ET/d2iL2UNbXUafGjxSnEU2moFcJ1CspV9GpBIZfhMwfa8bmfnMep8UBRbqcSshGnJwKYo3iFoP5n/3ZcmF7EL99ng+IOAx02ihsYcm5qUfDPGXcgArmMQZNIPUSfeo8V0/NRxFPFPaiXTLH41zNuHLvqk1yB7AslsBhLrWsHGcgj59ChAAAgAElEQVReQL0xXnk5bULuZGjUA7mMweMSj1cA2U4WLAtMBaOifRbxsp32ofta8PlDV/Hc8BgVyER0Q6PT0ChleGQrda8Qyj0tBtzTYuD1NewrB/XmhS+Qg1E0GzS8Ff9rMWhV+NzA9qI/L8uyOHRpVpLZXe6AXucaB/Q4dqseP/n5NLxLMTTopDcMgRAhZYeDTGNfZz3qqlViL2dN3F1Ht4gFMi+f7hqlHJ/YvxnHrvqoFyURVSqdwUvnZvDo1kZoVRSvKCemGjWshipRRk5PBoXvgSwEhmHQI2K2+27GvMsdLNb5w9Jhy16giT1xkRApGJ1agDsQxYC9NO6kSqEXMm/bHx/dvQk6tQIvHHHx9RKErOn0RAC+UIKGg5Qpu1UvSgHkDkTLskAGsqOar3qXEEmIN4RlNa65ELQq+bpjLTuaa8EwEOUCihCpGXJ6oJSXRrwCABp0aqgUMkyWY4Fcq1Hio3s24cVzHowv3xojRGg/dXqgVcnR103xinJkt+lxzR/BQiQp2GtGE2n4QnHRDujxzW4zIMMCF6YXxV7KLVxzYbSbqyGTre9YYrVagQ5zjSR3wwkRUjZe4cH+zvqidhLik0zGwFZXVZ47yADwiX2boZTL8MUj43y+DCGrysYrPHh0WyOqVOKcgiX8cnAH9aaFK4ImVzpYlOkOsk3cISx34vKG0JlnFtFh1Uvuv4MQoZ2dXMDUfBT9jrU7wEhJS522fAtks06NX77Phh++M4mZhRifL0XIu5wc9yMYSaK/RDJXJH89zcIXc1yLN1uZRiwaazUw69SS2nmNJFKYmo+uO3/Msdv08C7FMbtIP39I5Ro8Ow2lnMFj20trDkCryL2QeT+C/ZkDHciwwFeO0S4yEdaQ04NqlRx93Waxl0J4UletQouxStBibmVISJlGLIDszquUegiPz4UBrD1i+nZcpxPKIZNKlcmweHHUgwNbzNBXlUa8gtNq1GIxlhI0QpeL9wK5xajF+x0W/MvpGwiGE3y/HCEAgGQ6g5fOz+Cx7Y3QKCleUc4cVgOcU/OCvZ47EIFGKYNZ4pOoNsJu08M1F0I4Lo2DelyLt3x3kLc310LGAKOTwv37IERK3nHPY3ohhoGdpXcnlduEEGvktCBNPJ/u60Akkca3Tl4X4uUIwciYD/ORZMllrkj+7DY93IGoYBfg7mAEtjptWU8JtVv1YNns5C0pcM2FIWOAtvr8Yi1alQJbGnSUQyYVa8jpgUohw3u3lVa8Arh5zkOsHLIgBfLWplo8urUB3zgxIbnWQaQ8DTk90KkVeGhLvdhLITzjbqMLdVAv2+KtfOMVQG40QRo7ry5vCK1GbUEjZ+227EE9lmV5WBkh0sXFK3q7zNBpSiteAeQMCynnAhkAnn24A8FIEt877RbqJUmFSqQyeJniFRWDO6gnVM7UHYyUbQcLTkOtBk21GskM2XDNhfKOV3DsVj18oQQ8dFCcVJi3bgQxsxgr2TkAtRolDFplee8gA8C9m4x4YLMRXzk2jkQqI9TLkgo0MubDYiyF/hL9UCD50WuVaDNpBTmotxBJYimWgq3Md5CB7M6rFA7qpTMsxn3hvA/ocewSbVtHCN+GnB6oFTI8WoLxCk5LnRbuYFSU1xasQAaAZ/o64FmI4cc/nxLyZUmFGXR6oNMo8NAW6l5RKXoE6nfLHRYp1yl6uexWPcbnwliKiXOCnDMVjCKRyqDDXF3Q92+31EIuYyTVto4QvqWX4xUPdzegRq0QezkFE7PVm6AFcl+XGdsttXjhiAvpDOXBSPHFU2m8cmEGj+9ogkoh6D9vIiKHTY+p+Sj8oTivr1PuQ0JycTuv56bEPajHdbDoLHAHWaOUY0tDjSR2wwkRyplrAXiX4iV/J9VmrMJUMCpKzShoBcEwDJ7p68D4XBivXpgR8qVJhTh2xYclildUHLvVAID/2+grPZArZAcZgOg55DFvtkBury+sQAayF1Cjk/N0UI9UjKFRDzRKGR7Z2iD2Ujak1ahFIp0RZdiP4FtsT/Y0YZNJi+eGXfRhRYpuaNQDfZUS+zqoe0Ul6bHWAgDvt9HdwQh0GgX02tI7EZ6v+ho1mvUa0XdeXXMhmKpVqKtWFfwcdpsBwUgSU/PiZBkJEVI2XjGDR7Y2oLqE4xXAzc0IMWIWghfICrkMnznQAefkAk64/EK/PCljsWQar16YxeM7GileUWF0GiXa66sF2EGOVMTuMce+vPMqpo10sOA4lnfDKYdMKsEbE374QnH020t/DkCriL2QRakinrrXigadGs8Nj4nx8qRMHb0yh1A8hQEaDlKRuH63fHIHo2U9Yvp2DpsB1/wRLETFO6jnmgujo6GwA3qc7iYdFDJG9N1wQoQw5PSgSikv+XgFADQbqsAwEKWThSgFslohxyf3b8bImB9n3dJoRE9K36DTgzqtEns6TGIvhYjAbtXDsxDD3BI/B/VYlsVksMJ2kJd3Xs+LVFgGwgkEwokN7yBrlHJ0N+lEz1MTwrdUOoOXzs3g0W0NqFKV/hwAlUKGZn1VZUQsOL+2exNqNQo8P+wSawmkjMSSabx2cRZP9DRBKad4RSXi+1DZXCiOWDJTER0sOCsT9UQqLLkOFoX2QM7lsOnhnKSJeqS8nRoPwB9OlOxwkNXY6iqsQK5RK/CxvW14+cLMyillQgo1fNmLSCJdFpkrUpgdVj0Yhr+JeisdLCooYlFXrYKtrkq07K5r+WdD5wZ3kIFsr+yFaHLl75GQcjQ0Og2tSo6+7tKPV3BajdrKySBzPr63DWqFDC8coV1ksjGDTg9M1SrsbjeKvRQikhq1Ah3mGoxO8RPbmqygISG5HAJku+/ENReCWiFDs2HjFyWO5VaATp7+fRAituRyvOK92xqhUZZ+vILTYtTCuxRHLJkW9HVFLZBNNWp8+P5W/PidKUxT+x1SoGgijUMXvXiipwkKildUNIdVz+MOcrZAtlbAmOlcdqsBNwIRzEcSgr/2mDeEzfXVkMuYDT9XV1MNVHIZjZwmZeuky49gJFlW8QrgZicLbpNCKKJXE791oB0A8OVj4yKvpDR87sfn8LcvXRJ7Gev2paMuDPzzMXz75DWE4yleXuP1y15Ek2kaDkLQY9XDuxTnpam8OxBFfY0KWlVp9xXNF5dDFqOwdM2FC56gdzu1Qo6tFh21eisD0UQav/CF4zhyZU7spawLy7L45DfexMe+dhqHL80iw9NUuCGnBzVqBQ50mXl5frFwsTah41GiF8hWQxV+4R4rvn/ajUBY+B2KUhJLpvGDM25888Q1wW81FCKTYfGVYxMY84bwuZ+cx+6/PIQ/++l5jM8VN3M+6JxGfY0KD26m7hWVzmHjr9+tOxiBrcLiFYB4BXIsmYY7GNlwB4tcPdZsXISvAoUI4/S1AM5OLuBbJ66JvZR1uehZwqFLXrwx4ccnvnEGfX83jK8cG8dCpHjtExOpDF46P4PHtpdXvALAysFooXPIohfIAPB0bzuiyTS+USL/2MXy9vUg4qkMwok0hi9L/8r5zPUgvEtx/M1TDvzo2b14dFsDvnPqOh75+yNFu5IOx1M4fMmLJ3ssRbkNS0rb9uZayBh+ui5MBqMV1cGCo9cqscmkFXzn9Zo/DJYtTgcLjsOqx1IshesiHPghxXNizAcAOHp1TtQe3es1NDoNuYzB0T98GF/41V1orFXjL4YuYvdfHcIf/3AUl2YWN/waIy4fFqJJ9NvL706quUYNjVImeCcLSRTIWxp1eN/2RnzzxDWEeLoNXw5GXD7IZQwMWiWGRj1iL2dNQ85pqBUyvHdbI3a11uGfPrwLI3/0CH7/sS5cmlnEJ75xBg///fKVdIEfcocveRFLZiheQQAAWpUCnQ01RW/1ls6wmJ6PoqXC8secHh6z3Xfi8oYBAB3mjQ0JyWW3iRcXIcUz4vKhvkaNZJrFqxdmxV7OXbEsiyGnB3s7TGio1WDA0Yx/e3ovhn53Pz6wsxk/fHsST/zTMXz4Syfxs1EPUulMQa8z5PRAp1Hgoa76Iv8XiI9hGLTUCd/JQhIFMgA809eBhWgS33vjhthLkazjY37c02LAQbsFhy7OIpqQbswinWHx4rl3z4Jv0Gnwu49uwfH/8xF84Vd3oUG3fCX9l4fw2R+N4vLMUl6vM+T0oEGnxv1t1L2CZNmthqL3u/UsRJHKsBW5gwxkd16n5qOCxuDGvCEwDNBeX7wd5K5GHVQKmejjs0nhguEEzk8v4td3b4LVUIVB57TYS7qr89OLuOaPvGtnd0ezHn/zSw688dlH8cdPbsVkMIpnvvs2Hvrb1/G/Xx+DP7T+gUeJVAYvn5/B+7Y3Qa0or3gFp8WoFXyanmQK5F2tddjTbsJXjo8jnpJu4SeWhWgSo5Pz2NdhwoDdgkgijdcve8Ve1h2dnghgbil+x51dpVy2ciU9+DvZK+n/eGsSj//T0XVfSYfiKbx+2YuDdopXkJscNj18oThminhQb6UHcgVmkAFxdl5dcyFYDVVFnQamlMuwzVIr+G44KZ6T436wLLB/iwkDDguOX/WJ0mFlvQadHshlDB7f0bTq7xu0KnymtwNH/vBhfPk37kOHuQb/78uXseevDuP3f/BzONdxMXd8bA5LsVTZda/I1WrUwh2ICDroRzIFMgA8+3AHZhfj+NHbU2IvRXLeGPcjwwJ7O+vxYLsJ9TUqDDmlG7MYGp1e9yz4Hmv2SvrUH2evpN2B7JX0gTWupA9dnEU8RfEKciuumCtmEeTmeiBX0JCQXD3cQT0Bd15dc6GiHtDjOKx6nJ9epIN6JWpkzIdqlRwOmwEDjmakMixeOS/NmAXLshganca+znrUVavu+li5jMFj2xvxnU89iNd+/wA+/EALXj43gw98YQQf/N8j+PE7U0ikVt80GjzrQa1GgX2d5Rev4NjqqhCKpzBfxIONa5FUgby/sx52qx5fPDqONH143eKEyw+NUoZdrQbIZQye7LHg0KVZRBLSy2xzs+Af2daQV0usuurslfTR//EwvvTr96Kdu5L+68P4gx+cfdeV9KDTg8ZaNe5trSv2fwIpYdsttZDLmKLmkCcDEcgYFGVgRSmq1Sixub5asJ3XTIbF+FyYlwLZbtMjFE9hwh8u+nMT/p1w+fFguwlKuQw91lq0GrUYlOiZnNGpBbgD0bx3djsbdPh/fqEHpz77KP7v92/HYjSJ//avP8fevz6Mf3jl8i1tLGPJNF69MIvHdzRBpZBUSVdUrSJ0spDUnybDMHimrwMTvjBeOjcj9nIkZWTMh/vbjCv5on6HBbFkBocvSS9mcXoiAF8ogYECT9PKZQzet6Np5Ur6V+5rwc/OefCBL4zgF58bwU9+PoVAOIEjl+dw0G6BjOIVJIdGKceWhpoi7yBHYdFXQVnBg2jsVn3RDz/eyfRCFNFkGh0NxTugx1lpW0cxi5IzPR/FhC+MvR3Zlp4Mw6DfYcHImA9BCbaJHXJ6oJQzeHz76vGKteg0Snx832a89vu9+OYnHsBOmx7//PoY9v31YfzXf3kbb14L4OiVOSzFU2V/J5U7/+EWcFiI5D7tH9/RhPb6ajw3PCZo1kTKvIsxXPWGbrl9cn+bEWadGoNnpXfl/FOnp2iz4DsbdPjzD2avpP/0/dsxH0ni977/c+z768NIpDMYcDQXYcWk3HDjkYv1GeIORGCr0A4WHIdNj+mFGOaW1n94qFCuuezubicPO8hbGmqgVsgoh1yCRpbbu+X+LOy3W5DOsHjpvLQ21ViWxaDTg/2d9dBrlRt6LpmMQW+XGV/9+P0Y/u99+M19bTh2ZQ4feuEkfu/7P4dBqyzreAUgTi9kyRXIchmDp3s7cH56EUev+sRejiSccPkBAPs6br4B5DIGB3ua8Pplr6Ra42XjFR48uq2xqIdrajVK/Oa+zTi0fCW9t8OEfZ0m7GoxFO01SPmwW/UIhBOYXijOQb1KHRKSi8shC7GL7PJmhwkVswcyRyGXYUdzrWC74aR4Trj8MFWr0N2oW/najuZatJm0kjuT83P3PKbmo+gv8ibOJlM1/mf/dpz67KP4q/9ix5bGGvzm3s1lf3erRq2AqVolaC9kSf6JfnCXFU21Gjw/PCb2UiRhZMwHfZUS25trb/n6wM5mxFMZHLoonQMKJ8ezs+D5alaeeyX93U/tpngFWZXdlr1wKsahslgyjdnFeMUe0OPsaK4FwxT38OOduOZC0FcpYVrjYFOh7FY9zk0v0FmXEsKyLEbGfNjTYbrlc59hGAw4mnHC5curNRrfhpweqOQyPLa9kZfn16oU+MgDrfjP396P33vvFl5eQ2psRq2g46YlWSCrFDJ86qHNODUewNs3gmIvR1Qsy+KEy4897aZ3tTK7t7UOjbVqSV05Dzk9qFbJ0dddXrPgSWnZ2qSDQsYUpZibmq/sFm8cnUaJ9vpqQVq9ZTtYVINh+LkAttsMiCTSRR97T/jjmgvBuxRfNUrQ77Agw0IyMYtMhsWLox4c6KqHvmpj8QpyU6tR2GEhkiyQAeAjD7TCoFXiudddYi9FVNf9EUzNR7Gv0/Su35PJGBy0WzB8ZQ5LMfHHbSbT5TsLnpQWjVKO7iZdUYo57pZepQ4JyeWwGTA6xX+rtzFvGJ08xCs4DpqoV3JGxt4dNeRsbdKh3Vwtmc2id9zzmF6Ilf3BOaG11FVhej5a8LTBfEm2QK5WK/CxPW147eIsrszmN12tnIy4sjnsvXcI4A84LEikMjh0UfxuFiNjPsxHkkXPXBFSCLu1OAf1JpenN1V6xALI5pBnF+PwFnEIy+0WIkn4QnFeWrxxOsw1qFLK6aBeCRkZ88FWV4VW07svVBmGwYDdglPjfkEOka5l0DkNlUKG927jJ15RqVqNWqQyLDxFOluyFskWyADw8b1t0KrkeGG4cneRT4z50VSrQXv96u2OdrXUwaLXSGLc5pDTA51agYe2lPdpWlIa7DY95iPJlQK3UO5gBCq5DI06TZFWVrqE2Hl1+ZYP6PFYIMtlDHY019IOcolIZ1icGvevunvM6Xc0Z2MW58TdRebiFb1dZug0FK8oJqFbvUm6QK6rVuEjD7TiJ2enBT25KBWZDIsTLh/2dprumMWTyRj02y04esWHhah4MQtuFjzFK4hUOKzZg3ob3SWcDERhrauiA6HIDmGR8XxQj88OFrnsNj3OTy8IdruWFO7c1AIWYynsXSVqyOlu0mFLQw0GRY5ZvHUjiNnFeFmPfRYLNyxEqHpQ0gUyAHzqoc2QMcBXjo2LvRTBXZxZRDCSvOtVM5A9oJBIZ/DaBfG6WYyM+bAYS2FgJ30oEGnoaqqBUs5seJcw2+KN4hVANvrW2VDD687r2FwIKrkMLTz/mTtsesSSmZWey0S6VqKG6/hZePpagNcI0FqGnB6oFTI8SvGKorPoNZDLGME6Wax/DrBILPoq/OIuK77/phsDO5tRxcPuZFejTpIjGk9whxLWaAB+T4sBVkMVhkY9eOpemxBLe5dBpwc6jQL7O6l7BZEGtUKOrU21Gz5U5g5E0MNT28JS1GPV4+gVH1iW5aXLhMsbRlu9Fgqe+7raV+4wzKO7SbfGo6VhIZrkbfdMo5Shs0Gafw4nxvzobtTBrFPf9XH9dgv+6bWr+Nm5GXxsb5swi8uRXo5XPNzdgBq15MurkqOQy9Bs0AjWyaIk/gaf7u3Av781iQ+9cJKX5//U/s34k4HtvDz3Roy4fGg3V6NJf/fsIzdu8+sjE1iIJDc8tSdf8VQar1yYKftZ8KT02G16/PTsdMHFXCieQjCSrPgWb7kcVj1++PYUZhfja342FWJ8LiRIwdpeX41qlRyjUwv40H0tvL9eMZwa9+Mz336Lt+f/9icfwENbpLXJEUum8ea1AH71wdY1H7ulUYfuRh2GnB5RCuQ3rwXgXYpT9woetdRpBcsgl0SB3G6uwY//6z7MLhb/dOrXjk/gP89O448PbntXn2ExJVIZnJ4I4Kn3rG9HuN9uwZeOjuPlCzP4ZYE/7I9d8WEpVv6z4EnpcVj1+Jc3buC6P4K2Oxx0vZubLd4oYsHhhrA4J+fRpG8q6nMnUhlcD0RwUIAde5mMwQ6rvqQ6WexqMeDLv3Ff0Z+XZVn8wb+dxY/fmZZcgfz2jSDiqcyaUUNOv8OCf3ztCmYWYrxcwN3NkNMDjVKGR7Y2CPq6laTVqMVrAg1HK4kCGcj23+RDNJnG737vHZy5FsCD7Xc+ACC0s5PziCTSq/Y/Xo3DpkeLsQpDTo/gBfLQqAf6KiX2l/kseFJ6uPHIo1MLGyuQaQd5BXdQb3RqAe/bUdwC+bo/jHSGRUdD/n9XhXBY9fj2qetIpjMlMaq3oVaDx7bzU/S97/wsXrkwg3iqB2qFdA5anxjzQy5j8GC7cV2P73dY8A+vXsGLox58Yv9mnld3UzrD4mfnPHh0ayOqKV7BmxajFr5QApFECloVv3/O0v9E4NmjWxugUcowNCqNBuOckTEfGAbYvc6inWEY9NubMTLmQzCc4Hl1N8WSabx6YRZP7GgqiR8wpLJw5wsKPVTmXm4RR4f0bqpSydHVWJwhLLdzLU+26zQLk4W12/SIpzK4OksT9QZ2WrAUS+H4VZ/YS7nFiMsHh02/7pZpHeYabLPUCv4z/Y0JP3yhBN1J5dlKqzcBDupVfEVTrVbgka0NeHF0BunMxgYKFNOJMT96mvUwaFXr/p4BhwWpDItXLgg3bvPolTmE4hSvINKkUsiwzVIL52RhB/XcgQi0KjmM1et/H1YCu1WP0cmND2G5HddRot0szA6yfeUOA//TAaVuX0d2LLJUptEBwFIsCefkwrrjFZwBhwVvXQ9iel6YbgdANl5RpZTj4W6KV/CJ624jRKu3ii+QAaDf3gxfKI7TEwGxlwIAiCRSeMcdvGvPx9XsaK7FJpNW0D6Qg04P6rRK7OmQTjyFkFx2ay3OTy0iU8AF8GQwgpY6LS/dGkqZw6aHP5zAdJEnWrm8IVj0GsFuUbeZqqFTK2hgCLIXk4/vaMSrF2YRS6bFXg4A4I3xANIZNu+fhVyG/UWBdpFT6QxeOjeDR7c1oEolnXhKOeJ6IQvRyYIKZAAPbzWjSimXxDQ6ADg9EUAyzeZ91ZyNWVhwwuVHQICYRSyZxmsXZ/FED8UriHQ5rAYsxVO45s+/3+1kMEoH9Faxku0ucGf+TlxzIV4n6N1OJmPQs7wbTrLT6JbiKRy9Mif2UgBk4xVqhQzvaa3L6/s211djR7NwMYtT4wH4wwkaDiIAY7UKWpVckE4WVNUA0KoUeGRbA146NyOJqUonXH6o5DLc37a+Qwm5+h0WpDMsXjrHf8xi+LIXkUQaA45m3l+LkELZCxyPzLIs3IEIbHRA7122WWqhkG18CEsulmXhmgujQ6B4Bcdu0+OiZwmJlPif/WLb22FCnVYpmTM5J8b8uL/NWNB01n6HBe/cmMekAIXU0Og0qlVy9FG8gncMw6DVqKWIhZDe77DAH07gDQnELI5f9eE9mwwF3arZbqlFe301hkb53w0fdHpgqlbhwc35F/KECGVLQw3UClne7byCkSTCifTKoRByk0aZPahXzBZps4txhOIpdPI8Yvp2dqseiXQGV2aXBH1dKVLKZXiipwmvSSBm4V2K4fLs0pqDsu5kwJ7duOE7ZpFcjle8d3tjQYU8yZ+tTkuH9ITU190ArUou+hz3QDiBC57FvOMVHG5oyEmXH75Q8ftGc6KJNA5d9OKJnibeJ14RshEKuQzbm2vz3u282eKNIharsVv1GJ0q3kE9roOFkBELIJunBvK/w1Cu+u3NCCfSGL4sbszipIubJFvY+ZZWkxYOm573Q4cnXX4EI0n007RNwbQatbgRiBT9kPDtqLJZplHK8d5tjXjpnEfUmAX3obB3Az2F+x0WZFjwGrM4fMmLaDJN3StISXBY9Tg/tZBXpxou40Y7yKuz2/SYjyQxGSzOTs5KgSzwDnKrUYtajaKkBobwaXe7EcZqlegxi5ExH2o1Cuxo1hf8HP12C85OLvB6O37QOY0atQIHuqQ1YKWctRirEE2m4ef5rBUVyDn6HRYEI0mcWC5SxTDi8qFGrcBOW+EfCt2NOnSYq3m9ch4anUZ9jRoPbqbuFUT67DYDwok0Jnzr73fL3cKjAnl1xd55HfOGUKNWoEGnLsrzrRfDMLDb9NTqbZliOWZx6OIsoglxYhYsy2JkzI89HaYNTbjlulnwVewnUhm8fH4Wj1G8QlBCdbKgAjlHb5cZNWqFqH0gT4z58OBm44ZiC9mYRTPemPDDu1TcNkwAEI6ncPiSFwftTZIaz03Indit+Rdz7mAEdVolamgq1qq6m3RQypmi7by65kLoaKgRpaWe3WrA5ZklxFPSaG8mtgG7BZFEGq9f9ory+jcCEUzNRwvOH3NajFrsbDHw1qFqxOXDQjRJ3SsEdnNYCBXIgtEo5XhseyNeOj+DpAgxi6n5KK75IxuKV3AGeIxZHL7kRSyZocwVKRkd5mpUKeV5FXPuQIR2j+9CrZCju0mHc0XaQXZ5he9gwXHY9EimWVyeoYN6APBguwn1NSrRNotGxpajhgWexck1YLfg3NQirvnyb/O4liGnBzqNAvu3bHydZP1a6qhAFkW/3YKFaBIjY8KP2+Res9BDCbm6GnXoaqzh5dDhkNODBp0a9xXQho4QMSjkMuxors2r3+1kMEojptdgtxrgnJzf8GGZUDyFmcWY4Af0ONwdBsohZ8llDJ7sseDQpVlEEinBX3/E5UNjrbooF0wHHfzELLLxihm8b3sT1AqKVwipSiVHfY2a904WVCDf5qGueuhEilmcGPOhvkaF7kZdUZ6v396MN68FMLtYvJhFKJ7C65e9OGi3ULyClBS7TY/z04vrOqiXybCYCkZXdirI6hw2PRZjqQ1nAV1ecTpYcGx1VTBolTQwJEe/w4JYMoPDl4SNWWQyLE66/NjXUV+UuA00sngAACAASURBVI3VUIX3tBqK/jP92NU5LMVSFK8QSauxijLIQlMr5HhsRyNePj8jaON4lmUx4vJjT5E+FACg39EElgV+VsQr50MXZxFPZah7BSk5dqse0WR6pVvC3cwuxZBIZ2CjiMVdFWvnlfs76WwQJ2LBMMxK2zqSdX+bEWadGoNnhd0sujSzhEA4UZSoIaff0YwLnkWMr+O9v15DTg/0VcoN56RJYVqMWt6n6VGBvIr3O5qxGEvh+JhwfSDHvCHMLcWxr6N4XSE6G3TY2qQr6q2lQacHTbUa3Jvn6E9CxMZ1XVhPMbfSwYIiFnfV1aiDSi7bcA7ZNReCQsZgk0mcAhnIFvtXZpdEH5AhFXIZg367Ba9f9iIUFy5mccJVvKgh56C9CUDxhobEkmm8emEWj+9ohEpBZZQYWo1aTM9HeT0vRn+zq9jXWY9ajULQoSE388fFvRrtt1vw5rUgPAsbz+osxZI4cnkOB+0WyCheQUrM5voaVKvkGJ1cu53XJPVAXheVQoZtlo1P1HN5w2g1aaEUceiQw6ZHKsPiomdRtDVITb/Dgngqg0MXZwV7zZExH9rrq2HRF+/i1KKvwn2b6or2M/3YVR+W4in0O5qL8nwkfy11WmRYYHqevxwyFcirUClkeHxHE149PytY258Rlx8txqqi/0DmohAvjm68m8VrF2eRSFO8gpQmuYzBjub13UbndpCtBtpBXovdpse5qQVk8hjCcjvXXEi0/DHHbjMAQNG6cpSDe1vr0FirFuxMTjKdwemJAPYWcfeYM+Cw4NLMEsa8G+9UMuSchkGrxN4i3vEl+bnZ6o0KZMH1OyxYiqdw9Ar/3SxS6QxOjfsLHi99N+3mGmy31GKoCH0gB8960KzXYFeLoQgrI0R43EG9taZluoMRNNaqqfn/OtiteizFU7he4IGZVDqDa/6w6AVys14DU7WKOlnkkMkYHLRbMHxlDkuxJO+vd9Y9j3AizcvPwiftFjAMMOTc2GYRF694YkeTqHc8Kl2rif9hIfS3ewf7Outh0CqLUliu5dz0IpZiqaIeSsjV77Dg7RvzmNrArYiFaBJHr1K8gpQ2h02PeCqDq967H9ZxByLUwWKd7NbsBbNzHdGV1dwIRJBMs+gUeMT07RiGQQ8d1HuXAYcFiVQGrwkQsxgZ84NhgD087Mw21mpwf5sRQ6Mb+5k+fHkO4USa7qSKrKlWA6Wc4fWgHhXId6CUy/DEjia8emGW90MbXP6Yr9s13ECPjXSzePXCLJJpFgM7KXNFStfKRL01dgkng1HKH6/TlsYaqBWyglukueayAxzEGhKSy2HLHtQTa8SyFO1qqUOzXiNIzGLE5cOO5loYtCpenn/AYcGV2RCuzBYesxga9cBYrcKedopXiEkuY2A18NvqjQrku+h3WBBOpHHkCr/dLE64fNjapEN9jZqX52+rr0aPtRY/3cAH3JBzGlZDFXYudwIgpBS1mapRo1bcdZcwmc7AsxClDhbrpJTLsL25Fs4Cd165Fm/tIkcsgOwFVIYFLtBBvRVczOLolexYZb5EEim8cyPIS7yC80RPExgGBR/WiybSOHRxFk/0NEFB8QrRtRi1mKQCWRx72k2o0yp5vXKOJdM4cy1YlJGad9Nvb8ZZ93xBoxkXIkkcu+rDgMNStB7NhIhBJmPQY717MTc9H0WGBfVAzoPdqsf5Ag/qjXlDMOvU0FcpeVhZfuw27g5DYXGRctXvsCCRzuC1C/zFLN68FkQyzfIWNQSABp0GD242Ysg5XdD0x+HLXkQSaQzYKV4hBS1GLe0gi0Uhl+GJHgteuzjL2y23t68HEU9litrzcTUDK90s8i/2Xz4/g1SGpcwVKQsOmwEXPYt37J95swcyFcjrZbfqEU6kMe4L5/292Q4W4scrgGyusb5GXfBueLm6p8UAq6EKgzyeyTkx5oNSzuD+Nn577A84muGaC+NyATGLQacH9TUqPLDZyMPKSL5a6rQIRpK8HSClAnkNAw4LIok0hi/zM25zxOWDXMbw/oZrMWqx06YvaGjI4KgHrUbtSn6TkFLWY9UjkcrcMYfIHfqwUcRi3RzLLdJGp/LbeWVZFi5vSPQDehyGYeCw6Wnk9G0YhkG/w4JjV31YiPBTjIy4fNjVWgetSsHL83Oe6GmCjEHed4YjiRQOXaJ4hZS08tzqjf6W1/DgZiPqa1QYLOI0ulwjY37stOmh0/B/e7HfYYFzcgE3/Ou/JREMJzAy5kM/xStImXCscVDPHYhALmNg0WuEXFZJ6zBXQ6OUYXQyv+yuL5TAYiwleou3XHarHq65EMICTo8rBQMOC1IZFi9f2HhP/dvNRxI4P73Ia/6YU1+jxp4OE4acnrxiFocveRFLZjBAw0Eko8WY3cTgq5MFFchryMYsmnD4oheRRHE/MBdjSTgn5wWb5X5wOTc1mEebm5fPzyCdYVc6YRBS6jaZtNBpFHe8je4ORtFs0NAuUR4UctnyEJb8dpDHltvtSa1ApoN672a36tFirOLlTM5Jlx8sW9zx0nfTb2/GuC+c19/xkNMDs06N+9soXiEVN3eQqUAWTb+9GdFkGq9fKm43izfGA8iw4P2AHsdWp8U9LYa8PuCGRj1oM2mxo7mWx5URIpy1bqNTD+TC2K16nJtaRDqPg3pcB4sOiUQsgJsH9WhgyK0YhkG/vRkjYz4Ew4miPveIy4dqlRw7BRpC9URPE+QyZt0/C8PxFA5f8uLg8vcRadBXKaFTK6hAFtMDm42or1EX/YDCyJgPGqUM79kk3GS6AYcF56cXMbGOwzT+UBwnXH6KV5Cy02PV49LM4qqj5CeDUSqQC+Cw6RFNpleK3vVwzYWgVclhqZVOnKWxVoPGWjWNnF7FSszifHFjFifG/Hhgs1GwyXTGahX2dpgwNLq+mMVrF2cRT2XQT/EKSWEYhtdOFlQgr4NcxuCgvQmHL3mLmksbGfPh/jYj1ArhxtlyMYv1dLN4aSVeQR8KpLw4rAYk0yyuzNxazEUTafhC8ZVsG1m/9Q5hyeWaC+P/Z+/Ow9u6zzvRfw92EARAEgRJEKAWUiK1EJC8S3ZkKnUyWaTE7iTTJU2nnW5Juk3Smd7emW7T3rYzXW4nTZOZprfpkqbtpGlaO7Gd2E0by7ZEerdISpZEAloAcAEIgNh3nPsHeCCI4s6z4Zz38zx5HkuBgJ+Wg/Oe3+9dBp0W2U3n9Lo7djwZUMmO9tuw19G2o2Lv9cwt5xFYyoqWasg563PhZiyHS3Obp1k8MzmPXpsR9+8VtsMG2b6BLjOCCSrSk9RZXz+KlRr+5Qo/3SwiqQJmIhnRvxT6O8y4b2/nlhqlPzM5j0GnBYddVhFWRoh4fNwx+qqc2dBKsQdN0du+QWc72gzabY1q9kcysso/5njddgSWsoK1j2pVDMPgrM+FC/4YYpkiL+/JTZIV+174b470QadhNr0XpgtlvHAtig96XbJ7kCP1PORgPLejvtaboQB5i+7f24keqxHP8JRmccEfAwBRqnZXO+N14Z351IZHodF0EROBGM56Kb2CKI+n04yONv1dx+i3W7xRgLxdWg2D0X77lndec6UKwst5WQbIPo8dLIst7S6qzRlvP6o1Fs9d4mdoyAV/DA6LASO94m7EdFoMeORAN56Z2nhoyL+8E0GpUmvMEiDyMtDVhmKlhmianwe2ZhQgbxE3bvO7V6PI8JBmcX52CXazHkckKH7j0iw2KlD49qUF1FhQzhVRJIZh4HXb7yrEagwJoRSLHfF67Lg8n0JlnSEszQLReh2EHAPk0ZV0EcpDvtthlxWD3RY8s41uSOthWRbnZ5dwcsghye7sGZ8LwXh+w4LMpyfn4bKbcM8ApVfIEXfaJ0QeMgXI23DW50KpsvtxmyzL4oI/hpODDkkqYvvsJjywr3PDAPnpi3M40NOO4V753bwI4YPXbcfVhTQK5duFesF4Dia9Bs52o4Qra11etx2Fcg2zWyjU406w5DIkpJnTasTjx/vRK6PiQbnghoaM+2O73rXzRzOIpIuip1dw3nekD3ots25OdTJfxouUXiFrXEG1EL2QKUDehnv3dKLPZtpS/u5GbsZyCC/nRev5uJazvn5cXUxjZo1pYpFUAa/eiOMMpVcQBfN57KjUWFxZuH0NBBM5eDrb6N/9Dm2nRZo/moWGqfellqM/+oF78KFjdIK2ljM+F2ps/aRxN87PSpdqCAD2Nj1OHXSuOzTkO5cXUarWcIbSK2SLm3h6K8Z/oR4FyNug0dSfnF+8FkVqF8Ub5/31ooSHJXpqBoAPjPaBYbDmk/O3phfAsqCcK6Jo3DF6c1FZMJ6nEdO7sN9hQbtRt6VOFv5IBgNdbTDpxeviQ/gx0mvFgZ72XdfknJ9dgqfTjD0SPiSd8boQXs7j7eDdufPPTM3D3WHGPSL1ZybbZ9Jr0Wsz0g6yHJzxuVCq7i7N4sJsDH02Ewa7LTyubHt6bCY8uK9rzTSLZybnMdJrxUGRiyYIEZO7w4wuiwFTTUVlwQQNCdkNjYbB0X7bljpZ+KPy7GBBNlcfGuLCK9fjiKQLO3qPao3FRCAm2e4x5z1HemHQau66FyZzZbw0E6U5AC1gj0C9kClA3qZ7Bjrg7tj5uM1ajcUF/xIePuCQ/KI763NhJpLB1aYj5oVkAa/djNORElG81YV6yVwZ6UKFCvR2ybdSqFfeoFCvWmMRWMpiyCndJgHZnTM+F1gW+Pb0ztIspsNJpAoVPCxhqiFQn8b26HA3npmaR61pCuTzlxdQrrI446V7odwNdLYhRAGy9BimPjTkxZkokrntp1m8s5BCIleW/KkZAN4/6oKGwR3HZM9OzYNlb3e6IETJfB47ZiIZFMrVxhEd7SDvjtfTgVKlhmtr1Ddwwok8SpWaLAv0yNYM91ox3NuOpy/ubLOokWoog3vhGZ8L88kC3gomGj/39OQ8PJ3mRs90Il8DXW2YTxXWnIy6GxQg78AZXz/KVRbPX97+k/MFrihBwvxjjtNqxIlBB55uGrf5zNQ8DvVZ6cZFVGHUbUe1xuLyfArBOA0J4YNvCy3SuA4WlGLR2s54+/HazTgWkttPs7gwG8NIrxVOq/QdY95zuBcGnaZRgJ/IlnB+donSK1rEQFcbWBaYW95Zus96KEDegWMeOzyd5h2N2zzvX8Kg04I+uzzaB53xuRCIZnFlIY255TzeuJmgym2iGtzu0FQoidDKuFLaQd6dvY42WE26DTtZzEYoQFYCLs3iW9PbuxcWylW8diMueXoFx2rS4/SwE8+upFk8f3kBlRqLD9EcgJawR6BeyBQg7wDXB/LlmSUs50pb/nWlSg2vXo/LIr2C8/6jfStpFvN4diXgp/QKohZ9NhO6242YDCURTORgNelgb9NLvayWxuV2b1So549m0GUxoNNiEHFlhG8HetpxqM+67ZqcN28lUKzUZHUvPONzYTFVxBu3Enh6ch57HW04KsEgL7J9XN1IkAJkeTjr7UelxuL5bYzbvBhaRq5UlbT/8WqOdiMeHqoXKDw9OY+j/Tbsl7C7BiFiqgdzNkyHkwjGqYMFX7weO67Mp1GqrF2o549mcIB2jxXhrM+F128mMJ/ceh/aC7MxaDUMHhrsEnBl2/PY4V4YdRr89fhNXPDHaA5AC+m1mmDQaihAlotRtw17utrwzW30gTw/uwSGAU4MyidABupPzteXsng7uEzdK4jqeD0dmImkcW0xQx0seOJzd6BUXb9Qzx/NYqiHHsSVgDtx3M4u8nn/EnweO6wm+ZzWtBt1ePdID75xcQ7VGkv3whai0TDwdJopxUIuuDSLC/4Y4tmtpVlcmI1htN+OjjZ5HSu+72hfY+T1WS/lXBF18bntqLFAeDlPO8g88brXn6gXz5YQz5Yo/1ghBp3tOOKybbkmJ10oYzKUlFV6BYcLivd3W3DERekVrWSgq433YSEUIO/CWZ8L1RqL57YwbjNXquCtYEI2RQnNuiwGPHaoBw/u65J0ohEhUvA2tXGiDhb8GOgyw27WYyp893Qy6mChPGePufDWrWWEthCgvBKIo1pjZXkvfOxwDzrb9PjofR5Kr2gxe7racCtGAbJsHHHV83W3crT06vU4ylVWlk/NAPDHH7sHX/7xB6VeBiGi67WZ0LPSaorGTPODYRj4PGsX6vmpg4XicMM0vjW1+WbRef8SjDoN7t3TKfSytq3NoMNLv/Q9+OTYkNRLIdv0M+8+gOc+82ijZS0fKEDeBW7c5gX/EmKZ4oavveCPwaDV4IF98ilKaGbUaWHSa6VeBiGS4Nq90Q4yf0bddlxdSKNQvrN5vz+agVGngZseRhRjr8MCr9uOp7eQZnFhNoYH9nXJ9n7TbtQ1Ug5J6+izm+Cym3nd+acAeZfO+FyoscC3Nhm3eX52Cffs6YDZIM8vBULU7P59XTDrtZSDzCOf245ylb1jlD1Q74G8v9tCQYjCnPG5cDG4vGEngWi6iKuLaVmmVxCyGgXIu3Soz4pB58ZpFolsCZfnU7KYnkcIuduPPbIfz336UXqA5RGX2z25Ks2i3sGC0iuUhkuz2KhY78LKeGm5phoS0owC5F1iGAZnff145XoMkfTaYw7HAzGwLGTV/5gQcptBp6ECVZ65O8zoshgw3dTJolCuIpjIUf6xAg10teHYQMeGm0UXZmOwmXQYddvXfQ0hckEBMg/OrqRZPLdOmsX52SVYDFr4PB0ir4wQQqTBMAxG3fY7dpBvxLJg2foENqI8Z70uTIWTuBnL3vX/sSyLl2eXcGLQQek1pCVQgMyD4V4rDva04+l1npzPr3wp6LX0x00IUQ+f245ri7cL9fyReuA05KQhIUr0AW8fgLXTLG7Fcwgv5/Gug5ReQVoDRWw8OeNz4dUbcSym7kyzCC/ncSOWw8OUf0wIURmvx45qjcU78ykA9QI9ABjsph1kJfJ0tuGePR14+uLdAfL52RgA4GHKPyYtggJknpz1ucCywLdWPTmfn10pSqD8Y0KIynAT9bh+yP5oBu4OMxVDKtgZrwuX51MIrAyE4Zz3L6HXZqTTA9IyKEDmyYEeKw71We86Wrowu4TudgNGeq0SrYwQQqThspvQ3W5ojJz2RzPUwULhuHHNzzbdC2s1FuP+GB4Z6qYJdaRlUIDMozNeF167kcBCsp5mwbIszvtjOElfCoQQFWIYBl63HVOhJGo1FoFoFgeog4Wiuexm3L+3846anCsLacSzJUo1JC2FAmQefXDVk/NsJINouohHhii9ghCiTl5PB2YiaQSWMsiXqxjqoSN2pTvjc+HKQrqRc97of0yphqSFUIDMoyFnOw67bHh6cg5Ac/4xPTUTQtTJ67ajxgLfWCncoh7IyveBURcYBo2eyOdnlzDYbYHLTuPFSeugAJlnZ30uvHlrGeHlPM77YxjoMmOgiwYQEELUybcyUe/Jt8IAKEBWgz67CQ/s7cIzU3MoV2t49XqcxkuTlkMBMs+4cZvfvDiHiUCMRmoSQlSt12ZCj9WIW/Ec7GY9utsNUi+JiODsMReuLWbwD2+EkC1V6V5IWg4FyDzb123BqNuGL57zI12oUFECIUT1uF3kIaeFCpZV4v2jfWAY4Pe+fQUMA5ykWhzSYihAFsAZbz8SuTIA4GH6UiCEqNyomwuQKb1CLXqsJjy0vwuJXBlH+23oaKOTA9JaKEAWAJdmcajPiu52o8SrIYQQaTV2kKkHsqqc8fUDAKVXkJakk3oBSrTH0Ybvv38A3pWbAiGEqNn9+7rw4P4unB5xSr0UIqKzXhe+8XYYjx93S70UQraNAmSB/O5HfVIvgRBCZMFm0uPvP3FS6mUQkXVaDPjaJx+WehmE7AilWBBCCCGEENKEAmRCCCGEEEKaUIBMCCGEEEJIEwqQCSGEEEIIaUIBMiGEEEIIIU0oQCaEEEIIIaQJBciEEEIIIYQ0oQCZEEIIIYSQJgzLslt/McNEAdwUbjmEEAB7WZbd9cgxul4JEc2ur1m6XgkRzZau120FyIQQQgghhCgdpVgQQgghhBDShAJkQgghhBBCmlCATAghhBBCSBMKkAkhhBBCCGlCATIhhBBCCCFNKEAmhBBCCCGkCQXIhBBCCCGENKEAmRBCCCGEkCYUIBNCCCGEENKEAmRCCCGEEEKaUIBMCCGEEEJIEwqQCSGEEEIIaUIBMiGEEEIIIU0oQCaEEEIIIaQJBcgtjGEYI8MwX2IY5ibDMGmGYd5iGOYDG7z+MwzDLDAMk2QY5s8ZhjGKuV5C1Gaza5RhmMcYhrnCMEyOYZjvMgyzd4P32rfymtzKr3mPOL8LQtRr5bp7lmGYxMr98/MMw+jWee3HVq71LMMwTzIM0yX2egl/KEBubToAQQBjAOwAfhXA3zMMs2/1CxmGeR+A/xvAYwD2ARgE8BsirZMQtVr3GmUYphvAP678XBeA1wF8dYP3+jsAbwFwAPhlAP/AMIxTwLUTQoD/BSACwAXgOOrX8k+vfhHDMEcBfBHADwPoBZBb+bWkRTEsy0q9BsIjhmEmAfwGy7JfX/XzfwvgBsuy/3Xlx48B+BuWZfskWCYhqsVdo6gHuj/KsuzDKz9vAbAE4B6WZa+s+jXDAKYAdLMsm175uZdQv4b/RMz1E6ImDMO8A+A/sSz77MqPfx+AjWXZT6x63e8A2Mey7MdWfjwE4B0ADu6aJa2FdpAVhGGYXgDDAC6t8X8fBXCx6ccXAfQyDOMQY22EkLuu0TuuSZZlswD8Kz+/2lEAgVU32ovrvJYQwp8/AvADDMO0MQzjBvABAN9e43Wrr2c/gBLq1ztpQRQgKwTDMHoAfwPgr1bvPq1oB5Bs+jH331ah10YIWfMaXX1NYuXHa12T23ktIYQ/51APflMAQqinQj25xuvoGlUYCpAVgGEYDYC/Rv1p9WfXeVkGgK3px9x/09EPIQJb5xpdfU1i5cdrXZPbeS0hhAcr1+1zqNcKWAB0A+gE8LtrvJyuUYWhALnFMQzDAPgS6kUBH2FZtrzOSy8BONb042MAFlmWjQm8REJUbYNr9I5rciUHeQhrp0hdAjDIMEzzbtSxdV5LCOFHF4ABAJ9nWba4cr/8CwAfXOO1q6/nQQBGANfEWCjhHwXIre9/AzgM4EMsy+Y3eN2XAfw4wzBHGIbpBPArAP5ShPURonbrXaP/BGCUYZiPMAxjAvBrACbXSpFiWfYagLcB/DrDMCaGYb4XgA/A11e/lhDCD5ZllwBcB/AphmF0DMN0APgR3FnPw/kbAB9iGObUysPubwL4RyrQa10UILewlZ6pn0C99cwCwzCZlf/9EMMwe1b+ew8AsCz7bQC/B+C7AG6u/O/XpVo7IWqw0TXKsmwUwEcA/DaABICHAPxA06/9E4ZhmjtU/ACA+1de+z8AfHTlPQghwvm3AN4PIApgFkAFwGcAYOVaPgUALMteAvBJ1APlCOq5x3e1gyOtg9q8EUIIIYQQ0oR2kAkhhBBCCGlCATIhhBBCCCFNKEAmhBBCCCGkCQXIhBBCCCGENNFt58Xd3d3svn37BFoKIQQA3njjjSWWZZ27fR+6XgkRBx/XLF2vhIhjq9frtgLkffv24fXXX9/5qgghm2IY5iYf70PXKyHi4OOapeuVEHFs9XqlFAtCCCGEEEKaUIBMCCGEEEJIEwqQCSGEEEIIaUIBMiGEEEIIIU0oQCaEEEIIIaQJBciEEEIIIYQ02Vabt80kc2X8wfNX+XzLBq2GwY8+vA/7ui2CvL/aTQRiyJUq+J5DvVIvhRBCJPfrT03j0WEnHjtM34lyw7Is/vLCDXxg1IU+u0nq5RCF4jVAzpereGZqns+3bFjOlVCp1fBbT3gFeX+1+8PnryGwlMWr/7UHGg0j9XIIIUQyi6kC/mr8JtydZgqQZSi8nMdvfPMywok8fuXsEamXQxSK1wC5z27Cm7/6Xj7fsuEn/up1vHA1CpZlwTAUwPHtVjyHpUwRl+dTGHXbpV4OIYRIZiIQAwCcGHRIvBKyllvxHADghWtR/IrEayHK1TI5yKdHnAgl8ggsZaVeiuIUK1UspgsAgHPXohKvhhBCpDURiMFq0uFoP20WyFEongcAzEYyCCVyEq+GKFXLBMhjw/Wx2S9cpQCOb+FEHixb/+8XrkakXQwhhEhs3B/DQ/u7oKV0M1kKNgXFtKlDhNIyAfJAVxuGnBa6GAQQTNSfxh/c34U3by0jmS9LvCJCCJHGfDKPG7EcpVfIWDCeg7vDDHeHGedo04wIpGUCZAAYG+7BRCCGfKkq9VIUJbiSz/XxE3tRrbG4MLsk8YoIIUQalH8sf8FEHp5OM8ZGnLjgj6FUqUm9JKJALRUgnx5xolSpYeJ6TOqlKEookYdey+D9R/tgNekojYUQolrj/hjsZj2OuGxSL4WsI5TIYaCrDWPDTmSKFbxxMyH1kogCtVSA/OD+Lpj0GjpS4VkwUT+uMug0OHWwG+eu1buFEEKI2owHYnhwfxe1u5SpQrmKxVQRnk4zHjnQDZ2GodRLIoiWCpBNei1ODDroYuBZKF5/GgfqxZALqQKuLqYlXhUhhIgrlMghGM/jJKVXyFZ4uV4zM9DZhnajDvfv66TiciKIlgqQAeD0sBPXl7K4GaN2b3yp53NxAXIPANAuPSFEdSYCcQDAySEKkOWKq5nhNnVOj/TgykIai6mClMsiCtRyAfLYyEoAR7vIvMgWK4hnSxjoMgOoD3s51GelP19CiOqM+2PobNNjpNcq9VLIOriuS9w9i2sBS/cswreWC5D3d1uw19FGO5w84fpJDqzsIAP1L5zXbsSRKVakWhYhhIhuIhDDQ/sdlH8sY6F4DgatBr1WEwDgUJ8VvTYjxQSEdy0XIAP1AO6CP4Zihdq97VYwzj2NNwXII06UqyzG/dQthBCiDsF4DuHlPKVXyFwwkYO709x4iGEYBmPDTrw0E0WlSu3eCH9aNkDOl6t47Tq1dtmtRj5Xp7nxc/fvt2ZP5gAAIABJREFU7YLFoKXCB0KIanAbAtT/WN6C8XoP5GZjwz1IFSp4O7gs0aqIErVkgHxyyAGDVoNz1yiA261gIoc2gxZdFkPj5ww6DR4+QO3eCCHqMRGIwWExYLi3XeqlkA0EE7k7TjwB4F0Hu6Gldm+EZy0ZILcZdHhwfxcNtOBBMJ7HQGcbGObOnLuxYSdCiTz8UeoWQghRNpZlMR6I4cSg467vQiIf6UIZy7nyHTUzAGA363HPQAfFBIRXLRkgA/WpejORTKMnItmZ+kQi810/T5XBhBC1uBXPYT5ZwInBLqmXQjZwu2Zm7XvWVDiJpUxR7GURhWrZAJkL4F6kAG7HWJZFqKkHcrOBrjYMOS0UIBNCFI/LP6YCPXkLrdF1iXN6pQXsSzN0zyL8aNkA+UBPO9wdZiok24XlXBmZYuWuggfO2HAPJgIx5EvULYQQolzjgRi6240YclL+sZxxPZDXumcd7behu91AaRaENy0bIDMMg0eHnTg/G0OZWrvsSKMHctfdT+NAPY2lVKlh4jq1eyOEKBPLspgIxHBisIvyj2UuGL+7qJyj0TB49KATL16Lolqj4nKyey0bIAP1NItMsYI3blK7t51o5HOtcVwFAA/u74JJr6EG7IQQxbq+lMViqkjpFS0glMitWVTOGRtxIpErYyqcFHllRIlaOkB+5IADOmrtsmO3d5DXTrEw6bU4MeigP19CiGKNB6j/casIxvPr3q8A4NRBJxgGtKlDeNHSAbLVpMd9ezvpYtihYDyHjjY9rCb9uq85PezE9aUsbsao3RshRHkmAnH0WI0Y7LZIvRSyAZZlEUzk1iwq53RZDPB5OmhGAuFFSwfIQL1y9fJ8CpFUQeqltJxgIr9uegVnbKUymHaRCSFKw7Isxv0xnByi/sdyF8+WkCtV162Z4YwNO/F2cBnLuZJIKyNK1fIBMvXr3blQfO0eyM32d1uw19FGu/SEEMXxRzNYyhQpvaIFcB0sBtbpusQ5PeJEjQVemlkSY1lEwVo+QD7ssqLHasQLFCBvS61W74G82Q4yUH8IueCPoVCmdm+EEOUYD8QBACcpQJa9YHzjrkucY54O2M16avdGdq3lA2SGYTA27MTLM0uoULu3LYukiyhVa/Bs8mUD1APkfLmK129QtxBCiHJM+GNw2U3Y69j8e5BIa7O2pBythsGpg904dy2KGrV7I7vQ8gEyUG/tksyXcTG0LPVSWkbjy2aT4yqgPl3KoNVQ4QMhRDFu9z+m/ONWEIzn0dmmR7tRt+lrT4/0YClTxDsLKRFWRpRKEQHyqQNOaKi1y7aEtvg0DgBtBh0e3N9FR1aEEMWYiWQQy5YovaJFhBK5Ld2vAODR4W4AoHsW2RVFBMj2Nj3u2dNJhXrbwA0JcXdsvoMM1AsfZiIZhJfzQi6LEEJEMe6v9z+mASGtYas1MwDQYzXhaL+NYgKyK4oIkIF6v97JcBKxTFHqpbSEYDyHHqsRJr12S69vdAuhJ3JCiAKM+2Nwd5jh2UKaGZFWrcYinMhv6+9qbNiJN24mkCqUBVwZUTLFBMhjI06w1Nply4LbOK4CgAM97XB3mCkPmRDS8mo1Fq9cp/zjVrGYLmy5qJxzeqQH1RqLC7MUE5CdUUyAPNpvh8NiwAtXKYDbimA8v6UCPQ7DMHh02InzszGUqVsIIaSFXV1MI5ErU3pFi+BSArdzz7pnTwesRh2lWZAdU0yArNHUA7gXZ5aotcsmytUa5pP5be0gA/Ujq0yxgjduUrs3Qkjr4vKPTwx2SbwSshVb7YHcTK/V4JED3XjhahQsSzEB2T7FBMhAPYCLZ0uYnktKvRRZm18uoMZiywUPnEcOOKDTMPRETghpaROBGAa6zPBs8zuQSINrS7rVonLO6REn5pMFzEQyQiyLKJyiAuRTB7vBMNTaZTPcl41nkzHTq1lNety3t5P+fAkhLauefxyn9m4tJBjPo9e29aJyzthIvbicUi/JTigqQHa0G+Fz22mHcxON46od7J6cHunBO/MpLKYKfC+LKEitxlKuOpGldxZSSObLOEEBcssIJnI7ul+57GaM9FopJiA7oqgAGainWbx1K4HlXEnqpchWMJGDVsPAZTdt+9dy7d5epC8csoHfefYdfPjz5yn3j8gO9T9uPaH49rouNRsbceK16wlkixWeV0WUTnkB8kgPaizwMrV2WVcwnkd/hwk67fb/+g+7rOixGvECBchkA6/dTOCd+RTeDtL4dyIvE4EY9jna4LJT/+NWUKrUMJ8qbKuDRbOxYSdK1VrjwYiQrVJcgHx8oAN2s54GWmwgtMPjKqDe7m1s2ImXZ5ZQoSN0sgaWZeFfKYp56u05iVdDyG3VlfxjSq9oHfPJPFgW2+qB3Oz+fZ1oM2gpzYJsm+ICZK2GwamD3Th3jVq7rCe4jZGdaxkbcSKZL+NiiHYHyd0WU0VkihUYtBo8PTlHD1JENi7PpZAuVCi9ooXc7oG8s3uWUafFw0MOvHAtQjEB2RbFBchA/Uglki7infm01EuRnUK5imi6uKvxqqcOOKFhaOw0WdtMpH7d/eCDA1jKlHCejjaJTIwH6ql3tIPcOhpdl3Zxzxob6UEwnsf1pSxfyyIqoNgAGQBeoLHIdwkltt9wfTV7mx737OmkIyuyptmV9IqfODUIq0mHp94OS7wiQuomAnEMdlvQa9t+gTKRRjC+86JyztjBekxA9yyyHYoMkHtsJhxx2WiHcw2N46pt9kBebWzYiclwErFMkY9lEQWZjWRgM+ng6TTjg6MuPDe9gHypKvWyiMpVqjW8ej2OE5Re0VKCiZ0XlXP2ONow2G2hHv5kWxQZIAP1PNk3biaQLpSlXoqscMdVu8lBBuoTilgWeGmGuoWQO81GMjjQ0w6GYfD48X5kS1X8y5VFqZdFVG56LoVMsULpFS0mGN95UXmzsREnJgIxFMr0sE62RrEB8ulhJyo1FudnKf+xWTCeg1GngdNq3NX7jPbb4bAYaEIRuYs/Wg+QAeChQQd6bUY8+RZ1syDSmgjU7wUnBrskXgnZjt10XWo2NuxEsVJr/DsgZDOKDZDv3dsJq1FHOUerBON5eDrNYBhmV++j0TB4dNiJF2eWUKtRZfBmzl2L4uN/9grmlvNSL0VQy7kSljKlRoCs1TD4kK8f565FaHgPkdS4P4YDPe3osVL+cavIlSpYypR2nRII1AszjToNxQRb9PevB/HV125JvYwt+/OXr+MTf/06r/GIYgNkvVaDRw5049xVau3SLJjY+USi1caGnYhnS5ieS/Lyfko2HU7i5dkl2Mx6qZciKK5AjwuQAeCJe9woV1k8O7Ug1bKIypWrNbx2I067xy0mlOBqZnZ/zzLptTgx6KAAeYv+9MUAPv/dWamXsWUvzy7h+lIWGs3uNv+aKTZABuo5R3PJQuOmTfjL5wKAUwe7wTCgwoct8EcycNlNaDfqpF6KoBoBstPa+Lmj/TYMOS3UzYJIZiqcRK5UxcnBbqmXQrYhGOdavPG3qROIZhvvS9ZWrtZwM5ZFMJ5viZM/lmUxFU7C6+7g9X2VHSBz7d4ogAMAJPNlpAoVXo6rAMDRboTPbacn8i3wRzMYcrZv/sIWNxPJwKTXwN3Us7RerOfGK9fjik8xIfLEjRl+iHaQWwoXyPJ1zzo9wrWApXvWRoLxHMrV+sn7VFj+J8SLqSKi6SJ8Hjuv76voALm/w4zh3nYK4FaEeOpg0Wxs2Im3biVa4ilTKizLwh/NYshpkXopgpuNZDDY3Q7tqmOux4/3AwC+eZGK9Yj4JgIxDPe2o7t9d8XJRFyhRB4mvQZOnv7e9ndbMNBlxjkqLt9Q86n7ZEj+AfLkylTfUTcFyNsyNuzEq9fjyJUqUi9Fcrd7IPMYII/0oMbW83/I2rjRy0M9yt9B5lq8rbbXYcHxgQ48+TYFyERcpUoNr99I4CS1d2s5wUQOns62XReVcxiGwenhHlzwx1CsULu39fij9YmDTqsRUy0QIE+Fk9BqGBxx2Xh9X8UHyKdHelCq1hpHbGomxA7yMY8ddrOehrJswB+tP40rPcUiV6ogvJxfM0AGgCeO9+Od+RSuLdIIeCKeydAy8uUqTtKAkJYTjOcxsIsR02sZG3YiV6rijRsJXt9XSfzRDHptRjy0v6slUiymwkkc7GmH2aDl9X0VHyDfv68TbQYtpVmgnldkNepgM/NXKKbTavCug904dy1K3ULWwQXI6wWOShFY2XVY7/d5xtcPrYahYj0iKq7v7YP7KUBuNdwOMp9ODjlg0GooD3kDXM2Mz2NHeDkv64m5LMtiKpSEl+f0CkAFAbJRp8XDQw68cJUCuGAiD08Xf8dVnNPDTkTSRbwzTzuDa/FHMmg36tCzy+EscrdWi7dmTqsRjxzoxlNvz6n+WiTiGQ/EcKjPii6LQeqlkG1I5spI81hUzrEYdXhgfyedeq6DZVnMRuoBMpfTK+dd5LlkAbFsifcCPUAFATJQP1K5Fc/hRkzdrV3qLd74/bIBmrqFXKPCh7XMRjMYclp4fzCRm9lIBloNg32O9YsRHz/Wj1Aijzdv0fEmEV6xUq3nH1N6RcsJCpASyBkbduLqYpq66qwhmikiXahgyGlpBMjTMg6Qp1YK9Lweflu8AaoJkHsAQNWVqyzLIpTI81qgx+mxmXDEZaMn8nX4I1nF5x8DwEwkjb2ONhh063+tvG+0D0adBk9RsR4RwcVgEsVKDSeoQK/l3G7xxv896/RIPSZ4kdIs7uKP1FPlhnraYTPpMdhtkXUni6lwEjoNg0N91s1fvE2qCJD3ONow2G1Rdc7RUqaEfLkqyA4yUB/K8sbNBNKFsiDv36oyxQoWUgX1dLDY5EGg3ajDe4704unJeZSrNZFWRtRq3B8DwwAnKP+45Qi5g3ywpx0uu4lqk9awumZm1G2XdYrFZCiJ4V4rTHp+C/QAlQTIAPDosBMTgRgKZXW2dml82QjwNA7Uj6wqNRbnZ6lbSLOASjpY1Ccv5bZUiPjEcTfi2RK1BiSCGw8s4YjLBnubske8K1EwnofVpBPk745hGIwNO/HyzBI9qK/ij2bQZtCiz2YCAPg8dswnC4im5Veox03QEyL/GFBRgHx6xIlCuYZXrselXookhDyuAoD79nai3aijJ/JVbj+NK3tIyM1YFpUau6UAeWzYCbtZj6feom4WRDiFchVv3lqm9IoWFUzkBNk95pwecSJdrOCtW8uCfUYr4gr0uJoZr4zzkEOJPJZzZXgpQN6dE4MOGHUa1ebJhhL1YgSPQCkWeq0Gjxxw4NzVCHUoaMIVru3pUnaAvFkHi2YGnQYf9Lrw/OVFGuBDBPPWrWWUKjUaENKigvEc7x0smj18oBs6DYMXVFybtJbAqqmvR912MIw8J+pxqR8+N/8FeoCKAmSTXouHBh04J1KnhUq1hoqMjm5CiRy62w1oM/DXA3m10yM9mEsW7hhTqXb+SHbTwjUl4P7Ot5pK8vjxfuRKVfzz5UUhl0VUbDwQg4YBHtjfJfVSWoKcJss1isoF3EG2mfS4d28nnXo2WWvYU7tRh8FuC6bC8ttpnwwlodcyGO4TJoVR2XftVU4PO+GPZhvpBkL51tQ87v/t7+B/fOuKoJ+zHcF4nveG66s9utLujb5wbuMarivdbCSDfrsJFuPWHsAe3NcFl92Eb1A3CyKQiUAMR/vrkz7J+grlKn7tqWkc+bXnMBmSRxAUzRRRrNQESwnkjA07cWkuhUi6IOjntApu2NPqe5bP0yHLQr2p8DIO9dlg1PFfoAeoLEAeGxE2gMsUK/jFr13Ep/7mTSTzZVkVIQUTOcG/bNwdZhzsaccLKk1jWa1SreFGTC0t3jI40Lv1NjsaDYMPH+vHuWtRxLMlAVdG1KhQruLtW8vU/3gTl+aSOPvHL+PL4zdRrbG44JdHkXUwXk8JFDLFArjdw//Fa/K5V0uJq5lZ3XXJ67ZjMVXEYko+DxKNCXoC5R8DKguQB7stGOgyCxLAvXkrgTOfewlffzOEn333AXzi0SFcW0wjX5L+2KpaYzG3nBcs/7jZ6REnXr0ep9xSALfiOZSr7B35XEpUq7HwRzdv8bba48fdqNRYPDs1L9DKiFq9cTOBUrWGE4OUXrGWWo3FF8/58cQXziNdKOMrP/4QPJ1m2ewShla6Lgl96nm03wan1Uinniv8kQw0DLDXceefOxeETskoD/lWPIdUoSLIiGmOqgJkrrXLBf8SShV+8oMr1Ro++51r+Hd/Mo5KlcVXP3ES//l9I7h3TwdqLHB5PsXL5+zGQqqAcpUVNJ+LMzbcg1K1hnGZ7ERIyR+93XBdycLLeRTKtS0V6DU77LLiYE87nnqbulkQfk0EYtBqGDywjwLk1eaW8/ihP3sF//1bV/DYoV58+z8+incd7IbPY5dNAMSlQQq9qcMwDB496MRLM1FUa1Rc7o9msaer7a6UhSMuGzQMMCmTByjgdtEgBcg8Oj3cg1ypitdv7L7d281YFv/ui+P47Hdm8OFj/fjWp081vpB9K2MPp2SQ03W7xZvwO8gP7O+EWa+lJ3I0HVcpPMViNrr1DhbNGIbBE/e48dqNRGPHiBA+jPtjGHXbYTVR/nGzb16cw/s/+yIuhpbxex/x4X9//F50WgwAAK+7A7fiOSznpE95CsbzgheVc06POLGcK+OiDO7VUvNHM2t+j1uMOhzoaZdVq7epcBIGnQbD20jt2y7VBcgnhxzQa5ldBXAsy+JrrwfxwT96CbORDD73g/fgf37/cdiavox7bUY4rUZZPHE1AmQRdpCNOi0eHnLghatR1bd780cycFqNii8S8m+jxdtqHz7WDwD4xkUq1iP8yJUquBhapvSKJulCGb/w1bfxc3/3Fgad7Xj250/h+x4YaPS6BdAYtiCHNItgIid4egXnXQe6oWGg+tqZao1FYGn9mhmvuwOToaRs7utToSQOu2yCdohSXYBsMerwwL6uHQfIy7kSfuZv38Qv/sMkRt12fPvTjzZu8s0YhoHXbZfFE1cwkQfDAP0dwu8gA/ViyFvxHG7E1L0rWO9goez8Y6DewaLLYkDXyk7Udgx0teG+vZ3UzYLw5o2bCZSrLPU/XvH6jTg++LmX8OTbYfz8YwfxtU+exL7uu7+XRvvlFSALXVTO6bQYcGygQ/WnnqFEDqVKbYMA2YalTBELMijUq9VYTIeT8Lptgn6O6gJkoH6kcmUhjflkflu/7vzsEt7/2Zfwz5cX8UvvP4S//ckTcG8QdHrddsxGMsgWpS1YC8VzcNlMovXiPT3cAwA4p+IG7CzLNiYSKd1sZPsFes0eP96PKwtpXFmQPl+ftL5xP+UfA/Xx73/4/FV83xfHAQBf++RJ/MJ7h6HXrn0fsLfpsdfRJnkecqVaw9xyAQMiFJVzTg/3YDK0rOqOOrc7WKy9qeNtpI1K/wB1I5ZFulgRbEAIR5UB8thKAPfiFp8Yi5UqfvuZy/ihP3sFbUYt/umnH8GnTg9Bq2E2/HU+j10WhXrBRA4ekZ7GAWCPow37uy14QcVP5EuZElKFyo7SDloJy7IrLd52/vs843VBq2HwFO0iEx5MBGLweexb7smtRDeWsvjon4zjc/86i++9x4Nnf/4U7tu7+QOD122XfGLafLKAao0VbQcZqJ96sizw0ox671n+yNo9kDlHXDZoNYwsThi4NQjZ4g1QaYA83NsOl920pZyja4tpPPGFC/j/XrqOj5/Yg2d+7hRGt1g1yVVXSv2FE4wLO5FoLWPDTkwEYiiUpW9zJwW1FOgtZUpI5su72kF2tBtx6mA3vvH2HGpUSU52IVusYDKUVG16Bcuy+Oprt/DBz72EG0tZfOFj9+L//b5jWy5W9LrtCC/nJd1JDSbEq5nh+Nx2dFkMOKfiPGR/NIPudgM62tZOlTMbtDjY0y55PAPUd7GNOg0OCrwBpcoAmWv39vLMEsrrjINmWRZ/ef46PvTHLyOSKuBLP3I/fusJL8yGrU9s6bGZ0GszSpqHXKxUsZguiNLBotnYiBOFcg2vXN99t5BWtF7DdaWZ3UWBXrMnjrsRXs7jjVsJPpZFVOq1G3FUaqwqB4QksiV88itv4Je+PoXjAx349qdP4YzPta338MqgUC8k0pCQZhoNg1MHu3HuWlS1D+mzkQwGN9no8LrtmApLX6g3GU7iSL8NunXShfiiygAZqO9wposVvB28u7VLJF3Aj/7Fa/hv37yMRw5049uffhSPHe7d0efUKz+lax8zt1wAy4r7NA4AJwcdMOo0qn0in41kYNZr4bKZpF6KoHba4m219x7phVmvxZNvUU9ksnMTgTj0Wgb37e2Ueimiemkmivd99kV890oUv/zBw/jKjz8El337ASZ3Oiple9JQIgeNiEXlnNMjTsSyJVyaU2ctRL2ofOPvcZ/Hjni2hLmkdIV6tRqLS+EkfAL2P+aoNkB+5GA3tBoGL6wqJPvny4t4/2dfwivXY/h/nhjFl37kfjitxh1/js9jR2Api4xEhXq3eyCLGyCb9Fo8NOjAuWvqLNTzR7MY6rFAs0meeqvzRzKwGLRw2Xf3IGAx6vDeI714ZmqetyE+SracK+H3n7vSMilMXMHYREDYAULjgRiOeTpE6Z8rB4VyFb/5zcv44S+9CrtZjyd/5hH85KODO/7esZn0GOy2SHqMHkzk4bKb1y0mFMqpg/Wx02q8Z8WzJSRy5U27LnllMN8hsJRFtlRtrEVIqg2QbSY97tvT2WjtkitV8F/+cQo/+eXX4bKb8PTPvQs/fGLvHX0id8LrtoNlgUsSHVkFE+JMJFrL2LAT/mi2EaSriV9FHSyGetp3fZ0A9W4Wy7myqgtltuqZqXl84bv+lmlNNe6P4XP/Oosf+NMJfPKv38DNWJb3z0gXypgOJ1WTXnFlIYUnvnAef37+On704X345s+9C0f6d9/2alTi9qTBeA5uCe5X3e1GeN12VfZD9m/xJPBQnxU6DSPpA9RUuB6cCzlBj6PaABmo58lOh1P41yuLOPu5l/F/XruFT44N4Z9++hEc6OFnOkvjyEqqADmeh17LoFeCo/7TI9wTubq+cPKlKsLLedUEyLsp0Gv26LATnW166maxBVwAI/SOLF8mAjHoNAz+42MH8eJMFO/5w3P47WcuI5kv8/YZr99IoFpjcULhBXq1GosvvXwdH/78eSxlSviL//AA/tuHj8Kk33p9zEZ8HjvmkgVE00Ve3m+7gomc6CmBnNMjTrx5K4Fkjr9/l62AqyXZ7J5l0msx3GuVNEd9MpSEWa8VZcaAOs6h1jE27MTvP3cVP/aX9V3jv/2JE7zvPjitRvTbTZI9cQUTObg7zJu2pBPCYLcFnk4zvjJxE4sCNBc/2Gtdc0iL1AJL6uhgkSqUsZAq7KrFWzO9VoMzPhe+/kYY2WJF1W26NjMdrudJjvtbI0AeX2m99pn3DuOHHtqDP3j+Kv7s5ev4+pthfOY9B/GDD+7ZdcHNeCAGg1aj6PzjxVQB//lrF/HSzBLec7gXv/sRLxztO08BXAu3MzcdTuLdh3p4fe/NFMpVLKaKoheVc8aGnfjjf53Fr39jmve0RAbAh47146CAo5F3yh/JwKjTbDjXgePz2PHtSwtgWZaXk8Ptmg4ncVSEAj1A5QHy0X4b7t/bCXenGb/54VHY24QZCTy6UvkphVBcvIlEqzEMg4/e58Hn/mUG1xbTvL53jQW0GgbvPdy7rc4iYmg8ja/TcF0pGiOmeXwQePy4G1+ZuIV/vryIJ+5x8/a+SlKq1HB1IY02gxZXFtKIZ0s7mmIolsxK67VPjg0CqHf3+b2PHsO/P7kPv/XMZfzqU5fw5fGb+OUzh3F6ZOcB2bg/huN7OnjbSZWj6XASr99I4Le/dxQfe3CPIAHKUbcdDFM/9RQ7QA4vr3SwkGgH+fhABw70tOMbF/k/xaqx9fzZz3/sXt7fe7f80XoHi63krns9dvyf14IIJfKixxbVGovpcArf/8CAKJ+n6gCZYRj8w6ceFvxzfB47nr+8iFShDNsW+1HyJZjI4339wufqrOfT7xnGp98zzPv7Pn9pAT/112/g8nxKdjtG/mgWGgbY51B2gMxXi7dm9+3phLvDjCffDlOAvI6ZSBqlag0fe2gf/vLCDbx6PYb3j26vnZeYXr8RXzP1YdRtx9/95Ak8f3kR//3Zd/Cjf/Eaxoad+OUzhzG8zV22ZL6MS3NJ/Oz3HORz6bLz2OFevPRL70Y3z7vGzdqNOskK9aQqKufotBp85xfGBHnvT33lDVkM2ViLP5qFb4tDN5rnO4j99+SPZpAvV7e81t1SdQ6yWLhqS7ELH7LFCuLZkmTHVULySfRnuhX+aAYDXW2K3skC6i3eDFoN9vD4JanRMPjw8X68NLOEWEaaHEi54/7Nf+yhPTDrtbJPsxgPxKDXMrh/jUluDMPgfUf78PxnxvArZw7jrVsJfOCPXsKvPDm1rb//167HUWOhigEhQgbHHJ+no1EMJaZgQvweyGLxeuy4GcvJLr+5UK4imMhteaNjpM8KvVaaiXrcQxsFyAribfSWFPcflBQTicTSazOiu90oi6k+q6mlg4U/ksG+7jbec8GeOO5Gtcbimal5Xt9XKabDKbQbdTjgbMf9+zoxLvNCvQl/DMcHOjZMhTLoNPiJU4N44RffjY8/tAd/92oQp//gBfzpi34UK5u3shsPxGDQaXDPHuFbP6mB123HYqqIiAC1IxsJxXMwaDXotSqvf3wjt3tOXves60tZsOzWa2aMOi0O9dkkeYCaDidhMWixv1uc+ysFyCLoshjg7jBjUuQnrmBjIpHyAmSGYeDz2CW5SDdSrbEILGVFqbCV2mwkw2t6BWekz4pDfVbqZrGO6bl6kYpGw+DEoAPXFjNYkulue7pQxlQ4ueXOEl0WA37j8VE89+lTuH9vJ37n2St47x++iG9NzW84vWsiEMO9Cs8/FpNJq8/FAAAgAElEQVRUE/WCiXqLNyX2j29OTZCTxtTXbWzqjLrtmAyJP1FvMrSMo/120ZoOUIAsEp9H/N6SjXwuCXpKisHrtmM2kkGuJM0QlrWEE3mUKjVBAkc5KZSruBXP8Vqg1+zx4268cTOhyh7aG6lUa3hnPtVoH8l13XklIM+R7q/d2Fnqw4EeK/7iPzyIL//YgzDpNfjU37yJ7//TiTVP4ZZzJVyeT+HkYDdfy1a9Iy4bNIz4wVwwnpekZ78YOtoM2NPVJrtNHX8kC4YBBrexqePz2JEuVHAzJt73c6Vaw6W5VOPhTQwUIItEivyjUCKPNoNW1hXuu+F121FjgcsyGg26k6fxVnR9KYsaCxwQqGXRh47Vi86EqCZvZf5oFoVyDaPu+kAIr9sOi0Er237I4/5667V7d1hI++iwE8/+/Cn81hOj8Ecy+PAXXsZ/+vuLd7SNfOV6HCwL1QwIEYPFqMOBnnbRd5BDCem6LonBK2FHq/X4oxl4Os3bOn3xSjDfYSaSQbFSEy3/GKAAWTQ+98qIRhH/QXEN16XoVSgG7klSTkdWagmQZwVo8dbM09mGB/d14cm3wqIf48kZdwo1utKZRq/V4P59XbLNQ54IxHfdek2n1eDjJ/biu794Gj/16CC+eXEOp3//BfzRd2aQL1UxEYjBqNPg2IB03XqUyOvuEPUYPVOsIJErK7JmhuP12BGM55HIlqReSsPsDmpmhnutMOg0osYz3GeJMUGPQwGySLgdH1ED5HhOkdXAnF6bCb02o6yeyGcjGXRZDOhU6K49ZzaS2fax3HZ9+Hg/ZiIZvDPPbw/tVjY9V58iNdh0Qzs55MBsJINIWtyCqs1wrdf46ixhM+nxXz5wGN/5hTG8+5AT//M71/DuP3gBz07N4/59nTDqKP+YT163DUuZIhZT4uS3327xptx7lk/iybqr1WosAkvbD5ANOg0O91kxGRIvXWQqlES7USdq+1QKkEUidv4Ry7IIJfLwKPhpHOB2OeST0+WPZtRRoBfNYKBT2FZ2Z7wu6DQMnno7LNhntJrpcBJH+m13FKlwBXByy0N+daX1Gt+jn/c42vC/fug+/P0nTqLHZsRiqoiHhyj/mG9ce1Kxvl+5AFnJ96yjMguQ55J5FMq1HZ14ej12XAqnUKuJc8IwGU5i1G0TtYCTAmQReT120dIBlnNlZIoVxRY8cLxuOwJLWWSK8ijU80ezii/QA+ot3oT+fXZaDBgbduIbF+dE+xKWs1qNxaW5FEb7bXf8/Gi/De1GnezSLCYEbr324P4uPPnTj+CrP3UCP/6u/YJ8hpodcdUfxMQK5ho9kBV8z7Kb9djnaBO95et6/NEsAOxoU8fn7kC6WMGNWJbvZd2lvFKczM0/EAsFyCLyuu0IJcTJP2r0QFZwwQNQr6ZlWeCSDJ7I49kS4tmS4vOPuVZ2YjwIPH6PG/PJAl69Ia/dUSlcj2WRK1UbHSw4Oq0GD+7vwoTMBoaM+2O4b0+noKcMGg2DhwYd1N5NAGaDFgdFLNQLxnOKLirneD0dstlB3s001FERd8OvLaZRqtTu+u4TGgXIIhIz/6jRA1nBx1WAuBfpZgIqKdALxnP1VnYi/D7fc7gHbQYt9URGU4HeGjeJE4NdCCxl7+juIKXlXAnvLKR4T68g4vK67ZgSqVAvpPCico7PbUd4OS+LSaH+aAYdbfodPZQc7G2HUacR5VSc23H3UYCsXGLmH93eQVbucRUAOK1GuOwmWQTI3NO40gPkxq5Dr/C/zzaDDu872odnp+ZRqtQE/zw5mw4nYdBp1tzt4XoAy6XdG7VeUwafx45YtoS5pPAPXsF4XvH3K0C6ISxr4aa+7uShRK/V4Ei/TZTfx2Q4CatJh70OcTf8KEAWkd2sx/5uiyhFD8F4Dh1telhNesE/S2rcLofU/NEMjDoN3ArOoQPq/SiBnR3L7cSHj/cjmS/j3LWoKJ8nV9PhFA73WaFfY7T3kX4brCYdxmWSZjHuj8Gkp9ZrrY4r1JsS+J7FsiyCiZyiC/Q4R1dqCORxz9rd1Fef245L4SSqAteITIeT8Hnsop8uUIAsslGRgrlgIq/49AqOz1Mv1EsVxBvCshZ/NIv93RbRxmBKZTaSQY/VCJtID1/vOtANh8WAJ1XczYJlWUzPJdfNwdNqGDy0v0s2O8gTgRju20ut11rdoT4rdCIU6sWzJeRKVcXXzACA1aTHoNOCSYl3kJO5MpYyxV1tdIy67ciWqri+lOFxZXcqVqp3TA8VEwXIIvO57ZhLFrAkcP5RSOE9kJtxuxxij/JezR/NYEgFHSxmo8J3sGim12pwxufCdy4vyqZbidhuxXNIFyob3iRODDpwI5bDfDIv4sruFs+WcGUhzVv/YyIdk16L4V6r4Hmmauhg0cwng1PPWR5qZriuEkI+QF1byKBcZRvD1sREAbLIxMg/qtVYhJbVs4PMTdaRMkAulKsIxnOKzz9mWVaUFm+rPX7cjWKlhuemF0T9XLmYDtfHqXMT9NbC5ftKnWbxysouNuUfK4PPUx+PLGShXkglXZc4o247FlIFSYf78DH1dchpgVmvFfQBanJldoSYI6Y5FCCLTIz8o2imiFKlBo9Kvmy6LAa4O8ySjpy+EcuixoqXlyuVxVQRmWJF9N/nvXs6MNBlxlMX1dnNYnouCb2WwXDf+n/uh/tssJv1kqdZTARiMOu18Eqw40P4N+q2YzlXRigh3MlEo+uSSu5ZPhmcevqjGRi0ml3NStBpNTjabxM0npkOJ2E36yWZ6UABssi4/CMhd5AbIztVclwF3N7lkIo/svOG662k0cFC5J1yhmHw+DE3Xp6JIpqWvj2S2KbDSQz3WjfM6dWs5CFLPTBkPBDD/fs6YdDR7UUJfCKcegYTOXS26dFu1An2GXJytN8GhoGkmzr+SBb7utugW6PodztG3XZcmksJVqg3GZKmQA+gAFkSQucfcS3e1FARzBl123EzlkMyJ02hHndcNdit7B3k2UgagDgt3lZ7/Hg/aizwzKS6dpFZlsV0OLlhegXn5JADwXi+cWQttqVMEdcWM5ReoSAjfVbotYygwVwwro4OFhyLUYcDznbJd5D5OAn0eezIl6uNeyCfCuUqri6kG2mUYlPH45rMeD0dePLtOURSBfTYTLy/P3dcpfQx0824XY7puSQeOdAt+uf7oxm4O8wwG5RdtT8TycBm0sHZbhT9sw/2WnHEZcMXXwzg9ZsJ3t9/oKsN/9f7RmQ3qGAuWUAiV8ao27bpa7nBHBOBOD56n/gBxyuB+B3rIK3PqNPiUJ8NU2HhWr2FEnkcdlkFe3858rrteHl2SZLPLlaquBXP4azPtev34u69k6H6KRefri6kUamxkuQfAxQgS8LbNDDkMUEC5Bx6rEZVjV/l/kwnQ9IEyLMRlXSwWCnQkyqI/NTpIXz2O9dweT7F6/vmilU8PTmPj9zrkV0e+UYT9FYb6bWis02PcX8MH73PI/TS7jIeWILFoJVsx4cIY9RtxzOTc2BZlvdrv1ZjEU7k8W+O9PL6vnLn9djxj2+FsZgqoFeAOGAjt2I5VGssL0Xl+7vb0WbQYiq0zPt3zuQ2vvuEQAGyBLj8o6lwEo8d5v9LIZjIqabYgdPRZsCerjZBdznWU6uxCESzeGi/8nfN/NEMvudQj2Sf/6Fj/fjQsX7e3/fGUhan/+AFjAdisgyQtRoGh12b7yBrNAxODDokK9Qb98fwwP6uNYeZkNbl89jxd6/ewq14Dnsd/NZZLKYLKFXVU1TOad55fe8RcQNkPjpYcLQaBqP9wtQATYWWG0X4UqBvMQlw+UdC5SEH43lVFehxvG5pCvXmUwXky1UM9Si7QG85V8JSpiS7AJIPex1t6LOZJO8AsZbpcBIHnO1bPhE6MehAeDnfKNYVSyRdgD+apfQKBWo+oeNbo4OFyu5ZR1x2aBhpRk5zxdaDPBWVez31Qr1KtcbL+3EmQ0l43dIU6AEUIEvG67YLMkmnXK1hPplX3Q4yUL9Ig/E8EtmSqJ/rj/D3NC5nsyKPmBYTwzA4OeTAK4GYoP1ed2J6bntTpKTqhzyxkn9MA0KUZ7jXCoNWI0hRWaPrksruWWaDFgd7rIKP8V6LP5pFv90EC09dQ3weO4qVGmYi/BXqFcpVzEQykuUfAxQgS8brsSOaLmIxxW+j8PnlAmosVDMkpJnPLXw7orXMqi1AdiqzmObkoANLmRKvX/K7FUkVEE0Xt1SgxznY0w6HxSD6bvi4P4Z2o67R650oh0GnwWGXMBP1uK5LUh2jS8krwhCWtfA99ZV7gOfzVPzyfL11nFT5xwAFyJJpzj/iU6PFm0rGTDc7KlGA7I9mYDfr0d1uEPVzxTYbycCk18Ct0KPQ2x0g5JNmMbWDIhWGqechj4u8G/5KIIYH93ftuq8qkSevx47pcBI1nvvdBuN59NrUVVTO8XnsWMqUsMDzRtlGuGmofG7o7HdY0G7U8Xrv5YJt2kFWoUb+Ec/HK42RnSrcQbab9djnaBN9xr0/msGQ0yK79mB8m4lkMNjdDq1Gmb/PgS4z3B1myUc1N5sOp8AwwJEtFOg1OzHkwHyygJsxcfKQF1MFBJaylF6hYD53B9LFCm7Esry+byiRU+X9Crj94CvmwJDFVBHZUpXXoVYaDYNRt43XtNHJUBLd7Ub0idzhoxkFyBLh8o/4zkMOxvPQahi47NL9o5KS19MhwQ5yVvHpFcDtFm9KxTAMHhrswivX47zvku3U9FwSg92WbecKnhzsAiDebjj3UEEFeso1KtAJXSihzpoZoP7gq9Uwom7qNFICef4u97rteGc+hTJPhXrT4SS8bpukG08UIEuIO7Li8xg0mMihv8Ok2mNOn9uO8HIesYw444iT+TKi6aLieyDnShWEl/OKDpCBeh5yPFvCtZWJgVK7FE7uKAdvyNkOp9Uo2tjpiUAMNpMORyj/WLEO9rbDqNPwGsw1isoVmra1GZNei+Fe/jfKNsK1eDvA86aO19OBUqWGa4u7/+7MlSqYiaTh9XTwsLKdU2cUJRNc/tF8kr/8o2BcvcdVgHC7HOsR6stGbgLR+rGq0gNkbgdUDmkWsUwRc8nClkZMr8blIU+IlIc8Hojhwf0OxabfEECv1eBIP7/H6HPLedRYqK4HcjOfm/+Nso34oxlYjTo4rfxOQ/XxWKh3eS6FGnv7PaVCAbKEhMg/CibyqhoxvRpX7S/WkZVfoOMquVFyi7dmA11t8HSaZVGoNz1Xnxa40yruE4NdWEwVcX2J35zR1eaW87gZy+HESloHUS6v245LPBbqcT2QVX3P8tgRz5YQXs6L8nn+aAaDAkxD3etog9Wk4+UBitvg8kpYoAdQgCypRv4RT9PfCuUqoumiqneQrSY9Bp0W0Y6s/NEs9FpG8UeEs5EMtBoG+3ieoiVHJwcdsshD5nrO7jRtgSuYEzrNgnuY4PovE+Xyuu3IlqoI8PTQFVRxUTmHz53XrZiNZAQ58WQYpj6si4ffx1QoiR6rUfQR3KtRgCwhLv9oKpzi5f0aHSxUfFwFgLeLdCv80Qz2OSyKz/mejWSw19EGg07Zv0+gHugt58p4Z4Gf63KnpsNJ7HW0wW7W7+jX7++2oNdmbAzwEMq4Pwa7WY/DfZR/rHS+lZxQvjZ1gvGcqovKAeCQywq9lhFlUyddKGMxVRRs6qvXY8eVhRSKlequ3mcynJS0vRtH+Xc7mfO57ZgKLfOSf9QY2anCHsjNvG47FlIFRNLC95bku5+kXM1E0orPs+bc7ocsbGC5mem55I7yjzmNfsh+YfOQxwMxPLS/CxrKP1a8IacFZr2Wt7TAYCKv6qJyADDq6htlQkwpXI2rJRHqnuVzd6BcZXFtYefDljLFCvzRDLxuaQv0AAqQJTfqsSORKyOU2H3+ER1X1XG7HEJ/4ZQqNdyM5xSfl1uu1nAzpvzfJ6e/w4y9jjZJC/WSuTKC8fyup0jVpwMWG8WkfAvGcwgl8pReoRK6lUI9vr5b1V5UzvF57JgMCV+ox30PCBYgcwPQdnHCcHkuBZYFvB7pT6QoQJYYl3/ExxdOMJ6DUafhvTq11Rztt4FhhG++fiueRbXGCnZcJRc3Y1lUaqxqAmSAy0OOoSpRHvKlOW6C3u5uElzgOi7QbjjlH6uP123HdDjFy7Wh5iEhzbzuDiTz5cYpsFBmIxnoNAz2OoT5M/d0mmE363cVz0yuDE+TcsQ0hwJkifGZfxSM1ztYKH2i22YsRh2GnO2C7yDPRoQ9rpILtXSwaHZi0IF0oYJ35qXJQ26MmN5FigUA7Olqg8tuwoRAu+ETgTg62/QY7rEK8v5EfnweO/Ll6q5PJXKlCpYyJdWnBAK3d16Fbk/qj9ZrSfQCpbQwDNPYDd+pqXASLrsJPVbp89IpQJYYl3/ER1FZaDmn+gI9js+9u4t0K7gbxKBKAmSlPwg0a+y8SpRmMT2XgrvDjE6LYVfvwzAMTgrUD5llWUwEYjgx6KD8YxXx8tR1IZzgambonjXca4VBq9lVasJWiDH11eu24+pCGoXyzgr1psLJxr8xqVGALAM+jx1TPDQKD8bzdFy1wuuxI5IuYjElXKGeP5KBy25C+zbHALea2UgG/XbTtscdt7JemwmD3RbRJtGtVp+gx08O3okhB2LZEmYi/OYhB+N5hJcp/1htBp3taDNod73bydXMeOieBYNOg0MufjbK1lOvJckK3rPf67ajUmNxZWH7E/XShTIC0SwFyOQ2PvKPUoUykvkyHVet8AowhGU1f1QdHSxmoxkc6FXfEfpDgw68dj2OSrUm6uemC2UElrK7Tq/gnBRoOiCXf8x1/SDqoNUwGO23N3JFd4q6Lt3J6+Zno2w9wXgO5Sor/A7yLtJFplda3ko9IIRDAbIM8FH5GYxTB4tmR/pt0DDC5XSxLLtyXKXsAr1ajRWssbzcnRxyIF2s4NKcuHnIl3c5QW+1ga42uDvMvAfI44EYHBYDDqooN53UeT12XJ5P7erhsVFU3q7uonKOz2NHulDBzVhOkPcXq5bE3WFGl8WAqR08QHH9tWkHmTRw+Ue7OV65PbKTAmQA/z97dx7e1nnfif77YgdIECTBHaAWUqJkSaBly46lOLYUp4mTOFOniTzJzNxO9zbJZKad9rmdzrRzO522907nPtNJOm2T3jZN0naa6dhxmsZ2nDSJ7Ti2vEi2RVIbRVALCC4ACRDEvp77B3goiOKC/RyA38/z+GlEQeArlefgd973t8Bi0GF/j7Wsi7QYvnASkWSm6UdMe5fjSKRzO6pATyaPTq53mkWlI6Y3cny1K0e1pgNKkoQz7nz+8U4vCt6JXA4bEukcpioo1PMEYywqLyBf77UaGOJe7YE8VONNHXmiXjmnt+PefO2FXSUPTQyQVUDOP6okHeDWFD0eV8lcVcrt3sja03iT76zKH4A7MUDusZow3N2ylkpQLxe8IfS2GavarvHEsB3BWBpXFkrPC9zIjaUY5lcSOM784x1JPgKv5DPLE4izQK/ASK8VBp2mZps6bn8EPVYj2kzlTeYshcthw1VfpORCvfGZZdXsHgMMkFXD5bBhYjZU9g6PJxCD1agreyxtM3I5bFiMpDBfg0K9tYbrTR44undgi7dCJ4bzecjpOuYhj3srm6C3EXk3vFrBvryrfoL5xzvSXnsLWo26yk492QP5NnqtBof622qWFljPmhmX04ZsTsLFEtpkhmJpXF+KqSb/GGCArBoux2r+UaC8/CNPMA5np4XHVQWqscuxGbcvglajDj1NPpRlyhdBZ4sBnRW2G2tUx4fsiKaydRkDC+R7w7r9kao3yXd2WDDYWb085DPuJXRbjU2fg08b02gEDg+UH8yFYmmEExmeeK4jD2GpViqUTJLytST1Gmq11te5hM/eidXhSKMMkGm9W8Fceccr+ZGdvNkUOtTfBq1G1KR1jlyg1+wPJDu1QE8md2ioVx7ypbkwclJtpkjlpwMGKv7wLex/3Ow//7S50dVCvXJOV+QWb9xBvp3LaUMkmcG1pWhV39cfSSKcyNTtXt7XZkJXq6GkzalqDUeqJgbIKiHnH5WzUyVJEmaCzOdaz6TXYn9Pa02KHnZCizdJklZbvDX333MrXa1GjPS21m1gSLVGTG/kxLAdoXgal+Yr68oxvRiFL5xkesUO53K2I5XJYbKMvPa1rkv8zLpNOTuvxXDLU1/rlConF+qVEs+Mz4Qw2Fn5cKRqYoCsEnL+UTnpAIuRFOLpLHeQNzDqzF+k1SzUiyQzmAslmj7/eCmawnIsvaN3kIH8LvLZ68G65CGPz4RgbzGgr636Y1aPV6kfsvzn5bxm2plGK5ioxx3kje3rboVJr6l6HvJazUwd7+UuZzuu+sKIpTJFvX7Mu4xRR3uNV1UaBsgqIj9xlXoEequDBW8267mc7QhEU/Aulz+EZb1pBW42Sri6sLML9GQnhuyIp7MVD0YoxsTsCo44bDVJXei3mbHHbqm4UO+16SX0thmxt4v5xzvZbrsFVpOurGBuJhiH1aSDzcKi8kI6uVCvyjvIU74ILAYt+m3Vf/DezKjDhpx0q6/7VpZjKXgCcVUV6AEMkFXF5bQhmspierG0/CMPZ9pvqpJdjs2411qfNXeAsJNbvBV6oEaT6NZLpLO4uhCuSXqF7MRwPg85W2Yecj7/OIATzD/e8eRj9HIC5HzNDD+vNjLqbMfEbKjsa3QjckpgPa/ZUibqya9RU4s3gAGyqsj5R6XmIcv5XE6mWNzhQJ8VOo2o6pGV2xeFViOwq7O5A2S3L4KWOu86qFFniwEH+6x4bTpQ0+9zZT6MTE6qaZHK8SE7wolMUbs6G5nyRbAYSXK8NAHIB0GX58JIZUpLP/IE4+xgsQmXw4ZYKotri+UPYVlvWoGpr71tJvRYjUVtTsmppWoq0AMYIKuKnH9Uah7yTDCGrlYDLAZdjVbWuEx6LQ70WasaIE/5Ithtt8Cga+7LJ98WqL67Dmp1fMiOszcCSGZKa3xfiom1Ar3afUicWOvKsVjWn5fTM05wQAghH8ylsqUV6uWLyrmDvJnRKrcnjaUy8C7HFUkJHHXaiiqSH58JYY/dorqUm+b+hG8wa/lH3tJyHT2BOBy82Wxq1Jkfe1mtQr2d0MECYIu3QieG7UikczjvqV0/5AnvCmxmfU1PgnraTBjqbil7N/zM9BIGbCbsYjoXAWtFVaUEc/5IEol0jieemxjqboXFoK1agDztr28Hi0JHHDa4/RFEklsX6o17QzXdGCgXA2SVGXW248LsSkn5R/mJRLzZbOaIw4ZQPA1PoPJCvUw2h+tL0aYPkMOJNOZXEju6xVuhB/Z2QojqTaLbyIXZEI442mq+Y398yI43rgWQKbErRy6Xzz9m/2OSDXaaYTPrS9rUke/DrJnZmLbCISzrKdHBQjbqtEHaplBvKZKEdzmuqgEhMgbIKiPnH8mdEraTzUmYXWYP5K3IuxzVuOF4gnGks1LTTxBzr+46cAc5r91iwF19bTUr1Etlcrg8F65LDt6JITsiyQwmSsxDvuqLIBBN4TjTK2iVEGLthK5Y7Lq0PZejHRdnV0p+iN2I2xeBRgB7uur/7y3vCm/VAehWgZ66WrwBDJBVp9TxyPMrCaSzEvO5tjDS1wqDVoOxElNXNuL2rT6NN3lnh6urOYU7vYNFoRPDdpy7GUQiXf085Ku+MFLZXF2OGeUCu1J3w8+483nLHBBChY44bJhcCBd9XbCofHsuZxvi6ezaRkUl3P4odnVaYNRpq7Cy0vRYTei3mbbcnJKL+GrZvadcDJBVZri7FWa9tujdzlsTiXiz2YxRp8XBfmtVWr1N7ZAeyFP+CAxaDXNNCxwfsiOVyeEdT/X7IV/w5ndz6xEgd1uN2NdT+nTAM9NLcLSbufNHtxl12JDOSrgyX1yhnicQZ1H5Nlxrud1V2NRRuGbmyDatAMe9IQx1tcBqUleBHsAAWXW0GoEjjuLzj9YCZO4gb0m+SCst1HP7Iui2GmEzq+9iria3L4I9XRbotLxFyN61txMaUZt+yBOzIbQaddhdp+DzxJAdZ68Hip4OmMtJeP1agN0r6A5rp57FfmYFY3Dy82pLQ10taDFoS275ul42J2F6Maroieeow4ZpfxThRHrD3x/3hlQ3IETGTz8VOuKw4cJsqKj8I08wDiGAgXbuIG9l1GFDOJHBjaVYRe+Tfxpv7vxjYLWDBdMrbmMz63F4wIYzNSjUG/eGcHigDRpNfYrfjg/ZEU1li34QvzwfxnIszf7HdAdHuxkdFj0mijyh8wRjPIXYhkYjcMRRXIu0rcwEY0hlcorWkrjW5jvcWfPgDycxF0qobkCIjAGyCo06bUikc2vH+VuZCcTQ32Zq+p68lSp1l2MjkiTB7W/+DhaJdBY3AzEW6G3g+FAn3rm5XNU85Ew2h0tzK3Vtc3R8qBNA8bvhZ9j/mDYhhIDL2V7UvTWTzWF2OcGuS0VwOWy4OLtS9CnPRtY6WCg49VUOfjfqdCLvkI861VegBzBAViVXCb0lZ4JxOPk0vq2RXisMOg3GK8jpWoykEIqnm35n9fpSFDkJ2NdrVXopqnNi2I5UNoe3bgSr9p7Ti1Ek0rm6FqnYW4040GstulDvtekl7Oq0wMGTKtrAaJGFevMrCWRzEneQi+By2pDM5HB1ofyJem5fvshvqEu5zyx7qxGOdjPGN9hBHpsJQQjg8ID6CvQABsiqVEr+kYcTiYqi12pwV39lvSWV7CdZT1OrnTq4g3yn+/fk85Cr2Q95XKExq8eHOnH2enDbMcHZnITXp5fWdp2J1nM5bcjmJFyc27p14FoPZH5mbUveVa0kD9ntj8DeYkBHi6FayyqLy2HbcHNq3LuM4e5WtBjVWbDJAFmFNBqBw47te0smM1nMryTYwaJIow4bJrwryJUwhKXQreOq5g4cry5EIAQwtANyrUtlNenhctMZUDwAACAASURBVFQ3D3liNgSzXouhOj+QnBi2I57Oblspf2luBSuJDNMraFPyMfp2wZwnyK5LxdrdaYHVqKuoPemUTx1TX11OG64vxRCK316oNzYTwqhK848BBsiqNeqw4eLc1vlHs8sJSBKfxovlctoQSWZwbam83pJuXxRmvRb9baYqr0xdpvwRDHZYYNLXv29mIzg+bMc7nmXEU9XJQ77gXcGhgTZo61SgJ3tgb3H9kOXfZ4EebabfZkJXq2HbTZ2ZQAwaAfTbGCBvRy7Uq6Q9qdsfUcWGjvwAdaHgAWphJQFfOKnKEdMyBsgq5XLakNom/4gN10uzVixQ5g0nf7NpqVunAaW42cFiS8eH7EhnJZyrQh5yLiflR0wrkIPX0WLAwT7rtrvhZ9xL2GO3MKihTQkhVo/Rt9tBjqOPReVFG3XacGkuvG0a1EYC0RSCsbQqui7Jn72FhZzyz4oaR0zL+FOqUltVfso8HNlZkv09rTDqNGXnIavluKqW5L6ZDJA3d/+eTmg1AmemFyt+r2tLUURTWcV2UU4M23H2ehDJzMa74dmchDfY/5iK4HLYcNUX3vJkxROIsai8BC6nDalsDpMLxQ1hKaSmlMCOFgMGO823PUCNeUPQCOCQSgv0AAbIqrXH3pLPP9riidwTiEOvFeht8iP/atFpNTg80FbWDnI8lYV3Od70AbInoHzfTLVrNeow6rRVZWCInLOpWIA8ZEcyk8N5z8bXxIXZEMLJDNMraFsuZztyEnBxbovPLBaVl+TWRlnpn1lqK7YedbTflk894Q1hf49V1RMVGSCr1Fr+0RYXhicYg6PdXPfcxUbmctgwMRtCtsRCvenFHdbBore5/56VOj5kx9hMCNFkpqL3uTC7AoNOo9iO/QN77RBbTAeUv36CATJtQz4q32xTJ5HOYmElyQK9EuzqtKDNtPVG2WbcvgiMOo1qWjMecdjgCcSxHEtBkiSMzYRUnX8MMEBWtVGnDZe3yD+aCXAiUalcznbEUllcWyytt6Tbny/sU7Lhej3Iw2mYYrG1E0N2ZHISzlaYhzw+E8JdfVboFRrpbbPocai/bdN0kdemlzDU3YIenlLRNnrbTOixGjc9ofMus8VbqYQQGHW2l9Xqze2PYKi7VTU1M/ID1Lg3hPmVBBYjSVXnHwMMkFXtiGPr/CNPMM6Z9iXabpdjM1O+CDQin/rSzK4uRNBjNaLNpFd6Kap2354O6DSiojQLSZIwMav8LsqJITve2mA6YCabw5vXg9w9pqK5tjj1lIvKualTGpfThsvzK5vWCWwmP/VVPZ9Xcp/3sZnQ2ueviwEylWurYC6azCAQTfG4qkTD3a0w67UlB8hufwSDnc3f+mzKzw4WxbAYdLh7sL2igSGeQBzhREbxAPn4kB2pTA5v37y9IHjcG0KE+cdUApfThil/ZMPUI09wdQeZn1klcTlsSGclXJkvvlAvkc7CE4ypKiXQZtFjt92CCW8IE94QtBqBQ/3qLdADGCCrmpx/tFEni5kgj6vKodWIfKFeiUdW7h3QwUKSJLZ4K8GJIftaEFmOiVllJuit966hjacDvjYdAMD+x1S8UacNkpTPrV9vJhiDQatBr5XpOqVYa5FWwqbOtcUoJEkdHSwKuVYHoI3NhLC/p1X1G04MkFVMCAGXc+MjKx5Xlc/ltOHi7AoyWwxhKZTNSbi2qK7jqlpYWEkikswwQC7SiWE7sjkJb14LlPXnx70h6LUCI33K/nu3mfQ4ssF0wDPTS9jf04puq1GhlVGjObJF14WZQByODrNqcmIbhbPDjA6LvqQ8ZLnFm1o6WMhGnTZ4l+M4ez2g+vxjgAGy6rkc7bgyH74jP3CtBzKHhJRs1GlDPJ1dK7zbzuxyHMlMrukDR7W1BVK7e3d1QK8VZadZTHhDGOm1wqhTfhfl+JAd7xTkIaezOZy9HuDuMZWkx2pCX5sJ4xuML/cEYxxqVQYh8h2tStlBdvuiEALY26WuTR2Xox0AEE1l4XK2K7ya7TFAVrlR58b5R55AHBaDFp0tBoVW1rhuHVkVN+NeDhybPcViypf/GWOLt+KYDVrcM9ix7SS6jUiShAuzK4qnV8hODNmRyubw1mpXjrGZEGKpLAeEUMlcTtttE9NkHnZdKtuo04bJhTs3yjbj9kfgaDfDbFD+4bvQYcetnGOXylu8AQyQVW+zRuHy07gQPK4q1d6uVrQYtEUfWa1NJGr2ANkfQZtJh+5WHqkX6/iwHRPeEFYS6ZL+3FwogUA0hSMOdRSp3LenY3U6YD7Yl3fFH9jbqeSyqAGNOmyY9kcRLrgmIskMgrE0d5DL5HK0I5OTcLnIQj21Tn1tM+kx1NUCnUbgYJ9V6eVsiwGyyjk7zGi36O/oLekJcCJRubQagcOOjXc5NuL2R9DZYkBHk+/WX13IF+jxoat4x4c6kZNQch7yuMIT9NazynnIq23rzriXcKDXCjsflqhER1ZzSwsL9dZqZviZVZa1HsJFnHrmchKmF9VbbP3egz14cF+X6gv0AAbIqieEyFd+FgRzkiRhJhjncVUFRh35Qr10EYV6bl90R+TlutnirWT37uqAQacpuR/yhdU2R3epqM3RiSE7zs8sIxRP4+yNANMrqCxrp54FmzosKq9Mv80Ee4uhqDzk2VAciXROlTvIAPAfP3IIX/3Zdym9jKIwQG4A6/OPlmNpRJIZHldVwOW0IZnJ4erC9hP1pvyRpp+gtxxLYTGSYoBcIpNei3t3teO1a6UFyBOzK9jXra42R8eHOpHOSvjLH11DIp3D8SGmV1DpulqNcLSbb9vUWeuBzM+ssmzV0Wq9tamvTd51qR4YIDcAl8OGbE7Cpbn8kdVaBws+jZdN3uXYLg85EE0hEE2p9mm8WtY6WDBALtnxITsuzK4gFCs+D3ncq/wEvfXu39MJnUbgL390DUIAD+zlDjKV54ij7bZ7qycQY1F5hUYdNlz1RRBPbV2ot1ZUznt5xRggNwC5HYr89OgJcEhIpfbYW2A16jC2wRCWQtM7pUBvrcWb+gsn1ObEkB2SBLxe5C6ybyUBfzipmgI9WYtRh1GnDeFkBgf72po+555qZ9TZjmuLUYTi+YfGmWC+Zob1DeVzOduRzUm4OHfnEJZCbn8ENrMedl6/FWOA3AAG1uUf3dpB5nFVuTQagcOOtjuKH9fbMR0sfBGY9Bo4eARasqO72mHUadYmz21nbYKeynaQgVtT85heQZWQT+guFGzq8POqMrdyu7fe1JGnofJhpHIMkBvAWv7RajA3E4yh3aKH1aRXeGWNbdTZjkvzYaQymxfquf1RGHXNHzhO+SMY6mqFllOuSmbUaXFsd/H9kCe8KxACOKSiAj3ZQ/u7AQAPr/5fonKs9Zr3hlaLymNw8sSzIr1tRnRbjRj3breD3PxTX+uFAXKDcDlsuOoLI57K5p/GebOpmMthQyqTw+TC5r0lp3wR7O1qafrAUW7xRuU5MWTHpbkVBKOpbV877g1hqKsFLUZdHVZWmhPDdjz7b96DUwcYIFP5OloMcHaYMe4NIRhLI5rKsmamQkIIjDpsGN8iLTAUS2Mxkmz6E896YYDcIFwOG3IScHEuBE8wxuOqKljrLblFoZ7bH2n6YodYKgPvcpwBcgWOr7ZEe72IfsgXVFigV+jwgI3Hs1Sx0dVTz1s9kPmZVakjDhumfBFEk5kNf39qh6QE1gsD5AYxulqod94TyvdA5g5yxXZ1WtBm0m3aWzKRzsITiDX9zWZ6tS0QA+Ty3e1sh1mvXZtAt5mlSBKzoYRqRkwT1YrL0Y6bgdhazj13kCs36pQ3yjZOs5BrZngvrw4GyA2it82IrlYjfnDZh1QmBydvNhWTc7s3a/V2YymGnNT8Nxu2eKucQafBfXs6th0YIk8XU/MOMlE1yHnI3x6fBwD27a+CjYawFHL7IzBoNfy3rhIGyA1CCIFRp22tEIgXQHW4HO24PL+CZObO3pJr/SSbvOBhyheBViOwx97cf89aOz5kx5WFMJYiyU1fI6fzHBpQX4EeUTXJwdyZ6SUWlVdJT5sJfW2mTdMC3b4o9nRZoNMytKsG/is2EHlgCMAeyNXictiQzkq4Mn9noZ58XDXU1dw7q1O+CHbbLTDoeDuohNwibas85AuzIey2W2AzM1ig5maz6LHbbkE2J/HzqoqOOGwY26TVm9sfafqUwHriJ2IDkYvKAO4gV8tWhXpufwSOdjPMBvWMA66FKX8E+3hTrdio0waLQbtlmsWEd4X5x7RjyLvILCqvnlGnDdOLUUTWFeolM1nc3AE1M/XEALmByDebHqsRJn1zB2314uwwo92i3zCny+1v/tZn6WwO1xejTf/3rAe9VoP79nRuWqgXiqVxMxBj/jHtGGsBMneQq8bltEGSbg1hkd1ciiGbk3gvryIGyA2kp82E3jYjq4GrSAgBl8N2RyeLXE6C2xdt+qfxG0tRZHhTrZoTQ3Zc9UXgD9+Zh3xhbYIe849pZ3CtntCxqLx61gr11gXIO2Xqaz0xQG4w/+HDd+HTJ4eVXkZTcTlsmFwII5G+Vag3t5JAPJ3FcE9zF66xg0V1nVjth7zRLvLaiGmmWNAOcd/uTvzSw0N49HCv0ktpGl2tRgzYTHds6sj38qEmLyqvJwbIDebxow782CHebKpp1GlDJifhckGhntu3M57Gp3bI37Nejgy0odWo2zhA9q7A0W5GR4tBgZUR1Z9Bp8G///Bd6LGalF5KU9moPanbH0W/zaTKCZ2NigEy7Xiu1SEs4wWVwTvluGrKF8EAb6pVo9NqcP+ejrV2jIUmvCGmVxBRxUad7ZhejGIlkV772k6omak3Bsi04w3YTOhsMdx2ZOX2R2Az69HV2ty7fe0WA04Mdym9jKZyYtiOaX8UCyuJta+FE2lML0aZXkFEFZMLfeVdZEmS4PaxxVu1MUCmHU8u1CssepjyRTDc3QIhhIIrq73/9OOH8d/+6d1KL6OpyP2QC9MsLs3l03fYwYKIKrV+ot7CShLRVLbph1rVGwNkIuTzkK/6Ioin8oV6bn/zd7Cg2jg8YIN1XR6y/PDFAJmIKtXZYoCzw7x2X2EtSW0wQCbCrSmFF+dWEIqn4Q8nMcx8LiqDViPwrr2deG361kS9C94QetuM6LYaFVwZETWLUeetU8+1mhl+ZlUVA2Qi3OrXOT6zjOnVmw2ny1G5TgzbcW0xivlQPg95YjbE/GMiqpojDhtuLMUQiqXh9kdgNerQwwfwqmKATASgr82ErlYjxr0rcPujAPg0TuWT85DPTC8inspiyhdhegURVc2oI999aWI2BLc/gqGe1qavmak39nYiQr5QL39ktYxuqxF6rcBgh1npZVGDuqu/DW0mHV5zB7CrswU5ifnHRFQ9cqHe2EwIU74IHtzHbkTVxh1kolVHHDZM+SIY9y5jj70FOi0vDyqPViPwwJAdZ6aXOGKaiKrOZtFjV6cFr7oXsbCSZIFeDTACIFo16rAhJwFn3Eu82VDFTgzZcTMQw3cvLMDeYkBfG6eJEVH1uJw2vDK1CAAcElIDDJCJVsmFejmJNxuqnJyH/KOpRRxx2JgfSERVJW/qAGzxVgsMkIlW9baZ0NuWrwIe7mHDdarMwT4rOix6AEyvIKLqkzd1dBqB3XaLwqtpPgyQiQrIhQ98GqdKaTQCD+zN7yKzxRsRVZtc+LvLboGeNTNVx39RogLHdnfCqNNgiAEyVcFDI13QCODuwXall0JETabNpMeBXisO9fOEqhbY5o2owM++Zw8+eKQPrUZeGlS5T96/Cw/s7cRAO1sGElH1/dXPvQtGHfc6a4FRAFEBo06LvV3MP6bq0GoE9vVYlV4GETWpXnbHqRk+dhARERERFWCATERERERUgAEyEREREVEBBshERERERAUYIBMRERERFWCATERERERUgAEyEREREVEBIUlS8S8Wwg/gRu2WQ0QAdkuS1F3pm/B6Jaqbiq9ZXq9EdVPU9VpSgExERERE1OyYYkFEREREVIABMhERERFRAQbIREREREQFGCATERERERVggExEREREVIABMhERERFRAQbIREREREQFGCATERERERVggExEREREVIABMhERERFRAQbIREREREQFGCATERERERVggExEREREVIABMhERERFRAQbITUAI8UkhxCUhRFQI4RZCPLTJ6/6tEGJeCBESQvylEMJY77US7XRCiE4hxDdWr9cbQoh/vsnrhBDiD4QQS6v//VchhKj3eol2EiHEZ4UQZ4UQSSHEVwq+vkcIIQkhIgX//cct3mePEOIFIURMCHFZCPFjdfkLUNXolF4AVUYI8X4AfwDgEwDeANC/yeseBfAbAB4BMAvgGwB+Z/VrRFQ/fwIgBaAXwFEAzwohzkuSdGHd634RwEcB3A1AAvCPAKYBfLGOayXaaWYB/B6ARwGYN/j9dkmSMkW8z9cAnAHw4dX/nhJC7JckyV+1lVJNCUmSlF4DVUAI8SqAL0mS9KVtXve3AK5LkvQfVn/9PgD/U5Kkvjosk4gACCFaAAQBHJEkaXL1a38NwCtJ0m+se+2rAL4iSdL/t/rrnwPwC5IkHa/zsol2HCHE7wFwSpL006u/3gPgGgD9dgGyEGIEwDiALkmSwqtfexn5z1w+4DYIplg0MCGEFsB9ALqFEFNCiBkhxB8LITZ66j0M4HzBr88D6BVC2OuxViICAIwAyMrB8arzyF+f6210zW70OiKqnxurn7VfFkJ0bfKawwCm5eB4Fa/fBsMAubH1AtADOA3gIeSPa+8B8FsbvLYVQKjg1/L/ttZygUR0m/XXIVZ/vdF1uNE128o8ZCJFLAK4H8BuAMeQv2b/5yavLeU6J5VigNzY4qv/939IkjQnSdIigD9EPt9pvQiAtoJfy/87vMFriag21l+HWP31RtfhRtdsRGJeHFHdSZIUkSTprCRJGUmSFgB8FsAHhBDrr2egtOucVIoBcgOTJCkIYAb5Ap7tXEC+2Ed2N4AFSZKWarE2ItrQJACdEGJ/wdfuRv76XG+ja3aj1xFR/cmfuxud6FwAMCSEKNwx5vXbYBggN74vA/jXQogeIUQHgF8B8MwGr/srAD8nhDi0+rrfAvCV+i2TiCRJigJ4GsB/FkK0CCEeBPA4gL/e4OV/BeBXhRAOIcQAgF8Dr1mimhJC6IQQJgBaAFohhGn1aw8IIQ4IITSrtTt/BOBFSZLWp1JgtcbgHQC/vfrnfwLAKICv1/PvQpVhgNz4fhfAm8jvTF0C8DaA3xdC7Frt07gLACRJeh7AfwXwAoAbq//9tjJLJtrRPoN8+ygf8q2gPi1J0gUhxENCiEjB6/4MwLeQr4afAPDs6teIqHZ+C/n0xd8A8H+s/u/fAjAE4Hnk0yQmACQB/DP5DwkhviiEKOxQ8Unki+iDAP4LgNNs8dZY2OaNiIiIiKgAd5CJiIiIiAowQCYiIiIiKsAAmYiIiIioAANkIiIiIqICulJe3NXVJe3Zs6dGSyEiADh37tyiJEndlb4Pr1ei+qjGNcvrlag+ir1eSwqQ9+zZg7Nnz5a/KiLalhDiRjXeh9crUX1U45rl9UpUH8Ver0yxICIiIiIqwACZiIiIiKgAA2QiIiIiogIMkImIiIiICjBAJiIiIiIqwACZiIiIiKgAA2RqSJlsDqFYWullNLxQLI1LcytKL4OImlQgmlJ6CURlYYBMDel3n7mID3zuJUiSpPRSGtrPfvVN/Or/Pq/0MoioCc0ux/Gu3/8enp+YU3opRCVjgEwNZ2Elga+94cHCShJB7iJX5PGjA7g0t4ILsyGll0JETWZyIYxMTsLzE/NKL4WoZAyQqeH85Y+uIZXNAQBuBmIKr6ax/fjdAzBoNXjy7IzSSyGiJuMJxgEAP7y6iFyOp33UWBggU0MJxdL4m9du4FB/GwDAwwC5Iu0WA95/qBfffMeLVCan9HKIqInMrN6fA9EUxr08paLGwgCZGspfnbmOaCqL3/3oYQDcQa6G0/c5EYyl8YPLC0ovhYiaiCcYQ1erAUIAL036lV4OUUkYIFPDiKey+PKr1/HIwR4c290Je4sBM0EGyJV6aF8XeqxGPHWOaRZEVD0zwTgODdgw6rDhxSs+pZdDVBIGyNQw/u7NmwhEU/j0qWEAwGCnhTvIVaDTavCxe5144YofvnBC6eUQUZPwBGJwdphx8kAP3vEsYznGlm/UOBggU0NIZ3P485ev4f49Hbh/TyeAfIDsCcQVXllzOH3MiWxOwjffnlV6KUTUBCLJDIKxNAY7LDg50o2cBPxoalHpZREVjQEyNYRvvjML73Icnzm1b+1ruzrN8C7HkcmyuKxS+3pacc+udjx5zsPe0kRUMbmAerDTjKOD7bCZ9XjxCvOQqXEwQCbVy+UkfPElNw72WXHqQPfa1wc7LMjmJMyFmBZQDU8cG8TkQoTV5kRUsbUAucMCrUbgof1deGnSzwdwahgMkEn1/vHSAqZ8EXz61DCEEGtf39VpAcBWb9Xykbv7YdSxJzIRVU7ugTy4ep8+OdINfziJixxtTw2CATKpmiRJ+NMX3djVacFjrv7bfk++8XrYyaIq2kx6fPBIH775jheJdFbp5RBRA/MEYmgxaNFh0QPIB8gA271R42CATKp2ZnoJ5z3L+KWTQ9Bpb/9x7beZoNUIdrKootPHnFhJZPC9S+yJTETlmwnGMNhpWTv162kz4VB/G15iHjI1CAbIpGpfeNGNbqsRH7/Xecfv6bQaONrN7GRRRe8e7sKAzcQ0CyKqiCcQh7PDctvXTh7oxrkbQYQTaYVWRVQ8BsikWuMzIbx8dRE/9569MOm1G75msNPMHeQq0moEPnavEy9f9WOexY9EVAZJkuAJxjDYab7t66dGupHJSXhlakmhlREVjwEyqdYXXpqC1aTDv3hg16av2dVpYZFelZ0+5kROAp5+m7vIRFS6QDSFWCqLwXU7yPfu7oDVqMNLk5yqR+rHAJlUye2P4NsT8/iXJ3bDatJv+jpnhwVL0RSiyUwdV9fc9nS14P49HXjq3AxbMhFRydZ3sJDptRo8uK8LL11huzdSPwbIpEp/9pIbBq0GP/Pg3i1ft4udLGriiWODmPZH8dbNZaWXQkQNZmb1fuzsMN/xeycPdGM2lMCUL1LvZRGVhAEyqc5cKI5vvO3FJ+8fRFerccvXrrV6Y6FeVX14tB9mvRZPnWOaBRGVRr4fr99BBm61e+NUPVI7BsikOn/x8jXkJODnHxra9rXyDjIL9aqr1ajDh1x9eOb8LOIp9kQmouJ5gjF0WPRoNeru+L2BdjNGelvZD5lUjwEyqUowmsLX3riJx+8e2HD3Yb0Oix4tBi0L9Wrg9DEnwskMvntxXumlEFED8QRiW96/T450441rAdaOkKoxQCZV+eqZ64ilsvjUqeGiXi+EwCA7WdTE8b12ODvM7IlMRCWZCcbv6GBR6NSBHqSyObw2zXZvpF4MkEk1oskMvvLqdfzYXb0Y6bUW/ecGOy0s0qsBjUbg9DEnXnEvwrvMHG8i2l4uJ8EbjMPZeWeBnuy+PR0w67VMsyBVY4BMqvG1N25iOZbGZ95b3O6xLN8LOc62QTXw8XudkCTgaRbrEVERFsIJpLK5LXeQjTot3j1sx4ts90YqxgCZVCGZyeIvXr6G40OduHdXR0l/drDDjHg6i8VIqkar27kGOy04MWTHU2+xJzIRbW+rDhaFTh3oxs1ADNeXePpH6sQAmVThm2/PYn4lgU+f2lfyn91lZyeLWjp9zIkbSzG8eT2o9FKISOXkepDBDXogFzo50gMAePEKp+qROjFAJsVlcxK++JIbhwfa8PD+rpL/vHyUN8M85Jr4kKsPLQYtnjzrUXopRKRynmAMQgCObQLkXXYLhrpamIdMqsUAmRT3nQvzmF6M4jOn9kEIUfKfd64GyDd5VFcTFoMOj43249nxObZlIqIteQJx9FpNMOq027724ZFunHEvIZFmr3VSHwbIpChJkvCFF93Y29WCDx7pK+s9zAYtuq1GdrKooSfuG0QslcW3J9gTmYg25wnGNhwxvZGTB7qRzOTw+rVAjVdFVDoGyKSoH00tYtwbwi89PAStpvTdY9muTgtzkGvovt0d2GO34KlzTLMgos15g/GihjwBwIkhO4w6DV7i2GlSIQbIpKg/fcGN3jYjfuJeR0XvM9hhXquepuoTIt8T+bXpAFNZiGhD6WwOc6H4tgV6MpNeiweG7HhxkoV6pD4MkEkxb98M4sz0En7+PUNF5attZVenBXOhONLZXJVWR+t97F4nhAC+/hZ7IhPRnWaX48hJgLPIHWQAODXSjWl/lNNQSXUYIJNivvCiGzazHv/sgV0Vv9dgpwU5KX+DptoYaDfjPfu68NS5GeRy7IlMRLdb64G8xZCQ9U4e6AYAvMhuFqQyDJBJEVcXwvjuxQX81Lv3oNWoq/j95Jw35iHX1uljTniX43hteknppRCRysiF0oNbjJleb6irBc4OM/OQSXUYIJMivvjSNMx6LX763Xuq8n67VgNk5iHX1qOH+2A16fAUR08T0TqeQAw6jUC/rfgAWQiBUwe68ap7EakMU+RIPRggU915l+P45jtefPJdg+hsMVTlPXvbTNBrBXeQa8yk1+Kf3D2A5ybmEE6klV4OEamIJxjHQLu55I5EJ0d6EEtlcfY6272RejBAprr78x9OAwB+4aGhqr2nViPg7LCwF3IdnD7mRCKdw3Pjc0ovhYhUxBOIlZReIXv3sB16reBUPVIVBshUV0uRJP7Xmzfx0XscGGgv/Ua6FWeHmZXQdXDPYDuGu1vw5FmmWRDRLTPBWEkFerIWow737+nEi8xDJhVhgEx19ZVXryOZyeFTJ4er/t4cFlIf+Z7Igzh7I4hri1Gll0NEKhBLZbAYSRU9JGS9kyPduLIQxlyIdSSkDgyQqW7CiTS++up1PHqoD/t6Wqv+/oOdHxFDbgAAIABJREFUFizH0lhhbmzNfexeBzQCnKxHRACAmWA+sC12zPR6pw70AAB+yDQLUgkGyFQ3f/v6TawkMvj0qervHgOFnSy4i1xrvW0mPDzSjaff8iLLnshEO55833WWkWIBACO9rehrMzHNglSDATLVRSKdxV/86Boe3GfH3YPtNfkecu4bW73VxxPHBjEXSuCVqUWll0JECpN3kMsp0gNutXv70dVFTkQlVWCATHXx9Fte+MNJfObUvpp9D+4g19f77uqBzaxnT2QigicQg0mvQXersez3ODnSjXAyg7dvLldxZUTlYYBMNZfJ5vBnP3Rj1GnDu4ftNfs+NoseVpOOrd7qxKTX4vGjA/jOhXmE4sz7JtrJPMEYnB0WCFFaD+RCD+7vglYj8NKkr4orIyoPA2SquW9PzOPGUgyfOTVc0c2zGOxkUV+njzmRzOTwrfOzSi+FiBTkCcQxWGaBnqzNpMexXR3sh0yqwACZakqSJPzpi24MdbfgA4f6av79BjssTLGoI5fDhgO9VqZZEO1wnmCs7BZvhU4e6MaEdwW+cKIKqyIqHwNkqqkXJ/24NLeCT50chqbE8aPl2GW3wBOMI8fOCnUhhMAT9znxjmcZU76w0sshIgWEYmmEE5myhoSsd3KkGwDw8iSLf4txYynaUA8T/nASbn9E6WUUhQEy1dRfn7mB3jYjPnrUUZfvN9hhRiqTgz+SrMv3I+Dxow5oNQJPcheZaEeS6z7K7WBR6FB/G7pajXiRaRZF+fmvnsVvfmNC6WUU7Te/MY73/+FL+MPvXlF9txIGyFQzoVgaL1/14/GjDhh09flRk4/4mIdcP91WI957oAdPv+VFRuU3PCKqvkp7IBfSaAROjnTj5at+9ljfRjKThdsfwds3lyFJ6v+3kiQJb91chs2sxx/9YAqnv3gG11U8jZUBMtXMdy7OI52V8Jirv27fc5Ct3hRx+pgT/nASL1/lsSjRTnNrB7nyABnI5yEvx9IYm2G7t63cWIohJwGLkSQWVtR/arqwksRiJIlfft9+/Mk/vxfXF6P48B+9jL9786YqA3wGyFQzz47NYbDTjFGnrW7f09FuhhDcQa63Rw72oLPFgCc5eppox/EE4mgz6WAz66vyfg/t64JGgFP1tjHlu5XL2wgPE/IaXc52PDbaj+d/5SEcHWzHv/v6OD71N+cQjKYUXuHtGCBTTQSjKbwytYjHXAM1b+1WyKTXotdq4jS9OjPoNHj86AC+d9GnupscEdWW3AO5WjpaDLh7sJ3t3rbhXg2QNQIY94YUXs32xr0haDUCh/rbAAD9NjP+5ucewH/48EH84LIPj37uh3j5qnr+f84AmWriuxfnkclJ+Mho/dIrZLs62epNCU8cG0Qqm8M/sCcy0Y7iCcSqUqBX6ORIN87PLCPAB+5Nuf0RONrNGOm1YmxG/QHy2EwI+3taYTZo176m0Qj84sPD+Pt/9SDazHr85JfewH/+1kUk0lkFV7q6NqUXQM3pmbE57LZbcHigre7fe7DTwml6Cjg00IZD/W3siUy0g0iShJlgvCot3gqdOtADSYKqdhTVxu2PYrinFS6HDRPekCrzeGWSJGHCG4LLsXHK5eEBG5751+/BT53Yjb985Ro++iev4PL8Sp1XeTsGyFR1S5EkXnUv4TFXf13TK2SDnWbMrySQzCj/BLrTPHGfE+PekOI3NiKqD38kiWQmV7UCPZnLYUOHRc80i01IkgS3P4Lh7haMOm1YiqYwG1JvP+TZUAJL0dSWNUkmvRa/8/gRfPln7sdiJIUf/+NX8KUfXVNsrgEDZKq671xYQDYn4TEF0iuAfIqFJAHeIPOQ6+3xow7otQJPnuUuMtFOINd7VDvFQqsReGh/N3446efgpw3MhRKIpbIY7m7FkdVd2XEVF+qNFxTobee9B3rw/K88hIf3d+F3n7mIn/ryG1hYqX/wzwCZqu7Z8VkMdbWsJeLXG3shK6ezxYD3HezF37/tVX0TeCKq3Izc4q3KKRYAcOpANxYjKVyc44nUevI0uuHuVtzV3wadRqi6UG/cG4JOI3Cwz1rU67tajfjzf3kffv8njuDN6wE8+rkf4vmJuRqv8nYMkKmqFiNJnHEv4bFRZdIrgPwOMgB4uIOsiNPHnFiKpvDCZZ/SSyGiGqvmkJD1HtqfHzv94hXeS9aTO1js62mFSa9VfaHe2EwII71WmPTa7V+8SgiBf/HAbjz7bx7CYIcFn/qbt/DrT51HNJmp4Upv0dXlu9CO8fzEPHISFEuvAIDuViMMOg07WSjk1IFudLUa8QfPX8bzF+aVXk7RnO1m/Nv3jyj2YNfMMtkcPv/9q/jJE7vRYzUpvRyqIk8gjq5W422dCaql22qEy2HDS5N+fPaR/VV//0bm9kfRZtKhq9UAABh12vD8hXlIkqS6e5gkSRj3hvDBw31l/fnh7lZ8/dPvxue+N4kvvOTG69cC+NwnjuKeXR1VXuntGCBTVT07Nofh7hYc6C3uGKUWNBqBwQ4zbi4xQFaCTqvBZ987jL/40TW8cS2g9HKKkszk4A8ncfJAD47tru1Ndyc6P7OM//GDKSTSWfzmY4eUXg5VkSdY/RZvhU6OdOMLL7kRiqerNoikGUz5IhjuaV0Lho84bPhfb3ryHUWqXDBZqZlgHMuxNFwVDA0z6DT49Q8exMmRbvzq/z6P0188g19+33585tQwdNraJEMwQKaq8YUTeP3aEj77yH7Fn2DZ6k1ZP/3gXvz0g3uVXkbRIskM7v+97+Gpcx4GyDVwZT5/HPyNt7349Q8ehL5GH2hUf55gDPcM1u6aOXmgG3/8whRemVrEh13KnUyqjdsfwcMj3Wu/lrtDjHtDqguQ5dzozVq8leKBITue++WH8H99cwJ/+I+TeGnSj8994mhN/s68S1HVyOkVSgwHWW9Xp4VFelS0VqMOH3L14Znzc4in2B6w2iYXwgCAxUgKL3F8cNPIZHOYXU7UdAf5nsF2WE06/twUWEmk4QsnMdzduva1A31W6LVClXnIYzMh6LUCB4os0NuOzazH5z95Dz7/yaOYnA/jQ59/GV8/N1P1PtAMkKlqnhmbw0hvK0YUTK+QDXZYEE5kEIqllV4KNYgnjg0inMzgOw2UN90oJhfCOOJoQ1erAU+e8yi9nIb1+vQS/ukXz8CnQMurjcyFEsjmpJp0sJDptBo8tL8LL036VT0Io56m/VEA+QI9mVGnxYE+K8a96mv1Nu5dxsG+Nhh11c1Tf/yoA9/+lYdwqL8Nv/bkeXz2a28jlqpeAR8DZKqKhZUE3rwewGOuAaWXAoCt3qh0D+ztxGCnmZMAa2ByIYzD/TZ89KgD37/kw1IkqfSSGpIQAm9cD6imnZecxlaLDhaFTo30YH4lgSurJxE7ndzBYri75bavuxztGJ9R10Q9SZIwPhNa69Vcbc4OC772i8fxfz56AOFEBqYqBuEMkKkqvj0+B0kCHhstr0q12uQjP+YhU7E0GoGP3+vEK+5FeJfZIrBaFiNJLEZSGOmz4vR9TmRyEr75zqzSy2pIhwfaIARUc4w+U6MhIevJubZMs8ib8keg14o78m5HnTasJDKq2hi6GYhhJZHZcoJepbQagX/13n34yk/fD42mevVPDJCpKp4Zm8PBPiv29SifXgFwB5nK8/F7nZAk4GnuIleNnH880tuKg31tcDls3KUvU4tRh33drZhQyQ7yTDAGjQAG2msbIPfZTDjYZ8WLDJAB5HeQd9tb7ih2lYvg1PIABdxaSzUK9LZTzeAYYIBMVTAXiuPsjaAqivNkbSY92i169kKmkgx2WnBiyI6n3qp+wcdONTmfD5Dl1o9P3OfExbkVXJhVz4d4I3E5bBjzquMY3ROMo99mrktXkpMHunH2RgCROg2JUDO3P4J9BQV6spFeKwxajWpScIB8BwuDTqOK2qRSMUCmij03ni9qUlsLHnayoHI8cZ8TN5ZiePN6UOmlNIUrCxG0W/TothoBAD9+9wAMWg13kcvkctrgDyexsKJ8HrcnEIOzo7a7x7KTI91IZyW8OrVYl++nVulsDjeWYhjuabnj9ww6De7qt2JcRTvI4zMh3NVnhUHXeOFm462YVOfZsVkc6m/D0AZPtEoa7LBghuOmqUQfPNKHVqMOT55lt4VqmFwIY6TXutYbvd1iwPsP9eKb78wilckpvLrGI+dyjs0o360gPySkPj1379vdiRaDFi9N7uw0ixtLMWRy0m0t3gq5nDZMeEPI5ZQ/YcjlJEx4QxUNCFESA2SqiHc5jrduLis6Wnozg50WzARjyKrgRkGNw2LQ4TFXP54dn0OUx7kVkSQJk/PhOyZrnj7mRCCawg8u+xRaWeM61G+DRkDxPOREOouFlWRNW7wVMug0ePc+tntz++UOFpsEyA4bwskMri9F67msDV1fiiKczGDU0a70UsrCAJkq8tzYHAB1DAdZb7DTjHRWwoJKeoZS4zh9nxOxVBbfnmBP5ErMhRIIJzMYWTcg4KH9XeixGvEUeyKXzGzQYn+PFWMKB8hyp5dad7AodHKkGzPBONx+5YM/pcgB8lD3nSkWQL7VGwBV5CHLa6hVi7daY4BMFXlmfA4uhw277RtfrEraxU4WVKb7dndgj93CNIsKyR0s1u8g67Qa/MS9DrxwxQ9fmA+wpXI5bYr3u5ULoOs51vik3O5tB6dZuH1R9LWZYDXpN/z9/b2tMOo0qshDHp8JwajTYH+vutIvi8UAmcrmCcRw3qPO9AoAa0d/7GRBpRJC4PQxJ16/FsDNJf78lKuwxdt6TxxzIpuT8M232RO5VKNOG5aiKcyFlHu48KzWd9QrxQLIB+PD3S148crOTc1x+yMbFujJ9FoNDg20KX7CAABj3hAODbTVpctJLTTmqkkVnhvPp1c8prLuFbKBdjM0ggEyledj9zohBPDUW+y2UK4r8xH0WI1otxju+L19PVYcHWzHk+c8OzqntBxq6Hc7E4jBoNWgZ7U7Sb2cOtCD168FEE9l6/p91UCSJLh9kU3zj2Uuhw0XvCFF629yOQkXvCGMNmh6BcAAmSrw7Pgc7nba6nrEVgqDToN+m3ltp4OoFAPtZrxnXxe+fm5GFRXhjWhyIYwDfZv3P33iPicmFyKqyJdsJHf1t0GrERj3KtfJwhOMwdFhrvpwhu2cHOlGKpPDa9eW6vp91cAfTiKczBQVIEdTWVxbjNRpZXeaXowimso2bP4xwACZynRzKYaxmZBq0ytkg51m5iBT2U4fc8K7HMdr0zvvw7hS2ZyEq77wlgMCPjI6AKNOgyfPcpe+FCa9FiO9VkV3kD2BeN16IBd6195OmPSaHTl2emq1QG9fz9YB8qhT+UI9+eFNXksjYoBMZXlmPJ83qLbhIOvt6rQwxYLK9ujhPlhNOg61KIMnEEMinbujQK+QzazHo4f78A/nZ5FI77wj80qMOvL9bpVKT5mpYw/kQia9FieG7DuyUE/u3rHdDvJwdwvMeq2iD1BjMyGY9VoMb9JtoxHolF4ANaZnx+Zwz652OOtYoFGOwQ4LfOEkEuksTHqt0suhBmPSa/FP7h7A02/N4HceP7xp5Tjd6YpcoLdFigWQT7P4h/Oz+N6lBXxkdKAeS2sKR5w2/N1ZD2aC8boHqpFkBsFYuq4FeoVOjnTjhW9dxLkbQfTbTFV9b40Q6G0zrg22URO3L4IWgxa9bVvnfetWC/WU7GQxsVqgp2vQAj2AATKV4dpiFBdmV/Bbj92l9FK2tcuev4HPBGPY19N4s+BJeU8cc+JvX7+J58bn8In7dym9nIYxOZ8PkPdvcxz87uEu9NtMeOrcDAPkEsjFT+PeUN0D5Fst3uqfYgHkC/XwrYv4+Bdercn7/+aH78IvPDxUk/euRL6DRWtRwbvLYcPfvelBNidBW+c88WxOwoR3BZ+4f7Cu37faGCBTyeTuFWpPrwCwtsN9M8AAmcpzdLAdw90tePLsDAPkEkz6IhjsNKPFuPXHjFYj8PF7nfjTF6cwH0qgr8o7gs3qYL8Veq3AuDdU93vxWoCs0A7ynq4WfPln7oevBkOg/uj7U3htekmdAbIvggeG7EW9dtRpw1devQ63P7JlHUAtuP0RxNPZtbHojYoBMpXsmbE5HNvdgYF2ZXYPSrE2LIS9bKlMQgg8cd8g/su3L2PaH8HQNvl/lLfRiOnNfPyYE3/8whS+8bYXnz41XOOVNQejTosDfVZFjtHXeiAr2MHovQd6avK+r08H8PLUYk3euxLRZAazocS2BXqywlaA9Q6Q5dznRg+QGzc5hBTh9kdwaW5FlaOlN9LVaoBZr2WrN6rIT9zjgEYAX2dP5KKkMjm4/RHsL/KDeW9XC+7f08GeyCVyOWwYm1mu+7+ZJxBDi0GLDkvz5eS7nDb4w0ks1GB3uhLXFuUCveKK3oa6W2ExaDE+U/9WgBPeECwGLfZ2NfZmAgNkKsmzY3MQAvjQkcYIkIUQbPVGFettM+HkSDe+fs6raPP9RnF9KYpMTip6BxnIt9Sb9kfx1k3levs2GpejHSuJTN3vb3IHCzUWslVKDUNYNjLly7d4266DhUyrETgyYFOk1dvYzDKODNjqnvtcbQyQqSTPjs3h/t2dDZUnONjBVm9UudPHBjG/ksArKjx+VZsr8/KI6eID5MdGB2DWa9lSrwTyEXa9g6B8D2R1dzAq16GBNmgEFNl53YrbH4FWI7DbXnzbNJfThguzK8hkczVc2e0y2Rwuzq3A1eDpFQADZCrB1YUwriyEVT8cZL3B1V7IPLqlSvzYoR7YzHo8yQBuW5MLYWg1AkMl9EBtNerwoSN9eOb87I4cI1yOkV4rDFpNXfOQJUmCJxhTrINFrVkMOuzvsWJMZdMd3f4IdndaYNAVH7a5HDYkMzlc9dVvot6UP4JEOtfw+ccAA2QqwbPjcnpFn9JLKclgpwXRVBbBWFrppVADM+q0ePzoAL5zYR6hOH+WtnJlPow9dkvJvcdP3+dEOJnBdy/O12hlzcWg0+Bgf30n6gWiKcRS2abdQQbyO69KDmHZiNsXLblA2KXACYP8s9jII6ZlDJCpKJIk4ZmxOTywtxM9bY2TXgEUdLJgmgVV6Iljg0hlcvjW+Vmll6JqkwthHNhmQMhGju+1w9lh5ujpErgcNkzMhpCrU278WgcLBcZM14vLYcNiJIW5kDoK9TLZHK4tRjHcU9pUur32FrQadXU9YRifCaHVqMPeElJB1IoBMhVlciGCKV8EjzVgI3/5KJB5yFSpI442HOi1Mk92C/FUFjcCsbJaS2lWeyK/4l6Ed5mdZ4ox6rQhnMjgRp3ub7eGhDT3DjKgnkK9mWAcqWyu6AI9mUYjcHigra7pImPeEI442qBp8AI9gAEyFenZsVloBPDBw42VXgHcambPHWSqVL4nshPveJYx5QsrvRxVcvsjkCSU1MGi0OljTkgS8DQfQopyZK3rQn2KymZU0AO51g71t0GrEZhQSR6y219aB4tCo04bLs2tIF2HQr10NodLcytrnUAaHQNk2pYkSXhmfA7Hh+zotm49A16NWow62FsMmAkyQKbKPX7UAa1GsFhvE2sdLMpIsQDygdfxoU489daMqnJA1Wqk1wqDrn6Fep5gDB0WPVq3mZDYyEx6Lfb3tKqmUE8OkPeVESC7nO1IZXKYXKj9A/3kQhipTA4uZ3vNv1c9MECmbV2eD2PaH2247hWFBjst3EGmqui2GvHeAz14+i1vXdsnNYrJhTAMWg12V7DD+MSxQdxYiuHN68Eqrqw56bUaHOpvq1shlicQa+rdY9mo04ZxBYawbMTti6Kr1QhbGYNZRld3c+vxACV/j1HuINNO8ezYHLQa0ZDpFbJ8qzfmNFJ1nD7mhD+cxA+v+pVeiupcWQhjuKcVOm35Hy8fcvWhxaDFU+c8VVxZ8xpd7bpQj0K9mWB8LW2tmbmc7QjG0mspJUqa8keKnqC33m67BVaTri674ePeEKwmHXbbm+PngwEybSnfvWIW7x62w97aeOkVsl2dZniX49zxo6p45GAPOlsMLNbbwOR8GAd6KxsxazHo8NhoP54dm0MslanSyprXEYcN0VQW06vjiGsll5PgDcbhbNIeyIXkXVCl85AlScKUL4LhnvKuKSEEXA5bfXaQvSG4HLammbDIAJm2dGF2BdeXYnjM1bjpFUC+UC+bkxRr2zO5EMbnvjeJpUhSke9P1WXQafDRow5876IPwWhK6eWoxkoijdlQouz840Knjw0imsriuXH2RN6OPJSh1sHcQjiBVDa3I3aQD/RZodMIxfOQA9EUQvF0WfnHMpfThsvzK0hmajeAJ5nJ5gv0mmBAiIwBMm3p2fF8esWjDZxeAdzqhaxUq7fnxufw+e9fVeR7U22cPuZEKpvDP7An8pqrq4VA5XawKHT/ng7stluYZlGEfd2tMOk1NW9LJqep7YQcZJNeiwN91rr2EN6I258/FSh3BxnI93VOZyVMztduot7kfATprIRRR3MU6AEMkGkLkiTh2bE5PLivCx0tBqWXUxH5hu5RqJPFC5d9ODrY3tBpKnS7QwNtODzQhicZwK25svoBXE4P5PWEEDh9rxOvTQdwc4kFtlvRrRXq1bbV21oP5CYeElJo1GnDuMIT9aZ8cou38gdvyEHrWA1/PuQi0WZp8QYwQKYtTHhXcDMQw0cauHuFrN9mglYjFOlk4QsncH4mhEcO9NT9e1NtnT7mxIR3BZfmVpReiipMLoTRYtDC0V6dAOrjx5wQAvj6W8z13s6osx0T3hVka1ioJ28wOHZIgOxytCMUTyta4O32R2DSazBgK//ffLDTDJtZX9MUnHHvMmxm/dpgrmbAAJk29czYLPRagUcPNXZ6BZDfYXG0mxW50b14Jd/p4JG7GCA3m8ePOqDXChbrrZpcCGN/r7VqU7QG2s14z74uPHVupm6jlBuVy2FDPJ3FtL92x+ieQBy9bUYYddqafQ81kXdDa7nzuh23P4KhrtaKrikhBEadtpqm4IzNhDDqbJ4CPYABMm0i371iDu/Z11VW70U1Guw0K7KD/MJlH/raTDjU31b370211dliwPsO9uLv3/bWZVKV2k0uhDFSYQeL9U4fc8K7HMdr15aq+r7NZrQO45E9wdiOKNCTjfS1wqCt3xCWjbj9EeyrIP9YdsRhw5X5MBLp6hfqJdJZTC6Emyq9AmCATJs4PxOCdzmOx0YHlF5K1ezqtNR9ml4qk8PLVxfx3oPdTfVkTbc8cZ8TS9EUXrjsU3opilqMJLEYSVUl/7jQo4f7YDXq8NRZ7tJvZai7FRaDtqYDQ2Z2yJAQmVGnxcF+a92GsKyXSGcxE4yXNWJ6vVGHDZmchMvz1Z+od2U+jHRWYoBMO8OzY7MwaDV4/6FepZdSNc4OCxYjKUST9eur+ub1ACLJDB452Dz/jnS7kyPd6Go17vg0C3mU7YEqtHgrZNJr8ZG7B/DcxBzCiXRV37uZaDUChwdqN1EvlclhfiWxYwr0ZEcc+UI9JVJ8pv1RSBIw3FN+gZ5Mbr9Wi58PuRVeM7V4Axgg0wbk7hUPj3TBZm6O9AqgoNVbHXeRv3/JB4NOgwf32ev2Pam+dFoNPnavAz+47MPiDu5zPTlfvRZv6z1xnxOJdA7Pjc9V/b2bicvRjguzoZoMRJoLxZGTAOcO2kEG8juv4UQGNxRIz3P75Q4Wle8gO9rN6GwxYHym+vnU4zPL6GwxVK04Vy12fIAcT2VrWvVbbclMtqbNvgHgbc8yZkMJPNYE3SsKrbV6q2Oh3gtXfDgxZIfFoKvb96T6O33MiUxOwt+/7VV6KYq5shBBu0WPbmv1WxneM9iOoe6WHb9Lv51Rpw2JdA5TNSjUW+uBvINykIFbu6JjNQgst+P2RyAEsLer8h1kIQSOOGpTqDfuXcGRJpqgJ9vRAXI2J+GR//YiPv+9SaWXUrSf/NIbeNfvfx//93OXajb04pnzczDoNPixu5orLUDeQa5Xod60P4Jri1E8cpDdK5rdSK8VdztteOrcjKI9U5WUL9Cz1uRDUgiBJ44N4s3rQVyr8TjlRnbEUbtCPfnkrZnaeBVjpNcKg06jyMhptz+KwQ4LTPrqdA0Zddhw1RepaqGeXKA32mT5x8AOD5DHvSHMhRIN86F2cymGN64F0GM14ks/uoaH/98X8PNffRMvX/VXbf25nITnxudwcqQbVlPzpFcAQIdFjxaDtm7T9H6wWrTFAHlnOH3MicvzYVyY3Xk9kSVJwuRCuCbpFbKP3euARoCT9bYw1NWCFoO2JsGcJxCDTiPQX0E/3kak12pwV39bzacUbmTKF6loQMh6LqcN2ZyEi1Xs235xLt97u9nyj4EdHiC/MrUIAJgNJfC2R7k+h8V6djX/7ss/cz9e+XeP4LPv3Yd3PMv4yS+9gff94Uv46qvXKy5ieetmEPMriaYYDrKeEAKDnZa6BcgvXPFhf0/rjqr63sl+/G4HDDpNzdMAMtkcvvbGTXzqr88hFFdH0dr8SgLhRAYjVS7QK9TbZsLDI914+i1vQ6XF1ZNGU7tjdE8wjoF2M7RV6nHdSEYdNlyYXalroV4uJ2HaH6lK/rFM7jJRzbZ18sPYKAPk5vKqexF77BYYtBo8c179xR/PjM3inl3tcHZY0Gcz4dc+cACv/MYj+O+fuBtWkx6//Q8XcOL/+QF++5sTa8n9pX+PORh1GryvydIrZIOdlroU6YUTabw+HeDu8Q5is+jxgUO9+Pt3vDWpE8jl8sWzH/jvP8S/f3ocz1/4/9u78+A28/M+4N/3xQ3iIgiCBAFSEg9REgnIWu0pUquV5PXKS61Tu5t0pkmaTOJO0jpNp7XrJM1MJ01mmtZOMk4zzThu3XSaJm3jNJfF9a7XknZXotZab7QRqIMHIEriiZsgQNzA2z/Al6IkHjheAC+A5/MnRQGvRJB88Huf5/us4OJdj+DPU4rpCg7obfVmrM9nAAAgAElEQVTmcRuWw4nNww3yNLtVjzvLa4Lncs8HY03XXsGz2/SIJjOYC1SvvWdxNY5kJoc+ATKQeRa9EiaNXNA3UM6FMEwaOTp1SsEeUyyatkBOpLP4+H4IZw934OWD7XhrclnUm5rm/Ou4vbSGMfvjJ7sKqQSfP2bD33xpBH/9pRF85kgH/vdH8zj7u+/jp799HT+44yn4tIVvrzg9aIZG0ZhDZT1GNeaD8Yq31Fyd9SOT46hAbjJvHrdhNZbGpbvCZSJzHIcPZnz43H+5ii/92Q1IJQz+6KePw9gix9VZcRSKfMSb0EtCnvTpwx3Qq2Q0rLcLu02PVCa3+TURykKTLQnZqhInr3vhD7mEWBLCYxgGdqte0BacyYUw7A04oAc0cYF840EIyUwOI/1teOOoBStrCdx4GKr1Ze2Ijzd63b5z68Onug34vX/0KVz7tTP4ymcOYtYTxRf/58d45Xcu41sfuLEaS+36HD+6H4Q3kmy49IqtultViKez8Ed3/78o18UpL3RKKY7va63o8xBxOTnQjg6dAt8RqID75GEI//i/Xsc/+e8fYTWWxu/9xFF871++jNeGOnGirw1XXX5RzE9Mr0Rh1ipgUMsr+jxKmQSfO9qFd26viKa9RGwcNgMACFoExVIZ+KMp2JosA5k3YNZAIWWr2ofs9uVPq4VssQAAu82AWW8EsVT5+wBiqQxmvRHYN15zjaZpC+QJtx9SlsHzB9pw9nAH5FIWF5zibbO44FzGs/ta0VVAzqBJo8AvnRnAlV85jT/8yWdg0avwH96awou/fRG/+v+cuLtDg/745DKUMrahTz172iqfZJHLcXhv2otTg2ZIJU37LdaUJCyDLzxjw/szPnjXEiU/zqwngl/4k4/x+T+8hhlPBL/xxhFc/PIpfOEZ22YP6MkBE7yRJGY8wkd6FWvGExF8QchOfvxZG5KZHC44l6ryfPVmn1ENrVIqaDG3ENqIeGvSeQqphN1YwlK9WSWXN4pWtQzGFmHfdNqteuQ44I4Aw8R3l9eQ49BwG/R4Tfvbe8IVwNFuAzQKKTQKKU4PirfNwu2L4u7yWtEnuzIJi9ftFvz5L7yEt375JD5/zIq//vtFfPb3r+Anvvkhxp3Lm31q2RyHtyZXcOaQGS0N2l4BPMrwrOTK6cnFMPzRFM4caq/YcxDxevO4Ddkch78qIRN5IRTDV75zE6994wNMuAL4168exPtfPY2fHTkAhfTxqKfRgfzr62qN+3GzOQ6z3ojgK6Z3YrfqMdihxXdo9fS2WJbBcJde0I1p/GCzrUlbLID8yfztpbWqDYi6BR7Q4zlswkUB8o/RiAN6QJMWyGuJNJwLqxjpe7TdbMzRBW8kiR/dD9bwyrY37lwGwwCfHS699eFIlw6//QUHfvhrZ/Hrrx/G8locX/qzGzj5ny7jDy7O4nu3luGPJnHe0SXglYsP/wP+YaByBfLFKS9YBjh1sHFP4snO+to1eKbHUFR8ZCCaxG9+9w7O/M77+NubS/j50QP44Kun8ctnB3acB7AaVOg1teDqrE/Iyy/afDCGRDpX8QE9HsMwePO4DX8/vwqXV9g+20bhsOkxtRxBKiPMoB5fIDfrkB6Qz5iOpbK4V4ElLNu554sK2n/M69ApYdYqBGnBmVwIw6xVoKMBB/SAJi2Qr98LIscBJ/pNmx87e8gMpYzdjFITk3HnMp7bZ0SnvvwXoUEtxz99uRfvfeU0vv0zz2KgQ4PffXcGv/Rnn0Alk+D0YGMXdSq5BO1aRUWTLC5PeXGsp1XwW2Okfvz4s92Y9UZxc49Tmmgyg2/8YAYvf+0y/se1OXzhGSve+8or+PWxIwW9fkYHTLg+FxSsECrFND+gV6UWCwD4B8eskLCMYL3ejcZu0yOVFW5QbyEUh1LGol0j/JbEeiHkyeteVmMp+KOpipwgA/m7ME4hCuTFcMOeHgNNWiBPuPxQylgc63nUWN6ikOLMITPemlwRVcbmrCeCaU9E8ME5Ccvg7OEO/MnPv4CLXz6Fnxs5gF85NwiVXJiNPWLWY1RXrAfZu5bA5GK4ofu4yd7GHBYoZSy+8/H2Sy0S6Wx+2c/XLuMbP5jFqcF2fP9fncJ//IeOguYMeKP9JsRS2ZoOGM9uFGEDFTjt2km7VoHTg+34qxuLyAgcZ9YI7AJv1JsPxWBrVTdkUkGh+to1UMkkgrau7IRPsOgzC7ckZCu7TQ+3L4posvRBvfVkBi5fdHN7YyNqygL5mtuP5/Ybn+rpG7N3wR9N4qM58bRZjE9utFfYOyv2HH3tGvy7N47gZ0cOVOw5xKS7VYX5YLwij315mrbnEUCnlOHcUCf+9ubSY2tdM9kc/vzjfAzjb124gyMWHf7mSyP4w588XtLt1Bf72iBhmZrGvU17oug2qqo+u/Dm8W54I0lcEUnUnZj0GNXQKaWCDZXNB+PobtIEC56EZTBs1VWnQPZWJsGC57DpwZU5qHd7aQ0c17j9x0ATFsjeSAIznihGtrRX8E4faodKJhHNdDTHcbjgXMYLB4wwaxuzx6cWeoxqLIfjggfpA8DFu15Y9EocquLtZiJObx7vRiSRwffveMBxHN6+tYJzv38FX/0LJ0waOf70iy/gf33xBRztLj0iSaeU4VPdBlyp4aDezEoEB83Vf72fOWSGsUVOmcjbYBgGDptBsGJuPhRr2gSLrYatetxeClf8roXbF4VcylZsKHJ48w5D6W+g+L/byCfIjRtXsIMP3QEAwEjf0wWyWi7FmcNmvH1rBf/+c0M1j+ia8UTh8kbxMyeGa3odjcZmVCPHAUurcexrE+4WVjKTxVWXH58/Zm3qW5Ek70RfG7r0SnzrAze+fXUON+dX0dfegm/+1DN4bahTsNfIaL8Jf3BpFuFYGnq1TJDHLFQqk4PbF8WZw9W/YyKXsvixT3XhT3/4EA8DMWiUwv86U8pYqOX1+WvSbtPjv125h0Q6C6Ws9Na5cCyNSCLTtEtCtnLY9PjjiRxcvigOdeoq9jxuXxS9ppaKrfU2a5Xo1CnLegN1azGMTp2yoQ/v6vM7vwwTLj/0KhmOdG3/4n7DYcG4cxnX54LbnjJX07hzCSwDnBuqXHtFM+oxPspCFrJA/mguiFgqS+0VBEA+buvN4zb850sudOmV+NqbDnzhmFXwN94nB0z4/YuzuOb247O7LBKqhPuBdWRyXNUSLJ705nEb/njiPl7++uWKPP6/ONOPL39msCKPXWl2qx7pLIfplUhZdyn4geZmTrDg2a35/8fJhXCFC+R1HLFU7vGB/BuocjYDOhfDsDdwewXQZAUyx3GYcAXwUm/bju/MXhk0Qy2X4IJzuaYFMsdxuDC5jBd729Cubd7J4UrgC2Sh+5Av3vVCIWVxYpu7E6Q5/bNX+nHYosPpQ+ayTvF2w+e5X3FVv0CeXuFXTNemQB7q0uObP/UMPGvJijx+PRcAm+uRF8PlFciUgbyp19SCFnl+UO/Hn+2uyHMkM1k8CKzjjQpvtHVY9Xj3jgeRRBpaZXF3niKJNO751vH5T1krdHXi0FQF8sNgDIurcfziqd4dP0cpk+DThzvw9q1l/NaP1a7NYmolgnu+dXxxdOdrJaXp0CkhkzCCJllwHIdLU16c6GtriiQQUhiVXFLxolUmYfFib1tNBvVmPBFIWAa97ZWZti/EuTLy4RuZrVWFVrWsrFNCYOsJMhXILMtgyKqvaNTbg0AMOQ7oq3AqDP/m79biGl7ashOiELc3hvvq+Q1kIZpqSG/Cle8/PrHHyfCYw4JQLI1rG/3KtXDBuQQJy+C1oY6aXUOjkrAMbK1qQbOQ3b51PAzGqL2C1MTJARMeBmMVXYCznemVCPa3qSt2Ok5KxzAM7DZD2Xm388E4tEop9Krq9reLlcOqx53ltYoMeQOA27sR8VahBAveozsMxQ/q8W+6GnXFNE/QAjmSSOP/fPRw85aM2Ey4/ejUKdFr2v2049TBdmgUUow7a7M0hOM4jDuXcaKvDW1NHMxeSbZWlaCv08tT+Xi301QgkxoYHci/6b/iqu5WvVlvFIOU2CJadqsOM57IY1GDxZoPxWhAbwu7TY9UJodZT2U26vEZyJW+K9OmUcBqUJV0Gu5cDMNqUDV8fSJwgZzBr/7lJP72pjhi0rbK5Thcc/kx0m/ac3pcKZPg1SMdePv2SsXeJe7m9tIa7gdiGKtyP2Ez6TGqBS2QL055MNihpT49UhO9phZ06ZVVbbNIpLO4H1ivWf8x2ZvdakA2x+Hucul5t/PBGA3obVHOyWshXN4orAZVVdJT7FZ9SSunJxdWG/70GBC4QO4yqHB8Xysu1OjkdTd3V9YQiqUx0l9Yr82Y3YJwPI2JGuSLjk8ub7RXUHpFpXQb1QjF0lhLpMt+rLVEGh/fD9Uk6ooQIH87fXTAhGvuQNU2gbq8UXAcapZgQfbGL3EoNc6L4zgshOJ0grzF/rYWaBXSivUhu33rVevpt9v0uB+IIRwr/PdgOJ7G/UCs4fuPgQr0II/ZLbi7vLZ5m0Asrm30HxeaTHHyoAnaGrRZ8O0VI/0mtLbIq/rczeRRkkX5p8hXZvzI5DjqPyY1NdJvQjiersqmL+BRgsUAFciiZdEr0dYiL7mY80WTSGZyNKC3BcsyGC7x5HUvHMfB7YtWvP+Yx58C31oq/N9ye7E5+o+BChTIr2+0BbwlslPkCbcffe0t6NAVFmqtkErw6lAH3rm9glSmem0Wk4thPAzGcL7CES/Njj8RESLq7eKUBwa1DMfKiFIipFz8m/+rs9XpQ57xRCCXsNjfRsWTWOUH9Uov5vifj9Ri8Ti7TY+7yxHBa4OVtQRiqWxJa+dLsTUKsFBOKpBL16lX4rn94mqzSGVy+KiExR/nHRasJTK4WsXBl3HnMmQSBq8dofaKShLqBDmb4/D+tA+nDrbXfPMiaW4mjQJHLDpcqVIf8rQngj6zhl73Iuew6jHjiSCeKn5Qb4GPeKMWi8fYrXqksjnMeCKCPq7buw6g8gkWvNYWObqNqqKiACcXw+g2qpriDndFfrKdd3Rh2hPBrMAvnlLdXFhFLJUteoHDaH87dEpp1Yp9juNwwbmM0X5T1VfGNhu9WgatUlp21NvNhVUE1lPUXkFE4eSACTcehrCezFT8uWZWIhjsqM4vclI6u82AHAfcWS7+FJmWhGyP7+0Wug/Z5c3XTH3m6uWK2616OIsYOJxcCDfF6TFQoQL5s8OdYJj8sJkYTLj8YBngpd7iwrDlUhavDXXi3dseJDOlx+QU6uZCGIurcYw5uir+XCR/ilzuspDLU16wTD4akJBaGx0wIZ3l8NFcsKLPs5ZIYymcwEGKeBM9vpgppZibD8Zh0iho+dETeoxq6JRSwfv93b51aJVStFcxPs1uNWA+GEdoPbXn567GUngYjG2u3G50FSmQzTolnt9vrFmO8JOuuQIYtupLOpUdc1gQSWZwZabyty3HnUuQS1i8eoSWg1RDd2v5UW+Xprw4vq8VBnXj324i4vfcfiPkUrbibRb83UFKsBC/Dp0C7VpFScXcfIgi3rbD93YLHfXm9kXRb9bsGUUrJP40vJBBPf415GiCBAuggpv0zjssmPVGBe/RKVYslcEn86Gi2yt4I/0m6FUyXHBWNts5l8unV7x80EQbi6qkp02N+VAcuRJjsVbCCdxeWsOZQ/SGhoiDUibB8/uNFZ+bmNlYkkAZyOLHMAwcVn1JK6dpScjO7FYDplcigt5drmaCBW+4q/A7DHyBzP+dRlexAvncsAUsA1yo8dKQj+aCSGe5gvOPnySTsDg31Il373jK2ka0l0/mV7EUTmCM0iuqprtVhVQmB180WdLfvzyd355H/cdETEYHTJjxROFZS1TsOaZXImiRS2A10OliPbDb9HD5okX1pmeyOSytJmBrpa/xdhw2PdJZbjPusFxriTQ8a8mqF8h6tQz72tQFvYGaXAhjX5u6aWakKlYgt2sVeLG3DRcml8Fx1Qmu3841dwByCYtn9xlLfozzRy1YT2Xx/kzlTmXGncuQS1l8+jCdRlYLn+1Zah/yxbteWA0qHKRBJSIio5txb5Vrs5jxRNDfoQXLVu9WMCmd3aoHxwF3itiotxxOIJvjKAN5B+X0dm/nno9PsKjegB7PbtUX1ILjbKIBPaCCBTKQ79+951vHlEDvsEox4fLjmX2GsoYMXuptQ6taVrGe6lyOw1uTyzh1sB1aZXO8MxOD7jKi3hLpLCZcfpw5ZK5qvxghezli0aGtRY6rFdwCOuOhBIt6UkoxN08Rb7uytapgUMtKal3Zjtubb1vqq1IG8lYOmx6Lq3EEdrmbGlxPYXE13jT9x0CFC+RzQ51gGdRsWC+0nsKd5TWMlNh/zJNKWJwbtuAHdyvTZnHjYQgrawlaDlJlVoMKDFPaCfIP7wUQT2epvYKIDssyONFvwlWXvyJ37/zRJPzRFPUf1xGzTolOnRKTC4UPlS3QkpBdMQyzEZEmUIHsi0ImYTYz+qtpuICFIZv9x3SCLIw2jQIn+kwYr1GbxYf3AuA44ESRC0K2c95hQSyVxeUprwBX9rgLzmUopCzOUntFVSllEnRolSVt07s85YVSxuKlvtJ62wmppJP9JvgiSUxXYEiaH7wepIi3ujJcZDE3H4qBZYAu6jPfkcOmx6wnIsjBmdsXxb62FshqsHhns0De5TScf3NFBbKAxhwWzPnXcXup8N4noUy4/NAopDgqwC2BFw4Y0dYixwWBs52zG+0VpwfN0Cikgj422VuPsfioN47jcGnai5E+E5Qyygcl4jM6ULk+5JkVinirRw6bHnP+dUQS6YI+fyEUh0WvqknBVi/sVj0yOQ53i+jt3onLG61J/zEA6JQy9Jpadj1Bdi6E0Wtqga6J2kAr/so/N9QJCcvUZGnINXcALxwwCrIKVSph8Vl7Jy7d9SKWEm5L1cf3g/BGkpReUSM2o6robXoubxTzwTjOHKb2CiJOXQYVettbKtKHPO2JwqCWoV1bvWUGpHx2W35Qr9DDqvlgjBIs9mC35RdmlLswJJ3N4UEgVvUEi62G9xjUu7UYhr2J+o+BKhTIrS1yjPSbMO6sbpvF0mocc/51QdoreGP2LsTTWVyeEi7NYnxyGUoZS72sNdJjVGNlLVFUluWljTab04P0NSPidbLfhOv3goJvAZ31RHCwQ0vDqXXGXsBt9K3yS0JoQG83XXol2lrkZSdZPAzGkMlx6K/BgB7PYdNjOZyAN/J0PKQvksRSONFUCRZAFQpkADhvt+BhMIZbi9Vrs5jYODkpNf94O88fMMKkUWB8Uphs53x7xQrOHDKjhdoraqLHqAbHAYuhwvuQL055cdiio948ImqjA+2Ip7O48UC4bV8cx2HaE6H2ijpk0ijQpVcWdNqZSGfhWUtSgsUe+I16t8o8Qd5MsKjhCTJf/G73b+E/RgVyBXxmqANSlsEFgQrLQlxzB2DSyAX9QS5hGbxu78SlKW9Rges7uT4XgD+axHlHlwBXR0pRbBZyOJbG3z0I4cyh9kpeFiFle7HXCAnLCLpVb2UtgUgig4M0oFeX8uuR9y7mFlcpwaJQdqseM54I4qnS79S4fPkCubdGPcgAMGTVg2GAyYWnDzInF8NgmPznNJOqFMgGtRyjA9Vrs+A4DhMuP17qMwl+G3DMbkEincNFAdIsxp3LUMkkdKu+hvhInfkCT5Dfn/Uhm+OoJYaInlYpw7Fug6CDevzWsIM1vBVMSuewGTDnX0c4vvugHj+4TC0We7Nb9chxwJ3l0k+R3d51dOgUNd2DoFFINwb1nr7jxA/oNVuQQNXGU887urAQiuOmQKHau3H7ovBGkhipQATXc/uNMGsVGHeWdxqeyebw9q0VnD1sLmuJCSlPu0YBuZQtOMni8pQXrWoZPtXdWuErI6R8owMmOBfDWI2lBHk8PuKNMpDrEx/RdXuPU2T+wIBaLPbm4Af1yqht3L5oTdsreA6bYdt+6snF1c1/ZzOpWoH86pEOyCRM2YVlISZcAQDAiIADejyWZfC63YLL0z5Ey2izuD4XRGA9RctBaoxlGXS3qgoqkLM5Du9Ne/HKoBkSWrFL6sDJARM4Lt9yJoTplSjMWgVaW+SCPB6pLnsBCyEAYCEYg1zCwkxJJXvq0Clg0ihKXhjCcRzcvmhNB/R4dqse3kgSnrVHg3retQQ8a8mm6z8Gqlgg61UyvDzQXpU2iwmXH91GVcVuD513WJDK5HDxrqfkx7jgXEaLXIJXqL2i5rqN6oJ6kP9+PoRQLE3tFaRuHLUZoFVIcUWgNosZT4QWhNQxY4sctlbVnsXcfCgGa6sKLB0E7IlhGDhs+pJPkH3RJCKJjEhOkJ9OOuHfTDXTimleVRPAxxwWLIUT+GReuKnqJ2WyOXx4L4DRCpwe857paUWnTonv3iwt2zmdzeHtW8v49JEOWjQhAj0FFsiXpryQsAxePkgDeqQ+SCUsXuxrE2RQL5fjMOuNUHtFnSukmJsPxikDuQh2qx4uX7Sk4X2XCBIseEe6dGAZPPYGyrkQBsvk/6zZVLVA/vSRDsglLC6UWFgW4tbSGiKJDE70Va5A5tssPpjxYa3ArURbfegOIBRLY8xO7RVi0N2qRiSRQTi2+9fy4l0vju9rhV7VPJuESP07OWDCfDCOB4H1sh5nPhRDIp2jiLc6N2zV42EwtuvPO8pALo5jYwnLnRI26rl9+e/LPnPtEix4arkU/WbN5lppIH+C3G/WQC1vrgE9oMoFsk4pw6nBdrw1uYxcrjJtFnz+8YkKDOhtdf6oBalsDj+4U3ybxbhzGRqFlE4iRaKQqLel1TimViI4S+0VpM7wd9PKbbPYTLCgFou65rDuvv0tkkhjNZamAb0i8P25pSwMcXujaJFL0KlTCn1ZJbFbDZhcXAPHceA4DpOLYditzTegB1S5QAby/bsrawnceBiqyONfc/txqFOLNk1lhwuOdRtgNagw7izuNDydzeHt2yt4ldorRIPP+txt5fTl6XysH/Ufk3pzwNQCq0FVdtwbn2AxIIJhIlK6zWJumzgvAFgIUQZyscw6JTp0isdOXgvl9kXRZ9aIZjOl3aqDP5rEysZwni+ShN3afO0VQA0K5LOHOyCXsrhQZGFZiEQ6i4/vhyqSXvEkhskvDflg1rdnpuRWEy4/wnFqrxCTQk6QL931otuoEsWkMSHFYBgGI/1tuOb2I1vGnbtpTxS2VhVt/axzerUM+9rUO/Yhb2Yg0wlyUfInr6WdIIuh/5hn34hzcy6E4dwo+O1NGPEG1KBA1iikOF2hNosbD0JIZnKCrpfezZijC+ksh+/fXin471xwLkOrlOLkwcoX8aQwOqUMBrVsx6i3RDqLCbcfZwbNonmXT0gxRgfasZbIbP7CK8XMCq2YbhTD1p036m1mIFMPclEcNj3u+dcRKWIuaT2ZwVI4gb4abtB70hGLDhKWweRCGJOLYUhYBkcsdIJcNWOOLngjSfzoflDQx51w+yFlGTx/oDoF8lGbHrZWFcYnCzsNT2VyeOf2Cj5zpBMKKbVXiMluSRYfugNIpHM4c7ijyldFiDD4pUmltlmkMjm4fVHqP24QDqseC6E4gutPL5CZD8bQIpegVU3DyMWwW/ODereXCh/Um/NvDOiJ6ARZJZdgwKzB5GK+QB4wa5p2mVlNCuSzh8xQytiCC8tCTbgCONptqNo6RIZhMOaw4Oqsv6BNVVddPkQSGVoOIkLdrerN3rsnXZryQiWT4IUDxipfFSHCaNMoMNSlwxVXaQXy/cA6MjmOTpAbhN2288KQhY0EC7pbVhx+S2ExechuXz7iTWyte/aNOwyTC+GmXBDCq0mB3KKQ4swhM96aXCmrJ26rtUQazoXViqyX3s15excyOQ7fv713msUF5zJ0SmlVeqRJcbqNaiyEYk+9HjmOw6UpL0b6TTRUSera6IAJnzwMlZTVuplgQQVyQ+CLuVvbFMj5DGRqryhWu1aBLr2yqD5ktzcKCcugp01c/98Omx7B9RQC66mmXBDCq0mBDABj9i74o0l8NCdMm8X1e0HkOOBElYvPYasOPUY1vrvHCu1EOot3b3vw2lAn5NKa/beTHXQbVUhnucdWbALAjCeKxdU4zh6m9ApS3072tyOd5XB9rvi10zOeCCQsg14R9UqS0umUMhwwtTzVk85xHOZDMVoSUqLderu34/JF0WNUi67lcutQXrMO6AE1LJBPH2qHSibB+OTuhWWhJlx+KGUsjvVU94vJt1lccwe27efiXZn1I5LMYIzaK0SpZ4cki0tT+Xi307QSnNS5Z/e3QiFlS8pDnvFEsL9NTXdRGojd+vRGveB6CrFUlgb0SuSw6THnXy94gZjbuy6qAT3eoU4tpCwDKcvgUBPPHdSsQFbLpThz2IzvTa4gk82V/XjX3H48t99Yk3diY3YLsjkO7+ySZjHuXIJBLaP2CpHiI42eTLK4NOXBUJcOnXpxhLgTUiqlTILnDxhLGtSb8UQx2MS/KBuRw6bHUjgBfzS5+bHNBAs6QS4Jf9q6XevKk7I5DnP+dfSJrP8YyP+sGOzUYrBT29Rvimt6r/8NhwWB9RSul9lm4Y0kMOOJ1qz4HOrS4YCpZcelIYl0Fu/e8eDcUCdkEmqvEKMugwos83iBvBpL4e8ehGg5CGkYo/0mzHqjWAkn9v7kDYl0FvcD6xgwU4HcSDaHyrYUc5sZyHSCXBJ7EYN6C6EYUtmcqBIstvr6m0fx9TeP1voyaqqm1dorg2ao5ZKyl4Z86M731I301aZAZhgGY3YLrrn9CGx5N857f8aH9VSW2itETC5lYdGrNk9QgPzXLcfR9jzSOEYH8j8jrxaRZuHyRsFxoBPkBjPUpQPDPF7M8dtEqUAujbFFDqtBBWcBJ8gubz7BQqwF8pEuHY50NWf+Ma+mBbJSJsGnD3fg7VvLZbVZTLj80KtkNf1ijjksyHHA9yRguRMAAAqhSURBVG493WZxwbmMVrUML/VWN2GDFKfbqHqsB/nSlBdtLXIcbeIhBdJYDnfq0NYix9VZX8F/hxIsGpNWKUOvqQXOrQVyMI5WtaxqUamNyGF7urd7O3zEmxh7kEleze/3jzksCMXS+PBe8ZPVQH7qdsIVwEu9bZCwtcttPNSpRW/7020W8VQWF+96cG7YAim1V4had6t68xZjJpvDe9M+nBpsB1vD1xUhQmJZBiP9Jlx1BcBxhUVszngikEtY7BdZFBUpXz7v9lGSBZ+BTEpnt+nxMBhDOLb7oJ7buw6TRg6DWl6lKyPFqnnFdupgOzQKKS7cLK3N4mEwhsXVeNXWS++EYRict1twfS4Ab+RRf997017EUlm8Qe0VotdjVMMbSSKRzuKT+VWE42mcPUTb80hjGR0wwR9NYmrjZHgv054I+swaeoPfgOw2AzxrSXg34i0XQvHNgWVSGvs2vd3bcfuiom2vIHk1/4mnlEnw6pEOvH17BekS2iwmXPmT52rnH2/n/NEu5DjgnS1tFhcml2HSyPE8bWETPT6sfSEUw6UpL6Qsg5MHa/+6IkRIJ/k+5ALTLGZWIhjsoF/kjcixZaNeLsdhMRSHzUgJFuXgC2Tn4uqun+f2RUWZYEEeqXmBDORj0sLxNCZKWIM64fajU6dEr6n2fTwHO7QYMGs2hw5jqQwu3fXi3HAnnb7UAX571MNgDJfuevHcfiN0SlmNr4oQYVn0KvS1txS0djqSSGMpnMBBGtBrSEcsOrAM4FwIwxNJIJXN0QlymQxqOXqM6l37kAPRJEKxNJ0gi5woqraTB03QKqQ7xqTtJJfj8KE7gBP9baLZGz/msOCj+0F41hK4NOVFPJ3FmL2r1pdFCsAvC/nhvSCmPRFKryAN6+RAOz6aCyCZye76eTOe/CDRIA3oNaQWhRR97RrcWgxjPriRgUw9yGWz23bfqOf2rQOgAT2xE0WBrJBK8OpQB965vYJUpvA2i6mVCILrqZrFu23nvMMCjgO+N7mMcecy2rUKaq+oEyaNHCqZBP/3R/MAgDO0Xpo0qNF+ExLpHP7uQWjXz5vxUIJFo7Pb9HAuhjcHlGnNdPnsVj0WQvEdt+vyCRb91GIhaqIokIF8YbmWyOCqq/D4oWvu/C1CMW2n6zdrcahTi7+4sYBLU168PtxZ03QNUjiGYdBtVCEcT2Nfm1oUbTuEVMKLfW2QssyefcjTKxGo5RJYDVQ0NSqHVQ9fJImPH+QXdtHXunyOPQb13N4olDIWXXr6vxYz0RTIo/3t0CmlRS0NmXD50dfeIro1wGN2C24triGZyWHMQe0V9YTvvzs9aBZN2w4hQtMopDjWY9hzYciMJ4KBDi1FHTYwfj3yO7c96NApmnq1sFCGNgrknVZOu3xR9Jo09H0lcqIpkOVSFq8NdeLd2549++IAIJXJ4fpcUFSnx7zXNyLdzFoFnt3XWuOrIcXg++/OUnsFaXCj/e2YXAwjtMNtYCBfIFOCRWPjB/WC6yka0BOIXiXDAVMLnAvbJ1lQgkV9EE2BDOQH3CLJDK7M7D1dfXNhFbFUFidE1H/M62vX4NxQJ35u9AC9Q6wzJwdMONZjoL5x0vBGB0zgOOCae/slTf5oEv5oivqPG5xKLtn8GtOAnnCGrdtv1Euks1gIxWlArw6IqkAe6TdBr5JhfHLvNosJlx8sA9Gub/7mTx/HL57qq/VlkCKdPdyBv/rnI1BI6TYjaWxHbXpoldId5z74Ab1BinhreHx2bzcN6AnGYdVjKZyAP5p87ONz/nVwHA3o1QNRFcgyCYtzQ514944HifTubRbXXAEMW/XQqymnlhBCiiWVsHiptw1XZv3brp2e2di0RxFvjc++sTDERifIgrHbth/U4xMsKANZ/ERVIAPA+aMWRJMZvD+zc5pFLJXBJ/MhUbZXEEJIvTg5YMJCKI4HgdhTfzbjjcKglqFdq6jBlZFqOtFnglouwdGNgT1SvqEuHQA81Wbh8kbBMMABSkkSPdEVyC/1tqFVLdt1achHc0GksxxG+sXZXkEIIfWAH3LebqvezEoEBzu0lObSBPrNGtz5zXPUTiMgrVKG3vYWOBeePEFeh61VRWkhdUB0BbJUwuLcsAU/uLtzm8U1dwByCYtn99EgFSGElOqAqQVWgwpXZx+/Y8dxHKY9ERykBAtCSuaw6p+KenN7o+in9oq6ILoCGcgvDYmlsnhv2rvtn0+4/HhmnwEqOb0DI4SQUjEMg9F+E665A8hkH20xXVlLIJLIUP8xIWWw2wxYWUvAu5YAAORyHO75o9R/XCdEWSC/cMCIthY5vrtNm0VoPYU7y2uiWi9NCCH1anTAhEgiA+eWk67pFVoxTUi57E9s1FtcjSORzlEGcp0QZYEslbD4rL0Tl+56EUtlHvuzD+8FwHHACREuCCGEkHoz0m8Cw+CxtdN8xBsVyISUbqhLB4bBZh8yJVjUF1EWyAAwZu9CPJ3F5anHe+MmXH5oFFIc3YhQIYQQUjpjixxDXbrHCuTplSjMWgVaW+Q1vDJC6luLQor+ds1mH7Lbtw4AtCSkToi2QH7+gBEmjQLjk0uPffyaO4AXDhghlYj20gkhpK6M9rfjxsMQosn8HbsZT4QSDQgRgN2qh3MxDI7j4PZF0aqWoU1D0Yn1QLRVpoRl8Lq9E5emvFjf+KG9tBrHnH+d2isIIURAJwdMyOQ4XL8XQC7HYdYbofYKQgRgt+nhiyThWUvC7aUBvXoi2gIZAMbsFiTSOVycyqdZTGxkdVL+MSGECOf4vlYopCyuzPoxH4ohkc5RggUhAnBstIM6F1bh9lGBXE9EXSA/u98Is1aBcWe+zeKaOwCTRk4/uAkhREBKmQTPHzDiqsv/KMGCWiwIKdsRix4sA1x1+eGPptBnpv7jeiHqAjnfZmHB5WkfIok0Jlx+vNRnos1OhBAisJMDJri8Ubw/kx+MHqAoKkLKppJLMGDW4sJGbG0/fV/VDVEXyEB+aUgqk8O3PrgHbySJkT5qryCEEKGN9rcDAP7yxiJsrSq0KKQ1viJCGoPdpkdwPQWAIt7qiegL5Gd6WtGpU+KP3r8HIJ/ZSQghRFiHOrUwaeSIp7PUxkaIgPg+ZLmEha1VXeOrIYUSfYHMbrRZpLI5dBtV6DbSi4sQQoTGsszmAQT1HxMiHH6j3gFTCyQstYjWC9EXyAAw5rAAAK2XJoSQChrdKJDpBJkQ4Ry26CBhGRrQqzN10WT2TI8BXzrdhzeOdtX6UgghpGGdG+7E7aU1nB401/pSCGkYSpkE//b1wxju0tX6UkgR6qJAZhgG/+a1Q7W+DEIIaWhapQy/8bmhWl8GIQ3n50cP1PoSSJHqosWCEEIIIYSQaqECmRBCCCGEkC2oQCaEEEIIIWQLKpAJIYQQQgjZggpkQgghhBBCtqACmRBCCCGEkC2oQCaEEEIIIWQLhuO4wj+ZYXwAHlTucgghAPZxHNde7oPQ9yshVVP29yx9vxJSNQV9vxZVIBNCCCGEENLoqMWCEEIIIYSQLahAJoQQQgghZAsqkAkhhBBCCNmCCmRCCCGEEEK2oAKZEEIIIYSQLahAJoQQQgghZAsqkAkhhBBCCNmCCmRCCCGEEEK2oAKZEEIIIYSQLf4/uo8WxnLK4n0AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 720x720 with 9 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "db.show_batch()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torchtimeseries.models import *\n",
    "# Select one arch from these state-of-the-art time series/ 1D models:\n",
    "# ResCNN, FCN, InceptionTime, ResNet\n",
    "arch = InceptionTime                     # ✳️   \n",
    "arch_kwargs = dict()                     # ✳️ \n",
    "opt_func=Ranger                          # ✳️ a state-of-the-art optimizer\n",
    "loss_func = FlattenedLoss(nn.L1Loss)     # ✳️ this is MAE loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "InceptionTime(\n",
      "  (block): InceptionBlock(\n",
      "    (inc_mods): ModuleList(\n",
      "      (0): Inception(\n",
      "        (conv_layers): ModuleList(\n",
      "          (0): Conv1d(1, 32, kernel_size=(39,), stride=(1,), padding=(19,))\n",
      "          (1): Conv1d(1, 32, kernel_size=(19,), stride=(1,), padding=(9,))\n",
      "          (2): Conv1d(1, 32, kernel_size=(9,), stride=(1,), padding=(4,))\n",
      "        )\n",
      "        (maxpool): MaxPool1d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=False)\n",
      "        (conv): Conv1d(1, 32, kernel_size=(1,), stride=(1,))\n",
      "        (bn): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (act): ReLU()\n",
      "      )\n",
      "      (1): Inception(\n",
      "        (bottleneck): Conv1d(128, 32, kernel_size=(1,), stride=(1,))\n",
      "        (conv_layers): ModuleList(\n",
      "          (0): Conv1d(32, 32, kernel_size=(39,), stride=(1,), padding=(19,))\n",
      "          (1): Conv1d(32, 32, kernel_size=(19,), stride=(1,), padding=(9,))\n",
      "          (2): Conv1d(32, 32, kernel_size=(9,), stride=(1,), padding=(4,))\n",
      "        )\n",
      "        (maxpool): MaxPool1d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=False)\n",
      "        (conv): Conv1d(128, 32, kernel_size=(1,), stride=(1,))\n",
      "        (bn): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (act): ReLU()\n",
      "      )\n",
      "      (2): Inception(\n",
      "        (bottleneck): Conv1d(128, 32, kernel_size=(1,), stride=(1,))\n",
      "        (conv_layers): ModuleList(\n",
      "          (0): Conv1d(32, 32, kernel_size=(39,), stride=(1,), padding=(19,))\n",
      "          (1): Conv1d(32, 32, kernel_size=(19,), stride=(1,), padding=(9,))\n",
      "          (2): Conv1d(32, 32, kernel_size=(9,), stride=(1,), padding=(4,))\n",
      "        )\n",
      "        (maxpool): MaxPool1d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=False)\n",
      "        (conv): Conv1d(128, 32, kernel_size=(1,), stride=(1,))\n",
      "        (bn): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (act): ReLU()\n",
      "      )\n",
      "      (3): Inception(\n",
      "        (bottleneck): Conv1d(128, 32, kernel_size=(1,), stride=(1,))\n",
      "        (conv_layers): ModuleList(\n",
      "          (0): Conv1d(32, 32, kernel_size=(39,), stride=(1,), padding=(19,))\n",
      "          (1): Conv1d(32, 32, kernel_size=(19,), stride=(1,), padding=(9,))\n",
      "          (2): Conv1d(32, 32, kernel_size=(9,), stride=(1,), padding=(4,))\n",
      "        )\n",
      "        (maxpool): MaxPool1d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=False)\n",
      "        (conv): Conv1d(128, 32, kernel_size=(1,), stride=(1,))\n",
      "        (bn): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (act): ReLU()\n",
      "      )\n",
      "      (4): Inception(\n",
      "        (bottleneck): Conv1d(128, 32, kernel_size=(1,), stride=(1,))\n",
      "        (conv_layers): ModuleList(\n",
      "          (0): Conv1d(32, 32, kernel_size=(39,), stride=(1,), padding=(19,))\n",
      "          (1): Conv1d(32, 32, kernel_size=(19,), stride=(1,), padding=(9,))\n",
      "          (2): Conv1d(32, 32, kernel_size=(9,), stride=(1,), padding=(4,))\n",
      "        )\n",
      "        (maxpool): MaxPool1d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=False)\n",
      "        (conv): Conv1d(128, 32, kernel_size=(1,), stride=(1,))\n",
      "        (bn): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (act): ReLU()\n",
      "      )\n",
      "      (5): Inception(\n",
      "        (bottleneck): Conv1d(128, 32, kernel_size=(1,), stride=(1,))\n",
      "        (conv_layers): ModuleList(\n",
      "          (0): Conv1d(32, 32, kernel_size=(39,), stride=(1,), padding=(19,))\n",
      "          (1): Conv1d(32, 32, kernel_size=(19,), stride=(1,), padding=(9,))\n",
      "          (2): Conv1d(32, 32, kernel_size=(9,), stride=(1,), padding=(4,))\n",
      "        )\n",
      "        (maxpool): MaxPool1d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=False)\n",
      "        (conv): Conv1d(128, 32, kernel_size=(1,), stride=(1,))\n",
      "        (bn): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (act): ReLU()\n",
      "      )\n",
      "    )\n",
      "    (res_layers): ModuleList(\n",
      "      (0): None\n",
      "      (1): None\n",
      "      (2): Sequential(\n",
      "        (0): Conv1d(1, 128, kernel_size=(1,), stride=(1,))\n",
      "        (1): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "      (3): None\n",
      "      (4): None\n",
      "      (5): Sequential(\n",
      "        (0): Conv1d(128, 128, kernel_size=(1,), stride=(1,))\n",
      "        (1): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "    )\n",
      "    (act): ReLU()\n",
      "  )\n",
      "  (gap): AdaptiveAvgPool1d(output_size=1)\n",
      "  (fc): Linear(in_features=128, out_features=1, bias=True)\n",
      ")\n",
      "InceptionTime\n",
      "======================================================================\n",
      "Layer (type)         Output Shape         Param #    Trainable \n",
      "======================================================================\n",
      "Conv1d               [32, 10]             1,280      True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             640        True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             320        True      \n",
      "______________________________________________________________________\n",
      "MaxPool1d            [1, 10]              0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             64         True      \n",
      "______________________________________________________________________\n",
      "BatchNorm1d          [128, 10]            256        True      \n",
      "______________________________________________________________________\n",
      "ReLU                 [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             4,128      True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             39,968     True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             19,488     True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             9,248      True      \n",
      "______________________________________________________________________\n",
      "MaxPool1d            [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             4,128      True      \n",
      "______________________________________________________________________\n",
      "BatchNorm1d          [128, 10]            256        True      \n",
      "______________________________________________________________________\n",
      "ReLU                 [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             4,128      True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             39,968     True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             19,488     True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             9,248      True      \n",
      "______________________________________________________________________\n",
      "MaxPool1d            [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             4,128      True      \n",
      "______________________________________________________________________\n",
      "BatchNorm1d          [128, 10]            256        True      \n",
      "______________________________________________________________________\n",
      "ReLU                 [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             4,128      True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             39,968     True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             19,488     True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             9,248      True      \n",
      "______________________________________________________________________\n",
      "MaxPool1d            [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             4,128      True      \n",
      "______________________________________________________________________\n",
      "BatchNorm1d          [128, 10]            256        True      \n",
      "______________________________________________________________________\n",
      "ReLU                 [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             4,128      True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             39,968     True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             19,488     True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             9,248      True      \n",
      "______________________________________________________________________\n",
      "MaxPool1d            [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             4,128      True      \n",
      "______________________________________________________________________\n",
      "BatchNorm1d          [128, 10]            256        True      \n",
      "______________________________________________________________________\n",
      "ReLU                 [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             4,128      True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             39,968     True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             19,488     True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             9,248      True      \n",
      "______________________________________________________________________\n",
      "MaxPool1d            [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [32, 10]             4,128      True      \n",
      "______________________________________________________________________\n",
      "BatchNorm1d          [128, 10]            256        True      \n",
      "______________________________________________________________________\n",
      "ReLU                 [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "Conv1d               [128, 10]            256        True      \n",
      "______________________________________________________________________\n",
      "BatchNorm1d          [128, 10]            256        True      \n",
      "______________________________________________________________________\n",
      "Conv1d               [128, 10]            16,512     True      \n",
      "______________________________________________________________________\n",
      "BatchNorm1d          [128, 10]            256        True      \n",
      "______________________________________________________________________\n",
      "ReLU                 [128, 10]            0          False     \n",
      "______________________________________________________________________\n",
      "AdaptiveAvgPool1d    [128, 1]             0          False     \n",
      "______________________________________________________________________\n",
      "Linear               [1]                  129        True      \n",
      "______________________________________________________________________\n",
      "\n",
      "Total params: 406,049\n",
      "Total trainable params: 406,049\n",
      "Total non-trainable params: 0\n",
      "Optimized with 92695a7b8\n",
      "Using true weight decay as discussed in https://www.fast.ai/2018/07/02/adam-weight-decay/ \n",
      "Loss function : FlattenedLoss\n",
      "======================================================================\n",
      "Callbacks functions applied \n",
      "\n"
     ]
    }
   ],
   "source": [
    "model = arch(db.features, 1, **arch_kwargs).to(device) # c_out must be 1 for regression\n",
    "learn = Learner(db, model, opt_func=opt_func, loss_func=loss_func)\n",
    "learn.save('stage_0')\n",
    "print(learn.model)\n",
    "print(learn.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LR Finder is complete, type {learner_name}.recorder.plot() to see the graph.\n",
      "Min numerical gradient: 1.74E-01\n",
      "Min loss divided by 10: 3.31E-01\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "learn.load('stage_0')\n",
    "learn.lr_find()\n",
    "learn.recorder.plot(suggestion=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [],
   "source": [
    "epochs = 25         # ✳️ \n",
    "max_lr = 1e-1        # ✳️ \n",
    "warmup = False       # ✳️\n",
    "pct_start = .7       # ✳️\n",
    "metrics = [mean_absolute_error] # ✳️\n",
    "wd = 1e-2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>mean_absolute_error</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>8.241958</td>\n",
       "      <td>7.853509</td>\n",
       "      <td>7.854370</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>7.489723</td>\n",
       "      <td>7.580690</td>\n",
       "      <td>7.582139</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>6.972783</td>\n",
       "      <td>7.335127</td>\n",
       "      <td>7.336818</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>6.572676</td>\n",
       "      <td>7.087035</td>\n",
       "      <td>7.088806</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>6.233433</td>\n",
       "      <td>6.835031</td>\n",
       "      <td>6.836416</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>5.959407</td>\n",
       "      <td>6.741987</td>\n",
       "      <td>6.740504</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>5.755314</td>\n",
       "      <td>6.627546</td>\n",
       "      <td>6.627869</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>5.603128</td>\n",
       "      <td>6.478719</td>\n",
       "      <td>6.497616</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>5.475894</td>\n",
       "      <td>6.354661</td>\n",
       "      <td>6.393384</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>5.379145</td>\n",
       "      <td>6.393933</td>\n",
       "      <td>6.416293</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>10</td>\n",
       "      <td>5.292003</td>\n",
       "      <td>6.273191</td>\n",
       "      <td>6.292341</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>11</td>\n",
       "      <td>5.221508</td>\n",
       "      <td>6.126951</td>\n",
       "      <td>6.161241</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>12</td>\n",
       "      <td>5.161124</td>\n",
       "      <td>5.966456</td>\n",
       "      <td>6.029862</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>13</td>\n",
       "      <td>5.112274</td>\n",
       "      <td>5.779326</td>\n",
       "      <td>5.861611</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>14</td>\n",
       "      <td>5.067443</td>\n",
       "      <td>5.837413</td>\n",
       "      <td>5.888563</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15</td>\n",
       "      <td>5.026354</td>\n",
       "      <td>5.663989</td>\n",
       "      <td>5.717026</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>16</td>\n",
       "      <td>4.991047</td>\n",
       "      <td>5.448196</td>\n",
       "      <td>5.517660</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>17</td>\n",
       "      <td>4.958901</td>\n",
       "      <td>5.192246</td>\n",
       "      <td>5.267992</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>18</td>\n",
       "      <td>4.930110</td>\n",
       "      <td>4.935892</td>\n",
       "      <td>5.081102</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>19</td>\n",
       "      <td>4.904392</td>\n",
       "      <td>5.186227</td>\n",
       "      <td>5.259577</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>20</td>\n",
       "      <td>4.881000</td>\n",
       "      <td>5.017474</td>\n",
       "      <td>5.102685</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>21</td>\n",
       "      <td>4.859765</td>\n",
       "      <td>4.889932</td>\n",
       "      <td>5.011947</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>22</td>\n",
       "      <td>4.840405</td>\n",
       "      <td>4.816056</td>\n",
       "      <td>4.939561</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>23</td>\n",
       "      <td>4.822685</td>\n",
       "      <td>4.777628</td>\n",
       "      <td>4.891876</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>24</td>\n",
       "      <td>4.806410</td>\n",
       "      <td>4.839561</td>\n",
       "      <td>4.962036</td>\n",
       "      <td>00:00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "learn.metrics = metrics\n",
    "learn.load('stage_0')\n",
    "learn.fit_one_cycle(epochs, max_lr=max_lr, pct_start=pct_start, moms=(.95, .85) if warmup else (.95, .95),\n",
    "                    div_factor=25.0 if warmup else 1., wd=wd)\n",
    "learn.save('stage_1')\n",
    "learn.recorder.plot_lr()\n",
    "learn.recorder.plot_losses()\n",
    "learn.recorder.plot_metrics()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "preds, actual = learn.get_preds()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[8.2830],\n",
       "        [8.3600],\n",
       "        [8.2470],\n",
       "        [8.4015],\n",
       "        [8.4785],\n",
       "        [8.5241],\n",
       "        [8.5682],\n",
       "        [8.2123],\n",
       "        [8.1035],\n",
       "        [8.2057]])"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "preds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([ 0.,  9.,  2.,  9.,  8., 15., 20.,  2., 10.,  6.])"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "actual"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[[17., 12., 12., 12., ...,  1., 12.,  7.,  7.]],\n",
       "\n",
       "       [[12., 12., 12.,  7., ..., 12.,  7.,  7.,  0.]],\n",
       "\n",
       "       [[12., 12.,  7.,  8., ...,  7.,  7.,  0.,  9.]],\n",
       "\n",
       "       [[12.,  7.,  8.,  1., ...,  7.,  0.,  9.,  2.]],\n",
       "\n",
       "       ...,\n",
       "\n",
       "       [[ 1., 12.,  7.,  7., ...,  2.,  9.,  8., 15.]],\n",
       "\n",
       "       [[12.,  7.,  7.,  0., ...,  9.,  8., 15., 20.]],\n",
       "\n",
       "       [[ 7.,  7.,  0.,  9., ...,  8., 15., 20.,  2.]],\n",
       "\n",
       "       [[ 7.,  0.,  9.,  2., ..., 15., 20.,  2., 10.]]], dtype=float32)"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_small"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}