{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPython 3.7.3\n",
      "IPython 7.5.0\n",
      "\n",
      "numpy 1.16.3\n",
      "scipy 1.2.1\n",
      "sklearn 0.21.1\n",
      "pandas 0.24.2\n",
      "matplotlib 3.0.3\n"
     ]
    }
   ],
   "source": [
    "%load_ext watermark\n",
    "%watermark -v -p numpy,scipy,sklearn,pandas,matplotlib"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**7장 – 앙상블 학습과 랜덤 포레스트**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "_이 노트북은 7장에 있는 모든 샘플 코드와 연습문제 해답을 가지고 있습니다._\n",
    "\n",
    "-여기에서 발생하는 경고는 NumPy에 관련된 것([#10449](https://github.com/scikit-learn/scikit-learn/issues/10449))으로 향후 버전에서는 사라질 것입니다._"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 설정"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "파이썬 2와 3을 모두 지원합니다. 공통 모듈을 임포트하고 맷플롯립 그림이 노트북 안에 포함되도록 설정하고 생성한 그림을 저장하기 위한 함수를 준비합니다:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 파이썬 2와 파이썬 3 지원\n",
    "from __future__ import division, print_function, unicode_literals\n",
    "\n",
    "# 공통\n",
    "import numpy as np\n",
    "import os\n",
    "\n",
    "# 일관된 출력을 위해 유사난수 초기화\n",
    "np.random.seed(42)\n",
    "\n",
    "# 맷플롯립 설정\n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "plt.rcParams['axes.labelsize'] = 14\n",
    "plt.rcParams['xtick.labelsize'] = 12\n",
    "plt.rcParams['ytick.labelsize'] = 12\n",
    "\n",
    "# 한글출력\n",
    "plt.rcParams['font.family'] = 'NanumBarunGothic'\n",
    "plt.rcParams['axes.unicode_minus'] = False\n",
    "\n",
    "# 그림을 저장할 폴더\n",
    "PROJECT_ROOT_DIR = \".\"\n",
    "CHAPTER_ID = \"ensembles\"\n",
    "\n",
    "def image_path(fig_id):\n",
    "    return os.path.join(PROJECT_ROOT_DIR, \"images\", CHAPTER_ID, fig_id)\n",
    "\n",
    "def save_fig(fig_id, tight_layout=True):\n",
    "    if tight_layout:\n",
    "        plt.tight_layout()\n",
    "    plt.savefig(image_path(fig_id) + \".png\", format='png', dpi=300)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 투표기반 분류기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "heads_proba = 0.51\n",
    "coin_tosses = (np.random.rand(10000, 10) < heads_proba).astype(np.int32)\n",
    "cumulative_heads_ratio = np.cumsum(coin_tosses, axis=0) / np.arange(1, 10001).reshape(-1, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x252 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,3.5))\n",
    "plt.plot(cumulative_heads_ratio)\n",
    "plt.plot([0, 10000], [0.51, 0.51], \"k--\", linewidth=2, label=\"51%\")\n",
    "plt.plot([0, 10000], [0.5, 0.5], \"k-\", label=\"50%\")\n",
    "plt.xlabel(\"동전을 던진 횟수\")\n",
    "plt.ylabel(\"앞면이 나온 비율\")\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.axis([0, 10000, 0.42, 0.58])\n",
    "save_fig(\"law_of_large_numbers_plot\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.datasets import make_moons\n",
    "\n",
    "X, y = make_moons(n_samples=500, noise=0.30, random_state=42)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "사이킷런 0.22 버전에서 `LogisticRegression`의 solver 매개변수의 기본값이 `'liblinear'`에서 `'lbfgs'`로 변경될 예정입니다. 책의 결과와 맞추기 위해 명시적으로 `'liblinear'`로 설정합니다. 0.22 버전에서 `RandomForestClassifier`의 `n_estimators` 매개변수 기본값이 100으로 바뀝니다. 경고를 내지 않도록 하기 위해 현재 기본값인 10으로 설정합니다. 0.22 버전에서 `SVC`의 `gamma` 매개변수의 기본값이 스케일 조정되지 않은 특성을 위해 `'scale'`로 변경됩니다. 기존 방식을 사용하고 경고를 없애기 위해 `'auto'`로 설정합니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "VotingClassifier(estimators=[('lr',\n",
       "                              LogisticRegression(C=1.0, class_weight=None,\n",
       "                                                 dual=False, fit_intercept=True,\n",
       "                                                 intercept_scaling=1,\n",
       "                                                 l1_ratio=None, max_iter=100,\n",
       "                                                 multi_class='warn',\n",
       "                                                 n_jobs=None, penalty='l2',\n",
       "                                                 random_state=42,\n",
       "                                                 solver='liblinear', tol=0.0001,\n",
       "                                                 verbose=0, warm_start=False)),\n",
       "                             ('rf',\n",
       "                              RandomForestClassifier(bootstrap=True,\n",
       "                                                     class_weight=None,\n",
       "                                                     criterion='gin...\n",
       "                                                     n_estimators=10,\n",
       "                                                     n_jobs=None,\n",
       "                                                     oob_score=False,\n",
       "                                                     random_state=42, verbose=0,\n",
       "                                                     warm_start=False)),\n",
       "                             ('svc',\n",
       "                              SVC(C=1.0, cache_size=200, class_weight=None,\n",
       "                                  coef0=0.0, decision_function_shape='ovr',\n",
       "                                  degree=3, gamma='auto', kernel='rbf',\n",
       "                                  max_iter=-1, probability=False,\n",
       "                                  random_state=42, shrinking=True, tol=0.001,\n",
       "                                  verbose=False))],\n",
       "                 flatten_transform=True, n_jobs=None, voting='hard',\n",
       "                 weights=None)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.ensemble import VotingClassifier\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.svm import SVC\n",
    "\n",
    "log_clf = LogisticRegression(solver='liblinear', random_state=42)\n",
    "rnd_clf = RandomForestClassifier(n_estimators=10, random_state=42)\n",
    "svm_clf = SVC(gamma='auto', random_state=42)\n",
    "\n",
    "voting_clf = VotingClassifier(\n",
    "    estimators=[('lr', log_clf), ('rf', rnd_clf), ('svc', svm_clf)],\n",
    "    voting='hard')\n",
    "voting_clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LogisticRegression 0.864\n",
      "RandomForestClassifier 0.872\n",
      "SVC 0.888\n",
      "VotingClassifier 0.896\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score\n",
    "\n",
    "for clf in (log_clf, rnd_clf, svm_clf, voting_clf):\n",
    "    clf.fit(X_train, y_train)\n",
    "    y_pred = clf.predict(X_test)\n",
    "    print(clf.__class__.__name__, accuracy_score(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "VotingClassifier(estimators=[('lr',\n",
       "                              LogisticRegression(C=1.0, class_weight=None,\n",
       "                                                 dual=False, fit_intercept=True,\n",
       "                                                 intercept_scaling=1,\n",
       "                                                 l1_ratio=None, max_iter=100,\n",
       "                                                 multi_class='warn',\n",
       "                                                 n_jobs=None, penalty='l2',\n",
       "                                                 random_state=42,\n",
       "                                                 solver='liblinear', tol=0.0001,\n",
       "                                                 verbose=0, warm_start=False)),\n",
       "                             ('rf',\n",
       "                              RandomForestClassifier(bootstrap=True,\n",
       "                                                     class_weight=None,\n",
       "                                                     criterion='gin...\n",
       "                                                     n_estimators=10,\n",
       "                                                     n_jobs=None,\n",
       "                                                     oob_score=False,\n",
       "                                                     random_state=42, verbose=0,\n",
       "                                                     warm_start=False)),\n",
       "                             ('svc',\n",
       "                              SVC(C=1.0, cache_size=200, class_weight=None,\n",
       "                                  coef0=0.0, decision_function_shape='ovr',\n",
       "                                  degree=3, gamma='auto', kernel='rbf',\n",
       "                                  max_iter=-1, probability=True,\n",
       "                                  random_state=42, shrinking=True, tol=0.001,\n",
       "                                  verbose=False))],\n",
       "                 flatten_transform=True, n_jobs=None, voting='soft',\n",
       "                 weights=None)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "log_clf = LogisticRegression(solver='liblinear', random_state=42)\n",
    "rnd_clf = RandomForestClassifier(n_estimators=10, random_state=42)\n",
    "svm_clf = SVC(gamma='auto', probability=True, random_state=42)\n",
    "\n",
    "voting_clf = VotingClassifier(\n",
    "    estimators=[('lr', log_clf), ('rf', rnd_clf), ('svc', svm_clf)],\n",
    "    voting='soft')\n",
    "voting_clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LogisticRegression 0.864\n",
      "RandomForestClassifier 0.872\n",
      "SVC 0.888\n",
      "VotingClassifier 0.912\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score\n",
    "\n",
    "for clf in (log_clf, rnd_clf, svm_clf, voting_clf):\n",
    "    clf.fit(X_train, y_train)\n",
    "    y_pred = clf.predict(X_test)\n",
    "    print(clf.__class__.__name__, accuracy_score(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 배깅 앙상블"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import BaggingClassifier\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "\n",
    "bag_clf = BaggingClassifier(\n",
    "    DecisionTreeClassifier(random_state=42), n_estimators=500,\n",
    "    max_samples=100, bootstrap=True, n_jobs=-1, random_state=42)\n",
    "bag_clf.fit(X_train, y_train)\n",
    "y_pred = bag_clf.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.904\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score\n",
    "print(accuracy_score(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.856\n"
     ]
    }
   ],
   "source": [
    "tree_clf = DecisionTreeClassifier(random_state=42)\n",
    "tree_clf.fit(X_train, y_train)\n",
    "y_pred_tree = tree_clf.predict(X_test)\n",
    "print(accuracy_score(y_test, y_pred_tree))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "from matplotlib.colors import ListedColormap\n",
    "\n",
    "def plot_decision_boundary(clf, X, y, axes=[-1.5, 2.5, -1, 1.5], alpha=0.5, contour=True):\n",
    "    x1s = np.linspace(axes[0], axes[1], 100)\n",
    "    x2s = np.linspace(axes[2], axes[3], 100)\n",
    "    x1, x2 = np.meshgrid(x1s, x2s)\n",
    "    X_new = np.c_[x1.ravel(), x2.ravel()]\n",
    "    y_pred = clf.predict(X_new).reshape(x1.shape)\n",
    "    custom_cmap = ListedColormap(['#fafab0','#9898ff','#a0faa0'])\n",
    "    plt.contourf(x1, x2, y_pred, alpha=0.3, cmap=custom_cmap)\n",
    "    if contour:\n",
    "        custom_cmap2 = ListedColormap(['#7d7d58','#4c4c7f','#507d50'])\n",
    "        plt.contour(x1, x2, y_pred, cmap=custom_cmap2, alpha=0.8)\n",
    "    plt.plot(X[:, 0][y==0], X[:, 1][y==0], \"yo\", alpha=alpha)\n",
    "    plt.plot(X[:, 0][y==1], X[:, 1][y==1], \"bs\", alpha=alpha)\n",
    "    plt.axis(axes)\n",
    "    plt.xlabel(r\"$x_1$\", fontsize=18)\n",
    "    plt.ylabel(r\"$x_2$\", fontsize=18, rotation=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 792x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(11,4))\n",
    "plt.subplot(121)\n",
    "plot_decision_boundary(tree_clf, X, y)\n",
    "plt.title(\"결정 트리\", fontsize=14)\n",
    "plt.subplot(122)\n",
    "plot_decision_boundary(bag_clf, X, y)\n",
    "plt.title(\"배깅을 사용한 결정 트리\", fontsize=14)\n",
    "save_fig(\"decision_tree_without_and_with_bagging_plot\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 랜덤 포레스트"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "bag_clf = BaggingClassifier(\n",
    "    DecisionTreeClassifier(splitter=\"random\", max_leaf_nodes=16, random_state=42),\n",
    "    n_estimators=500, max_samples=1.0, bootstrap=True, n_jobs=-1, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "bag_clf.fit(X_train, y_train)\n",
    "y_pred = bag_clf.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "rnd_clf = RandomForestClassifier(n_estimators=500, max_leaf_nodes=16, n_jobs=-1, random_state=42)\n",
    "rnd_clf.fit(X_train, y_train)\n",
    "\n",
    "y_pred_rf = rnd_clf.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.976"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(y_pred == y_pred_rf) / len(y_pred)  # 거의 동일한 예측"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "sepal length (cm) 0.11249225099876375\n",
      "sepal width (cm) 0.02311928828251033\n",
      "petal length (cm) 0.4410304643639577\n",
      "petal width (cm) 0.4233579963547682\n"
     ]
    }
   ],
   "source": [
    "from sklearn.datasets import load_iris\n",
    "iris = load_iris()\n",
    "rnd_clf = RandomForestClassifier(n_estimators=500, n_jobs=-1, random_state=42)\n",
    "rnd_clf.fit(iris[\"data\"], iris[\"target\"])\n",
    "for name, score in zip(iris[\"feature_names\"], rnd_clf.feature_importances_):\n",
    "    print(name, score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.11249225, 0.02311929, 0.44103046, 0.423358  ])"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rnd_clf.feature_importances_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(6, 4))\n",
    "\n",
    "for i in range(15):\n",
    "    tree_clf = DecisionTreeClassifier(max_leaf_nodes=16, random_state=42 + i)\n",
    "    indices_with_replacement = np.random.randint(0, len(X_train), len(X_train))\n",
    "    tree_clf.fit(X[indices_with_replacement], y[indices_with_replacement])\n",
    "    plot_decision_boundary(tree_clf, X, y, axes=[-1.5, 2.5, -1, 1.5], alpha=0.02, contour=False)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## oob 평가"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9013333333333333"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bag_clf = BaggingClassifier(\n",
    "    DecisionTreeClassifier(random_state=42), n_estimators=500,\n",
    "    bootstrap=True, n_jobs=-1, oob_score=True, random_state=40)\n",
    "bag_clf.fit(X_train, y_train)\n",
    "bag_clf.oob_score_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.31746032, 0.68253968],\n",
       "       [0.34117647, 0.65882353],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.08379888, 0.91620112],\n",
       "       [0.31693989, 0.68306011],\n",
       "       [0.02923977, 0.97076023],\n",
       "       [0.97687861, 0.02312139],\n",
       "       [0.97765363, 0.02234637],\n",
       "       [0.74404762, 0.25595238],\n",
       "       [0.        , 1.        ],\n",
       "       [0.71195652, 0.28804348],\n",
       "       [0.83957219, 0.16042781],\n",
       "       [0.97777778, 0.02222222],\n",
       "       [0.0625    , 0.9375    ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.97297297, 0.02702703],\n",
       "       [0.95238095, 0.04761905],\n",
       "       [1.        , 0.        ],\n",
       "       [0.01704545, 0.98295455],\n",
       "       [0.38947368, 0.61052632],\n",
       "       [0.88700565, 0.11299435],\n",
       "       [1.        , 0.        ],\n",
       "       [0.96685083, 0.03314917],\n",
       "       [0.        , 1.        ],\n",
       "       [0.99428571, 0.00571429],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.64804469, 0.35195531],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.13402062, 0.86597938],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.36065574, 0.63934426],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.27093596, 0.72906404],\n",
       "       [0.34146341, 0.65853659],\n",
       "       [1.        , 0.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.00531915, 0.99468085],\n",
       "       [0.98265896, 0.01734104],\n",
       "       [0.91428571, 0.08571429],\n",
       "       [0.97282609, 0.02717391],\n",
       "       [0.97029703, 0.02970297],\n",
       "       [0.        , 1.        ],\n",
       "       [0.06134969, 0.93865031],\n",
       "       [0.98019802, 0.01980198],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.97790055, 0.02209945],\n",
       "       [0.79473684, 0.20526316],\n",
       "       [0.41919192, 0.58080808],\n",
       "       [0.99473684, 0.00526316],\n",
       "       [0.        , 1.        ],\n",
       "       [0.67613636, 0.32386364],\n",
       "       [1.        , 0.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.87356322, 0.12643678],\n",
       "       [1.        , 0.        ],\n",
       "       [0.56140351, 0.43859649],\n",
       "       [0.16304348, 0.83695652],\n",
       "       [0.67539267, 0.32460733],\n",
       "       [0.90673575, 0.09326425],\n",
       "       [0.        , 1.        ],\n",
       "       [0.16201117, 0.83798883],\n",
       "       [0.89005236, 0.10994764],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.995     , 0.005     ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.07272727, 0.92727273],\n",
       "       [0.05418719, 0.94581281],\n",
       "       [0.29533679, 0.70466321],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.81871345, 0.18128655],\n",
       "       [0.01092896, 0.98907104],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.22513089, 0.77486911],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.9368932 , 0.0631068 ],\n",
       "       [0.76536313, 0.23463687],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.17127072, 0.82872928],\n",
       "       [0.65306122, 0.34693878],\n",
       "       [0.        , 1.        ],\n",
       "       [0.03076923, 0.96923077],\n",
       "       [0.49444444, 0.50555556],\n",
       "       [1.        , 0.        ],\n",
       "       [0.02673797, 0.97326203],\n",
       "       [0.98870056, 0.01129944],\n",
       "       [0.23121387, 0.76878613],\n",
       "       [0.5       , 0.5       ],\n",
       "       [0.9947644 , 0.0052356 ],\n",
       "       [0.00555556, 0.99444444],\n",
       "       [0.98963731, 0.01036269],\n",
       "       [0.25641026, 0.74358974],\n",
       "       [0.92972973, 0.07027027],\n",
       "       [1.        , 0.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.80681818, 0.19318182],\n",
       "       [1.        , 0.        ],\n",
       "       [0.0106383 , 0.9893617 ],\n",
       "       [1.        , 0.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.98181818, 0.01818182],\n",
       "       [1.        , 0.        ],\n",
       "       [0.01036269, 0.98963731],\n",
       "       [0.97752809, 0.02247191],\n",
       "       [0.99453552, 0.00546448],\n",
       "       [0.01960784, 0.98039216],\n",
       "       [0.18367347, 0.81632653],\n",
       "       [0.98387097, 0.01612903],\n",
       "       [0.29533679, 0.70466321],\n",
       "       [0.98295455, 0.01704545],\n",
       "       [0.        , 1.        ],\n",
       "       [0.00561798, 0.99438202],\n",
       "       [0.75138122, 0.24861878],\n",
       "       [0.38624339, 0.61375661],\n",
       "       [0.42708333, 0.57291667],\n",
       "       [0.86315789, 0.13684211],\n",
       "       [0.92964824, 0.07035176],\n",
       "       [0.05699482, 0.94300518],\n",
       "       [0.82802548, 0.17197452],\n",
       "       [0.01546392, 0.98453608],\n",
       "       [0.        , 1.        ],\n",
       "       [0.02298851, 0.97701149],\n",
       "       [0.96721311, 0.03278689],\n",
       "       [1.        , 0.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.01041667, 0.98958333],\n",
       "       [0.        , 1.        ],\n",
       "       [0.0326087 , 0.9673913 ],\n",
       "       [0.01020408, 0.98979592],\n",
       "       [1.        , 0.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.93785311, 0.06214689],\n",
       "       [1.        , 0.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.99462366, 0.00537634],\n",
       "       [0.        , 1.        ],\n",
       "       [0.38860104, 0.61139896],\n",
       "       [0.32065217, 0.67934783],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.31182796, 0.68817204],\n",
       "       [1.        , 0.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.00588235, 0.99411765],\n",
       "       [0.        , 1.        ],\n",
       "       [0.98387097, 0.01612903],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.62264151, 0.37735849],\n",
       "       [0.92344498, 0.07655502],\n",
       "       [0.        , 1.        ],\n",
       "       [0.99526066, 0.00473934],\n",
       "       [1.        , 0.        ],\n",
       "       [0.98888889, 0.01111111],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.06451613, 0.93548387],\n",
       "       [1.        , 0.        ],\n",
       "       [0.05154639, 0.94845361],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.03278689, 0.96721311],\n",
       "       [1.        , 0.        ],\n",
       "       [0.95808383, 0.04191617],\n",
       "       [0.79532164, 0.20467836],\n",
       "       [0.55665025, 0.44334975],\n",
       "       [0.        , 1.        ],\n",
       "       [0.18604651, 0.81395349],\n",
       "       [1.        , 0.        ],\n",
       "       [0.93121693, 0.06878307],\n",
       "       [0.97740113, 0.02259887],\n",
       "       [1.        , 0.        ],\n",
       "       [0.00531915, 0.99468085],\n",
       "       [0.        , 1.        ],\n",
       "       [0.44623656, 0.55376344],\n",
       "       [0.86363636, 0.13636364],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.00558659, 0.99441341],\n",
       "       [0.        , 1.        ],\n",
       "       [0.96923077, 0.03076923],\n",
       "       [0.        , 1.        ],\n",
       "       [0.21649485, 0.78350515],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.98477157, 0.01522843],\n",
       "       [0.8       , 0.2       ],\n",
       "       [0.99441341, 0.00558659],\n",
       "       [0.        , 1.        ],\n",
       "       [0.08379888, 0.91620112],\n",
       "       [0.98984772, 0.01015228],\n",
       "       [0.01142857, 0.98857143],\n",
       "       [0.        , 1.        ],\n",
       "       [0.02747253, 0.97252747],\n",
       "       [1.        , 0.        ],\n",
       "       [0.79144385, 0.20855615],\n",
       "       [0.        , 1.        ],\n",
       "       [0.90804598, 0.09195402],\n",
       "       [0.98387097, 0.01612903],\n",
       "       [0.20634921, 0.79365079],\n",
       "       [0.19767442, 0.80232558],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.20338983, 0.79661017],\n",
       "       [0.98181818, 0.01818182],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.98969072, 0.01030928],\n",
       "       [0.        , 1.        ],\n",
       "       [0.48663102, 0.51336898],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.07821229, 0.92178771],\n",
       "       [0.11176471, 0.88823529],\n",
       "       [0.99415205, 0.00584795],\n",
       "       [0.03015075, 0.96984925],\n",
       "       [1.        , 0.        ],\n",
       "       [0.40837696, 0.59162304],\n",
       "       [0.04891304, 0.95108696],\n",
       "       [0.51595745, 0.48404255],\n",
       "       [0.51898734, 0.48101266],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.59903382, 0.40096618],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.24157303, 0.75842697],\n",
       "       [0.81052632, 0.18947368],\n",
       "       [0.08717949, 0.91282051],\n",
       "       [0.99453552, 0.00546448],\n",
       "       [0.82142857, 0.17857143],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.125     , 0.875     ],\n",
       "       [0.04712042, 0.95287958],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.89150943, 0.10849057],\n",
       "       [0.1978022 , 0.8021978 ],\n",
       "       [0.95238095, 0.04761905],\n",
       "       [0.00515464, 0.99484536],\n",
       "       [0.609375  , 0.390625  ],\n",
       "       [0.07692308, 0.92307692],\n",
       "       [0.99484536, 0.00515464],\n",
       "       [0.84210526, 0.15789474],\n",
       "       [0.        , 1.        ],\n",
       "       [0.99484536, 0.00515464],\n",
       "       [0.95876289, 0.04123711],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.26903553, 0.73096447],\n",
       "       [0.98461538, 0.01538462],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.00574713, 0.99425287],\n",
       "       [0.85142857, 0.14857143],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.76506024, 0.23493976],\n",
       "       [0.8969697 , 0.1030303 ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.73333333, 0.26666667],\n",
       "       [0.47727273, 0.52272727],\n",
       "       [0.        , 1.        ],\n",
       "       [0.92473118, 0.07526882],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.87709497, 0.12290503],\n",
       "       [1.        , 0.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.74752475, 0.25247525],\n",
       "       [0.09146341, 0.90853659],\n",
       "       [0.44329897, 0.55670103],\n",
       "       [0.22395833, 0.77604167],\n",
       "       [0.        , 1.        ],\n",
       "       [0.87046632, 0.12953368],\n",
       "       [0.78212291, 0.21787709],\n",
       "       [0.00507614, 0.99492386],\n",
       "       [1.        , 0.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.02884615, 0.97115385],\n",
       "       [0.96571429, 0.03428571],\n",
       "       [0.93478261, 0.06521739],\n",
       "       [1.        , 0.        ],\n",
       "       [0.49756098, 0.50243902],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.01604278, 0.98395722],\n",
       "       [1.        , 0.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.96987952, 0.03012048],\n",
       "       [0.        , 1.        ],\n",
       "       [0.05747126, 0.94252874],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.98989899, 0.01010101],\n",
       "       [0.01675978, 0.98324022],\n",
       "       [1.        , 0.        ],\n",
       "       [0.13541667, 0.86458333],\n",
       "       [0.        , 1.        ],\n",
       "       [0.00546448, 0.99453552],\n",
       "       [0.        , 1.        ],\n",
       "       [0.41836735, 0.58163265],\n",
       "       [0.11309524, 0.88690476],\n",
       "       [0.22110553, 0.77889447],\n",
       "       [1.        , 0.        ],\n",
       "       [0.97647059, 0.02352941],\n",
       "       [0.22826087, 0.77173913],\n",
       "       [0.98882682, 0.01117318],\n",
       "       [0.        , 1.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [1.        , 0.        ],\n",
       "       [0.96428571, 0.03571429],\n",
       "       [0.33507853, 0.66492147],\n",
       "       [0.98235294, 0.01764706],\n",
       "       [1.        , 0.        ],\n",
       "       [0.        , 1.        ],\n",
       "       [0.99465241, 0.00534759],\n",
       "       [0.        , 1.        ],\n",
       "       [0.06043956, 0.93956044],\n",
       "       [0.97619048, 0.02380952],\n",
       "       [1.        , 0.        ],\n",
       "       [0.03108808, 0.96891192],\n",
       "       [0.57291667, 0.42708333]])"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bag_clf.oob_decision_function_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.912"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score\n",
    "y_pred = bag_clf.predict(X_test)\n",
    "accuracy_score(y_test, y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 특성 중요도"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# from sklearn.datasets import fetch_mldata\n",
    "# mnist = fetch_mldata('MNIST original')\n",
    "from sklearn.datasets import fetch_openml\n",
    "mnist = fetch_openml('mnist_784', version=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
       "                       max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
       "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                       min_samples_leaf=1, min_samples_split=2,\n",
       "                       min_weight_fraction_leaf=0.0, n_estimators=10,\n",
       "                       n_jobs=None, oob_score=False, random_state=42, verbose=0,\n",
       "                       warm_start=False)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rnd_clf = RandomForestClassifier(n_estimators=10, random_state=42)\n",
    "rnd_clf.fit(mnist[\"data\"], mnist[\"target\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_digit(data):\n",
    "    image = data.reshape(28, 28)\n",
    "    plt.imshow(image, cmap = matplotlib.cm.hot,\n",
    "               interpolation=\"nearest\")\n",
    "    plt.axis(\"off\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_digit(rnd_clf.feature_importances_)\n",
    "\n",
    "cbar = plt.colorbar(ticks=[rnd_clf.feature_importances_.min(), rnd_clf.feature_importances_.max()])\n",
    "cbar.ax.set_yticklabels(['중요하지 않음', '매우 중요함'])\n",
    "\n",
    "save_fig(\"mnist_feature_importance_plot\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 아다부스트"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "AdaBoostClassifier(algorithm='SAMME.R',\n",
       "                   base_estimator=DecisionTreeClassifier(class_weight=None,\n",
       "                                                         criterion='gini',\n",
       "                                                         max_depth=1,\n",
       "                                                         max_features=None,\n",
       "                                                         max_leaf_nodes=None,\n",
       "                                                         min_impurity_decrease=0.0,\n",
       "                                                         min_impurity_split=None,\n",
       "                                                         min_samples_leaf=1,\n",
       "                                                         min_samples_split=2,\n",
       "                                                         min_weight_fraction_leaf=0.0,\n",
       "                                                         presort=False,\n",
       "                                                         random_state=None,\n",
       "                                                         splitter='best'),\n",
       "                   learning_rate=0.5, n_estimators=200, random_state=42)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import AdaBoostClassifier\n",
    "\n",
    "ada_clf = AdaBoostClassifier(\n",
    "    DecisionTreeClassifier(max_depth=1), n_estimators=200,\n",
    "    algorithm=\"SAMME.R\", learning_rate=0.5, random_state=42)\n",
    "ada_clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "plot_decision_boundary(ada_clf, X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAxYAAAEYCAYAAADBK2D+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzsvXd8XMd57/2dU7aiAwRAggBBUr33ShXLkmVLspplx1eWJV/39765N70p96a8b17HuXmTOImdxInf2I6cRLYs2WqULKuLapRESmInCAIEAaIQINF2z+6eMu8fZxdcAAtgASyABThffVYAdufMmXOW+/x25ikjpJQoFAqFQqFQKBQKxXzQlnoACoVCoVAoFAqFYvmjJhYKhUKhUCgUCoVi3qiJhUKhUCgUCoVCoZg3amKhUCgUCoVCoVAo5o2aWCgUCoVCoVAoFIp5oyYWCoVCoVAoFAqFYt6oiYVi0RFCvCyE+MFSjyMbIcT5QohjQogvLfVYFAqFYqWi7L9CsbJREwuFwqcMKAdql3ogyw0hxNlCiB4hRPtSj0WhUCjmgLL/c0AIcboQ4jkhRCytAX8hhAjMcEy5EEJO8bhiscauWDiMpR6AQlEMSClfE0KsklIeW+qxzIb0yl+zlPL6JTr/14FvAqNLcX6FQqGYL8r+z+ncNcBLwE7gY0Aj8B2gEvjqNIeWpH9+BOif8FprgYepWALUxEKhSLPcRKVIuBv4InAe8IWlHYpCoVDMDWX/Z82vAmHgHinlMIAQIgj8qxDim1LKg1McV5r+uVVKGV+EcSoWGRUKpSgKhBDXCyFeE0LEhRBHhRD/IoQoy3p9gxDi20KIPek2rUKI35jQx8tCiB8LIX5HCNGbdq1WCCF+IIR4SwhxUbpNXAjRJoS4L+vY5nT7eyb8fZsQ4k+FEJ1pd+/PhRDVWccFhRDfEkL0pV9/QghxU/rY6/O47sx5Pi+EeFwIkRBC/Dz92iYhxE+FEO3pvrcJIT6WfRzwAHBdxpWc1W9UCPGX6XEnhBDbhRC3TTOOP5nGPd081XFSyo9JKR+b6ToVCoViKpT9X5b2/3bg6cykIs0jgAdMeS78iYWtJhUrFzWxUCw5QoiPA78EduG7VP9P4CYg+wvrnwINwO8C1wA/Av5aCHHrhO5uBO5PP24HMsZrHfALYDNwA/AM8AMhRMMMw/tL4OZ0f58FLgf+LOv1bwNfBv4gPfau9Nhmy7fSx14P/E36uR8CnfiegBuADuBx4bugu4BzgceBd9O/nwsghNCBp4HP4N+vjwKvp4+9borz/0NWHxMfXXO4HoVCoZgRZf+BZWb/hRAacDbwfvbz6clCC3DONNdaCgxO87piuSOlVA/1WNQH8DLwg/TvGnAA+LcJbT4CSOCy9N9NOfrpAv5qQr/JiW2BH6T7+lrWcya+cfu19N/N6Tb3TPj7IBDOOu53gePp31cDLvAHE873H+ljr8/jXmTO80yO1yZex2nptp+ccG0vT2h3f/o+nDnh+VeAzQv0nv4J0L7U/7bUQz3Uo7gfyv6Pa7ss7T9Qkx7L53O89iLw5DTH3gHYQC/+xG9/+jo2LvW/TfUozEN5LBRLzQXARvwVk2y24LtULwOQUnYACCGqhBDXCCG+ii9KlROOeyPTdgKDwPcyf0gpbfxEsXUzjO/bUkor6+89QEXaTX9degw/nnDMv87QZy4envhE1jU3CSFuxs9ngMnXPJF7gJeklHsmPP8q6fs5ESHEHwkhnCkeM90jhUKhmAvK/vssN/sfTv9M5HgtkfV6LvYAn8YPl/oo8L+BC4H3hBAbpjlOsUxQyduKpebU9M8tQoiJr2lAFYAQ4kL8ihNXAEP4ohBgcjjf4SnOs1dK6U54boSZjfSuHMeQPm4d/qrNxHO2zdBnLiaNWwhxL34IwClAD/7KHswcwngqcLoQwpnYZfqRi39ifOhBNkdmOJ9CoVDMBWX/fZab/U+lf+b6Dmnge0xyIqXcj++lyPCmEOInwHvAX+BPOhTLGDWxUBQLN+O7RifSl44pfRF4B9+9uw/8ZL0c7b0p+p/K0E1laPM9zgUmGnDJ7Bk3biHEjcC/A/8v8JdSyr708/n2/TPgj/M9ebr/vnzbKxQKRQFR9j+78+K3/8fxx1yT47Ua/BK0eSOlHE4nrd8+m+MUxYmaWCiWmvb0z6SUMqcxEv5uqBXA56SUR7Neasg6finoxP8M1QPdWc/PlBCYDw8Au6WUv5N5Io9EwwztQPlU91OhUCiKhPb0T2X/x1PU9l9KmRJCtAJnZD+fThw/DfjJHLqtRiV1rwhUjoViqXkPPwnvKxNfSJfM0/F3RHWB4azXrsd3ES8lb+Gv2kx03X62AH2X468KZTPpHuGvlk38HD+BX4Lw1ImNRVYJR4VCoVhilP3PzXKw/08Dtwt/74oMtwLR9GuZczaIrDg3IUR9jnHVAJ8Eni3g+BRLhPJYKJYUKaUj/HrkDwshXPyKGkngYuDX8ZP7tuAbz38SQnwXP9HrfzJ5185FRUrZKoR4BPimEMIGPsQ3rJkku7m4xDO8BvxF+t68BdxJ7tjTQ8DdQojP4McjbwX+P+BLwEtCiD/Cd0uvTY+rD/jNeYxLoVAoCoKy/1OyHOz/X+OXwn1YCPG/8atkfRv4sZRyB4Dw9wp5CPhfnCjT+6QQYgd+uNYRfA/HH6fH9tcFGptiCVEeC8WSI6V8BL9CxBnAk/grLvfjG5khKeVW4OvAtcBL6dc+g18KcKn5Mn698f8beB4/ufCL6desqQ7Kg2/hG+nfx6/xvhE/Dnki/4wvHD8Efg9wpZQp/GobP8M35q/h10YP4CfpKRQKRVGg7H9Oit7+SykPp89ThZ8D80/4G+R9MatZPzCKn3ye4X/g52H8PfAG8Ff4+4pcJqUcKtT4FEuHkHI+k2qFQjGR9CZEL+OvEt0EfH+Kpn8qpfyTRRqWQqFQKBYYZf8VJzsqFEqhmAdCCCEnz87vAtqklF3pShfvTnG4qsKkUCgUyxRl/xWKySzZxEIIYQK/Bvw5/u6NkzaIEUJcjO8GzK55/IiU8q8WZ5QKxYw8JoR4F3/zIR34OPCr+C5ypJSDqEoXCkVeKF1QLDOU/VcoJrCUHouv4Cc3vTVNmxrgUSllrmoICkUx8AR+zOjv4wvLbuABKeW/L+moFIrlidIFxXJC2X+FYgJLnmOR3uTmn6ZYmfoc8A1OJP48D3xTSjmSo+1Xga8CRKOhi08/vWnBxqxQzEQ83o4QJtm7yUopkdImEmleuoEpCohkcHAUy3KxPQ0t4FBdXkIoYI61cF2PwcERrLiH7QkIOAQDBnXVFTl35hoZGiVhuaQcHRG0KS+LEgoGcB2X2NAoiYSOKzyMkKSsPIpjpbCGEyRtHWk4CB1Kq0swTTNH7+Np3dbaL6VcVcAbUjCULihWIkoXTg5s22ZoKEYyqePqLsGQoLqslNhInNHRFClHQ5oOhg5hXSeZMPB0FzMkqKooQQiNRMzCiiVJpgwwHUJRk0g4RPz4KElLYkvAdDADBmXVha0iP19tKPaJRRhISCmlEKIKv0pCWEp513R9Xnzx6fKtt/5hQcarOLnp799CT89mUql+AoEa6utvoaZm06R2O3c+iOdZmGb52HO2PYSmhTnnnG8s5pAVC4Tnefzd3z1MW1uQ7qSG3tjDF265lusvPAspJW+9/gEvP7+Llo5yUquOIiIJggGT26+/mI9efi6aNrko3w///hHaDxp0xU30xiPceuNlXHfFefR29PL8fz7P3r2ridf1UFaR4tRQCQf3hjhmJKB6EE2H+o31bLr9KiKlkRnHf1fgrveklJcsxL2ZL0oXFMsJpQuKbHbtOsjjj7/Mrl3rcNe3saoM1ooI+/eXMBQdRqsYRtc1LmpajX1kgN27mrCbDxGtgt/+2qcpjYZ5+bFX2L+9h/0dtYiNbZxSGUHrFxzoKMeu60OEkxgBk4tvvIDTLzl93GR1vsxXG4o6eVtKaWX9fkwI8VvAYSFEOPs1hWIx6O/fQkfHDzGMEgKBOjxvlI6OHwJMEpH6+lvGXtO0EjxvFMcZpanpU4s+bkXhGR2N8+STW+g4rNEd16D+KJoQREP+XlFHOvt4+/UPaW+vI1XTix5Ncsq6er7yqZsoL5n5S/9M1DnQ1ykYsHW0+uMEo0E23XUVa09ZO+++ix2lC4piQumCIpvW1sM8++xWDrZXk6rpQTNtqpImnX0BhvQUWuUwVeVR7r76Qj58bTdt7VUkq/sQZgpDD6MLwQevfkDLBz20dpXj1XdSKjyCQ0la2hqx64+gRVLUbajn2rs2EY6Gl/qSJ1HUE4sc6EACfwMdxTIm3xWepe4zm56ezRhGydhqk66Xjz0/8TyZv/3x9BII1NDU9KmCjkex+EgpeeedXTz77Pu0tJeRqLYQDccImAa3Xn0hF57WDEAqZadDHDS0gEc4YvLA7R8pyKQCQJPguTpCA2EIzr7qzJNiUjEFShdWEIW240oXFItBLGbxxBOv8O67wxweDiDrehCmy6qKUqK2pN8z0AyboKFxdrSU5368h46hALKuF2G6VFaUcOumC3j6e5vZtzvIMUOHxk40HVbXVaENxPCkjjAgEDW55o6ri3JSAUU2sRBCVOMnQ31BStkihPgs8KyUclAIYeBXCnlISukt6UAV82I2Kzyz6bO19ds4jgU4WFYvsdjBefU5EV+Y6sY9p2klpFK9OdvX1GxSgrHC6Ozs5aWX3qO1tZZEbR9aqcX6Nav41Xs+xqqKwsa55kIIj3rTJjEaons4CHV9CAFmaOaciuWK0oWTh0Jrg9IFxWLxwgtb2b27j8O9tbChnUBQcPumS7jx4rP54b/8fKxdjafR2zFCR3c9bGzDDAquv+oCbrj6fJ787pMcatUZ8Dy0VceJREPcduc1lHiCbY++toRXNzuKamIBRIB1QCYAMQy8IITw8CuFvAL80RKN7aShmFZ48qWj4yEcZxhNCwEm4OE4w3R0PFSwsQcCNXje6Nh4ATxvlECgpiD9K4qfRMLG8zykNNBNj2g0yH//9M3UlJcu+Lld12XNml66e1ZxpGQE0diJpmtsOHcDG8/duODnX0KULhQJy00blC4oFot4PIHnGQgDNBPOO2Utd153KQlrvCPVADxPR5i+t/mUDXV84iOXYqdsXMfFdYNohoduCG646RJOOaWRnv2dS3NRc2TJJxZSyuuzfj+Mv1tl5u/vM/WulYoFYCG8CROZ7QpPPlhWF2CiaZmVWx3Pc9PPFwYVH6tYKFzXZesr2+nrdegdikJdHwgwTd9Ejw6O4rmSnu5qhsuGEBVDRMsifORXrqdmzcr7AqN0ofhYjtqgdEGxVAgtv2Tq6ZKutQImZC8mSz6xUBQXC+FNmMhyXeFR8bGKhaCzvZvnHnudPfuCDIY0RONhNENw9unruOicUwBIJW0AbNsEw0YIuPTjl6zISYWiOFHakBulCwrFeNTEQjGOhfAmTGQhVnjC4bVYVjuep+HncrpAknC4uRBDHkPFx568jI7GefPNHfT2RUgEY2AmEcLE0MeXjY2NxnnnzZ309oWwAjEwUwhhTGoHcOzoIJt/+iL791YxVDqIqBimvCzCvXd9hFOa1wCQiCc4uKON44NREoEEIuS71nVDX/iLVijSLEdtULqgWAxaWjpobz9KT185suooEolp6Liuy9tvfMBAv+RYXIfV/Us91EVBTSwU41iMFaOFWOFparqPAwe+g+fFkdJGCB1dr6Sp6b5CDVsxSxY6Hnux8DyPt9/eyXPP7WD/oVJSNXFEwwDBgMHtV19MRUkU8CtGvffObl549n32HyolWZWuGBUw+PhVF1BZVjKp73jMwnM8HCeAFnIIR0y+fO/HWVNXjZSS/dv28+YzH7C/rZREZRLWHgckmqERLYsu8p1QnMwsR21QulCcrBRtGBmJ8fjjr/LueyMciUWQ9V0Iw6OuqpwrNqzjn/720ROe6CbfE11eWkpihdevUxMLxTgWK1600Cs84wVpeRurlcBixGMvFlu2vM9LL33I7v1r8Na3oQVcTl+3mq/d+dFxSdtvv/EhLz//Prv3rcHb4Lfb2FTPF++8gZqKfJO7BYbueyL2vruXt599j117GnCb2xEBG5kMIkwHw9TRdeWxUCwey1EblC4UHytFGxzH4aGHNrN/v6Rz1ERb200woHPntZdzQdMaHnv4OXbvriVW04cojVGZ2bvilZ0c7o8gy4+B8NByeLKzSVlJOra3cOxYiFQgBkYKIfS8cziWAjWxUIxjOceLKnd08bAY8diLxbFjQziODrrECMC6tTU8eP8dk5Lujh8bxnV1MEA3Ye3qSn7ngdvnvCPq8MAItq3hARguUgpEMIlm6JxywUbKqhe+vK1CkWG5aoPSheJipWhDMmmTSKRIpUrQwinMgMbnP7GJ6y88i727DuI4LrYdQAvZlJUE2NSwhhce2cWhYxG8uj5EwKGsNMpNmy7OfQIJQ+29PPvkhxw4WIpVmUI0Hkczdc689Myi3cMC1MRCkYOVbIhXigu22FmMeOylIhQwZ54sCAjk024aBo8eT2+yl+5S9yirqeTau6+mur56zv0qFHNlpWqD0oXFYyVrQzQUzPl8nQXb3zrG4cEIorELw9C4+tKz+cQNlxIwTaSUtO9uJzbsMhQzobafaqnRt+0we1sa8Na3IwIOVfVVXPupTZRXl+c8T7GgJhZFhDJu+THX+zTRBWtZnezb9w1aWyuJRtfN+n6r92tqlmN1l2IjYSXHJhUAmqFxx9dvQ9Omd50rVhbKzuTPXO6V0oXF5WTTBg3QJSQTQUQ4hW4KPnnjZVxz+bkAjBwb4ZWfv8bu9216UxFkYyfC8CgPR3CHDKQOmimpbKjgti/eMq/FqsVCKVSRkDFunmel4w4tOjp+SH//lqUeWlExn/uU7YJ1nOPYdj+g4zjWrO+3er+mp77+FhxnFNsewnVdbHsIxxmlvv6WpR7assCKWcRHLIZHwshoDIREaEJNKk4ylJ3Jn7neK6ULi8tK1wbHcTh44DAjowFsI4nU3XGvCyEoKfHDmAaPDvL095/mvTc0erQErO4lGNa5+ZYr2bChIX2A/yMYDi6LSQUoj8WsWMiViJUSdwjFcZ9yjSHbBWtZ3QhhIISOlKmx/vK93yvp/VoIlms89pIjof3DVnZuOUDLoTKSVccQlQMgJKEijqk92Vkom7fS7EwxaoNltRMO+zvXK11YeFayNgz0HOMfN7/Lvv1hBkMC0dyObmiYRu4FoaH+IeyUQyIRRtT1EY4Gue/+W1hVW8V7Hct3MqomFnmy0JUMVkrcYTHcp6nHoI+5YP3SgwHAQ9cDOfuZ7zhOdlZCPHY8btHbe5zRWBAZspDInO0SiSRHe48xOhpEBi3EFO2ykVJy5HAvliVIeC5Sd6i14d3nD9DaWYVs7kBoLtIKIyJJQpFQoS9PUQAW0uatJDtTrNqQSg0BnZSUrFO6sEisBG3o7u4nkXCIJwyoHqLCFWx/dRd796/BXdeBCNrUVJRy3y1X8+ZTb87QmwAhEBpEVsACkppY5MlCr0SslLjDYrhPU43BceI4zmi6lYHn+cWkM5slzeZ+r5T3S5EbKSXbt+9j8+b32N9eilUxgljTj2kaXHv+mePa7Xh/P89tfo/9bSXEyy1EQz+GqbPpgrOm7P9Y/yDPPvoqu3a4HHUjyOZ2NEMS1UIMWiG8kI1ueJRJnaGuRjj14GJctmIOLKTNW0l2pli1IRisJZnsxbYrULqgmIl43OKpp15n6ztH6RgsQ67pRJgulWaI1HAAz/TQQy6Na6r5vf96F07KYaZpxUpDBezmSSrVj6aN3+DKX4kozE6KKyXucDHuk2V1c+zYdo4de4djx7ZjWd3j7tNUYwBJU9MDaFoY0wwBHsFgDYZRMev7vVLeL8VkpJT89Kcv8Mgj7/JhSwXWmiNoZTHWN6ziz776aa4+9/Sxdk/97CV+9si7fLCvgviabrTyGE1rqvnDr3yKq84/PWf/bfsP8/C/PMWbr4foi4xAfR+RaIDP3HEtZSURIB1WKyCIxliQraIoWUibt5LsTLFqQyi0FsOoVLqgmJGBgSG++93HeOGFGO0JCQ3dBMLwyU0XcmZzA0IIBAIhoKq8BEPXGRwYJpV0sBIGBFb4znhplMciTxZ6JWK5xh1OjFcFbcFXbKT0P7jZP7OZ7r3KdsGeGPvs7/dyfb8UM+M4Ll1dfcRiYYgkMAKSj1x0Fvd/4tpxyXNSSg539DI6GkZGkugBj6svOIPP3Xot2jRJdu0HOkkkBEmpo0VSVFRG+a2v3kM4FKTj1Z2LcYmKArKQ2rBc7UyuXIrFWM2fqzZEo+s455xvTBi70gXFeI4cOYplJRkZqUfb2Ec0GuAPPn876+pX8ejDz41rKz2Plza/ydYth2nrrcCt60YEbUqjUdY11E1xhpWBmljkyWLsOrrYcYfzTaTLFa9q24OkUkNEIgtzn3p6NhOJ1I+5sgFse2icOz3f92q+93slxIkqZsL/grKhoW7aihz+KpVgfUPttJOKE+0ZW9kqK40QDgWJj8axkzZWMogMJCGPPA3F0rPQ2rAUdmY+2jBVLkVFxaUMDr4DFLc2KF1QzIQQvv02DY3aynJs2yE2GieR0JFmCh1w2/r5ZYdDHzpaYxeGrnHxeadxx8evJBT083dGBkewU+DigXDJBBG5jos1PEoqpYNus9y0QE0s8mSlrUQUIpFuulwGTQsvyH3KJzlupb1XihWOlOx6exdv/2IXLYfKSVYNoJVY6JpOyNWXenSKGVhp9ma+2jCVLsTj7TQ1PbBg90lpg2IpaG/t5Jebt7Jvf4TBsI3WMECJ0BGOx9BQGVpzG6GIyRc+fROnrvdLyMaGY2x5/A12bBulO1aOXN+KZnqsWlVHrOcYWx7fyoF9YQZDHjQeQjM01p6ydomvNH/UxGIWrKSViEIk0k1lyCE25lYuNPm605eb90dRPFhWEtt2Sdk6mFPHxFpWEsfxsB0NDGdO59KAcF+M53bu4/BwFNZ2IXRY3VTL+kCIvt2DoM2tb8XiobThBNN9wV/I+6S0QbHQjI5aOE7GwyCpsDx+8qM3ONBViWzsQjNdVlWV8Ynzz+DDLR8C/r4VpiFoqK8GoPNAFy8/+gp796wiXhNHrIkTCBhsuu5CosNJXvn+GxzoqMZr6kSYDtGKKFffcSWrm1cv5aXPCjWxOEnJtyzedEZxKSpgLEZI2mxZ6DKKisVBSsm77+7h2We3s7+9nETlMURpnGAgQGNt9bh2H2zfxy83b2NfW5nfriRO0Aywtq56mjNMJgTotkf/QCWi7gh6UOOyy89Ethxl/7tJ+r0gsrkNTYOataqyjGLhma82LFVlJKUNioVidDTOk09u4Z13j9M5XIFcfxDTcCmRAbqPl0JpHD3kcfm5p/L5266jbW/HlH21vL+f0WEdyxNoJQnKqqLc/4XbCIcCvPidxxkcjOKWxNGDDuvOamLTnZvQ9eXluVYTi5OUfIz/TEZxKQz5fF3ZC7F6pDZFWv4kEikefvgXfPhhnM7hCLKhC82QbFxby/9x143UVvrvaTKZ4rGHn2PHh3EOD0XH2q1bs4ov3XUDddUV055HSknCsnBdDamd2JFVAGgCTUj6Xm+lo7WO1OpuRDhBMBTk0o9fzMbzNi7gHVAofOarDUv1BX8+2qA2OFRMRWtrJz/96Yvs2VtNrDqBaBggaOrcfOn59O1s83PvhEDXBOef1oyh6ySTKVxP4klASCZW9hMC0DSEJqipLae0LIqTsv3XMv9pgrWnrl12kwpQE4uTlnyM/0xGcaniVefqyp7v6tFU4qM2RVr+dHR0093dz9Gja2FtB8GQxmdvuoIbLzl3XNJ21+FeeroH6OtbAw2dBEIad95wCTdefv60yd0Aw4MjPPfz1/jwfYueRAjZ1A4aaNqJ4zzHY2QoRDKYQoskqW2q5aOfvYFAKLBQl65QjGO+2pAJg12KPIa5aEMhvApKG1YuW7fuZHDQIO5paKVxqiujPHj/nURNk3/d2T6ureu4vPDU67yzpZP2ozW4zQcRpk15eRXBgLk0F7AEqInFSUo+k4J8k+GWy8rLfFaPphOfqVb4QLBz54MqtnYZIMeKbvgrRaGgwaVnbJw0WZDS9zoI4bcLBnUuOnNyu4nseG8vLz+zjX0HqkjWjSIqE4QCJldfcDp9H0zc/M6vGaVpgqYzGtWkQrGoFEIbThZdAKUNKx0p01Wg0otAGxvqWFVRRjxmjWsX8uC9Z96mpaWS4aiHaOzCMDQuveAMbv/YFcvS8zBX1MTiJGYm4z/XWNmFTlabS//9/VsYGvKTqTQtSDi8mmCwJu/Vo+nEJ9cKn2V1I6XAMCIqtlbBe2/sYKA/RCJgo0cTrF1TzZc++3FG+4d4cdLEQqFYWhZCG4pRFwAsqx3bTgAOQpiEw6sxjMq8vQpKGxQAVVJjeFBnJGkiGkYpLQvxlXs/QUP9yZcbpyYWiimZS6zsXN3K+YrCXPrPHAOZFQOPePxQ+rrMvJIKMyt0yeQAyWQ3rptCSgPDCOdc4TPNCgwjrGJrlwme5yGlxJMyHRNbYKQfZ6sJ3xNx6QWnU1oSYeToEBKJRCzMeRWKBWC22jCfcKN8tGE+upNKDQEamhYEXOLxQ5hmjHA4vyo8ShtWNq7rZXm0T5DxXgMgPATp3SY0ga5Dc2NtzkmF53jpjRtXrr1XEwvFlMwlh2IubuXZiEJPz2Zc1yGZ7EBKGyFMDKNs2v4zY4pGQ8Ri7YCOlBrx+GFCodq8kgoDgRoSiU6SyX40zUSIAFImcZwE/f1bJq3wbdv21XTp3ROo2NriQ0rJO+/s4tln32f/oRpSqzsRZpJQsITwAoYgSSnZv30/b2x+n5aD1VirjiBCcTQBmtQW7LwKRSGYrTbMNdwoX23I9O95NrHYbqS0AUFHx49m1KtgsJZksh/w8LXBIZXqZf36L+V1L5Q2rExGR+M88cSrbNs+QtdICBo6QEjqKsvpONTNUz/bwv6WKMcDcUTVIEIYTEzSzsaKWbzx1JvsePc4h4ciyLUdSOFRXlm6eBe1SKiJhQKYelVotrGyc0lWm43oWFY7qdQQmhZAiADgkko3yvOvAAAgAElEQVT1AVPvN5AZk2n6HotkshvPSyKloKnpgbyur77+Fvbt+3P8nQc0wEUIMM26nOPMhAo4jpO1iiUIh+tnPJdicTh2bIhHHnmBD3dI+twArO1C0yWnNtbz9bs+StDMlWyXtco0Rw+DIaHzrT282RKgJxVCNnQhDMmq1dXUx1IcG1ATC0XxUAhtmGsSc77akEr1I6VOMtmJEAZCBPA8G8tqH/tyP9W4QqG1GEbJmJ0Gf7Eq32tT2rDy2LZtD888s419reVY1RaiYYBAwODWKy/COB7je//4Cm19EeSaHoTpUlkWpbm8nM7WeM7+Dnx4gNeffo+WlgriVSlEw3EMU+eKK89l03UX+o1klrpMn7JX9KiJxTKk0LGqhay1PZfY29mIjm1bgIfnxfE/hgLQ0s9PPSbL6sRxhvG8BJAu6yaieV9XTc0m2trKse0EUqYQwiQSWTdlLG59/S20tn4bxxkGzPQ4k6RSw9MKnWLxeO219+nsHOXoYC1iYzuhsMGXP3kdl511So6kbcneXQd55qm32ddahbWqBxFIYBoRwsHZeTaq0BjpT9EzUAMb2zDDOjd8/DLOvuBUXv/2E4W8RMVJRrFqw1zz9fLVhkCghqGh7YCLlEl8e2sAwRm8IhrDwx8CMh3uIoEkrivzttNKG1YW8XiCV17ZxpEjEaxwHK00TlN9Nb/+K59goKufp9/aTefhRljXjhmSXHfpOdz90Sv4xU9fAiZPLOyUzbYXt9HTFSYWTKCVxaipreDTv3IjlZVlAAz3DbLtZ69zqDVMv+cgVg0gNI1QNLTIV18Y1MRimbEQG+4Ustb2XPIyskVndPQgqVQPvltacPDgP7Nhw1fH2mqawHXtzF+AC7hoWnjK/iORZoaG3ku3t8eeF8Kc1b0Lh5sJBq2x+wRg20M5xbGmZhMdHT/CdS1AppMCG9E0U8XSFgmO4yClhtAFmi44tWkVl5996qR2oyNxnnjsRT7YnuCIFUKu7kYzPFbXVvKlu28gEgrOeK7sGF0N8Dzh1zHXobq+nHMuPK2AV6Y4GSlmbZjr3hYnVvdtYrFDQGYBKTTuS7hv47emXzuhC7peQSrVn7Pv/v4t2PZgOmxKckIbdEyzVmnDSYrruniexPN0RCCFaercdd0l1JSX0tvRm86r8G13NBrgjusvxTR05BQ5E146R0NKHU130HWNqzedT2VlGZ7rseeF99nzahutXWU49b2IUJJAKMSlH7+INRvWLO7FFwg1sVhmLMSGO4WstT2XvIyM6MRiXXjesaxXDLq7HwUYm1x4nsRPwhb44mEAMv18buLxdoLBNSSTXWP9+n24GEZJ3vdu9uLoUVZ23rgyc67rqljaoiW3//mN196nrfUYR/rrYWMbwaDG3TdewfWXnoM2Q5nZ2Gic5594nUPtGkdiGtT3AIKJhyVHLXY9+TadbYIjMR1R1wtCEIzMPGlRKKC4tWGuex7V19/CgQPfwXUH8BebxkZGa+u3x/qOx9vxV/8d/EmCARi47iiBwLqcfff0bCYcXo3nrSIW25d+ViBEkJKSddj2kNIGBeBvWjcVju3w2ivb2bujn4M9ZcjVXUg8SiJTL3Zm7H9/ew+Htu3j0KEGnIZOtIhN0xnruOqTVy7rMuNqYrHMKPSGO/39W0gkjmJZXWhaaKwMaz5u6qmYbV5Gpu2+fX8y7nlNC+J5Lj09z4xNLEwzTCqVQtMCZCYHnpfCNKf+EGfiaFOpAYQIoGk6nufieXFGRw8hZZydOx8cFzYwXUhBvuI4V/e/orhIpVK+Z0MDTResb6zhhsvOnfYYKSXvv72LV5/bQUt7GcmaOKLBwjQN1tZVM3Q4dqL/UYuX/voZWtvLSVRbiDXHMEydczadw/qz1y/05SlWCMWuDXPZ28Jf3X8Iyzo64RUTx7HGvvj7Nr55LMcCdDzPBpLU19+Ss+/s3Lt4PJTO2QPPizM4uAPP8/P2JoYnKW1QZAg6Hv/53afYty/CUEQgGjvRdcH5Z23k1o9ePuPxbsrxcyukjtAhGDW5/JbLlvWkAtTEYtlRSIOUcZ0HAhUkk714XopYrA3bjqHrRl7VkhYGHd8LEQMCSHkifyIcbga6cZzhsXjWQKB2UmnAbOOfSBzF82x0PYDn+ZU/PM8CUkipAxE8zxpbcQKmDSmYTVLfXNz/ioXF8zzefnsnO3Z003akDK/+CEJ4RMN5egfySKzr7TrK1lffp+1gLcmaPrQSi8aGVdx/z43sfX0nQ4fb/IYSkkNxOtqbSVT2I0otqldXcf2nr6OkomT6kygUWaxUbXDdkfRvfj5dJg8CnLEwJ//aLXR9XXqfiBRCCEKh5pylaXt6NmNZR7CsXqLRpjFt8CcTKSBIpjx5tu2fKdxMacPyZmQkxuOPv0rHYZMB14HyQTQhCAcDHGo7wsu/3EbboSpSNT3ohk11wuRQb5BBPYVWOUR5WYTP33MD6xv97yPWqMWWJ96g67DO0ZSE+mPpjVWX98RhJtTEYplRSIOU7TrPVMVwHAvbHmT9+t9a1FjPnp7NWX9lquK4QBIhTpRjy1x/MNg07vqzV6UmGn/XtUkmj6Dr5XhefGwlK/PPPxptGAsfyIwjn5CCmRIl5+r+Vywc3d1HeeyxV9i5S6NfGNDYhabD6evq+dyNhXtfUkkb6UmkNNACHuFIgM/d9REqy3NMFiS4noEIeATCJtfctUlNKhSzZqVqw4nCHJmJBfhhUe7YpClz7YZRQmnpWVnXft+4vrK1IRRqJJFoIxZrRdcr8bzj+JOKQPpcHtFoM5pmjNn+fMLN8kmgV9pQXHiex1tv7eC553bS0lFKqiaGqEwQNA1uveJ89mzdzXtbj9MxZCLrehGmS3VZlIgjOOrpaIaLYWrccfMVrG9cjZSSPe/s4a1nd9ByqIxklYVo8D3RV111Hus3NCz1JS8oamKxzCikQcp2nQeD1QSD1WNxnnPpb647Yvf0bGZoaBuZ0CY/iU4jE1NbX/+Jsfb5XP9E419Sso7RUbDtQQyjEteNpzeoiRCNNhAMVgPjwwZmCinIN1FyLu5/xcLgOC6PPvoiLS0aR23Q1h6jJBzkS7dfx//z23fwib+dXIGjJHqcj93873QcKUfW9iKFRzCQqwztzEysNKVQFJKVpA3Z7U94rD3GbyomxhaU8r32bG0wzXJ0XSceP4zrDhEON2JZhwGBpmnphOzqcbkPM4WbzSaBXmlD8dDScphXXtnGgQMNpBoPo4dTnN68mtf//UEe/L6HFbNJJg0IJkATNDW6/NUvdvOT742v4pex8d1t3Wx7cRsHWtaQXHMELWqxuqGGuz71EcorVt6+FRNRE4tlSKEM0kTXeTI5QCzWAbiTcg5mYj47YhtGCUJE0tU5kvgC4k8qTLNuXFWoTH/TjSuX8Q+H16LrJhdd9M8A7Nz5IJ43vopHItFJKuVXCcm4yDOTjokhBQuRKKlYWFzXxbYdXDeKFnAwDI17b76KS87YSE93iI2nnAi5S6Ucurv62bevivqzgsiGI2iGR0NtFZ+58eq8z/nii58n7obRDI+tP6rE0DViI+egc5QLLv/JQlym4iRmJWjDxPaJRC9S+iGsJyo3BQmH1856EWeiNgSDNWNlYS+66LvT6sK2bV8dC6uNRk8khGdrg9KF5UkikURKiecZaIZHeVmI3/4vt/Hk30aorWlnyLSJJUwoiVNRGkEmmjH0fVP2l7RSSM/vT5gekZIgv3LvxwiH/cWr/3HbRfR3h3CSlxAf/BiWFYBgikjlILd/bf9iXfaCsWQTCyGECfwa8OfA56WUD+doI4D/C/gM/lL2NuBrUsrYxLaK2ZPtOrftBInEIcAjFFo/LucgH4M4F4OafUwksoZ4/FDakyDw3dEOnufvVDqbmuzZ+1Zk786dnYcxMWzA3zn1CIHAGjQtNOYid10H0wxNCikodKKkYmFxHIdXX32Po0clA5YGdYMIAaG09+HgwTCH2k94LJJJGykrcFwB9UcJBU3uuekqrrn4rBkrQWVjWWWU1HahByVrmwMEAyYD3YO0H6go+DWuBJQuFAdLqQ0T20cijcRiB/C92GWcyK8YLbg2TKcL2WG1AKHQ2knhZkoXlh/9/YO8+eZOuo6U4ZQPIHQHXQsgBCQSKVoPNGLbATwB9EkGNA3pBfjkDadx0w0axxICVg8ihMA08vtK3d8don7dCIOd/WB76GEBwSSx4Vo0bflvkLqUHouv4Ps135qmzQPALcAFUkpLCPF94C+AX12E8a14st3HiURLemOfRoLB8Svz+RjsmQzqgw/+C//4j49TUnKiepNtD/Pkkxdjmv7KkW0Pk0odSb8qAInrDiBl2azE7MS+FQHARMoUqdQRqquvzHntqVQvqdQggcAaSkr8laiMizyROEwweN4kt7qq6rF8aG09zM9//jp79oUZCgvE2k50Q3D5Wadw3ilNADi2oLzaBTLBFkkcR8O2I+im4J6bLue6S85esms4iVC6UAQspjbM1D4YrCGZHMRxetLPLJw2zKQLmbDaVGoQTTMnhVwpXVg+2LbDiy++w6uvtnGwJ4pTewwRShINh7jzygv58UPPMDBwBkm7GjOUQAc0IYgEA4wMJ+loC9NieWhNnWiGxsXnnsqp62fOnfBcl0QsTt/+XkbjBjLkQMBF0zQiJVGC+RYRKWKWbGIhpfwHACHEJ6dp9ivAd+WJIMu/BV4gh4AIIb4KfBWgqam2sINdwWTcx/7KT924utq58gqmipOdyaCOjMT5/d+/l9/7vXvHXs+4nX0jD44TA8L4wpEgEw6VSHQRiVwE5CdmmX0rXHcE102h60F0vSZd63zytQNj158h20V+zjnfmHSO+SRKznd33ELvrrvY/S8mnZ19PProi+zeXUuspg9RGqOmopSv330Dpzfmt/mQAExTn7FdBiue4P13dnO0P4DUPND8zR41AclEAmvUwnFKkRX9IDyEYIqtlU4+lC4UD4ulDRPJ1d6vDBXEMMI4ziC+90InleojEvFLPxdKG6bTBZgcVpvNUupCofpYir6XgmeffYO3326n5VAtbGjHMCWXnX0a9954JT/+t6c5sF/D8QRoEgGEQgEMTSMRT+B5YVzDRascprqqlPvuvoGmBt++xEfi7Hl3L31Ho9iREdDtdJl73xOx96UPSAyfixsMQDQGGoRLIlTXV9F3aPlPKqD4cyw2AK1Zf7cCVUKIcinlUHZDKeU/A/8McPHFpyudniWTY2r7iccPI6UfUxuJNDM4+M6UcbIzGdShoRjnnrth3DknH5PA3+Aok7idKS0YJ5kcwDAq8nIpp1L9CBHC84YBf1M9wwhNuQNrruv3j5ta/OaaKDnf3XGnO/7EeOYnTIXevXcpGRmJ4boejmOihxxKSoP80RfvorI0WvBzSSnZtX0/L2/ezv72UqwKG8wUQoOayjKGjg4xNJAgljRwDRtRc4xQOMBqR9I1c/eKEyhdWEQWWhsmkqu9lHFAS5cLh0zJ2cyf+YYbzVYblosuzNTHiTHNfTFrJekCwODgKK7rV+PTTLjwjHX8t7tvZGQkRippY9ulY5OKcChI0DRwbCd9tF81rLQsxK9/+S4i4SCe57H77T28/dwuWg6VkqqJIxoGMAMGV206f6zEbHxw5EQXGgSjAWobV5ZHq9gnFpntlTNk3tXlH4RWZEyOqW0DNEKhdXieRXf34+n9InLHyc5kUIeHY1RWjq+GMPEYIQJImdkYT+JPLvx49mSyG00z8nQpayQSbWhaOL3pkUsi0QoEpozJnctKUz7JghNXeVKpgXkl900Vr9zR8SMyO4nPx/Cv9ORDgUDPEcNqGB4joxpI/8uG4wTxPIEw/GRRTcxsct5/axev/XI7u/etwW1uRws6BAImp21Yy2DPAINDDvGkARELkqWce9GprHKgfWsflkiB4dfeF5qqHjUDShcWkYXWhonkau97tTU0zcTzDE5UiPI/n/mHG81OG5aLLmTaLpQ2rGxd8O1t0JxY7U8QjAwyMlhLwhLYCYHjaDhOCKGn/BYCDMM3O9te2s77r+5lz4HVyA3taKbH2nV13HHndZTlKjOeJh9tWW4U+8SiE2jK+rsJGAUGF3MQK80FmIuJMbUQHFcVCcC2xy0GTlolms6gDg/H+Y3f+Da/8RvfJhIJcd55G/nN3/w0V1114pj+/i3s2/dneJ4LmEAm0iGI41izqMku8b9j+AbD81JkhGgqw7oQdcVzrfJY1iFCofXjqo7MJrlvqnhly2qhtPTUeRv+kzX5cMPGBI2Nw3Qf6WdwUJJEQiDF6PFVNDfUcu6pTTP2MdA/iOPoSB30gKR+dQVnnh2k53CIwaMutq2RcoFUilX1CbQ9R3mnJcpQVCLWdaAbGhvOPZXSypVfjnCeKF1YRBZaG6Y6Z3b7d965l1RqKG3Ls7VBYNtDC6YNy0UXYGG14WTVhU2f/AdeePR3OW2DRWwwwWjMwA0kQXcZHazj2svOI5CekAwdHcJ2AmBKdFPSuL6We++7JWeZ8ZKSQQaO1UAySmrUxE36HvTK1YlFvb6FoqgmFkKIauAJ4AtSyhbgIeDLQoiHpZQp4L8Dj0kpF82lXWgXYDGK0cQxGUY50eiGcTG1EETK5LjjZpOU9p//+UfU1JSh6zq9vcf53vee5uabf4fNm/+Ca645D/DvZ0dHM4lEN77hj+C/1R5CBGhqeiDPeyUJhdbhOEdx3RT+qlYAIXR0XZ/SsBa6rnj2Kk8y2Y9ldQMOiUQLuq6PJULO5j5O5ZoH39BnMxfDP5fkw2L8Nw1+ct6+fe0MDwdImQnQHcDMaeirqkfZvs0lkajG1T3QHERcsGEj/PYDd8yqElSGYCDAD5/6AIBH/u6ndBw06bJMtLVdrEZn5wfNOOs60II2pdWlXHv3JmrWrCyXeCFQurB0LIY25EM43Ax04zjD6c9vpjy5jqaFZ/Flf/basBx0ARZWG1aSLgD09R2nr2+QY8cjyOgwEokm/HDW1v2HiY16jCR0KI9TXjLIoQPVJJMVEExCooRgwOTss3U+ctX5OXoXICAUCk65d9GnP/N3tByuRWxsY81ptdz42RsX9oIXmaKaWAARYB2Q+df7b8ApwFYhhAPsZpErfxTSBVhIMSrUhzbXmBznOIlE57ha3cFgOclkL7Y9NKddXevqKsf9/od/eB+vvfYBDz303NjEAqCp6b6x8Yw/T76TiowRtIhEzgHg+PFtSCnQ9ROuzsVYbcms8vgxyYcQwgBCQIJY7ACu6+YsZTsdU7nmw+GGOVcjyf63BIJUaphIJD/Xf7HG3ra0dPD446+zZ1+EoahEW9eBYWhcfe5plEYmb4T3rb//BU8+9hK7dzeSauogUunx2/ffTkNtNZnVzcIhcVwNV9MQAY+y2hLu+NrtK6LM4AKhdGEWfRXqy9xiaUM+ZOxeMNg0Z12A4tCGhdAFWFhtiMUO4TjHCQTqCIcnl9nNdVwx6oJtOzz//FZee+0QB3ujuHV9iKBNaSTM5adt4KF/fZIPP7DpTYWRjYfRDI/Pf/FbpI4Hae2uRNt4iA3NtXz9vtvUhqfTsOQTCynl9Vm/HwbWZv3tAn+YfiwJhXQBzleMMh9yy2onlRoiGKwdq6U91w9trjH5Rq+PQKBizEBpmsHq1XcSj7cXzCWcSjlUVU2fdzGX80w0sP4XQ5tg8ET0xGKUAMys8lhWN0IYaFoAEHie74qfqpTtdGTadXT8CMtqASAcbqCi4mIGB98BZleNJJcA+KEFFhAjEKihouJMeno209Hxb5O+rBRj7O1bb+3guee2sXtfPW5zB1rApq6qnK/fdQOnrK2f1N7zPDoOdROP66SEDbqDJnSi4ckTkKmIxyyOdg8wPBJEhmLkW+dJCAiEAmpSMQGlC7P7DPmf4x9hWe1AkFBozaz3m8hnXIulDRMpVEhSMWjDQugC+PdoeHg3PT3PIKWFEGHq6z9BWdlZc6pUla0N0eiG9J4efUCScLh57PidOx+cNJEtRl1wHIcf/OBJdu9O0TkcQTR2YegaV517GjdfeBaP/+R59uypZqRiGFE9TDgY5FM3Xsbgvi72Dh7Hz9CDkkh40qQiNhxjsH+Q4eEgMjyKnML+J0biDPUOMTwSQoZHESu0HuCSTyyKnULWpZ6PGGV/yG07AWgkk/0YRslYrOtcPrRT7VINSTQtXBChaGvr5mc/e43Pfe4m6uoqSSRS/M3fPMKuXe18//u/P6n9fF3PE0UoFFqNbQ+iaQau6y7IilouMiLmV7sKAnZ619bmsQpXuUrZ5odLaempY0IxOPgOFRWXzlrcpxIATQtzzjnfmHHlqRhjbw8f7iWZNJCmRA+6NK2p5k++eE/OpO2uzj6efOwVdu82OWYEEOvb0Qw4+9R1lJdEZjyXlJId7+7l5Wc/oKW9lETlMGJN3K8EcvFZC3F5iiKgWHQBTmhDItEHBNE0nUSiE11vxjBK5vxlbjG0YTYUIiSpGLRhoXShv38Lg4PvUFLSPE4XysrOoqnpgVlPyiZqQzS6Lj2hnFkbilEXYrEEQ0OjWFYFojSGGdD47I1X8rHLzmPnBy3YtkMiEUaLWpSUhPifX7mHyrISHt8/dc0+z/PY+cYutj6/hwMdZaRqjiKiCQIBkwsuPH1cu4Nv7WHnL/emK0YNICotjIDJ6RedPmX/yxU1sZiB+dSlnsh8xGj8h9xB04KARzLZTTBYPecP7eRSggPEYh2ASzgMTU33F8CYl9PXd5zrr/81hofjhMMBLrnkDF5++VusW1c3cwcTyMfdP1GEThyzeGKY6X///r9CSl+Mw+FmgsFqbHtozvGpU00G4vH2WQvSTAIw08pTsW0IZVkJ+vsHicUCyEACIaCyNJJzUrHj/X08+9Rb7N67GrvxMCKUorwkwudvv4bzTm2e8VxSSp559CU+fLeXg0cqkc2+67xp7So+d9dHqK4sn7GPufKnt13G8e7JHpXK1Qn++KmtC3ZehU+x6AKc+IxCL5oWQNP8/Idkspto9Mw5f5lbDG0oJPmGgS21Niy2LvT0bOacc75RkEXHfLWh2HQB/LyKZNLDSupQlkQIQW1lGZ7n0d11lLil42g2aBJd1yiJhIjHLIaOjxCLmRCKj/NEeJ7Hc//xS1o+HKatrwzWdaLp0LR+NbffeS2l6ZLmUkre/s+XaPtgiLbedDsDaptruebOTUTLClf6vFh04aSZWMw19nQ+daknnm8+YpT9IRfCxK+2qKcT0Ob+oZ1cSvAQ4BEKrZ+3Kz1DaWmEb37za3zzm1+bcx8Z5hq7WegEvHw5Ebp0Im9kpiomi+klmEkAZjpXIb9gzQcpJe+/v5/Nm99lX1spVsVxxJoYpqFz1Tm5V4TaWruwEgaOBnrIoa62nAe//KkcZQenPmdP51Hi8TAykkQPeFx58Znc9fGrFzz+9nh3iNWnxCY9332g8PtzrHTmog3zCc2ZeL7MPhAwt89Q5jOarQug4bqpeX2ZWwxtKBTzielfCm0odl2A+WlDU9P9RaELAPG4xdNPv8HbW/voGCxDNhxGmC6rKioo0Qz+5TuPsXu3wYAeRGw8iGbAmRsb2LN9P688+yH720tJVB5DVMYxTIMLzznFv34rxWDfICOjpYiSBJoJV206l2uvv2Tc+e1EipG+44zGSqAkgRaAs688k4tuuKjgOlEsunBSTCzmm0g0W8Mz1fmamh6Yk0sSxn/Iw+HVxOOHkNKvcpNtkGYrkicM3EPpet4AYXTdGFuJKFRcZCESCwsVu7mYFStm+yVkMbwE0+XrZAvATOdaiHKMs0VKySOPPM/27X20HilHrutEMzzWra7hv919E6urK6Y8VuDnOSAEVeXRvCcVk/sRCCFoqK+ZVixk+n+Oo4GeQu25vbQs9hfSXOebaxhjhsxnNKMLnuevpoKYsy5kru9EzL6/qVcwuIZI5MSXyWLRBqULhdUFv5CHhm37FZxzTQ6mO18x6ALA0aPH+dGPNrNnTynHQx6ioZuAqXPz5RewsbSMx/7jRfbsr8Ne24kIpSiLhrn31qs5/H4rT/90B229ZcimLjTDo7HB90TXVI33RIv0/zRNUFtXnXMcY+3SOlFVX7Wik79PionFYicSFdolCeNXjwyjEtOMkUr1YhhlY+X2gGlFcnqj6eHHewbRNIjF2tPnyr3b9WwN8HwEPPtclnWEUKhxVvW+c60Q9vW9hOfFkdIlkehldLQtr7HMldl8CVloL0H2exEObwQ6SSZ7cd0U0ei6cQKQz7mWyhuUIZFI0tXVx/BwFEosjIDko5ecw+du3jSnMrELivSQEgZsAesOoekaDac1LPWoTlqKRRvmEsaYIfMZ9fPt1pJMHsFPsF1HU9PngbnpQnbM/shIC6CTTB7HMAamDb9dCm0YGtoGRIhGG8ZyDpUuzF0XMu9DKjWE48TJFPKYjTYstS4AdHb2YlkJRkZWI+p7KCkJ8L8euIs1q6r48UObicVMHMNDD9usXV3N73zhDuyEzbbNWxkaLoWyOEZAcs0V53LbRy9f0ZOBQnJSTCwWO5GokOcbv4Kgj33Iw+HVrF//pXEf3J07H5xSJGFqccmInaaFAC9dpSL3btf5VB/JJSyZc3ieTSy2O12DXNDR8aNZiU4i0UsicQhdN8YEZLpVmVzGsrv7UQA0LTIWPuC6x2ccy+R+H8Ky/MSucHgtTU33FcSQLrSXYOKXm5KSddj2iaS8bIpl5Wk6YrEEtu2SsnWIxtE0wWmN9dNOKlzXY3Q0jm3rSN2ZU3UOK5bAcVySKR3M5LRtk1YSzzOQEiQSEYkTKS9h051Xsbp59azPrSgMK00bhJCUlZ097sv8fHXBNMvHcvr8yYWf15dtk/KpWJjpM1cFIdd1SCY7kNJGCBPDKJtxcpdt24Xw97XILIhNHN90xxZSF070XXhtWGxd0PXydNnxybpQiPMtNFJKBgdHSaU0XM1BCEnA1KmrrsBxHFKcCwAAACAASURBVOLxhK8ZuoMQgjWrKgmaJsMDwziOi23rEHHQNMH6xvpJkwornsDJ6E7InnIcqXgS1/ZI2jqErSnbrSROionFYicSFep8ueuIT127ezrRmm5lLnPceFe6wPPG73adT/URgAMHvjNp1UcIDcMoI5nsRAgDIQJ4no1ltdPfv2Va96/nOVjWYVw3ld4wzyUW68AwKmZclcl93W763gTSrXQ8z8OyOvN+X1pbv43jDOPvAguW1c6BA98B5r9B1kwu6Mw55mrAZ/vlphhWnnLhuh5vvPEhL7ywi5aOClKrjiIiCcLBEE11U3/WOg/38ORjr7F7j8nxAIhG33Nw5vrGvM4rpeSDd3bz6rMfsv9QGcmqY4gSi2AgQEN9ble4a7vj/q6sr+TWL986YaMxxWJzMmjDfHUByAq/1fC85KTw25kqFnZ0PEQqNZzTG5CZjGhaACECgEsq5Zc1nQ7/y3sfQvSmdcEGTGKxLjTNWHRdgP+fvTePjqu887w/z7239iqpJJVWa7O8YTC2sQEbMIshkGAMxJBMQjbSTZruJDOT3uac6aRn5u0z/b7pnpnu6e43fbrf9HRPh2yThSXgmCVgMHjB2HiVF9myNksq7aqSarlVdZf3j6tdVVJJKskG+3tOApTv8tzr5z7f57d9f4vLDbOlrY5ef6l4YaH3W0yEwxFeemkfx45HCap5mHVNCMWgpryY1kvt/PrlQ5xvcBF2xxGV/UiyxOqacg698yEH917k0uVCUqXdCFcCl9NFeUnh2LUN3eD0gdN88FYDjZcLSBX3INwqDoeLkonHGQaXDp7l9JvnudTmH1OMsjmcFJYWphv2xwbXhGGx1AWmubrfbGH6qZEBSws7PWnNtGiMkt1ox0+rE6g6rdt1NuojbW0/QtcHkSTXJK8P2NC0AUDD6o8rY00/x4yeqeHh8xjG0OiIGV2sIZmVlyTdcy8UXV170LQ4kuREkqzxGIaEYcRy1iBrphD0QpHrzdSV6LAajcb58Y9fpb7eIJhwIqo6kGW4aUUlzz52P35f+mK1t9/8gIPvXuBCaxF6ZTvCruHP8/D0Y/dy04rZDYuEmuTln77B2VMJOiIezGWdSIpJTVUpX9y1nUK/b9o5qUQK0zTRNBkUy7OVH8hfkFFRUK6mLcgrKFfnfc1rEdcCN6hq74ic6XhTu7nwgiznj3FDLHYZ0xSTul1PjoikVyyMxxuwouGuadGAZHIISGAYSaxO2g4s50pm725T0/eJx62aQNMc5QUJq2YpNms37sXgBVhcbpgtbXWhWAwj+0pww5kzTfzqV+9xrqEYtTSKKIjjstvYde/tmD1D/PB/76exMx+zssMq4i7w8dQn7+DE3hOcPyPo1hyI6svIimDNyio+9+g9+Eakx2OROL/58W9oOANdCQdUdSDJsHxlBTsfuwfvyHGJqMr7P36LS2dMgqpr7LjSujLufvwu3L7Zpczng6uFF64Jw2KpQ3a5ut94h85+EongiMdeQVFcGTajQ6RSYWA6aVme//SLxtT6DY/Hltb7lY36iJXrqmEYEcZJwgFMVSrQAA1ZLh3x0k9HX9/+CUaFjEUaFvkI4WTTpu/P+g7TLZYWCVna4RYZGUAKlys7j7U1Xo1xI8can2mmMj7LTJhrCHqhyOVm6kp1WG1p6aSvL0R/fw2itgmnS+Zrj21ny40rM56TSCQ5e7qRgQEPuieG7NTYeEMtz+x6AJuS3VIYvNxNX9cAfX2VsOwydqfEow9u4c5bb5wWKjdNk6b6Zg7++iiXWouJ5oWR8iIAuHyueT87cF1SNke4FrhB11MjdRdM83RnywuS5EWSbDidJRl5ATIrFlrrNtN4wUqpHY3mmVjcoQEgy+k3/n19+wkGfzX1jQB2QCY/f/2s6+Zi8AIsPjfMlLa6UOTayL5S3PDhh2cJh+0kJBPJo1JWnM93vvI4Qjf5waGX6B/Ih7wIitNgy/rVfOmRe7hY38zQ4JDFJysu4XArfOmJ+7lxVc2ka3c1BxnqD9PfV4NY3oTdJfPop+9hzQ21k47rb+km0h+mr68aUdOC4pK4c+cdLF+3fNGeG64eXrgmDAtY+pBdLu5ntwdGul32IUk2hLBjmgk0TaWt7UdpN6OWtyR986JMi0a2ZDeb+ojfv5ZweHRiWwuzYcR45ZWn0XUZIUxk2SINw5Dx+frZvv017PbJH+8orNQqGYt4TCxdBQPQsNmy25ilWyxl2Y+uJzAMHcvg0UfGK2ZMy5r4HuLxbkZzji3oCCHPy7uTyWsYj19K29V0PpjqOVqICs1EXKkOq8lkCsMwMUwTIYHDobCubuYNgK7pGIZpzUVFR5IkbrlhedZGxUQIAUIS2B0yN66unmZUaCmNvb94m3PHhmgb9GBWBBGyjpGyITlT5BXmzfmeucK41vlNa6/YIK4ifNy5weutIRKBZDKEJNkWjRcyKRZaEfBRjPMC2HjrrccJBuvw+foRwkSSLCPD4xlkx45zaZ99NOUW3MDEqEYC0Ckr2zHr+5udF0bTsHR0fZj6+m9ntf5+1LhhsXgBrgw3mKZJIpHCMARIJpIkWFNdRr7Xw0B/GEM3MA0ZZANZlti8tg5FtpyWJiCEpdrkcMisWj5dVGOsEk8IGFn/a5dXZBjMSK9uSaDYFSpWZDjuKsHkHhgL44ZrxrC4WjCX0GBZ2Q4aGr6LtRhLWAsU2GylxOPt5OffMul4SfIC0XkVWmVDdrOpj1gLvuX1GRkRoKHrCoFAB7FYAU5nBMNQMAyJgYFyIJmRCCwPjwfLG6VhGQACELhctTOOdabnXrnym8Bojm4rQnhwOCpQFGdWHpWysh1Eo01o2tAICQEkkOWCrEhtKtJ5z+Lx9pF82viCvT2ZJC4z1erMBUtd/JpKabz55ge8914rTd0B9JoWJFsKv8+Hw5Z+OTNNkwvnW9jz8mEaGr1EvEOI/GFsio2SKdKBuUJPey+9bT309FRAeQeSouOJexga9COqgrNfYBExqnXeejoxcxL7dSwpFpMbXK5KZNk2LcqbS16A9IqFNpsfTQszlRcgSU9PNYFAFx7PILpuxzAkJEknHC4Gzqa9n8ULlnohuDGMxMj1wOWqzWpNm4kXurr2EI22ommD2O0VuFyVWa+/HyVuWExegKXnhu7uAV544R1O15v0IUHlZYQElcVFnD5xgdd/fYQLzX5i+f0IXwSbzU6g4Mo5eK42TOyBsVBuuG5YLCHmGhoMBLbR3JxPKqVimkmEsOF216AoBSSTvbPmQ6YjqnRGx86d6xga6qao6CROZxhVzae/fwN5eaXs3l0/aTxgLbxCRKepj7S1PYfTWYuqNmN5bKyFNZWyo6o+Uiknuu7E6RxCllMYhkWKDQ1eGhpOpHljdqw0KhVrqjqxjAyDcHgtBw6cwOVysn79SpQZvM6ZyNHyqLgnSdeO/j7Twjr+vOPKH4oSwGbz0tb2HF1de+bkQUrnPUsmu3E4SnPi7VlMz9FSFr92dPTyi1+8yZmzbgZdBlJVB4osccfNq/nKw3ePeJ4mI5lM8coLezn+YYjWARdmRRdCMSgL+PntXfdTW1E8pzHouoFhmhgmIDIrSZmGYfnARjxbkgRO1cXQ9b4V15EGS8kNc+GFYNCBx9M2iRskqY7nn1cnjQXGN+lTFQuPHXs2LS+ARG/vCgKBLiKRIpzO6AgvCMLhEjL1eLHWnNRYtMZaM61oRXX1l7J+55l4YbRuxDBK5rxmTucGHbAhy2Is0jKXNXcxuWGxIwpLxQ2mafL220d4550LXGzzo5UHEc4kXpeLpx64nfb6Zt4+EaMj6sassGriKkuLeOaJ+ykPFADWum6a5sianXmNNnQdwzRHxAIyH6ePHGfAyPWuLZna64bFEmI+H7LLVYvDEZ+0+U2lwrhcy9A0K187XT7kXIhqaKibW2/9KZrmRdPcKEoby5ef5ejRp6aNZyYPlrWQxJHllcTjwTFJ2WDwRoQIkp9/iWTSzfBwEZKUIpl00dVVyG9+05r2esuWLWf9+h7AjcMRQ1FiWIbIOk6fzgNakWWduroTPP74XaxcmX0uLCzMozLxPYy/a9fY38VcGzDCZO+ZohTgdFbOa2xTsZieo6Usfj1w4AS9vTph1Y5cMYzP5+QPn9rBiorMRZitzZ20NgcJBiuhqg2bU/DwXZt55J5NSJbLMysYhsGJw2d4740zXGgNkCxvR9hVnA4vbpcz84mmCYaBoQl6TQ2pbABJlvDkX++QfR3jWCpumAsvBIMObrrpNIHAnknc0NcXoq9vYNLx8+EFp7OcYPBubrjhEJJkEosFgCQ2W4rh4dKMG9DRNcfhCJBIhBlNhyov/3TOUmxywQ0T3/V8eGHicYvBDYsdUVgqbgiHIxw71kB3tx89fxjZrbF+ZQ3feOJBGs+3cKSjl66uGlhxCYdT4rMP3cHdm29EEgJd0zn87nEOvt1IU0cxWnULQtHweXzIE/hB13VOvneKI3sv0tReMnJcCpfbPcmpaeg6jQfOUP/WRRovF6NVtiKUJA6XD8V+7Wy3r50nvQownw/Z7a6dUKjmwOHIR5IUqqufBjKHsOdCVEVFJ0eIwwsw9s+iopNA9nmBE1OlfL4bMYwIsdggzc1bSaUGqKsLYbcPI8sJdEMmNFTE+5dX0R/oSnu9/oSXoeY1rKtswiMbRCMu6tvraOmrhNFzDIne80W0dxzk1s35rF27nNF099LSQipm8EjnyqOS7bse9QBORXl5gt27J5ON5THLjbdnMT1HcylGXahCiKZpmKaEkEGS4Yaa8hmNCuscq67CRCBkcLsU7rvtpjkZFaGBIX79s72cPSPoMZQRhQ+T5TVlfOHT23HYp3frNg3Tim4YltdKSCaiMITT6+KOnVuoXFk5/UbXcc1iqbhhpp4W6b5Fn+/wNG5IJNx0dWXf2yEdL1gbzC8RjVZz+fIjVFX9Grt9iGTSw/BwAJAzpg5NXHNGa0ZyrTaUizUzN7xQP81oyxU3LHZEYam4YSzaYAqQTWyKxENb1uNy2K313xxZ/yXweOxs27QWSQh6gv38+hdvc+6Mg0GHQFS1IyuCm1bX8pmd28Y4YqB7gL0/f8c6ziZD1WUkRbBiVSWP7NyGLFvHDfWEOPLzfVyotzOgyFDVjiQLylaWs+2xO68pafHrhsUSYq4f8mjnU7u9hFQqjGkmSCS6J3lm5iKnl4mofL4gXm8biqKiaU6i0XJSqXyczjBzMSwmLiSqGmRgQHD69BpaWwMoShuBgIeSkkEMQzAwWMLRU3fjvfU8BSOPvyyvlZWBi7jtcWJJF419q+gYWsmB9slKP6PHv/5Pf0Ak7LO8wqbMcz+UUSRwu0Ps2PF9vN4Et9wS4NFHt+HxTC/2zpVHJdt3HQw6WLlyuuxbY6MzbRfYUOjIgscGi+85yiYPeyEKIR0dezlz5qd4vSHW3OQjVbKMy5FyivKnS7uOwjRNTnx4jt+8eoILLSMRBlsCh92LfY6eo2MHT9EdjNIbKkWsaMbpVvj84/ewbs3yjJ1YQ30hTGM8VC6ESfUNNdy9axtKhlqQ67h2sVTcMBdeGE2BssQ5XCO8UICuu+ekcDTbBlPTKgmHl5OX14KixDHNEtrb7yUQ6B171nSbzpnWjdk267MhF2tmLngB0ncIzwU3LEVEYTG5oa9vP01NzzMw0E7FMh892ipCCT+yrJDncvL+/pO8/eYZLraVoFdZkQiPyze2Zh9+5xjdnRoh1YcoD+LxOPjKZz/BytrJe54T+07QG0wRiuUh6rpweuzsenI7y+smF3dfePc0/cEUg7F8xPJu7B4b9zx5N8tWTC8C/7jjOsMtIeb6IY96PFyuyaHuWKxl1ntlS1R9ffvxeGIjRoULIVL4/U1EIuUEg+uyeq5EIklbW9dIf4pqhoc/z969pznf5EP1h6hYfoLyyhY6g7U0t67F4YjjcqmsroPv/r6VEzvUt5/etveQlCKE5ME0oqyqbKO4+l7yMiwuH/xwOTetGaals4doLGGl7xqCoVAxPfn99CTsdL4R5+LFF3jggXUERvIpAaqrS6cR3mgh5FzrJBbq+fF42tIW0eVKoWOpJTXTYb75vMeO/ZLu7p8SHnITM104vRHuLDhDyl7Dw/dtSXtOaHCIl36xl9OnDXo0G+ayDiTFpHZZMc/segCHbXqEYSZomoFpCg7sexr1bQcOh40TPx8v+isuT/CvuyfXCBm6MfkiQrD29huuCqNiXOvcMX3ndR1XBEvFDXPhhcrKQgxDxjBkhNDw+5sIheqQZe+cvdqZNpi1tWfQtBNcvlyLrt+ILMdwOGIUFw+PjWM+m87ZNuvZjBcWxg25iAhkKrDOBTdcDbwwev+5ckMw+A7nzv0joZBEOJqHwzfM1g0fkNe1ibs3PsEbL+6j/rRCv6yMRA5gzfIKfuvx7UgjhkUqpSGEBJJAlgU3rKqcZlQAaEkN05StWjkZ6uoqphkV/37nJtrOriepaiRSCtgTOJw2Lrxru2okYGfD5B4YC+OGK89y1xDm+iEvJAcyW6Lq6tpDOPwUJSVBIIVp2jFNDa+3g/7+LwK9Ge9hmianT19k9+4j9PbarFAkkEjK9Jv2sQ3dH/7H38UuayR0F5Iksay0EJ9TQ5Z/yhd2vkx30EltdTey9B1SuuWFLirq55vf+gdCXa9lNCwA8r1N3Lf5CIrcz+Cwm8bWG2iMFOPxC+JqFNUT43RXEe3PNeFyWsohQkAgkGTHjlvZuPGuBefDLtTzU1R0Mu3CGou15EyrfKklNadiPnNZVRO0t7+MqrqJaQ6EM4GQfVSUuHE5m7BnKNjfv+8Yl9si9ISKoa4Zl0vhC49sY8u6VdMiDJ/buZ7u4PTNRmm5ys92n5o8nnge3orLeDwOaqrtY7+3ZrlZuVowSnS77GfS63lex5JjqbhhLryQSDzL8HAtfn8TqZQNTbPh9bbicBTNS+EoHf7sz/4Qw5heJyJJLnbufBmbrRNJ+g667h3789LSNv7dv/vlrOuZw3ERn+8wDscgiUQBw8NbgJuzHttCayVyERHItOnOFTdcaV6A+c3lpqYXiERMhob84ImSNJwUub08eluU+tNtdFxOMhD3IGq7cLtt/Nant7N+VQ1CCJKJJO++/gENZ4Zp6XNjVrRjCpO8CU1Vv7pzI72dDiLDUdToGhKaBEoKZ94gm//u5Wnj6Qs6KSlrJRJKEVXt4I3i9rkYDKaX0r8aMdEAWig3XDcsRrBUHSLn8iEvxOORjqj8/rV0de2hre25sWdMJvtwuRTOnduG19uF3a6STJaiaW7y8krJZFgMDIR58cV3OHEyRXfSgZkfHhc+8OgIewqX08lnP7mVMuMEkr0YRR73FGt6Ci3ZaxkVK+PULjuPqgYQWEWHHZ3lCMmDlsxs2Hg8rZQHXiOpeUhqJeT7otyx8RiStpr/8R++ypH6i/zs1UPE5D4GkgroIzmOJnS059H5z6fZuOEcu3bdtyCFjIV6fpzO8IhU8DgWUkQ301y+Ep1QYX5zWdcN7PYYw8P5CMnqG1FVXIDT6SSV7Ml4XiqZGq/FUCRWLS9l682r0x7bHXRSt3J6h9+mxgmpc+Z1JadrGR8nbsi0VgGT+iJEo634/XFaW9cSDFaMcEMMMEkkNhEI5KaT70ybymDQwb33WrzACC8AdHZWzpqK5fG0jRWdq2oARYkRCOzB45m7vPR8uSEXEYFcFlhfjbwA853Lg6RSI84dAYpNoixQgZbsITUaYZBBVgSb1i5nw+paABrPtvDmK4dpuOglkpdCVHWgKBJbbrmBB+/eNHb1jhYJh9IIpoItL4lN0UAS6PEq7r1v8yK8hY8XrhsWXLkOkbNhoR6P9MpFU59R4lvf+n5aj9G6ddM/IE3T2bfvKG+/3cilTh9aSQhRlMCmKMiyZVkIBDevXsVTO7bhcjpoqy9CN6Ig+8euYxpRFHvR2H8nEwUoShRd82Y8ZiqKi06R1DzomhcB6JqX5MjvQpRy+82ruXl1Lb947SDHzzdhmpbMoa6baO4eelQH/9d//RP+y3/xsW3bl1DVIiTJMj6Kivr5gz/4UdYL+EI8P6qan7Miur6+/TQ2/j2GEcM0dVS1m0ikeezPr9Q8n+tcDocjvPTSPhwOL8KWsKq1EciyjGlEsKV5N7quc2j/CepP9dAczMco60Sgk+d2z2vMhmFw7GA9Z0500tThB7vVPViZQ+H3dXy08XHkhqlrVbpn1LRBvv71/4bHM+5xHeeFh3L2HLNtKhOJAhQlNlY8DiDLsVnXxoUKkkys0Vi9+svE40WARFFRP3/8xz+dl3rgfJCrAuurlRdgbnPZNE3OnGmit1egaSaGU0VgIAsbujZMKCRz8cIQrb1ezJJOTEx8HiuiHI+q7HvtfVpb3AzbVaT8CIGiPJ7+7CeoKB3fZyTiCWJDUQynjCbpCJuGzaawrKqE3nYv9iliHYauo0ZjiESSmGoDZxwwkWVppLvKtYfrhgXjHgnDSBGNnh2Tw2try175YjGQyxzITF4XTYvPKFs7Ec3NHbz44nuca3AScjKmorBhTR1PPbINj3vcyytNSDnxl32K3rYfkoKx+glDi1BUvWvsmHi8gqqqPUjCIJn0EQ6JacdMhcsZRtM8kxSiNc2DyxkGLC+Py2HnK4/fx5ceu3f8XmqCn716gA/PXELVXHh8vTgcYXy+XmTZj6LIdHaW5Vxzu7w8kTbHt7a2Luu/g9nQ1vYjdH0QSXIhhA2rc+wgbW0/wm4vvCJdsiH7uazrBgcOnOCtt85x8bKPZTfVsnXtSSTNTqCoBKctgaZFWTbl3bRf7uKVF97j7Fkbgw4ZUXUZWRGsW1XNEw9snfN4tVSKH/79i5w9Y2NAsfJ0EQZej5Nlpbnv0XEdVyeuVW6w20tJJHqw2/2LKhU626ZyeHgLZWW/wGZrRJZ1dF0mGl09ayqW0xlG0yY7FDTNnbUgycQajcJCFVluQtO8dHaWAblVT8rEC+XliZwVWF+tvADZz+VQaJiXXtrH8RMx7EXr2LrpEE7NhpBdVBQodHe18/7hO7kQNUd6HAluvWkFn7rLikRomoZhGJimguRIYLPLPP7JrZOMCgBdG+2zIhDCRAhBSVkhDqedqehv6+bDFw4RGdiA7hfgjoEkcHnd5BcX0BPN/fv6KOC6YYEVbjRNmUSiHSEUhLBjGCni8Rb6+vZfcQLJxf0zhVR1vRtZ9jE8fBGwOrRO7bwZicR45ZX9HDk6SPuQHbO0G2HTKfB7+erj97G6dmbVg9EaiVDXa2jJXhR7EUXVu8Z+dzkuUOA/TWR4GU7nIHZ7hLz8Tjz+22asr3C5DXq7fWN1GQA2eZh8f2zasRMNHY/LyW8/8QD33nYje/9JQdhTNLStYdPawySTkEq5kOXhnBPpTGokfX1P52STEI+3Y3WlHV0EZQzDGPndmDYHUikVVb3IsWPPZh0Cn2/YfLa5fPlyFy+88C5nz9sYsEmIqnYux0pZHr+HLSuCYISQ5QBly56kcMJ1UimN3S/u41KjwgA6UtEgeV4Xv71rOzfWza23CYBhwlA4SvMl6NdBKh/A43ZQVJBPXbV31vMHewY5/+FFjLH6bSuVSlauHbnBjwuuVW5wuSrRtBCRSAumGUcIF2VlD+f8ebPZVAphqapZ/25mlZnodht0d+dNqs2Q5Qh+/9x3esPDWwgE9oz8l0EqFc4pN8ysUpUbA/Jq5gWYeS7rus7+/cd5660GGtt9pIqHEYYX++Wb2X7zAPkOlc4OlVOn7uRMbyVScT+F+V6+9uT9rKwqByAaibF390E6OuwMGCnwhZGEmCYVHo/GObD7ECntFoRhglOzepymiVJryRTHXjpA2yW71fLRkURWZAIVRTi9H626u1zjumGBFW4cGjqDEMqED8/AMByLarUvZV5jupBqPN5OMhnG5yslP/+WMW/IKEzT5MiRs7z22nEutOaRKIwhlsWx2WQevPMWdtyzeVq34+/995/xf3/nX3j+N/+NO+/dMPZ7XmBbRiOhwHdkLKUprloLQW+Pl2Ts7Rmf6T/9ZR+9bT9EUryTIiHF1V+edNxQ3/4Ro6YfxV6Ev+xT5AW2saKqnPJiP+58la4+iWPnYGX1eXzuEPF4GfH4vRQW3pndC54DspVPzPX8mDoHEom+kW64jqxD4IuRGqKqCfbs2c+hQ720hVwYJd0Iu0aBz8PXHruX9StnLoDTNH2k06kTyZ5CsUl89qGt8zIqADANME0MQ0FSTGRF8Mgnbqdtn5G2ULu4PGGNI6Vx9M2jHHuvnZ6wQlWp1Z1VIFBsCk7PtU02H0VcCW5Y6nz3TNxgGEl8vtVjnvJQ6Ah9fTcuinGR6Zo+32Hi8VKGh1eM/RaNeunq+v6M4/jLvxykre1vJhVcW8bA02PHZPueE4lV9PXtwOc7jMvVjyS5Fk096Upww9XKCwBtbUFefPE9zpyzMWiXxrIkbl5Zxe88+hXyvR4GB4Z4552XaW8vQHIPY7fLfGnn3aysKsc0TU5+cJZ9r5/mYouPRFEMURHHZle4/64NLK+29humaXLuyHnef+0UF1vzSAJ2ZxwhCQqK8nG7p6/dekrH0Ax03YbDEyIyWIo338tg13hko6A8N7VIHzVcNyywwrHh8DFM0wHoWMSRwulMXySWi496qXN304VUk8luHI7StCFQTVvLiy++w+l6QR/jkp01lSU888QDBPx50+7RcKaF53+yl/Jl2YeIS8tVBgecROMVwLiRUlTUi5bsn/Hc2SIhMCpjaxkfkr0Y3YjS2/bDsfMFgtJiPwV+Ly2ddg6cXAYGRHrL+dGPj3LixAs88cS9lJdnbrQ3F2T79z7f+eFyLSMev4xhyFgSiQaQwuWqmjYHYrHLgITHU40sy1mFwBdS5J4Jr7/+PsePd9AaLIG6Fmx2ePD2DXzm/tszKj+NIplM8c5bH9DTLeiPS1AeQgiB0zE9bJ0OpeXq5EJtTKLDwR2l1wAAIABJREFUURzOED0RBcp7ECOeramSslNxav8pzhy+xKWWCrwrmxmpOEdIAlm5XpfxUcRSc8OVqOmYKzcsVZSmvDzBwICTeHwZ1lpmoaiod9bi7dkiIXN9z4nEKhKJVTQ2Olm3bm2OnnAyFpMbPoq8EInE+OUv93LxopsBJYlUNEi+18Xv7XqAdSNOI1VN8O7eI/R0KwzqOniHx9ZrgKaGNg7t/ZALDZUkq9uQXAlqqkr54hPbKZzQB6m9sYNjb33IxQvLSFRexunrIxldRn6+BzWs0B62jgtkMBS2PfIPyBX93L1rGzVrPzpKUIuF64YF1sfY1lZDPN6FaSaRZTsuVy2SpCBJ1qZjlDDi8RaSyTAORwlOZ+W8F/6lzt1Nt9AqSgFOZ+Wk40zTTU9PMz/+8Rs097jQS3sRjhQel5OndtzJ5ptWpm0Kpmk63/raX/Hn//Pr/P7v/HXW4/rJ7lO01T+HbsSx2caLu1OpELKUuXB7FDNFQsAyOiTFO35t2U9q5PeJ59ltCqurKwgNR2jr6gdFo6+gj3ePldLW9iZ33VXBJz95B44pG9Z0jZg8njZWrXqHb37zb6dtLrJdgOe7UFdXf5lLl76HpsWBBKCgKHlUV3952hwwTR2nswaHo4hEoo94PIhhWN73TGkeuVQpGUUkEkPTFITNRLLBLWtq+MJDM0eKTNOk4Vwzr778AQ2X3ER8+og3S+KuDTdw04rsohUTJWWbL17mzV8d4nyDm7BnGMM/hCQLNq1bxY2rq2e9ViwSt4hbASFbIgaSJIDp38vVgdUrZz/m2sZs3DDRkACJVCqEy1U+b6Ng4nc//k2qXLjwV3O6zlyfcfTeM3HDQr/zuWL37nrq63+YQY42O2XEpdwIT8XVxA3z5QWwIhixWCemGaO+/ttpjeXF4AVVTWIYOrpuR3LFcDgUfvfT97OursqSuj9xgTf2fEhDs494fhJROYhik9l++zpWVFm1MGo8gWEaVvTZppOX7+IbX3lkWhfsREzFME0Mw4awGXzy63/D17/5b7BdBX2HrgwWxg3X6lubhurqL0/Tqh7No5zoIUilVEAikehDUbxjH99cF6Qrkbs7daGtr//2WAjUNGF4OEp3dwdDw04aVQ1R1YksS9yx8QY+89AdM3qB//a7P+WW29Zw130b5zyubIq75wst2Y9knxxt0FMJEupFmo59nU0bv8Vrrz1JODx6TBFQA6Q4+Pwfc9fn/4KmIS9dvwpz5swv2bnzdtatWzFmXE1txORwXCQQ2ENr68q0m4tsF+D5LtSTSWK613TiHLD+/uMjxNGKEAqjUaNMG6JcqZSAZRycP99Ca2s/XX0+zMI+TExcs0QbhsIRdr/0LieORwmqdsyyIEIxKAv4+dqTD1BdNvexdHX08saL+zh/roRIYQ9SXpTCAh9femI7NZWl047/6s6N9E7ZNESG1qGIXjZu/Rlw9ZoT45hjp8BrFJm4we9fO8lzPDR0CtNMYRjF2GzZeXqnYvS7n/xNOjDNxKJGLmbihlHkWswiGyxWd+hM62s8fon6+m+zdesO9u9/kGBwBbHYeGNVm81k5851WXXvvpq4YT68AIzNQ6uuxY1hxNPOw1zygvWMKQ4ePEVfn0JY08EdRQhw2G3094V45cV9nDqp0avbMSusbIqqsgBfe/IByoosJ+Jg/xCnjp4nGMxHzx8ASUeS7NMco8OhYc4daaCry4eWNwCyNuI4yLyC65pO46EzhPshFJehPAwI/vEbjxIf8k87vqBc/cg0yrOwMG64bliMYKbQaX39tyd4CDQkyQEYJBJBHI6ieVnmV6quYyJGF21N0+juVonFoqCkONlXh/APUxLI55knHqB6ljSg08cbeeGne3njg7+f1ziySWmaLxT7ZKnbVKIXbSR/VLIX841v/RNqXGAq64knJvc6uHC+lrLSArrEIDFvjOOtAYL/6zi3bLR6X/j946HU0WZMRUUn0XUFWV6eNoyc7QK80B4m2cyf0b9/Ve2x+j4IARh4PJZHNt08zBXRDw1FeOmld/nwWIRg3IVZFkRSDMoDBTx616a05+i6wfsHTrJvRDEqVRxBFKg47TYeu+92Hthyc9oiu2ya4MUicQzdQNNsSE4dt9fBN57eSUF++mLt3qCDmimdffuCg7Q2TieV6/hoIxM3TPUcW0X6tjFegLl7bUe/+3g8OIEbUkiSC0XxLjk3QG439HNFLtWvJmKmmkOHI87Xv/46XV2VPPDAK/T17SCRWDV2XLbdu2Fykz6ns4dEohBd914RbpgrLwAjkQrrd49n2dhcnzoPczVfTNPk3LlmXnnlMOcbXUS8OqLaikRvu3kN7eda+T/7mmjqGpG6dybwOB189qGt3LHxBiQh0DWd9/cd4+DbTTR1etFK+hGuBC6ng533bx3jiKd3bqCt0SA2lCSRuh1sKRAm7rwhvvvz/dOiGqPoaQ5y/MX3uXjeRdgloLoNSRHUrK3h6PP5lK+cLhAw3tH62sB1w2ICMn14Ez0Eo1JtIKPrlq79fCzzuebuLgYKCu7g7Nm9qOoh7HYdZJn6zhqC0Ro+89BmtmfYqE1EMpniW8/8D777//5bPF7XjMfOhNlSmuaLqdGQRKwdkHB4Kq2GfbKfRMJFacGRaYaFXbHxn37vs7x79AwffvgCaza8h8ehEg3l84MfnGL9+k8Cm8c8UZZWugAkCgubSCT6cDgCkzYX2S7Asx2Xizqf0eMbGv4CMJEkGy5XLQ5HEbqup90QzUT02YxJ1w0OHjzJW2+d5WKbj+SYcaCw697beXjrhoxzbt+bH/D+wYucayyFumZku8naFZV89bHt5Hsz96rIqgneFAhAnsFjNRVJNYEaUdG0PMz8ARDXm+p9nJCOG6xGo+OeYyFsmKY2xgswd24Y/e4NQwUcQArDSI2kXy1dKlIgsI2hobN0db26qKpQ2Y4l1/fNpq5E160+GD7f4UmGxVRkWvemNunLy2vBZlOx28cL0XPNDbnkha6uPZhmDHDj8Syb0VjOpqZltnGN9i46djw2EonuQhqJRH/jiU9Q//4Z9h9o5GJbAOpakBW49UarX5bHZUWP2y6188avDtHQ4CTkNMeKvdetqeMzO++aJInfekHHZe8g5XBiK4yCBB6vC0OtZd3NXWnfTX9bDx/+fB/nz5YRK+lCeGN48j1s23UnZTVlPP/nV3+ceilw3bDIAhM9BC5X+UhoUANsc5Kem/pxKUoATRvOWNexmLh8uYs33/xXCgtPE00VoCJw2lLcWDnAjgdqKavaMPtFgL/+8x+zeeta7nkgvZf5SmNqNARTR3ZWY3OMR2FSuhu7oznt+ZIQbFoeYpl5me6QxMCwE4c3QmXeMd7d76S393O43QfHmjFpmgvQ0HU78XgQhyMwaXORrQdutg18rgo8A4FtdHWtn5bHPNOGKJNKyWxjuny5mxdf3MeZszYGFBlR1YEsw9q6Zfzu4w9Q4JvZqzM4MEQqJSPsJpId1tRV8K0vPDKn5801DMNgsHuAUJ9KNCGjKylEoB+7w4lszH7+dXx0MdVz7HKVE402YpoOS6UsS6/tVF7w+29DVV/FNBNIkmvM2E+lwkuWitTXt59Q6Aheb+2iq0JdCWRbV2L1vsjs6Jtp3SsqqpvUpC+Z9KIoKl7v+KY1l9wAuWt0N7rGj6ZFZcMNmQzA2bhB13UOHDg51rsoVTyMKFBx2W08uX0LD92+HkmS2B8aRtNsYDeQFcHmdcv52q4HACvi/NYrBzh+NET7sB2zZFwS/6lP38uKmum9SwzNwFAkywkkgc/roqKqhPbGzM7UeDiKoRukUnaEU8OZ5+DR330Eh8uR8ZxrEdcNiyww0UOgKAXYbFGSyW4UJS9r6bl0HxdoyLIbt7tsweHmbD0VsZjKq68e4P33e1m38QJRQ5BAxq4oVC5bhtueJBneC1X3Z7zXRPnWQ+/0caZe5eVf7Bv78+GhGF/+9H/mxpuX88q7fzPnZ8k1JkZD2ur/FN2Y7L22yTGSiYJ0pwKWUaI486it8hOIxmnu6CVhRqhdewynJ0hx8SlAoOsekkkvLS21dHVV84MffJ3e3pU4HDHa2+8lL6+U3bvrs/bAZTou14WHuQhjzzQmr/c2Xn31AIcOddM64MIo7UHYNfxeN888di8bV9XOev2e7gG6ugYYGPRgesOYmLjTNCyaD1IpjYvnmgmH7SQUFeQUCAUxW4dt06SntYfIsEYsoYA7DmoeGzatZuOqak6+dCAn47uOqxNTvxtJsiHLBdjteVmn7aTjhVDoCGVlDxMKHRmr65hv74T5erDnu8YstVTuQpBNXYmixEjMwA0zvSef70m83jYURUXTnBw9eh+KoqKqXv7kT1YhyzEcjhiJxCaef15NO6Zsxz46/lwXpC82N8TjK3nhhXenycmuW1HFs49tJ99rOZs62nvo6xlmIOQFn7X+j0YpkokUz/9gDxcbJDoTClJFFzabzPa7buETd98yTRIfoD/Yj67pJJMKKCnr+eSZ1/tUIknnuVZCISeaIwaShhC2Wc+7FnHdsMgCUz0ELlc5y5c/My8VqOmdr2NIkmtB+aPZeItN0+T48QZeffVDGpp8xAsSbMkPEUm4KA7ks6ykEElIaLrN8uxnwFT51r/5nnusd8To5v22VV/hb//XH03qY7HY+MIMefQ/maD8k65Q3OGIMzh8W8ZrTywA93pc3LSqku6eZmxmkK988S+IxfJQVSfRaD5CSDQ3r6Wq6gIlJe3IchHDw1soLy+hsTE3Xo1cK3DkIo85/Zg8hMOX+au/ep4LzR5iI8odNkXiE7fdzGfv34p9FtWNVErjnd8c5uD+yzR1u9FLexCOFF63iwduXz/3h50Ek8Zzzbz58hEaGl1EvAaipg1Zkbht/Wq8abTLp56vaRq6LoFseb28PhePPLqNnkudCxzbUiCVutIj+Cgj3XezcuU3c8ILsVgL1dULa5q5kMjmfNaYKyGVmw3SqTOBJWc7sQh76iZaliMoSoRQ6J6M1073nuLxXjSti/Xr/QwMlNDbW4qmuUkmHaRSCnl5A9TVtZFIFDA8vIWGhpuADxf8nIuhzLRY3GAYbnp7m/jXH7zN5ZALo6RnvHfR4/exfoWlwpdQk/zm1QMcPtRN64AHo9TqceTP83DPphsBq6mdGldR1UIk9zA2h8xnd97N5punp6+lEik+eOMIJw50ElW3IbwqyAayIuMvnC6hD9beqf10Myd3H+fiJR8xv4aoaUOySazZvAab/eOogbEwbrhuWGSJheZ5ZvroIcq6df/PpN/n6vXJxrv05psfsH9/A+culWIub0Gy6egij9U1Xjye8bCmaURR7JllXrOVb11MpGt41x28ndo0efQtU/Lo0xWKxxKbaGm4adq5pSOa1VMLwPVkPz5bDwDbtr2MybgCkKYJNE3G69VpaXl2xtzc+WKhChzZNmFa6Jj6+jrp7zc4ebYQs64VyaZTXRbgG098gopAZi8gWIv5hfMt7Hn58MimX0NUdaDIElvWr+Fzn7pzVvWo2aDGVPb84gBnG8rQa1sRdo3iQD5f2HUf1RUlM55bXJ6g9ZKbUF+AREJCM4F4kuq6GIZh0NsUJBqRUU0NlCRXp0bUhcYrPYKPOhaLF5LJ7owph9lyw0Iim/NZY5ZCwnU2pHs/weDmSepMo5hahD11E+33Rzl69Cmi0clS0+UjDTFh+nuKRJrQNCvV6Y473phwlo3nnvsjysou09z8u3R0PJ6Lx52ExeSFXHKDYRh0dDQzPOykZciGtKwTmyLz0JYNPLnd6l1kmib1py7y+u6jXGj2TnJK3XfbOh6//3YcNhumadLa1EE0CtGkBEVxBODzTE8n11Iae37wKo3nNNqHXWBLgmLi9/soKS3IGKG+8O4pzr59nvMXyzCWtyJsGv5SP3c/cReFJYWTji0oV9MWan/0GuUtjBuumGEhhLgX+J8jY0gC/9Y0zfenHLMZ+A1wYcLPvzBN86+WbKA5QrYf/Xy8Ptl4Kjo7e1FVO9j1saLXW2+/if7LPySVCmUt85pOvlVInklRjiMXn5vlbcwfmRreeTxPAjNvBkcxtVD8n5+PAuNScBMNl7b6IuzuWqKhI1MKwHUAbDaJFEDKSqhXFBO/vxtdd+Bw7EfXDTRtTU6efRQLCU8vllcx/ZiitLZuBLuBYjdZu3wZ/+GLjyKl6YMyEUNDEX790rscPxYlGLeK+IRiUBrw88yu+6mtmFuzwulN8Cx4PB0kkjKGDLLdpLyigG898wTybClQwL/uPoFhGPzy756ntclBMCkhV3Wxfet63vpeOxfP2hlU7LC8BUmB8pVVeP3pFaauYxzXeSHzZnCu3+5CPNjzWWMWw2M+F2R6Px7PLrLlhomb6H/+Z4Dekf+N36Oraw/Hjlmbb7e7llDoCDD6rBMjlTKjPAEpPJ5BbLYYgcBrJBKHGR7eklPH09XIC+nGpaohJCnB+fNbkLwJHA6ZZx69jzvWWeIpA/1hdr+4j5MnU/RodsyKTiTFpLK0iGeeuJ+KYmszHx4c5vUX93H6RIrupAuzug1JMagoL0krEa7GVGJDMdS4H8kTxeULIWnLSUWddDSNHze1Cd5gRz+JhA3TZiLZDUpqivnkVz6Ztp/XR0tSdvFwRQwLIYQfeAHYaZrmISHEfcCvhBDLTUuGYBQB4HnTNH/nSowzl8j2o5+P12cqOSUS/USjbYBOff23ycu7n1AoQiRqB2cMMPHnefAXb0MSc5N5neq9h9mjHLlEpohJcdEp4BNzula69CmPp5WN6wW/9Wx8zHCJho7g8d9GMtYyVgBufToGYGADsE0IHgoIR/JAj1NQ8AqDg+TUuFhIeHqxvIpTxyRJfi5f3khT8xpwWXOuwOee1ajo7Rng/zy3h/PnCwh7k4iKIZw2G49uv41PbFk/q0pZOkxsgjeKeEzlhede5cJ5uzU+YeL1uLIyKkbR39VPQk0RjfnAN0SeIeg+eIGGC+Ukl3VYEodeF3c+tpXKVeNFoX+283YG06TtffS0znOL67ww82Zwrt9uOqNFVdtJJkMcO/bsjBGP+awxue5lMFdkej9FRSeBB7O+TqbUqdraM3znO9PrYfz+24jFWiYYUA5AY9yoYGQ8GoqSQNMUZDlBILCHvr4dwM1zf9g0uBp5Id24Egknly5toLFlFZR3IYSgMM+Lpuns33eU995u4lKHF63MkpN1Oxx85qEt3HXLWktOVtf5YN9xDrzdSFOHj1RJGFGk4nQ42PmJW9m6+ca0PNPX2U8ioRNTFchLcM/T/50nPnM/K1dlbqaqRuJEBoaIRO2YzihCmLi8rrRGRS7wceGGKxWx+CTQYJrmIQDTNN8RQgSBB4BXJhwXAB4SQhwe+e83gb8wTXN4SUebA6T76P3+tXR17RmRLbQW+Wi0FU2LI4SGLNtxOMpRFP+MXp+J5JRKqahqK2DgcCxncLCXtrZ/oKv/Lno8foRbxeGwc2vdEG31fzqWThSo/mJWqUyL2cwuG2SKmLicYeJzVOHpDjqnpU9VBH5DV7BqmuGSjLVQve7PAasAPBY+i0Ue42Fxmw1SQGS4iEJ/D50dqxHCJD//MP39uY1afPWrv0cw+K1pv0/NGx7FqKctHD6GEG7c7gocDovws8mdzib9IhDYRmHhnRw8eJrjx3eTX3CR+x46ioGBJEF50WEa6t+muOxhCjPMtZ6ufpLJFLG4C6k8iMdj5z/+9hOUFOanPX6uME2T+mMNvLPnBBdafKgFA4jyGIpN4faNN2R1jVQixeE3PuDE/iCtfX700i6EI0WBzUEyKpNCQjg1fAEPjz37GMqUOpLBoPO61nl6XOeFCQo/9fXfnvTNzZUbpnuK20kkOrHbK2b0Sk/93qurvzLn/gdXovdFpoiJ0xnGmAM3TG1sN379poz1MKPpzAcP7sI0NSTJNSLQMmk0xGKFeL1dhEJWfYDPd5hcGRawtLww8fxsuMFu38jrr/8zun6csmVdPFjWQG/UR02xynDHQX5z3ODEyQ00DJeN9K4Q3LJ2BV/csQ3viExse3Mnr794kPMNDkJOMVbsvXZ1DZ979J5JcrKjiEfi7H/lIKePhukYzsOsbEPYdQoL/SyrTB/9Ng2DS4fPU//GOS62+kgW9SMK48g2Gys2rEh7Ti7wceGGK2VY1AGXpvx2aeT3iXgB+IlpmqYQohD4HvAcMG0XK4R4FngWoLo6u7DnUmNimDVd6PHSpe+haYOAHSEcGIZBNNqCwxHA6Syf8bpgkZOqXkQIG7JcRjCYIhyWkJwObtjwIU1ntrF6xTI+d4+LWO/P0aekEwGzGheL2cwuG2SKmMTVfMiBSJDdMUhKnyxNNzXVy1/2KeLRZkxtaNr5NhTsjghD4RJ6eyuQZZ38/GEaG52TcnMXikzkl65508S5Bm5MM0Us1gowTQ53pnNnC5G3t/fw4ovvMBi6zE0bjpMwJAzTIJAXwWFTcNhKMYwYHW1WqtxU48I0TXq6B1BVGU1oCGEgy7YZe1TMBaZpsufnezl9vJemYD5mTTuSYlBVUcwXn9hOIAvjJTYc47UfvMqFcw66DRmpqgNZkdh4yxpKNGj+sBMQCAEur2uaUZELfFy8WmlwzfMC5I4bphotyWQIu70Cr7cGSO+VXkhKTC4KfReCTBETVc3HngNucDrDI3WR45i6+S4re5hg8HkMY/oNhYD29tXY7TG6usoAA5er/yPJC1PPn2muWL2LTnHixK+prDxJShYM6zby3RFqAj2kUn7aWj2YwuDmzQeINq0latzAb+26j1XVFhfHonH2vnKQ40cHuRy2YZaOyMnme/n8p+9lZe10OVmAwZ5B3vjRGzScyyfk1hAV/dhsClvv2shdd29I2wTPMAwO/2QvzaeGaO7Jg+p2hAIlNSXc8+ltePKvzk3+1cQLV8qwEEyNE1ru30l5CKZpxif8+4AQ4o+Ay0II18Q/G/nz7wPfB9i8ec1V350qXehxeLgJkLCibFYTPoBEoofa2mdmvN4oOX344e8wOCjR368TS0mY9gQCgd+j8o2nPsm6VTW01f/pggqwF6uZXTbIFDFBqptWqA3jBdjZIpkowCbHJv02NdVr9Nl7235CKj66Dxq5t5DZ+fg/oWnQFS7BaRgoQgBBdu68c/y4JcTEuebxLCMabcE0ra6qkmSb0auYTYg8kUjy6qsHOHiwi9Z+Fzs+dYyEKWHgZllRCEnyYCKha3243OsA6O16dZJh0dczyMsvvMPp0yZ9uKCuCWGDVTXlsypHTUWmTtslpXE+9eDPiETc4FWR7SZ33nYjn37ozqxD2wNdA8SG4wwNFSFVXsbuUtj15HZWrKzixCuH5jTO+eLj4tVKg2ueFyC33DDRaLHSn2augVhoSsxiNLPLFpkiJpJUl3ZjPdcNvarmz5rqVVf3LMBIY0HrN7u9AsOIsWvXz7C6s0v4/TeTSoWRJBfr1q2d24PmCAvhhannQ/q5MrF30Z33NRAXkDTs5HndFLojaEmBrseJ40HIAqEr3H/zILfd/TkUWcY0TU4dPce+105xocVHojCOWNaPzSZz350befCeTWnlZEfR29FLQk0wPOxBlAc59LNv43FUc+IFhX+ccFygXOXvdh8DIBlVGeoeZGgoD/JiyHa46a4bueW+WxYtBSoXuJp44UoZFu1MT4ivBn45y3kyoDIx/+QjinRhW4tDZdzuGuLxIKaZRAgbNpszq8W6t3eQjo4E8bhJ1LAjnHEkISgptFNaVEntKstTlU0Bdq6QrQxstsgUMZlagD1fDA7fhsNxataC9lHjarSYXNd1jGQvmBE8DohKebgccexKivfP3ELX+SgXLrzAww9vYtOmG5Z0gZo410a7p0ajHZhmbNY+LDMVZJqmyenTF9m9+wgNl3zE/ClE1SA+dxynq4TK0hIi4QHAhkCMdSQWkpdU0lLVSqU03nnzCAffa50gJ5vE43LyuYfvZMu6VXN+V5k6bTc22EmmdBJJq+eEJKCmomRefxdCWP8nKxKB4pkVrq4ja1zzvACLww2QXQ3EUhVgZysBOxdkiphYPSIWLufa378BTbPSm2ZK9aqre5a6umfHPPq6rqFpKtYUBUUpm3dfklxiIbww9fxRjM4VVU1M613k8UVQUx5qy4qIDEbQkklSuoRsSyFsOh63k+qyCmRzEEWW0XWdV376JmdOhmjt82FWdSApJtWVJXzxifsp8vtmfD7TNBkaGCKZkDFkDYRJIupn7c1JEMlJx7Y3jkfF1aG41eMipSA8GkISBCoCV7VRcbXhShkWvwL+Wghxs2map4UQtwM3AHuFEAeAr5qmeVEI8XngNdM0Q0IIBfgu8EPTND9SPW3T5SGmW+RBQQgrDDma5zjq1Uh3LcvBJwADuz3A0NA62tvrKC09j1NOoGFnRWU+NkmlsGLH+F3sRSTUThL6MKaeRMh2JNmHw1mW82dPV8cA02Vg54LFjJi8/datGKkbiMX9Vt2Gmk9v/3q8eaVpDaG8wDZiQ+cJBV/G8kZZZOmxxfCUVXKgcRVtcR/Gsg5OdRYR/N9nOXr0PLt23UdJydJsSKfONYejaKzD+1Sp4+mQGBo6BZgIYcPlKkeSbAjh51/+5WVOnEzRnbRjLuu00opKiqgsq8Nu05AlCSFsgI6JhCxb6QGmEcFmL7LkZH91iPON7klysrffvJrPP7xtwXKyozBMk8G+MH09Tk6d95Mq7kG4E7icTpaVL01h6XVkhWuKFyB33DD1Om537UhB8fh1s6uBEITDpxDCHKvjkCQl5wXYc0nZmQsWM2Jy+vQafv/3X6Ko6CROZxhVzae/f8NY49N0YxkaOksw+KuRXyxu0LQ+NM1LdfXTV7S/x8J4AdJxgxAK8bgzbe8it7MYvzNJT8cwsYRCSSkoso5h2lleWUqBz4OmhZEky8gZGozQ09VHKFQI/jA2h+D+bRt56J7Ns27yhwaG2fd83NCkAAAgAElEQVTiu5w7qdGd9GKuaEJSDBRb5p4TuqZz/p0TnHmnheZgIVpJEOFM4nC5KSwtzHjedUzHFTEsTNMMCyE+C/yLEMLEcsfsANxADTC6orqAt4QQBtaubR/wn+d73yvRFTRTHqLff9skmTrDiKAoLkxTjBDG9IV/4rVMUx4r0nY6l6PrMZLJX6Pra3j/6DbW3XqIPE8ct7MKf9lnJ23E7e5aYuFjWEUJBqYWQtdCxLQEQ337r1ia01KjtFydZuAMh23k+/PA/gmrGNwOxeXQkobwRmVpY+FTgILDU4nNYUWCUqkQsuTit5/6I9advsjPXj1IXO6lJ+7k3QMV9PS8zMMP38att944dr3Fmp/zLars69tPKhXCNFOADdPUiEYbSaVcnDhxK4dPCbSSEKLIUu74/INbuW/TjYT6K8bqKBR7CUm1GZCQ7dWkUiGS6hAtrcs5ePAInTHnmJxsSVE+X3tiO7UV06UCM6U3lZaraVWfRhGPqXR39DE0LJPQnWgjxX6rV1byuUfvJW+O9RtqVEXTQTMAMb997Fy0ztPlzXY0eBjscnDjtoF53f9qxbXEC6P3zQU3TL2OqrYTDn+I3V6By1U5dt3q6qdnbLrX17efZHIISGCaDjQtiqadAwQu1wr6+vZftV20c4ny8kRaA8c0TcrLS4AHMQyw26G8nGmNTyfOJ1XtxeEoweOpGfvzUYMw23d5tfHC6JjScUM8buf9w9s41W4bk4mtKS3i4Q030HBygNLSE2iygukSqCkbhR4Nh7cYp9OBpoXRJ2QHqGoCXTPQdAkUHUmWWFFTMc2o+OrOjfROiH5FozHiQ0mEsoY7v/BdRFECh93B/Q9u5vRLPhCTU50BTN1g3/+3m0tnbfSYdqhsR1IE1Wur2bpjC/8/e+8dH9d53nt+31PmnOmDDhAkCHaKIimxiZJIdVmSJVmWbMf22rGVbHJ90xw7e+/Gye7eTZy7n9x780nxJpvs9Y2zrnHsWBZtq7qosqiw94IOAhiUATDAtNPP/jGoRAchiXL4+3wkAoNz3jkzc+Z93ud9nt/vp89plro0mG9suNbjwnvmY+H7/ivAdHbHyycc83Xg60vxfIslpV3tl3ohzqpr1vze2DnTTfwTxyoULiNJOuBjWb309VWQy/nEylpp7a2k/dweNm1YwZ2bH55yTVa+FSmwDM9KMal7wO0l2VjsPPy3kFxMV4G4b8f0RntXYqKfxijMXJH4pmoVY61lQgh2b13P1vUr+dt/fo62zhSW6jA8rHP8+MWxxGKh9+dMwW+6vuHFkiq7u58nGKzB8yowzSSOY1Ao+AwNaRxsqR9T5NixYQ2//uid2PmjNJz9P7CtFAIJ1zGQhIcerKe4/vMYGnY4dXwNb55YhVWWQtQYaAGFR+/axQdu2zqj3OtM7U3T+VMAuI5Hb3c/6QGTvC2DboDiEE+E+cRjd7BhzcwSg9PBMi3efOEtTh7qoa2/BG9lC0JziMVLCE9jxjQbFkKkm65vtrc1hGXM3Ff8fsb7JS6MnnstxIYzZ/63SeO4bgYI4DjDyLI8qfd98+Y/n9VQLxSqxvPKRuTKR/WzfQwjSWPj38/rfXm/Y6ZWrB07dsx57pX3U6HQiWn2oCiRsXajhbSWXYtxYfScibGhUMjjOD4DQyFOZksR1b2EdY3H92xjoDHJm6+8wLKVDch6nhLJxRcqidJ1JErWY47IuI+2NUdLbuONV45y6OUGmjpLsSuTReluPURF2VSBjb6kxsq1BkbOoK8zhXBk1KhLNpNACtus31DPBx/dQ2iW5MCxbApDOfoHViLWNKGGFO75+N3UrJpZNOedwHxjw7UeF+aVWAghgkADReH+db7vmxP+9jXg14FP+77/vXfkKpcAiyGlzfdLPVuAWaiz6pVjTxx/oiSc61oIEcDzHAwjR09vBW5AJhZLEwhKPHrXLu67beu0r8ux+gkEl2FY/aNXRHHh5+O7g/S2/8u/icTiajDRT8OWdDxcBAEsswdVq5hC+A7qGutW1tCeTFG/ooGdN5wiHs9x+vTb1NQ8suD7c6F9yAttEbh8uZu+vjZ8Pw7IWFY56bRDwfGIJtKI0iHKE1E+9/g93LCyloHUATrbv4WihFEClfgjBnm1dZ+ltHwvXR29/GTfa5w7qzKgmLC8E0mGDauW8euP30tJdOkIZrZt09bcwfBwAFf1QLcQkiASDvLHn/84qrKw/ZS+zj5e+t7LXDgfZzjiIFZ0oSgS23bcwD3370JR3t3JXA265NPKlJ2t98Ld9f0eGxZLVp5PbJgr8Viq2DA0dGrkXI1gsGaEy6SO7CZPHnc2jF6PqsojicU4hJBw3UHa27/zniUWvu9z8uQlksn+uQ+eBbIscdNN66l6B9pbrryfJEnH8yxMMzmWWMylwjfxnrGs/msqLriuy5Ej58nnx2PD8HCCbLYMSzaIRTPIIZudG9dwY0mCQz+7gEMPO295G9NVGCjEWV4ZIhb0qax7jMSVyoKtSZ7+zj4uXNAYDIzLyW5ct4JPfOguItNs4ni+T8/lXoYHLQq2VNxEEj6yovCpzzzIyhkUo64cw3XBF4AEiqZQueLaVJGbCddSXJhXhPV9vyCE+BPga8DvUHRGRQjxX4DfAH73Wg0co1gMKW2uoFMMLt+mUGhDCA1NW4bnFSYFmKs1DZpJEs73QQgP07RwXRkX0IMmDhG+/HufoCQ2s8vvmGQroyXBUYdQCdBwCpfndW3/ljGRAK8GqzHz7fi+D56JbaenJXwvry5nVVknO+vOYhaCpFJRHKeNXO5rCGETDk9W1VwqHfHRY3K5Nlw3j6oGCQbrpz02ny/w7LOHOHy4j5tu0lDVLKYVxPPBCzjo0TwFO8gTd+3gsTt2jCly9HW/gKKEJymNAfR0PcuhAw5vjShGedW9iIBDPBLiVz90Bzetr1/Q+65rl0hED6Npg5hmCb3JqcZXWqCX7q7KIklbM5FlifLSKDXr3QUnFQAXjlxkOO2TMQOI2hyhqManP/MwFe8SR+ZKbNozQLIxzFeOvv6ePP9EvN9jw2LJynPFhubm/0Ey+SOK0sMarmtPSTyWKjaMKkSBNxYbwEaIcY7SfMadfD15xjl8AkkK4HkehULHvK5tqdHd3c++fa9y5izk81NJ3wuBED6vvtrC3r11fOADu1GXUBL6yvspGKwhl2vBcQpFgY9Z2o2mS1YLhTZ0fdXYfQZL6y+xkNjQ2trFvn37udSgsuf28djgCoGvGeiBArYf4XMf3Mupg+fZ91KBwYDEo3cdxfQUNL2EDcsqUBQFx04z2P3iWGJRyBu89MxBjh/p53JaH5OTTcTDfOLDd7J+1fIp1+77PucPX2CwZwPhula27zxFNDJEJh+jZ3A7yc5NU5KK8hpjElHb932M4RwyfZxtLcVb1YRQHCKJOJK8cDPW9xLXUlxYyDfqG8AfAH8shPhH4DeBPwL+xPf9f3gHrm1JsZhJfLagMzoJFApdgIPvOxhGA563DE0rGwswV2saVExcehFidCJx8H0F8PE8A0myyecT6FoOXbXJiHtmTSpgXLJ1HC7ggRQu7jsylRg1yicYNdRLVD80bVXjShWoliadhoshFNVn1erxVpaFysAuNWZSq2ppDs6rFWqin8Yor8LMd4AvkKXgtN4et2xeRySfoatXx0TG11z60hFMM0tp6RCBQBpZHq9yLIWO+OgxnueM6ODLWNYQkJx0rO/7HDt2gRdeOMbFlihGiQkDNdy65hyubWA6KrpqEw/5rF33JPX1t0y6FttKoQQm7PD4kMtAeqiJnzyTGVOMUhWJO3du5on7bkGbhUg3HXTtEtXlL2I7YQyjHEXJUbf8ddKpvrEA5fs+v//F/4+LZ9M0dZcgrWpj44Zl/MYnPrig55r02kwbzxcgQJIFFVWJWZOKA69+BvNlDUVT+OnfxsYeX4ie+GgPbefFML2t44FQDbps2vPe99BegW/wPo0Ni13czxUbioTd4rzq+yaWNYwklU7aaV6q2AAWYON5GsWQbgEeilI+52J2IiZeTxH+yLXNzkOa7yJ2ohJUU5POxYvFcVXVp7Z2CMdxKSnp5bXXjl4x/jCHD3fT2hfErexDlDrThah5w/chk4mQ3NfH6dM/YOfONZOqjn/0R48zOFhUHNL1wJjXQXNzcFrC+URMJUSXY9u5Irdsjnaj6ZJVITQMo4tQaPxeWyp/ifnGhlyuwLPPHuDtw/1cHtLwK/s4OSE2uI5KMGBTEZMZGriF575/ZiQ56EWoLrGwSWVJHZFYbOT6XTIZF5zLvPXacSzT4sRbLSOGpSaidhBVlbnz1pt48K4d01aEB7oHeG3ffi6cEYTj3Wzb/jqGESRTKCUec6ksO0g2PbXddVRSFiDV2sPRp17n4vkSMrFB/HgGWZFYv2MDO+7f/o6oQC2F18SXH71lSlyAYmwoqbo2hPHmnVj4vu8KIf6IogPqj4B7gb/zff/P3qmLW0osZhKfLeh0dz+P5012XgawrC48z0NVix/61fQxplIHKBTaAA0hAgjh4Xk2xdW/SyaTIJ/38CQfW/E41Xoju3bdNue4owverot/MXL9PhBCIONjoAYnfyEn8gnmMtS7UgVq9OfWxiAvHb12zLtmUqtqadLn5YdxpZ8GkoqqV1BR95lZ28h0pcCalavp7h8i1Z/B1wsM2QqBvEqh0EU8bhKPV+J5uSXRER89ZpSTI0kqnmfhOMNoWt3IfXwj+/a9wqnTPn2eil9blPUrsJHmwQgrS84TCxeIRWpYu/ZjlFXcMeVa1JHvCnIC07JJdqbI57PkrQj5kfFqq0r5zY/cx7KKxbUgJKKHsZ0wjlNMnB0ngmkGx3a+MkNZfvaj/Zw8nqfb0PCXt4HkU1W+uOfLZ/Ic+MlBTh3JksyF8Fe24guP0rLYrOeZhRiRmstoYZXqleNBcSF64qM9tD2tIfSIM/a4kX3PaHEz4v0cGxa7uJ8rNhR555MtOTxvgEzm/NjvSxUbJCmM543KmbqAQk3NEyOqUPMfd+L1FJcGDkWBD2Uk7tgEr4gNC+EATFSCGv3XNG1OnvS4777/SqFQrLD88z9Pvi7DERQSxRZEWZYor4hf1aKvkDcZljNkIzmONlfQ0NiNNGG4piaNWKwbAMdxSCQiVFQkaGrS5+QvTHc/ybLCqlX/Yc73f7pkVdOWYRgtMwq6TIf5tvfNJza0tJTy4ovHRxb9BqJ2gIAq42lbxmJDImIR1Cs4f2oZbxytwygZQNQOElBl7t19Eyur2vB9E3yf7FCW7uQgrmNj2wFee60L14OMNh53VtRW8OknpjcstS2bIz8/wvEDnbT0BnGr+lhVfx7TCCHkBBFNxnXBEoLKslPAzHN/48EzDKZksq6EiGcJJYJ84FP3k6hMzHjO1WIpvCYGkzqK5k+KC3BtxYYFXYnv+88KIY4B9wHfAyZ5xwshNIouqPcBlUAS+Hvf97+yNJe7eCxmEp8t6LS3fwvTHJr2PMfpRlXXTnruxfSldnc/jxAaQkhIksy4KZJJf38lTz3zcZyaboRelM78+EO3sXvr+nmNPbr4TTb9v/hOAXDwPYFQYlTUfWrSsRP5BMCCDfWWAkvthzEd3j6YwHGmlj+ne47FOpCPVjpWVJVTURKjpaMXzxlmwAhz5uzN7Fh/lny+kYqKujnlCOfTxjF6zCgnpwgZ37eAEH19bXz3uy/Q0hPGqexD6CaRoM6vPng7e7ZumDOAD6UzNDVcxrM2Ay9gmMMM9LvYnocWKXAmuZJQSOWjD+xm77YbkBa5IKiqMRgc0MkVljHe+gFlZX04Vj9vv36Cgy9foKE9il2RRZQYaAGVh+7ezh27Ny/4+RqON3Do+aNcbCzFqMghlhVQAwp779jG7tun5y69EwjoLoUJAcMxBcnG8HvSNzsb3q+xYbGL+7liw0jpdwo8LzNJXWmpYkMxuQjgeS7x+I1jRm0Lxej1pFIHRty+C3heAVBQlBh1dZ+Zch2L4ah4nk9f3yCpVI7hXCVd+hDE7OkPlnyE5FNekeDJj9xDbfXk3fonr1AFGkVFjck3nz0xzXN7vPbWaV585Th2TQ9D7uQ531VsbL246WTbKm1Hq/F9Bc+DfH78e1denuOv/uo5JEkwPLySWCx8VcnidMmqquoIUY8kBec93nzb+2aLDZYVoL+/iR//5Dwpxhf9a5ZV8OiOzWB7wBZ8/1FOn2jg9GmfPs8dO27lsgp+44l7qSpLkE45JJu/QU9/P4ODKkIx0MIGhxs3ka7qKlafZJewrvP4w7exffPaaeNOT3sPrzz1OhfORRmOuGOJ5vIVLfQkb4BJwh8ey2sbmC2xcO2R9m8hELKgur7qHU0qlhJXxgUoxoZrJS4sKLEQQnwcuHnk14zvj3pLThqvG3gAaAa2Aj8VQiR93//+1V7s1WKhk/hsk0R39/MUCp2znH31Jq+WlULTlmGaHXgegDxWJTl9ehtOfBgpZHPD6hX85kfvJ6gvrPd08uJ45hand9NQbyZcjR/GdElJw8UQvT0anu+NKSlkM8V/21t1ArrLrXuHZ32OxfhpTKx0KHKYtcsjZLMmPztdT5sdo/3Hn+HGG9q5775tlJffPOtYo4HIcewR0ywbEOh6zZRjZLnYIz3Kp/F9mba2FoaGdS7lPaQVnSiy4NbNG/jsB/cSmuNecl2XQ/uP8/orl0j1hYAgNTU3s2r1WfRoDtvWONl6A+XVe/nCI3cQCV2d4/j3nz1Fy5lv4XkFFHV88s8N95DstHju+XYGFAlWFEnh69fW8snH7l6wnCwU9cyPv36C3p4ghYCBFClQWV3Cxz/5AWJztBkuNa6UDrxWemivxPs5NixmcT93bGhl+hggzdvFejZMHxtcwKS6eqoS4EIx+fXN3OK0GI5KNlugqytFJqNgy4BiIUUs4vHwtAtKSRLs3XEDd9yyGWka1bhRVaAr0TaDJ4YkSdxz203s2LyWx+9eR7pvstFaZqAaq1BaJPQ6Kp4VpPhZCtrawsiyQ01NCxcvlvGNb1xCCKioOM7992/m9tu3LDpZnDlZXZjfxcQExTRTFApJPM9AiMCkpHb62OBgWdDTc5nhXJi++ABCs4gGdT52x3Y6z7bx1LdPYxjjbayDpjy2KRXWdT7x0K3cOrIp5TgOp08EaDqzlpLyJsKJNDlT58zlzWSldZSNdM+uWVnDhx+8DX0W76Jjrxwj1SMYdiREIkMkHuRXPvkBYv1/B94PkdTx9lTPHgQpBOycMk4hk+f4T97g0gmDZD6Av7wdIXyiJbMb7l1LmE5SNtkYXpDi4DuJeScWQogHgG8D+yjq0P3PQoi/8X1/rLbr+34O+E8TTjshhHgO2AO854nFYjDTJFFd/TBDQydnOEtjKRKL4he/gCyPu60C5PMJOjvXI6JDqKrMg3tuXnBSMYr5LI4n8glGcaXq0VLCNCw6WrpGJrvRx24mN5yd5liZhrPNs47X3nIzy+smJkGC9tblGAUByIQjHgOp8YmykJdxbIP2hkYSiV58t4rH7qwil1s5adz6+rP8yZf/YE7eyUSM/v0/f6mcQl4aM+BrbIvh2C66ZHHjDf8Fx3FnHQeK92Bj49/juoMU7zkB2Nh2eiyAFI/5Gq6rAQN4noltuwwPh7F9m5Op1Uglw1SVxvntJ+5jzfKpHhJT3s+2JM/u28/5CxqDAQlK+qivuExdbTN6sEDODNI+fAMf+9CTrKubW5FjviipfojeMW5QmIHeLkxzmANH9jBQ2ocIOMSiIT7x2B1sXFu36OfxfR/f8/F9CUkCWRbcsvvGdz2peL/gemyYjPHYcGW/swKERsxNrw7TxQYhBLpev2SqTfNZIM/FUUkmUwwMFCv7hrGF9vYeBgctCo4EemFMved3f+0RVq94d6U9Y9EwOivYfdvkpOTgLwRF6xSQgwWG+mGU1OE4ASTJQQ9mqE/00jewgpNnbqe/bzn/8A8Suu6TSIRZt+4SX/7y/7IgKeLRv3/pSyXk89KYAV8uV5zL5utKPpqgmGYay+plVPUxEEhMalOrqHiQpqZ/wvdHY4OBadoMZeJ4ismZwVUoQZfbbtzAmkiIQ8+do6kzilORhZLxNhyhOiiyYNfm9Xzyg3vGNqXaGjs4/Op3CQYbWL56mJytcbjtBuLVd/DZz+4hPIc3hOd6dLZ04ZjFSpZpWPi+hJCLXLf1G+qoqSnHVh/Cav9OsUYoxcAbBidPoO4jk8fzPJrfPM+Zn1+goS2KVZ5FJAwUVWHrnTex+fYb53xvr2N+mK/c7G7gaeAg8GmKmuIfpeh4+vgs5ynAXuAvrvpKrzEUXTU/TDL5ryOPSBT7UQWaVr4kTqWjE4SiRIhGN5HLDZBO93Lk6K2kg8OI2BCypBJdoI7+QnEln8D3ctOqHn3q0a00XAzR3jp5wtCDPpXzIBX5vs/FU028+twReronJ0oDqQ/i2sNTzhkaCvDdr86U4M18rmVWYdk6+MVEwp2wjlfVPMuXN5BI9DMwWI3rSOze+T2SqYcomMVWs6B2CeGcwvUKc/JOrkSsfC/HToz4ZYwY8A1aXeTyBtlUNRdaS/G8Cwgh2LPnZmRZmpEgWSRw5ic45dYhSQrd3c+j69v52c9yJJMbqK29RDQaQFFsLFtn0Ahypmcl3bk6PnH/Dj54280zekiMIp83+NmzB3n7cN8k5Y4NNT3sXHkWw1Yx3AhVZQo3rGqhItQMLF1iMUrQHux+kfxwJ9mMy9tv30Or0FFCDrdu38SHHti9KOWnUQz1D/Hq0/tpadLY9+znMK0wSPD292JogWLyWV5jTCIB/lvG9dgwFbPHhvg7EhvGd7d/9arHXsx1wORd9srKD/H97/+CY8f6KBRUnnnmt+jqiiHLIXx8ECAE6EGP5XXeu55UzIVcVsb3QZYnbyZoWoEVKxpJJAboTq6kMh/n3rt/xLHTt9KTWg6GRjx+HEVpxbazC/ZGKS/fy4kTO1i71hgz4CtyZ2Z2JZ8uNtTVPcmlS38FgKJoaFoNmlaGbQ/R3f086fRKfvzjJJZ1I6tWXSAaDSDLDqYbYtjWOJusxxQ38rmHtnD09bP86JJOWgdRV5R/jYR0RotL0XCMX33kTlaPbErlsnl+8ZODJC8fZ93GY5i+IGNphIM+j+zooHJtcM6koudyL6/tO8jlFvBGWtUcV5CyVPyaThA+sViRl6CWF3mlVveLYPVAoIxA3UfGHgdId6Y4+vQhGs6p9MsyrOhEyFC1spI7nthLOLYwufOlIGH/MmPOCCyEuAF4DrgEPD6iU94khPgn4LeEEHt83z84w+l/CwwB31qqC76WMNrHWlQAAQigaXEkSVnScnRX1zP09DTS369y5OyttDhRRNkgsWiIzz52FzUVpfNWbZovrhwvnNiFdYWRzXTE7WjcnmLSMpSW2XLz5F0h3/dJdQ9gmcUqjOd6vPXqMU6fcOi1A/iJ9CTlD0e2MRVrynU6sk0ynJ71tUx3ri9s8IsTw5Xd0OHwALatYlmjbrc6lhOmJHp4LLEoiR6mpzuxZLyTVcsqaUumyPZ72Cs6OZtK0P0vbZw82ciDD0bIZvdNS5AEn3h8K0IIDMPCccB1HRznMn/9109zqS2KmVgGqfFWNhGwkWTYWF/Df3vyPsqm2Yk3TYue7v6xwltfzwAv//zkmGLURHLetqqf4Pt1k1qUrpQTXCokyveSKN/LpTPNnH91P+3tKxGrWwiFA3zw3p2LTio8z+PYq8c59kojjR0xnOp+TDtEvKKPyqoSYgkbRHHnbKJcIcDvP7qdVFLHGL4JI2eTzZSSyZYhKx6+PX7fLaT/dSHu3Ffi3Qp612PDzHi3YsNcvfxL7do80yJ24nVY1i18/evJIuE3YUM4x7AZRtZyeK4KAmRJEFAUzIJMRc21pXIW0N2xttgrEYkMYtsqnhdBCIEkJZClEm5c30DPYC0EC6yoP8PwYCWtrVmqqlQCAQ3XNWlufppYbDeBwMLU8GbDTOT5uron0fWKEQ6Fi217OI6B68pkMs089/whOjMB/LJyjp3cA4Ame6iqi6rI7N26Hilj8sz3ztGRCeBXFjeRyhNRfu3xu1m/srhhZBRM+nsH8R2PjtYkPV19HPj5OS61RXngoXOYCCwvSHVFnIqKBK4zRLb7ecLlt0/7ekzD4o3n3+DUGyna0zpeaT9II0FI+IiAjaLK3HrbFnbftmXsPLX8tkmJxChs0+LMT49w/mAPbalxuXM9rHPro7dQt6FuUUIA8yFhTzcPJ5tDJJtC1KyZ7P79bsSGdzMZmjUKCyHqgJ9RDAAf9H1/4rbvnwFPUtxx2jPNuX9FcUfqXn+0h+eXEKtXf45YbNOSTt4TUV6+l7a2Mt544w1On6nDX9uMEvS455atPHbvLlRFWZBq03ww3Xi59OE5FY+AMV7CRLQ2BieRn4fTGX7xo/00nM/i2MXdCM8X9AzpOJVpRNBEVRWUCTrSQgZ5mvlYyBBMzN52duW5vg/l9edJXtqGJLtoqkWhEGU0kwkELGw7gOMUW2IAHCfMpQt1nDhZ7OPcuGEz/f1V/Kc/jlJW1s8X/+O/XBXvRFUV1tZV42RlIpEAOZGmv30F7e0mf/qn1Zjm/47rjicAspwlkcjxB39QztBQD93dDvbIe6mqeUxT53SPAsu7kGQfXVMRI68vpIf5zEN72ba+fsqk6vs+Z0428NPnjtA/oFBV2Ux9/QU0zWDFmiBDsVra07WsXFbBb37kPipL4zQc+ybKFRwc6R3i4Pi+T9OFVl557jANLaVYlUmEYqEq4TFPjcWgu7Wbi29foK2tBqemCzlko2kaq9cuR1Jmr+SkkjrL1+YZSg6SG8ij6hkIZzFyy/jK0ank0fngaib6pVAemQvXY8PceDdiw9Wa+C0Esy1iN2/+c3p7B3n66aK6XIoikVcoPpqmImSoXnUJSRYsqyojHg2DELQ16tMSrN9L7No7zMvPlaFpHiXTB1sAACAASURBVKGIR29SHdvkCgQsHGdyNf3o4d1IwmW4Yxum7dAcHWKgv5pv//MX+egTf4cs+4BDKNTH3/zND3jkkd3ceOPqJZEz/dKXSkinp48Nn/98CV1d7aTTCp5XfC5VzZMrhOhQ8ojaAgFVRpNkSm2QciGcvAIIjr/SQ9+QjlmaRywrHvfgnu08fMd2FFkm3XeAtks/IJvuJpsN09q6kZ6e1eTyAQZGKgLhcA5PSrC6tprAiFeIJ0Vwp4kLvu/TcKKBQ8+foKE5SqHERNQOoCgSijo+r5dXVvLhx+8kUTK7Mp/v+3ScbuHks8dpaIqOyZ3LisS6HevZef8OlCX0L5kO083DNWtzV82TW2xseDfiwihmfWd9328HpooBF/+WZLQR8QoIIb5CUf3jXt/3r76h9BrHYsla84Vt2/i+YGV9AzduO0A8bLJyWRuFtItavnfJVZuWcjzX8XAcm75k8TZovtjGoZcbaL4cwy61QBnpQRIgVqSQFcGNG1bxxKN7CE8g/HbsD5NKTuV01N9k8OUvPTnrNZz41zLq1o7fqq7nkUwOkBSA8PGExyg5D6BQCBMMZsjnoxhGCFU1cJ006fROgsFib6ks+8Rig1RXd5HsqsbzvEXzToLaJUqihwlog2Bt5KGdPvsOgieB70tYVoCqqmEgO4F34pFOBzh+vJRYrAHTVTGEjK5aoNocv7wGUTlANKTz5MN7uWVTUWnDth36U8UKT0/3ZAdb23J45edvceKEQ58TYOXas6zdcArTVhmwFPRQntvXnefeis3cuuOJscCoBMrwvBzSBA6O9w5wcDLDOX7+o9c5dTxPV17Dr+5GKB4V5XE+/cQ9V2V05dgOnu/j+zJC8dFDAbL9lRx5deqx3tz0l196XI8N88M7HRtGMV0lYbGKTTNhpvG6up7l8GGJ/fvbaO4J4Vb1IjSboK7zkYdvZcfmtZz81zJWrl36BQwU1Z+mI2pXjMjAuq5Lum+IqXoCRTlZ05imTdf38X0fzx2Zb0dOHY0NmUyEgFbcIXZchaDuEI76hJGRFZ9IZIBMpgxTtcEHXS0waGgcbQnQ+Y/H2L7tHPfdd8uYR4OqKpSVTZXR1bQGotG30LRBLGsjqVTzpM8un5dGYkMGzxt9fR4DAwEOHoyxbFkLrixjoaCpFlLA5mLnKkIRg5U1FdyxYTVvvX6Bi21RjPgQKMWddAkfrXaAkAK1VWU8fs8tJKIhBnoHyQ68SarzO6TTGjk7ihbMs3bTYVKyycDgMkTAIRLWSZQtJxpk0rzse1nkwGQPoHQqzWtP7+f8KY8+NzCWkFZVl/HER++itGxhSk35dJaj+w5y6bhFj6Xh13YhZI9EVYI7PrKH0sqrd1xPNoWm+EgAk1qq/y1jyVM2IcTfUtQxv8f3/XdPNuiXHKWl7Wy68Tym5JG1Q7heYawqsdSqTfMZb7RVyiy04doFJDlIfd1fEtTqKJjr8X0YSg2RSg0zkKrg6195CYCBoRDpgAR1l5EUgTrB/CYSifIrj93B6pVT+/L/aQG7Wr/x6M2kJpT82pqCNF0Mo6geK8dM+sJomgDJwbE1in3QxSBSKIS5cfNB6ledpqNjLZpaQJZydHauwDSLwaS9vZ5ly5pxrEFMo4SO5kZKyxRq1v3GvK8TiklFTfmLWCOmb7JkUh54hfrSepryKh09lXieIJE4gSR5WFaAwcEqXFdhaGgVz726jhWbh9lc10hCz5GzghztXE+PUcc929fyqQdvRw8E6O/bT2vTU6QHkmSzQS5f3khvb/2ka3FdQedguFg10k22rGnA8hUsoSOrgnA4zrJyGUU5RVEEqIhRUvV///vPMjQYIZHoQNfz9PSspTNZRThWxfevQhLYdT2OHzrN/pfO09gexarIIuIGWkDhgbt3ctetW6ZVi5kvejv6eOunR2hrS2DEUwjVRJI1HFeQSDhTjh/qn76VoeHcCsyCjOsJkD18V+WLO+683nc7guux4Z3BTJUE180RDK6ZdOx8XJtnwpUKUL4P2awgnW7iB/t6yURcVt38NptrW6mIe8QTYWT7LbqO+6yq+1O0kdiw1Jit4pFsTfLaj96gr9vlh099nnx+sjfCYFqn5ZJOScnkHNfzLFzHZmhw9LtejA2GEWbLlkNsvPESl9tXo8h5NDVOX2r8dfX3b6S05AJBvYCk2miyja46nGpZDzW9JAsaP3tN5+zZl5GkURNCn02bSvjwh+8cG0fTGigvfx7HiWAY5UiSNaXi5HmCkpKTCOFOig2ZzCr2n1/LTZ7BmupWSvQChhGisXErZm89VQLMRpt9J9oI11zm/odPE9YLFKwgrV2rKfSuwsrHEALcnMsPm98Yu67tO36K6+sYQkZoDhY6wvfZuqaB7ou1bN+ykQ994Fac4TWk27/FP3zlSXp7lhMO9RCPd5PPlTKU6cMXq3jy37/CsddbaO6O4lT2InQTXdO4/8FdbL15/aKqOmdePEzXpQzd6VLEqjZUTWbHAzvYsGNu+fT5wnME4fhUeeTMDLHh3IHSsRZxxxR8cUfxc/5ljQ1LmlgIIVYCn6coh9Ey4UPc7/v+4q1vr4OamkYMQ8cM+KiyhKqOVxGWWrVprvGGUwf4yz8VDA58mlgsie8LhPBob1tOaVk38RKHlsYYwxkZWwhMoNkd+RKWF3s1Y7EwH3tsL6vqxkl7iiIv2udgIlJJnboJ0rSjP7c3Bvnx0TfHHv/wjltZsaaA78PRNxSsfHFxallhCs5GUv0WdXWXGM6EefnAhxlMVyHLxYVmb99yHF9i1dpjRKJDDAzrnDm7nsqmHHs/0D9WZpUVmWh8KoehqsagtTFIfV07bbk12G5R6q6srJ9wsJRb13bT2r+cFTe/RedbVSiKiW1rCOFQU9PM8HAJB958kJt3P0tHtobl4g5uv2kLAVXhTkAWAlmWyWcKdA28REfbNxkekhkuRNGCBerXH6Hbc2jtmaB0JTzE8kFkRbDthjWsX3EMJVCPLMtFAykh4bn2lIR1lEcxlA6xefOrmGaITKaOyqpBamu/x9tHPrnozzLZ0ctPn97P+XMqA4o8Jie7bvUyPvnhu4stFYuEZVi8+cJbnHyjh7aBIF5Vsf82Eg3yyKN7+c4fSvT3TJU/9Ke3KMCxVDQtj+0ooDi4jhgrff9bx/XY8M5hpkqCYaTQtIW7is+EiQpQtu3Q1ZUilxsma0bJVnezqvIyA6fu4cevrkGWbSoqmhD49PSt5XLbcu68+0W6UlxVcmGZFoXsVNnxK+G5HkdfOc6pwwN0DOn4sWEG83Ei8clJVWW8h+xQFXs+9SeTHn/5u18eOzbVuR7PKS4YDSNIKrWC9jaT2tpLGEaC/fu34zjjrTm5zDJsG8orBlhXq4FcRSB2L9WbtvLMweM0NXZh112mNR+EUQ8NW6L9FYemph+jaZtobKygrq6VXG7tWJtTWVk/nqdz+fJPEGILqdQBAoFyZNnAMPRJseGNtz/Ah3f+ggsnd3KiaQNKYMIGyQhXDB/q1l3glvWnsZwA8UgNofww8RUX2D8Qoc2vnlbcclcwS97UkYQgEY9SU12K77n4dh97H/vMWEuqNsKj6OutYf26twgFUwxnavF8hXC0hf6+FKfefINLuSrE8iIpfOOm1Tz0yO3oi1S5BLBNG89TELJAUgT1m+vYuHPjose7El9+9BYcSyLdM/kaJdljpuWLZcgERwztDJSxlqRf1tiwpImF7/ttTKLcXsfVorW1izfeOEf9KhtDSAg8ZFH84o5WEcrrPj0v1ab5Yi4VqHT3i3QnP8eGDYcRwsH1NSRMamsbaWnZjKpcoK3jXlBsED56rB+1pLjTL0mC23Zt4YG7d46Vgd8rlNcYXG4qljOrJqhWldcY/NOzXUA5UE5zWxf/+uOv8dRf1BAt7Z00xqGT95EZqETZ8jp+XidwyaHh7C9G+mpBlmH1xjLufWwP4QmeCqOck+Zjf4oUqECRVXzfY6B3iI6WYXQtxzIzxJZlrZwfKqGjexWl8RQB1cBxNHJGhP7hKmrK4zy+eyvHD5zjG0dfmvZ13nzzT3G8AAUhE44NE9VzBBSHe29+g9cbbNoHl48dG49G+fXH72Zd3TJazrxWbHGS5m5xSpTvxbJ66EttGXPGHns/y04Bc8vYToRpmLz6wpscvWLRH4uE+JXH9rJp3cq5B5kGvzbBTCuXyVHIbsE0FbSybu767F+wfdcm7r5353j5fobZ7FM7xsmHbRfDJFuDGIZE5H2iSPvlR28Bbrzh3Xq+67HhncNMXhKyHMJxsmO/z9dVfCYUFaC+wdBQlp4eC9v1CIQMznStIhCUuXd7mn98dQ2VVVlKE+cRwsH3IR4/wcULN9PetgbXa6etfdxgsqJmbrVAKPbMnz98gbd/cYp8bu7byPMEnb2RMaEJWREguaBOsysguSglk4mveqyfbKZ85OeB0YtACw+xcdd3qY0b6LoLtKMqH6aion3KsEfe3stzP75v5LchBPuprFHZ+dCtvHDoFNnA+HN6rocVzXI2VcKWrf83N26y2b37RcLhWmRZxrIcurr6aGkxCAaTPPXUs9x0008ZHPw/CQRXUZboRZaL/I+CEWOgbwUtfREqarupdQO47vRLvW21rWAGEbZOzusnoOcIhQ3u3f0KrzXupH1wardjztEJB31qa2oIBotzqe1lULSyKTy3cPntXGraTiLehizFsawgYOF5Gvl8iNU3HuP82T0kSqI89sSdLF+xsDgxEb7n0fTWBXpahujoS+BXduPjE9Bn9saYiPmSmweTenEmu+I2dEc4jqPViM6LYXpG2qWM3HhicS1j/D24uthw7XiAX8ck5PMFnn32IG1th6lc1kA42k8wIpE1E9TUFFuFRqsIi3WBnglzjedY/dhOCEUxsJwiD8J1FbZt+wWlZR189rN/RsdgFed7VlGz+n7uvH0rilzctVZVZUy2873GfNurVq9cxh/+3sc58I0SVtRP3Ulpa9ZZs6aG5pZurGAnTdkwmBIHf/z7mPkEuBL6/+oRjuoENIWKqgL//VsHcFwP14/iFNIYXpC+ZJpsTiBpNrlciGRe4nbVxLSC5AoxcoUY+CB7EI8PUBrS2VVazrPfP8flIRU/Mr1z7TYtj2FpRAIFymNZNFUDNIKYfHhnJ/FldxIuuRWASCg4VjWa6BtRdPbN4c6SsAb1IQQ2JYnzKIqB4+hks9UE9SFmSyxc18UojC8wWhs6eOX5ozS0RMnHLcTyQRRFYs+uzTx87y7+3RM7Z3Ta/cYcn+lEM61UxwDDAYNsXiObK2H5qgo+8OCtY8eqAZ94WfE9TacCeBParbtai/e9FvSIJBzMggQITFPH8wV4yuRdwncR81ENKQYPc36ruuu45jCRU2EYfXieTTg8nmx7XpZweOUY12KhLtDTwTDWc/LkWjTtEoFQHksEON52A3LJLv741+4m13AYxw0B2UmxIaAU2LDxNH/whf8AGGjxm4hWPzyjMpDv+xh5g1FKRG4ox/6fHOTCGYmUp+Hr87hthQcjvfrVteU8+ugeTj5dwrL6qYvMrtYwn//ixyc99vkvtk763XU93jx0iqNHLuI5HpdzQcgX45jpQ86durVv+tCQn/x44/EAXZcvcOSlL1GwKsbWppbtMpzLoQYH2fOxv+TgqUqqqx1isVZkOcrQkEHBkFCDFoO5IG2Gy45QDsvWKGRjdGQT+CNcwVhsEE92WF01TF+yip6AAer0c9EuLc9gLkxIyxGOpXE9ge1JaAGPD+3sILTsTrSSWyedYw6uwuz9PkIxcFwF38viO1kidR/BLJi4I7wUx7J5+2dHSKc2Eo8lCWgGAdXCsgMMpCvp668iGixw5z3buG3PTcjy4ttZh5L9/M4DN9HXsxVX+CDbIEDRAjQdEux6YG558IWQm4MRFz3iMJzSxirYo590b2sINegSSjjYhZFEywNjxClbDb43RIz5xoWatTnaTl9dbLieWFyj+MlPXieZPMLqtacwhU86F6EyViAazCD7Q9i2OamKsBAX6PlI0840nu/7WIYOfgbDUJBlA9vW0PUsul5AC5iYjkIk6POhHR0k1kBoiY3FruRQjKJYaXhnFEYkSaJqmUVX29TXUrPc4HOfeYTzDW388JmD5NUiAc60okQqu8ATYKngSXi2zcUzVXzzb19Ell3KyqKsW9eCZWnYtko4ZBEIWFxquJn6oItvBqmuvExP78qxiUsL5BgYqEbQwxtvDdIbMBC1/ciymLaHNOdqhHSPmhIHSdKRJQ3Ps0EKo+oxjPTL1Cy/d8p5E30jJiaYo4+nUwdG/la8j2T5Y5SWXsK2Q1hOEFnYlJZeoqenctr31Pd9Lp1p5tUXDpPPjQfgVL9On6vhL+tCUnxqqsr41Y/cTVVFkXQ3k9Pu6z8r4ZEdu6c8Pl3Ckc/kyWVNCoYKARPM0KwykJ4H8shs6bkQihSDQz4rc8v9RRL8/mfKWLXqHJl8EMJZtLAKVM845juFX8ae3esYx5WcCte1Mc0uAHR9+aTKxELJ49ORwCORW3jhhYO88UY3bQMb8apKiy2DkSCffOyOseqhEShBkYtzn+PoyMLG94uKf+WlLYADIojrGQy2F1WGr0wuBnvTvLZvP71d43KchiHTngriVvUhNAt5FvW3/d/5Q4xskRwsJFE0wdM1TuwzEGKywtAohJAIz8ML6oGHbmPbjo3se/pV+vuGGDVA1OMpstlSuIIcrkVTiPgEsTTh48WHaclEuNweobK8hdHlqCagIgKpdDlCdfCWd3G4dR23rz8NvoPnqUTCNoGASVPTeuo0D7cQoqry8iSuXCCQIz1YTVTP0pQsHZOJnWnRnnM19HCBmJ7B9SQQCsGgQFUCKHocN/0S0eX3TDonGr6HnK6R7X4e1+pDDpQQqHqUt1+1aTm7D88TlJW1UVNzkbhmcdutyykp7cV2NEwjiKza1FRdJjNcQXntalbfvG3O934mOKbNmZ8d5fzBJN09dxEp7wLZQ1ZkSqtLCUUljv28YqyKMBFLwW/wPZCU4mfo+wLfAz3iYGQVbrp/vGX4yHOVk35/L/BuxoXricU1imy2QG3tJUxLw1IgGqsgkgAz34FrdKBpWxZVlbgaadrhdIaf73ud/u5aVC3PcDZOb3c9+XyUSCSNaYYZGi7la1/9W2IJj9/6wj+Q636B0Aw7U6NYaKJwJYdiFO2NS2sU6Hke3gQH7K8+fXjGYx0L1q1cxn/4rScYGCgGk+P/GidWZtKbGsIPWJieRGfHBmwzyFf/6cvgS8hAeWU7N246xCc+/V8ZLoQ4c2ELLcl6QGA33sDjv/JXGLaGaanowQKaavPm/vvYtesXDMg2kuSyenkVn/jgHoRxgkL6JXxnAKGUEkzcB9RjDv4AuzAMnoaHje9Z6MGVc8rCjvpGXIl06gC97d9GViIogQo8L0eiJIkkWfh+sfzr+yBJFolEOw3H/holUEZJ9UMkyvcynM7w032vc+qESU9Bx5uwCymiRfJ4UNN49IFb2L1t47y4N44tTZtwTFSM8VyPnvYehgZtjCvcf++5Z9ek82TFI9UdGNuKmqgGlU6pJMqnrxAtBtcNl65jvriSUxGJrCSbBctKI0nqoisTUxOWDBcvfpWTJ1/jrWObycftK6qHOwmo48l4tPphdC2PrGQ5+MYHiYb7QAgcWyGfj/Htb38Bnxj//gvPYQOZCX4Gju1w5OWjHHutnZaeCE5Yxsfn0DOfxzTjIBxAEAxphMM6pVUF/vxfDkx5DSeeqmfDDcVdZ0VREJIL5LncOK1IWTEX8MGx5lddLIlH+fVfe5ShoeyYytSTv3qSc6+epPl0nr6CijexxSkdHT9Z8iBUQEQz+IpN1pXG5pb+7jV4jorrKrz01b8E4YLw+UXdOb7whd8hEiwwVAhy9txWeps3UVM9wJnWdTz+K3+FaWmYtoqm2mgBkzfPbie6+SVEyCCgKNx5203cfONqnOxR/KGXwU2DnEDE7wX+J/z+p8DJAIGia6HvIAcqwQ/hmH3Y1tR5LhDbRWmsOF82n2nhjW+MyMQGVVbWNrO1/lgxZrkyidIehOQRUG3wNFxPR4g8VRWXkb1hcsd+r2hqV/3QtF4U039uPp1nWznx7DEaGyPkYjaoFig+0USURFUCSSrGDM8WSy6zqgZdBrs18MFzJ8em4VSAgD4DEW+BeL/GheuJxTUG3/c5d66Z7u4MpWUmnnCpDGUJSinsQhBFW4YkXOo2/1+LGn8xUrKu63L49RO88UoDzR0x7MpyTl/cyc5ghlSqlrKybkLhDI6jYxTilFd1kOyqRkhRvHkoUy11olBeY0x7bvk8TWh83+fSkUuceOUUpjn/CcL3fTzHHassDKfuRNh5fFvD1w2E5OF5KpJqFysZAK5ErhDh4IGPs27DaYSAAD5ry3N0CpcOp4S3L29kU00rsViGnKlxvHkdLcOVlCNRXp3mgft2ceftWxnqP0hv8ilUJYKkVuN5OczBp6is+wxEP0vnpa+Ab4AUQg+uRNEqcOz0okj+g90vFpOKkftIkhPEomk6O9dj20ECgQK+L1DVOGVlvWPJR2/7tzh/uplDLwdo6oxiVw4jSk0UaXxyFkKweeNqPvrw3mkdWtuag3S0Tv18bWv25MN1XYYHMziGTcGWEcECQhKUl5eQDyQoL58sg7h8TYFka5BgxGWgNzBWsbBNispPE+AULAQ2vb3VuMIDM4iRC+Db4XmZH72bGuPX8f6GZaXwfZlcrh3ftxFCJRCoRNcr2L79fyx63CsTllxOob/fIhxpIV9RhdAtqqtL+exH7qG6YqpkZ7j8dlo6akiU/Bzb1DHVBPhQVdVIMqkR0HRa21YDICb4GVxuuMzrP36TSxfDDIU9xIoOJAkQAtMKEylPoqoKtcsr0HUJsLjcGKW0ND7lGiRZQZ2h8lheY0xKMFzbppDOoYd6eea/7Vv0+2ZbgtbOGFa5gahMI8tT56GiH1LxP99nxOxt3MLF81QU3cC3tOKuO8Wkp719E89dHKnC+gIhfPxlSZr7yqgcvoFO3aVmWSOxaJaCGeJy61YChWWs00F4QRRPJnmoGfPCK6xZdRzbKVbGVbUbtefrNLVsA9ayYV0PipLH8wLkjShWt4GqpnEchZ8+/8OZX7wPvb3BSTKxWzacxUTGRkZRZMrLeum4vAlZcrFtjUDAwPMT1FS3gKKBFANvGKv9OwBzJhe5dJZj+w7RcMKk2wzg1yQRioesKGR7tzDYqdJ+dvx4x5I4d7CUTXuWzohx054BTvyiAisvj1UsXEcUEw1v6ucvKf6iTO3er3HhemJxDSGdzvCjH73G8RMFkkaQ9VtNyiMZPFQCqo6Hi2e0oATrF/0cC5Wm7WxN8nd//l1++uJZBof6QYZIaQRFvo9jJ36b9qYYquaxpv4IsuwgT+jl9L0M0hWa1e8GrqYdarBnkAP7DnDhNKTsIL66+F1p05XI44JuIQDfk8bsMlS1OOkWDBukPGY+QcPAaNATKAWd+pph1u+spn7no8gjRPe2y910NxxDquqj39AYaqvAe+4CIlNg+bLnpyz2oZgErBpJREerDJIUxrHTs3ImZkOx/WnyffTII08DJpHS4k5WPn0az7MQioYkb8YyNPq6exka3s+FofsRK0aUQNat4P6928bK9UFdo2wWAyTHloiWjd9nkXCSivJLtLWuZkX5M6Qyu6ZVnnEdF3wfz5MRsg+SYLh7M+nOALYpTSJkl9cYlNcYtF0M41OsVkxrw+hDunOATL9BMCJx17/7Q0TAIRgLcvfH7qRyxfRtYNdxHYuHhGG0IElBhAgALobRQvAq4gJMJYE7joNtBwiHc0gBj1hJkP/4uY/MWj0Mxao48OZv0HQxjKYVN2XqVx4mEhnCsieqDGbxpQQvfufnnD6aJZnT8auTSIpHvDTKQw/tJhwJcfypBMvqFQJaYEbFnfni754t9tmbOYMTz7zJhSNDdKQ1fM3kbPf8CL7TwVccxIpOJNmnemUle+/dPjZf9zcmaXr1NKn+KMNu8VgAPAXsCW1Zo/tXPsiqTGhE6MMxdWJlMTIDw/iej2doCM3Cr+6lx9AZOL6DFZ1ryHoefTkdfwZ9hEfXNpDOahj2SIwxVHQVYqUNPHvgYdoHZW7dchTTDGDYAXTVQCgWb57bQutc702sWGHWdZ177t9OnfM2yDUgyQQ0lTU1PwDHBuEhlWwvvtzBk+B7SOpIG5RcggdY3S/OmFh4rsul109z5pWmkQ3O4qaUGlDZft92jv8kTk+rOkaSjoa7qS5voKtjObFEmqjmkDHXzv5aJuDcwVLsgow9QR4WINVZ5Pc5phiJC8X3XJL8KdWLUdSsyV+VKd77DdcTi2sE+bzBN7/5LMPDnWzbfYY7YsNEA1kUTaBKOoLiostHQky/xAHm5k/MV5q2kDd49dlDHD+c4tWDl1l911oq1txCPB4mlHX5xl/+Lj9qCvPrDz5I3doCQc1joLOPoUwFbx3axeBAOb/95P9DZ+cmTCvOyjXFisQ7yYOYDrZlc/RnR2g92z7BQGhmZIYVWnvDIyZP1lg5dQw++MX/TcGUhxS7aADoAxIIqShHp6oKq9euAAFGwSLZ2QeSg1TdPzaQ60FjOkbfCzkun32V8vJWaqouUaoVeGJ9kDNd9ZxJVeOs6ORSOkbPD1N89KPNxOMriEzwE5rY6vT/s/fe4XGd55n379TpAAa9g50Ui9gpNlG9WhJF2nJiObG9TtbefElsJ9kvZb27yedkN1dyJRvH3uw6zjrZ2IntOLFIyWq2usTeO4lGgOhlBpjB9Dnt+2MwgxlgUAmSEs37uiSSB2fec3Bm5n3e532e+76n40xMhvFcCm/l4xNM8fTEIKme4wThodMglJJMhACIDLvo7+kgkRCJ6XY8RQGEkgBFhS5+Yfculi2snfzi08Dt6qWh/jha0o6m2RGkONWlE2UtLcvi6vmr2O115wSkHAAAIABJREFU+IdLU2TGqEksomK3g6tQo3bJWF93V4uT7588xPMbt1O7JMqxN0syvIqAT0XXIBaWiMeg66qbhFaAzTuA5DBZtnE5Gx/ekNfZdbLSdm+rM+/O1I2CtyrOtfO2uWs63sFNRTb3IRbrIbUSTc9NAikfnqnnt3z8iex2qfFyssFgDDCIxB1Ygoksy9O2JKY9JZ7ceE+mLfHy6VocjjDHj64mFCrmV57/EywLenoWY1hW6rsomDg8Qb76f37G1u1jvjSyLGOboapPGlosSaBnCF0b23wY8Zfzkz/9EQDJhEBbZyFxbwyh1o8gMqekxRr9n2BZyJZIlaDibg9w8TvvZM6JROx0DpVjVAyALctkXtRByeqtHKVAKDaF2qU1mcOJiIu9v/Es/R39fLDvINGRCJZlYlkCOOLUrDvKXeWdFNoS6KZMd/cSfP5UbBkWdXyjGYurIEQo7kCQxlpOEwh4XCHECj8dZiFC53JWVbdT4BohknBwpnM5HWYhYkWukWo2Goq7WVXTTmmBQVFZPWr1IpJ9FWBGERUvZsKX6hMmltrNjw6AoqQq5/aFuYOJBTDqsWIaJq2HLtJ85AqmPsphMKGjo4CATYS6zpSL+7Iadjy9Dcc4jozH1cei+hMkknYSSTuypLG49Ge0+pg0uRhPbo4GZGSbhbNQnzA3f/3k+3xl4y4G2p3Ys9SeRnwKhiaiJYScsWZStc4XG7obXQz32+a12jIVxp7B9cWGO4nFhwShUBSns4UFCy4Q0SXCSQelRTEUDDB1LMFAkFQktR6B/O05M+FPTCcla1kWF0818v6rZ2hq9xD3Jlj/3F3IisTOrat45L6NRMMxvvXHP6CjtTdz7VhiKU3Ni1mw4DzFRf34feXEEvWoDg+6keslcTNgWRYdl9s5+NJJmltdREUnljB9YoEzhlDXjSQJrFmzlPXrl2Z+5O8a5MqBy3R12okJBlY6GIkmFIQQJJOCQhfrN61AVWWa37QI+heNk9VVsTv1zJrA7lBZuLgGNIVPfyYl5x+Nxnn7jRMEhBGCnggha5hVVeeIJVSGIwXY1QSraq8QCji4Fi1F8AYZ8YQZ9HuIRfpwBqNUVJciKfIEedjJOBPZyE4kLER0LYDNUZnVzvQ93EWbCQdSnBNdi6PHOwAThAqSyWGwOkhqIuGYm2BolPAu69g9USKagwd3ruWx+zdOkCecCWTZJBpOReLykh6G/BUkdRuCYGKMSt2Weo7TOZpYGIbOi99+mSvnLdY9dAWrzIcoW5zY91+JBQTAQotLHH8z9ZwUu0FpZTLnmjaHSTSculfVbiAIIlULYljxdh556O/oSYpItX2se2Atf/bx3yYRnji12tw6pTWJnCCVNk5KRiXOvjlWAVIcxnUHk+n6c/eoFy9f1wXu4KZgPPchFusGBExTRxBAklRUtQGYXG1mMhM9GDNbS8vJjoyE6evLlZNVbBKP7pwDydayGBiswNDiVFa2UlzcRzTmIjBcQSjspXzReRDB6bRhxRewfefaOTyhFLREkvhImN6oj2hCHlNbAGJJkbPXRnfrZS2jGFVeW8q2+9Yiy7NbCsVHIrS/f4muFpNgUkQzJTqsPNmJMwp13YBFYWkByzctAwEuvGkR8S3InGZZSqqlzZ7/Payor2DPb+zG3+PHHJWnM7o+oCx0mVhCZjipYLcs6hou0R10ca1zCUWCwLplMRp23UWh1YjXSoKcxfnQQ5hCKZ/+7EQbGRXYuRGyI4USPYUceh/JCmIIhehqLfZ4J4JchGIvyrQzCUWbsAInMJMhSAyQCnZS6s94GwgN4GgAeVzLmjkCaglDnYOc3HeQ5ks2AqYLa9REEMGEshQZ3VnoZMcz26heNGam662K093oSvlElPQy5C9H0+3Ikomm20nobqo9J2icJLFIcxbGz5vJuMSZN8tQ7QYrd+bOyYrDyKg9Aah2C00wWX2/Py8H4pcqH5lVbNCTIiMD6oTY4K24PkG/qWLD10++f92x4U5i8SFCWVkT8biNhCUj2AwU7ICOINtwFawGQNMCSGL+xflM+BNTSckODQzzsxfe5+J5Cx8KVk03omxRU13K83sfoKS4kGH/CP/rv/+AZasaWLk+19U1Eiml7dp9AAwMKCnJtxuAfByKZEJDEjr48V/9KwCmadHVqdIXt2FVpnowZ7IrJQDFJYXcv3klvSdbuNB2CEj1yPZ0KfRGPFgVAxPGUxWFex9Yz+YtKzO7bf/3g+YJ4z+/cTv+fhvH3sytEGkJkb/49Y/xjdFy/dJlDRw+eJZDB86zesV5EggkJBHBoZEgJW26+q6LtJ/djhVxIDoSXBiqZOuCKwSCEqGRKHZ7AtWm09W1gbff+GH+31cQWL1hBRt3rEGSxAmk7OjIBUwrTl83aMlUYJalOL1dBxkcXE5ZWSMedz+WBfFEEfGYjXCsDsUxgqzGkRQTu2eEhJ4iF3ocFhVLPk3Vwi3TvxmToGFxLLMburjmdWLxUgQEDh64h56eSsDE6RjiWoudSCiCRD+Hjkno5YOZkv3Dj23i/Iv1qDYLhzs3mMfCE5OdtTuGc/7d1eLkn44f4M1v7KPrqj0lay6A3WknEZbxlEzmypobENLGSfGQlLPzFc8TfGaLj2p/7h3kYjz3QRTtmGYSWVYpyMSFIOIkcSHfGGnTvL6+VzOJRSKxjHPnlqLaGlEdY3KyYtEm/uCzD1BUMLvPjRZPMNDtR0948A2V4QuWoiftWFgoagJEkBSRyqoS3G4nXa2z2GSwLM6+fIS+5u7MoVjEREssJhgpHSV7j8HmHUCoHvMgstlt3PfEZpavXjRBSc+yLAYud9L2zlmM8aRuKyWjGgooXBssQCsfQCxKpFpdJ7lVSZLY8ew2Fq1elDn2Fx9cyjnnKxt3UbUkwqWDxTmLyHQbTnozILu10ha6guCqwQxYJIdCJCwT0Fm96izXTDfDSZnglXKG+q+wdHk5dbWn0LUghqEgSRqykqCrZyOlBFmweWpXas13mKT/VbA7QaxJJQGhd8BWjugajWWj7UxWtB21/pdINn0jdVy0g6MC0VaKqQ2D6EStfJxkxz+RiOuMDCYQrDiykqCzawPNjQfoGHZglqe8i7JvS5IkVm5dxdr71mbazdL4w5ePZZ7j2prXiMRLAZELp9cQHCqmu2MRTsdQZv6brIqQnjfHVyNieebkfJs/vS2uSYnVs40N6WraRy023EksPkSw2WKEw95U+wyAXAZ6N5YeRTe0aY3vZsqfGC8lq2k6H7x2hKMfdNDW58SoHESwJXDa7TzzxD2sX7MUQRD40i/+d376wgEWr6ijqPSHPLftLq5dddDamPowakmBSFjC5TZmVVqeLdk6u5UqMhLh0EuHuHgyTG/AwbGLo18MwcIqHkLwJrCpCtu3r6W0bCLZbzwk08J37hpH/qWR7mEXiOldFQureBihOo6qymzZtpryilECoyBQX185I8nCdN++asutOrkKtRxlLEkS2blrPWvXLSNx4QAGpSm3vTQMA9EaoqDIRUiMYmkC7T0LAYHV1e24pCThsIe+qw2cbVxGTBvbHWqoaWX1sou4nFEiUSdH3uvl0pkWHttzL/HgGClbSyRJJJLomoAkhBkIpXkPDjyuCD87vApYxSc/9kNCERcgpbTU7VFMRNw2Gbl0L4XJw4hmEFEuo2Lhs7jLdkz7nGaKZMKLLEcwdDcLF13jj/70m2haED1mceCDMzQ1Ogk6QxjeESRJ4K5Vi3nsyW0zdnYtrYrTlUdRZqZCADOGkBu49NFS+kxK6DNBujKSHjvVM3zzDPLuYO4Yz31wOKqIRNrQ9RiGYczI+G4yE71ksp9EIsnrrx/i4MFervmXY1Z6M3Kyn3xqJ6uWNUy56MzGZ59ax2CPjatNKo0XKrCsSkBC11VkNYndaWFoMnY7xC2ZJUtqEca3m2ZhPOEaQI8nEc0O3nlxhIg1thCy1CRbP/fHCJJJcUURm3asHiezmuqTF0SRuoVV2GwTW6ziwQiX9h+h5XScvhFnqhqdB6YrjFDblRJRitlATSJIJnaPgwV3NSCOXtfusrNq26rMZtNkSLegpNtv0ki34eRb8InJIVDL8FZIeLxukvEEGCYyQyy3L6P5VAtmTQ/tIRfXDq+iocaemfeDYScXmjZyrWsJpSeusuRkCxsfsSHHD0DSP0GlKdn3OshORGWUMzmaRJAIgKs+66ZS7UxK6TaSHf8MagVidtxKtzt5NtLVfwUhdBhRSRKJjt5P92KswuCoqaHI0g3LqFww9rktqS7BU5RVdZkE8YQXRY6i6W5Wrz/PQE8Fv/21P8EUHSRWz731Nhsz8YW4XoiSiaGJE2LDfF7jRsSGO4nFhwCWZdHR0Us06kBSEmhCuuHSiyhqmFoAcwbGdzPlT2RD13X2/d9XaLqk0TXiRKhPtQGtW7OU3U9sx541+X7jh/+JkUCYf/j6Pv72z59lx8OvU79kTKf/0Juphfb2h4cyf58JZsq5iIWjHHvtKMN9Y7vHw36d1o4CkqURhAZ/jhumKAosWlzDnmfuxe2euDhMROOcef0EQ92+zLGRYZ3Wax4SxTGE+qFM7yuAKAjUN1Ty9LO7KChwp3Zx+l5PTcRtJWgzkMv7xsunMn3745FvAespcCEWVI32rI4lRqmdn2p+7Tc/wXvvnOLEsUsYepJ2fxXt/hoQLEgqiH3l1HojmB4/vZjUFXezdfElEppCSFOwucOs2nCQo8d28A/ffJcHH2whqbmAMPGYhbtARpSSSIqWkmYFbEqCiK5AXUq95F/+5XeIRQpJJF0ZR1KnI0GxN8rXvrMc+MSUz2S2KKtKZCRkfb2PsLD2PeIJJ8UlPhKxYcJDA5w4sZ7TV1Wsyj5E2aCo2MMzz87e2TVdQcoHc34UBYGU4VK2znlvi2teyX7pygiQahWYBxOkO7g5yOY+ANhspWhaBE0LzNj4bvwYAIYRIh6385d/+a80tnqIFmkIdakF3faNK3nq4S05crLTwbIsutsEVLmF8koZQ06CrNPfvgZZktj+aABBFDn2ZgmyIiEmhCmTChgjXANEg2FO7T9M4+kYfXGJcNkggjz2JUzx1xR2PrKZ1Run3oEHiIeiNL16nMhA2mvCYsRn0NpZgFYWRqgfnOCubI2Kbwip0xFEC5xJJFlm3f1rWLl1JfLQMZS+1xGTQ5hqMdpQBLM012RuPNI73Okd95nAVIsRzTBIhciqjKzKoAUxxSq2br+HFVuW8/4LB1LxsiDCNRxca9k0NoAI1HczmFDw9I0QbjyEiQ3DVJHETpSev6Gr5wiB0EJWL71KXHMBozHAaafAbgPimQcTHYkQG+pDN2QOH9zPqcP/D+FwAZo2FtcUJYrbPcJdq16ita2GSMHDCAXhMZpQXQ+CAN5KL7v27qSoNIswOANkErTeR1hc+z6JhBPdcFJW0YGoR0jUPzOr8abCzZB7LSjVJnhiTFURmQtuRGy4k1jcYgwPj7Bv37ucOZvAVrqGrRsOY+kKsupGFaNYSFQu/MqM/Cqm40/kQywcJxyMEI0VIXgiyKrIkw9vYceW1XnPLyhy8+U/+mX+7i+O0Ne1nwVL/8Ncf/UZw7IsPnXvUrraZBLaphwfIpt7mB2f+RNEyaJ+QSUr72rIxILKihLq6ycalFmWRdupZk69dp7W9gLi5tjEZzqiUNuDKFlU15WxavWizHglZV4aGqoQBCGVVHT8E8hOUCtmJZc3W6TLxiZkpPnQo6j1e5FlmYce2cKGjStou9qFZVkYhkHTlU46Owcw63roCLlxdlSzoiHA5hX9KM4S3KS4CLF4kmjEx6q7T/HyoVL6fR4URSORsGEpGlbSRok7BqJCdVURAnEkTILCY+ytXwPAn79yPzu2/4hE0ommuXDaEyhKhKMnfhHonfT3mivGm91FfD5G+l4lEuyhp03gxNl7aIkXIVT3o6gy23dO7uyazZ1IQ0uI81+RuIM7mCNS3IcUH0IU3aMJgszChb8zY6+K8WPE4wGGh/s5dnwTZztsWDU9CJJJRYWXz+59kMrymW8MAYSDYd7ff4Dhobuwu8f8YVSbiqoqiIKAMM2u/WQwDYPmgxc492YLV7s8aGUhBG9KDWjdlpV8748+zYjPAwLY7Aon/y31fS6pivFnL5+YOJ5p0nn4Co1vNNLeUUiCrPnfHYG6LgTBgpgd4mOVTcsdQVA1BEmkenE11YuqEEURURSpXVaL0+NE9B3B0fEDTNkFahmiGcbR8QNiMG1yMVtolY+krgUgusEM5yyei0qLePrff4yuli4igfzJykDXIO0XOli58hTBuI1EJF0NsGGzRVHkZo6fXkWZdywuAEhBDbMYXIUWSsTP0ECceDSOKBkcPb2Ja11OTp/eyYP3/Zh40kUsbsdhj2NXI7z93sfRClqhOvWZKyovZNnGpZlE0FXkonZJ7YyrZNnIXnCLvh6UvjcyCV6s8lOzeg/G8yfmu4p8O+NOYnEL0dHRxw9+8DMuXS4hWhpFMF0onXdz/xo/RY4ksuymqPITMzbBm4o/MRn8g8MkEibxhAQFCQRBoKR4TOrTNE0EYaKjsyDICELugky1G4SCCh0tDgwThv2pj5csm5lWp+kWbJZl0XyqiZbTzaPa3xCPJGlr+i84igaQVS21Gz+K8FA57gI7zzy9neXLGqZ7RAQHAhzbd5Cm8zBoKVjVXSnp0fTvBTicdh57cit3rVw46eQ2WWl4Krm8uSKnHJ3sT5Wp6/fmXMdbXIC3eGXm31u23s3gwDAvvvAeA8IQUWeUSy0NbNoQwBJLKasee49jURd+fztiYYQLw1VsXXwJ7AkSuoIgqdhtXiRHMSIJJLUYd+UvsSzL9FA3FtM/9DFKPcdxeoZJJrz0+O4lEqlnqsTic0+tY7B3YktSPqfsqeAs2cbBn0ZoPBugbcAD9V2I7gR1DRU8s+c+CqZwfh/PnYBU5WiqSsX1YHz5XE8IxJFRHJMTb+/g5xvp5CGl6DSzCsVUY0Qi3fT2apw4uYPGWBFCVT821cYzj2xm68a7ZmRGmYZhGJw/cIGjbzXS2uVBF3SwxxFFkbJyL0XeAoK9JpGgkmlpMkwI+hUk2cxpcyqtimOaJq1HLtN9sT1DWoiFElxtdjCsilDXhSgL1C6u5tFnduByO/n73y1j8d3pCvDYRmtPixPLsug61kTfubaMM3Y8kqS9ycGQJKeSiDQ5OC0cqslYuoTgSIBzbDxRgLKGSnY9uxNXYf5edKXvjVRSka4uS4WYo8cT85xYmKVbiY2OLSYHMdViEvXP5CyeBUGgbmndpGOs2LyCVVv9eM6+SSBiS5HNR5HAxOMOYDV05cYFTUGWNeKaTOf5lZSU+BDlJFFL4kLXEq5Jdmjowv/TUk5dvoclDZcpKvQTihRwqfUe/COlCFWDyKrCxoc3snzTxOrSfBjDmaVbr+uZj+dPzHcVeTxup9hwJ7G4hejo6CORMIgn7IiuOO4CB1/45f8Xzwx69SfDeP7EZIhF47z3yiFOHfPRGSjEqulEUAyKigqpqx4jiPV0DPLVL36d3/mTz7Fm0zIsy+K733wRXeulvDpXTWLTziAdLQ5ePHlkTvceGAjwwQsfcOU8DEfVjAagIUgYsga2JJIkUeR1Z4KfpLv58peeQ80j7ZkNXdM5/+ZpLnzQSVufK8MjcTjsrNu4FGW05G+3q6xdtwx1EpOlDJL+VKUiG1lyefMNpXTbrBOWsnIvv/LF3by0/10uXWjHtCWJRNy4hBFg7D2WlSSV1Yv4zHMP0T84jMZiijiOTQrjKaymoPLJjEPueER8h1hYP0h16ZXRhOLRvB4S+TDYa5vWKXsmSMaTDA8ME4m4wBVHUuGe7au5/4FNU+563Sj+hM2tj5LxJh4fHxSzA+hs5QmnQ3agSgcp4CMZqH7eUVq6c9ZO2pON8f77Jzl37jxNLXUIy1pxFzj4vS9+ApdzdnGnr6Of9/YdpPGymln0I1g43Q6qqkpRlNTG0/qdATpbnPzg5KEpxxvu9vH23xyk6ZLKSNyWUd0zJTElmKEYOD1OHtm9jYbF0/fJm7rOkf/9GlcvSgSiWeOJcmo8VcPudrBoWQ3Bpl6G+yFqguWMIjgNSmpKqMuSfi2tLqV6cfWUc0qa95B70I04A6PYueB6F88AJVUl2PyLUIPDJI2xjR6JKAZu1j1wN3A3AzRQyUkKzBH8wwqnOxfQ3rUI67INoWgEURaoWVLNurWp1usrrwxx17IWXI4IkVgJfYE1xKWFODwONj6ynsVrF0+QiU3jVgpP3EjuxM9LbLiTWNxipOaoVGO63aZcV1IxE1iWxaXTTbz36mma2jzEvUmEmmFkWWLn1jU8cv9GlCzpvZqGcp76hfv4k9/6Fp1tfUiyxOIVddQt3I/DOT8kKF3TOf3mSU5/0ElbvwujfAChTMvtbxWhyOuhosKLlFVSjwXtE5IKQze4/MF5/B1jKiCBgRAtTW6CLhNxlEeyas1SHn1ia14S37RQS1ItSVKWAeCoXN50uGmEYFI7VuUVxVy+2A62BE1Na9i0+RCJ2DCSWpAyqtLDFNbvpap0EdwFsBH45WnHjvgOEej4LpL4n4nFS5HlSF4PiRuNcCCMltSJJ2SEggiCKFBRUTJtKf1GVSX+qe+NGZ97I/t0s8eeTe/2Hdy+ME0Tny9IIiGBkjIts9uVWSUV8WicQ68c5eyxQTqHHViVKQlQd4GTopJC6usnzm1TQUskOf/acS4fGuCa34FZOYCg6DnzvySJrL1nJdseWJ9XGtaIJ4kMBse8ikyL8KCXkycVjIrBcfEk5QZV7nHRUODBd9hPb783411kc9i554lNLFw9ecV6MmTzHsYOhjHVmbWW3QxCcD5olY/gSv4AhyxktVVBrP4TrC1NSwCvBT4FgCscQ//JIcThXixnAleRi53PbqeyIdV6LPqOcHbJUewOg7hegd0RZfWSA7T6XMSCVazenr/V+sOAGzkn/7zEhjuJxc8RhgaG+dm+D7h03mTQVDNystXVpTy/535KSyYSpQRB4LnPP85zn3885/ivPLWSjjw7y7NZHFuWRWdjJwdfOkZLi4sRp4k46iGxdGkdRV535h7OvlBIddX0wa+vtYdj+4/S2uQgmhz7eGvqmLOrt7iAZ/bsoqZ27q7IU/EepsNsFrRfempDjlpUGqVV8RmPU99QhaoqxMt8nAsVEz20g01rz1BcGcBVWE1h/d5JKxJTIdz3KoLsRjfcCITzekjcSOi6wal3TnHq3Tau9nnRy/sQ7EmcDidV1aU3/Po3GvPRDpD9mokLljsGeT9P6O4eZP/+dzl/QcIvKbCwDUmCRYtrpn8xqfm66XQTB187S/PV0U2p2tSm1OZ7VnPfAxtoecuYoOIE+eOCZVl0nb/K6ZfP0DKOPL5oRQMuT2q+lySRVeuX4S2ZqOpnaDqJUITBZh/xuJyRfDURSFhg1PUgyQJ1S2sxgxGC3WEiYRuE3AxpKj7BQi9MeU2IksCiNUvY8sQm1LlsNjE972E6zPR7PZ9zA8ysrSobDreDhz/1ED1tPQR9QZZtWIaUpf6k9L1BIvFraHqKXK+NxoZqzwmamLtXyR2k8GGPDXcSi1sEy7IIBEJomoglaSmy2A3E5TPNvPPyUa40lZOs7kVwxHDY7Tzz+BY23L1s1jszc3HP7mntoeVUE6aVmmyiIzGaLibpjdqxKlKL/uKSAvbu3UVtde6i/zt/IAPJPKOmEAvHOPHSYS6fCNETtmGV9yPIWWU9AVRF5t77NrBl2+pp5f+mw0x4D7NFjsrUqNyfr3fmClKToaamnF/9D8/y0r736Lw2QIvhoOWdJykxVJat1CiuiANvp85dVM3yGSiqABjJYQS1jOKSoVH/CMj2kCirmj/RIcMwuHD4Ar6eMRfYP/vD5/H7V2NIZkq3XgC73c7C5SLe38r/+ZyPRO1mYT7bAfIFmzsGeT8/OHToHG++eYbLzWXotV0Idg2Xy8Enn9rO6uULp3398GCA9174gMvnrcymlCBbVFSXsGfPfZSMbkp9c4bfochQiJP7DtJ4VmcgqabI47JJcbmXJ/beS0m5d8rXW5aFr7mbSy+dIBLYiOG1EJwxctwkEm6KygvZtGk5vg+aaGt0EXAYCEUhBE84YycoiFBQUsC9e3ZQep0bErNdoM8Uou9IDhHZHPkOVUsmVsevp1VoLm1V1QurqV5YPeG4mBzCVRRjoCe7Xbgcp2Pohldf5rLonu9E7Ubjwx4b7iQWtwB+f3BUCUpjQHNiLWpDkCzqr2MHfTpcOdtMOKSiSSaiI0FlVTFf/MzTOOxz25mZDWLhKIdeOsyFEyMMBV2ZuV+zZPSyUW8IRebCq/8NPVbGB/+Q+/ryqjj93TbaGyd+aRxuneajlzj908u0XPOQKI4iVMdQFIklyxqQR/t8bTaV7TvX4vG4Jiwuu1odGLqIpJjULopljk+32JwL72EyTKYy5XE9B1z/7ntBgZtPf+ZJmhqv8drLh4iKQ/iTMkfPlmMzxwx73I4mlh9v5P4991JcOXX5XlK9GGaYL/3Hfx77PbQgkminYnV+d9O5oLe9l/dfOERTk514VhVqwO/FXd4NgoWiyFTXlGF3ilMmXWcOeBHz+HB1td4cR/g7uINbgfPnmwkG7eg2HdGhs7Chgi88/8S0crK6pnPi7VOcev8abX2uLJNJG488voW77146q00pQze48t5ZLr7TxtUeD3p5AKEkgWqzsfPh9XllYn/vqU34e8e+n6ahEx+JIdHA5kcuMhJzExgpA3JFRmyuJKsKC7n8g9bUZtOo9LTL66a8bqxVsqymjGUbl/HHz2ydsLjsbXWCAFWLcjd3plpwzgfvIRv5lKYW175PWF5DaBIX6VsNUy3mV778v8dI7DAqhTt/PhKTYS6L7ksHinNsojKvaZ1dW98dpHAnsbjJOH36Cq+8cowrzaUkKn0IzjgOm429j9xrNHrlAAAgAElEQVTDjg03xq/KMEyS8SSGIYJkIogCddVlNzypME2TK8cuc/xnl2hp95AoiSLUZiktCBaiKLBgYRV7n93F539cRn2e3fmOFicVNQnqFw8TCUbQYmM74d3XPLz+j+34BAlquxEli5r6CnbvuY/CwvxqQL5ee04VoKfdgbtQIxaWco7PpipwvZhMZaq85BwJHpyXawiCwPIVC1i0uJa33zjG6VNNmLU9xK2xYBzTZIZOlNHX8Tbrdtaw+ZFNKJMQ2d2VTxLo+O6ovLE7w9dwz6AdDHL9KMYfB4hFYhz6yWHOnQjQFVSxygdyq1BKEkEUCPSsQrBs+K+lDicTIs9v3J43MTR0EXfhROfTYB5C3Vzwd1/eTTIyMRH8sO583cHtD13XSSY1DEMG0UAQYVF95bRJxcjQCD/9/hs0XbQzKIoIo22qK1ct5rEnt2OfZfwYaOvl1L4jtDQ6CNjJKDwtXt7AA09tw+HML9rg73VQvSQKlknMFyY0EEJDYSReAFV9uAqHqV3Wi80kh2PR0+bh+FsmydIoQk0MWZW4+961rN6+OmNgl418C9KBdicWTDh+M93r8ylNJRJOqr0naPyQJhbX2xI2X1yT/OZvE+djUxdw5YkL+YjWc8VHrSpyPbhliYUgCPcBfzV6D0ngNyzLOjLuHAH4GvBJwABOAV+0LOtDzUL0+Q6MSgP6UNVSKiufzCh6HD9+mUDARkI2EF1J6qpK+c1Pfwz3LBU5Zorezj7eeOEgl68oDCsmQk0X4mhiARD1HSLS9xpmchhR9eKqfALnHPrtx8Pf6+eDFw7QdFFKLfrrUov+sgovBZ7Ugl0URDZuWs7ypfXTjJZS+Oi/2kd4BHR9LChEYgK+Ij+CLYnD6eCxj23lW7/7CV7/5kej3SWDSVSmHPYgiXk0YgNQFJnHntzO+k0rOPD+abRkyhwnmdTp7hzErOuhJeihb/8QLRf2s+vpzTTcNdF9N83LCPe9ipEcRFK9uGfB15hMUtayLC4fb+Tw6+doafMQL4kh1PiRFYm6+sqMH8VBh51FS2o43W3D4R5LOCygdkn0piaGaYR8HhasuTVqJrcDbue4AFPHhhuBlpZOXnzxIJcbHQQdIYTqEURRoq5q6ipoxHeI3sYfsWxhH8WeYi4EK+mL1/OLn3p01iaT8XCMMz85SuOJIF0jNqyKUbJ3oYtHn91B7YKqacfQowlGuv2EQxKaIIAzBgk3lYsqcLsdGL4h/FEFK2uuTOiQTMedhnJ27dnJXzz/AH8/yeLuw4p8SlO64cRuu3qL7mh6XG9L2HwttPOZv8GtmY9vpdLVzcYtSSwEQSgCXgCesizrsCAI9wMvCoKw0LKs7C3rzwJPAussy4oJgvAPwJ8Bv3HTb3qG8PkO0NHxj8iyG1WtwDTDGWOisQCSKtmKIqxYVDPjpGLEd2DUo8KPrJZQVPn4pNKy8ViC9149xOkjPjoCdszyAQRVx+Nx8twz97J8SR1R3yFGOr6HILsQ1DJMM8RIx/cA5pxcaAmNkz87wZmDPVwbdGaUNpwuB099bBt3rVgwq9K5ZVrEQhHiYY2oQwRbHNTsPtokktNg7boVPPToFlRVmVCRSGOuC82b0pc/icpULF4I81hYyuZxuNQSnr4/1y28r/mnDDX+GFUIE4k5uHB5Hf/6bZlVay9TXj/Wqle1oJKaxTW4SrfPifidhmVZtF9qx983WsmyoLO5iysXBHwoWLWjAgO1Zezeez9FRZ7Ma1/8SzeyMvF9voOPJm7nuAAzjQ1zG3d8slJcvJ39+9/j+PEeWvsKUhwGxcDr9fDpPfexqG7yxXzEd4jhju+CZhGLO1HUJFsXX+SCT5lVUmFZFm3HGznz+gVarxWk2lRrYsiKxIbta9iya20O4Tcf9HiSeDDMYNRPNCmBPQGiiSRLuAtc1EUsYsM6FAopU75sJNzYPTa2PbWF+uX1CIIwr4u7m7UDnU9pSpaixBNT81Bmi/E8Dq3ykUwiMNXPJr3veW4Ju4OPDm5VxeIxoNGyrMMAlmW9KwhCL/AQ8JOs834B+FvLstIzxl8Db/EhDiB9fa8iy26U0bKlNDoZ9PW9el3BY8R3gMGO7yHKbkS1DMOMMDiaBGQnF5ZlcflME+++eprmNg+xogRCzRCyLLF9y2oee3BTRk420vcaguyi6YrEd77VyOWLYUQRGhZ+nf/8P8tYvWHpjO/PsizaL7Zx6OWTtLS4iRRoqdK5LLJh3QoeHV30z3xAiIWiDPUECAXKMSVQHFEQBWx2JaMcqEVV/t2vPk1F5eQyr2cOFKHFpUyLzLVGFz3tDmwOM69B2njMd6KSD5OpTOniwnmTpp3OLVzzHcYTeglPbSHDEQ+a4WPrxkMcuRzhvfc34pLHtNg9BR2sXO9m5+4duDxzew4BX4B3X/iA5gsGsdhYgA6ZSiYhddgdPPrkFlatXjwnJ9ZbjUsHi9FiElpWGR5mtvi4VdKTtxC3bVyAGxMbJktWhodHaG310d9fCsV+JLvFtg0rePax7dMu5oM9rxAc0hn2K8Q0B5aaAF3mrsr2Gd9XsG+YE/sO0nxBxI8yRvauLeWJvbsoyNogyAfLsug718alV84RHtmMVWKCIwmigKfIgx2Bnv4EJ0/IGFLK40iURBQ1yyk5aecTX96DPInH0fg2GUi1PikOY4I5Wj7crB3ofG1FlVWdnDi3h2Ak91pznRumcgwHbpqb+M8L0nEByIkNM01KP+yx4VYlFouA1nHHWkePT3VeK1AsCEKhZVnB7BMFQfgC8AWA+vobR4KeDqldo9xdHVF0k0z2Y1kWhqGPGoBazEYHKtD3OqLsRlFGJWGlIrTR4+nEIhqJ8dq/vs2lMwl6og6s6tRub1VVCc/vfYCycXKy5qiqz998/TKf+uVq/vwbXkzT4NvfaOLX9n6N99u+O6PFXDgQ5uD+A1w6HacvYcOq6kVQTMpKi/j4x++jsmJ6b4dsGJrBUK+f0LBOXBcx5ZTbtmJTUgRdx9gWfpfhnDKpANDiEg63kWmR6W134HAbRMNTB9ibiclUpv70x0FgamOpmWI6t/Dsn5fYoKigkMHeTlYvbOZaqJygmWpBsoBAzEb/2w46W1/m7m0LcLhn18oXHBrhwuEerva50cuGocg/JjMvmkiSyOq7l/LI45P7jKT9QJIJMee7ZHNM3jsmKSaxPO+7pMxPv5mvq5Bgf5bPSkhClMA0c/u0Z7L4uN36bmeA2zYuwNSxYa6YLFkJBt/CNNcAIkgWiiqyZf2KKZMKy7JoPtNMrOcqvqFCDMVAsCcQBQFnYTFlHn3a+9GTOhffOMmlA920DzhTHhK2JHaHgwc+tpmlK6f3hogOhbi47zCtZ42UYpSsgaKjqCper5vEQJihiJSRk0W0cBW5Ka70IopjY/e2uCZNKmBim4wF2N068fCHi3qar61o9x8leLq0HWifl2tM5RgO3DQ38fnAXMzfRMUilud9F5X5U+vsbXUy0J7ahEvHBciNDTNNSj/sseFWfYMEYPy7rAPjGVXjz0vPbBOYV5ZlfRv4NsDGjctvrHbrFFDVUkwznJngAUwzjCAU8nd/t58LFwUGDaC6G0EQKPNO1ObOBz3pRxzXZymILvQsR8+rV64x0D1Ib98CWNqK6pDY/cQ2Nq5dlnGqzoaoejHNEH/9t6uQ5dTPLTPM489U88/fvcRAj5+Kmsl7cQ3D4PwH5zj5dgtXuzxo5SGE4gQ2VeGhhzazefPKvNedDKVVcZrOCsRCBgmtGGQdBBNHQYDi0gL02FJ83RNfM1sodoNYWEJLpBSETCNF3pUUM6c6cCMM66a8rxmoTOWTpJ2xMtV0buHjfi7JMuXV9RSEuqn0l2S4GLphMBKIkHR2cWnQS9eP/Mjy7L5ySU0k5DYR67qQJIFCrxtRSH2t7U6VQ//8B5zbX8D3v5b7uuz2s/Sfz2+cKMl79qCXcEDm+Y25bVqqzaS8ZqIM7ny916YhYveOLcDiUQlRtjATH71qyy3AbRsXYPLYoKpzV30bn6xYFoRCFoHANS417yLqHUBwhVFkOwXuySuLQX+Qd174gMtnDbbscKN4QpiGDZvNRnVdKYoQAXHyWGVZFj2XOzj9k5M0tziJeAwY9ZBYcfdSdj22BdU2dcXa1A3a3jtP4zvttPV40Mv9CCUJHIUBrGQDybBBR7eFbjhB1rAV+CgsK6D+LoXwUB394ygHs929Ve0GsbCMnhAyCzzDAISJCz5vVTxvG9SNwkzaiubSrpR57XSO4TfRTXw6TNeCNp3526WDxUQDck4FGWt0gZ9H/Wu+kE0QT8cF4LaMDbcqsegCHh53rB74tzzn1Y87JwwEbtytXR8qK5/M9M2KohvDCDE8PMi5cys5fEpBr/KPyvXZee6xrWxdu3xG48pqCYYZAWms6mCZEeQsp2dTN1PbyYKAIFg4XSqb100+vqvyiVGOBRiGB8sMYekRmtpW4S3toqRi8h7Ovmt9HNh3mKYrKgFVHFP4WFLLs0/vxD1FEMsHf5ePpx/9Q5prVYbkBJQMI0rQsKiKp3ffi8fjAppnNeZkWLcz9fHpanHy/ZPzUw24GRjfyvSdb+4hFnJwtauWUGTsazIp92M6t/BJfm73VPD5f787c8iyLJoar/H6K4eJiEOMGAEwZ+kLIpqIkklRsYdnnt01oXf7p/+zYMbtZ/mczMMBGVehlneMeXvPLQus3MqjKJk5O1+WAYYl5LrIzwE/J4oit21cgImxwTTD6HqY+vqPz3nM7GRF1w26ugaIRkOE4x6i1b0IkkllZTGf2fsAhZ6Ju6GWZXHqvTOceLuJlo5C9Co/F3zVbF9yheKCQjyFZWCGQI9im0TtLRqMcGr/IRpPx+iLp2RdBdmkqLSAx/fuorxq+or1UFsfF/cdo7XRQTBLMap+eT0PfPNluj+4xtXOArSylJKirMqsvf9uVt7zNHt+ffaeSvmwcmeq/am3xcXXT74/7fk5C9NbjOxWpu/8zW8TCTiw2aK0di0gOBobppTIncYx/HrcxOcbs2lBy9cyFA3IOAv1nDGqlkRm/L7PFdlVkUxcgNsyNtyqxOJF4H8IgrDGsqzzgiBsAVYAbwuCcBD4nGVZzcD3gF8VBOGHlmUlgd8EXrAs65buPE2FdK9sikzXj2UV0ty8gnPn1qJ7A0hOjbsW1/P5vQ/htM/c3LCo8nEGO743Ku3pwjIjmHqYkvo9GIbJmUPnOfjWZVo6yqjf9D6rFzbj9SQZvHAlr9JTRg1Ki0J8CBQ7iqOeYesB/vq//SP/6S++gCxPLJvHo3GOvnyE88f9dAbsWBUDCIqBp8DFs7t3sHjR7DSqtXiS068d4+LhQTqGHJgVKZK5y+3gY8/sYMkMFKPGY/xCM90qM1WLzI3AfJO+x7cyDfQ3UF/fwjrvG7T5PpU5bzLux3Ru4TN1E8+WrX3nzWM0NXbOsrEvpQi2dv1Stu1Ym1F5mivyPct8VYz5xMhAgJP7DtLe6sBnGggVgwiCSHl9iPpVY0aOZ94swzEP7RU/J4oit21cgImxQVVLqa//+HVx77KTlWBQJ5kMggAX/LWodpHdj21j64a7EAWBiO8Qob5XMZLDSKoXT+WTaNZKrpy4Qm9PMbo3wMLaNjYv7qPcq4Lhg1gYHPXY8ph/moZJ88ELnH+rmdZOD1pZGMEbR1Fltt6/nnX3rJzWiDQRjnHlJ8doOTFCd5ZilKvIxZbtqwgebuX0OzrD8qiyoAyVi6rY+cx2nHPgds2lTWa+cSMWgtmtTEF/KeXV/ShymPXeN2j0/SIw9VwxnTzs9UjH3krke575qhg3A1WLopnrpuMCcFvGhluSWFiWFRQE4Tng74WU5bROSuXDCTQA6dT4u8AS4JggCDpwiRkQ9GKxLk6d+sJNkfPLh+wAEon0U1bWR01NISPREhRV5qn7Ns4qqYAxgnZKFWoQWS2hpH4PSX0lL31rP5cvyPgliQXrjrJ1ySVEyU1J2SJMMzZB6SlbDUp2LcpUKgz7/Xzl+R+x+5ceZM9ncjcOf+WpdXRdFYiMxEgkN2DJKbdjuyfI7/3Ny9x/34a8ichksCyLa+eucvLl07S0eYgWagi1wxz64e8jmVW4PC6Ofj/rmc5iMT7+vOtdZObbEU8fnwrzTvrO08qk6S4cdt+MXj6VW3gqCdqOx/Uc5SXncNiDxOKF6OLCUZ5HLtItWTsq/Oyom2VL1kcYpmly+e3TXH6vndauAvTKfgR7EtVuY8sTGznxY4WpHOLvYHLc7nEBxicXPvr6Xs05fj3jGUYHuq5w9MRWOmwKZVUutm9cCYwpPQmyG0EtwzDDDHd8F9mzB8u0sCyJBdXX2LboMoXeCnBVZTYWbHm+2/7OAU7uO0zzZZWXX/8SiYQLBFBtKp5CJ6f+TaakKsafvXwi731blkXX8WYuv36Rq9c8JIpjKZERRWLVllUUBmO0/qiFzlFFw4P/9h9JxkpwF7pQVIVX/0dqnNkuxqdrk5kN5kqgvRELwXytTJruxDXD2DCZPOwffu5LDPfaKXT9IpUl53Hag0TjhSTEBXzlx/6J93Ed7Vh3cPvglrGULMt6B9ic50e1WecYwFdH/5vN6PMq5zdbZCt1SFIpktTB+vUHiTbfTZ85dwJhQenOCfKyL37vdXo6NfwxJ8KCPu6ub8frraCgKH0dGzopBah0YpFWg5KziOAjUZPf/uS3WLVhC1/9yy/mXCMwEKDlfBK7zYfsNJGLNBAt7HYVIbmAhx/K9zZOjrA/xNF9B2k8qzOg27CqexBki/KKYlz2BSxYngRyF+TXo8A018QgjekSmskqE11XHfO7a56nVUmRIyRmITs4GY9jLAkqJcGDKe8MNf3cc1uHplOXup0xdG2AjhPNdFyrQq/qQXRq1C2vY8fu7ah5SObpvu14SOLEK2PffVG2+MrGXbdbK9N143aOC3BjJGeLi7dz5YqHd9+9QHNHAclSH4IYw+kc27wK9b2KkEXyRiokFkvSf+Vfabv6LCPOILsWXSKuK3jtRSmSt+TFABKj4g4AiWicc68eo/HoUEbGPJF0UVjhp6K6BIc7bT+SpGeSOTvUN8yFfYe5ekHChwy1PQiSRXltKWuW19P7bivH29wZRUNJFhGsSlbcYyGKSbIT9+tZjF+vss5Mvrf5qhPdjS6G+2yZ1qv5QL5WJkWenSRtPh7HWBJUwgj3MzIaF1LPLbdtaCplqTvJRS4ycSEigUkmNtwuceHDJX8wTxAECUmS5k3qdbbIVurQ9SSG4SCRMFjd0Ehf2/wqk+iajmWJCBKIskBZoYnLk9vTKogezCyiVVoNKo1E3OD3f6ubkhKLP/0/v51R7TAMg9NvneLMe21EotsRXEmQTCRJoqLSS1GBm47WmX+EDN3g4ntnOP9OO1e7PeiVgRTfxGbjwUc3sW79ct7+djowzR/mw2tiqramySoT1xrntxQ5vlVJlcKocoSOwM1dHE2nLnU7w9ANLMvCQkKQQHUqbH5sUyapyF6s9F51YmpjDbSyLdWpo9qNnH7uO7g5uNVxIX3N+ZSc7e3188ILb3PhooRPlFLcBAkWLKzil3c/kDnPyJrzTdPE3zvEsD+KrCYYKhlEUHUKXHHchfXYs2XBs8QdOs+2cvqV07S0FhL1ji36HS4HC5bUwDRtT0ZSp/lnp2g52JujGGVz2Nm0azXauS7O/aCLQUPNbDYVVxWza+9OTu5zIYrz274yHwu36dqaJnPzTsvczhdyWpmwUOQwNjlMa2DHvF5nKkylLPVhVI+62ciODfGolIoNJiDkxoY03+OjjNsyscjG9cr5zQX5ZAWTSRVX4fSeCTOFpukcefMEbc0ROnwurIpewMIQCrHM0DiSdwhRHdu5SKtBIRWhaSa//1uNKIrJ1/58eU47U1djF00nG2lrrwNFA9misMBFZVUJ0jRBZDz6W3s5tv8IzU0pcp5Q34UkCyxbsZDHP7Ydp/PmKWzA7PkPN8PLYjKM3Wtuq1JLy0pwLeHAW5vRsgJV2q/jhjmNT6cuNQNM9/yvt8qUfn1XqwMjy6ldUswb+mwma7U4+2YZ9tGe2nyyhndwc3Er4gLMv+TsK698QFubwWDMjtjQj8Op8Knd97FqWUOOrKukejGMECMhGX9vgEhUQXYkiZgSksNk0+aVlFe1A+NM5kbFHaKBMBffOEVneyFRVxihIIK3rJCnPvkAZ/a5QJy8MmtZFgOXO7j00hlaW3MVo5asWUKNonLtJ81c6ytALx9AsCdQ7TY2P7qBJeuW3HT/mtlwIG51f/vYve7KtCt1tdfiHyimeImNo29tyfHqmK1fwmwwrbLUDDHd87/eSlP69b2tTkx97LMlKje2WpAvNmTHBbh9YsPt8VtMgeuV85sL8skKqmqSSHx+FqHtzZ28uf8QTU0ugi4Tsa4bSRJZv2YJNcuXEuv9ATqpSkWaP+HKIuCm1aB04Oxpi6OHAhQUiHziqSYE8dOZ837rvzyf2p21UhOTLItUVZXmaIVn49ef2sDAuAnBMk1koYeHHthPd8iGVZ4i5xUWuXlmzy7q6ivn5ZnMFrciUTh70EsiJqKNLvzTmG6Rm3uvY61K7R0l6HKMSFBBsY0R091FOrVLojfud5lOXWoGmO75X++ifyop2uzrfFRwvcE0f7Beddc83NpHErciLsD8S85qmoZlSaMVa7hv62pWL18w4TzJdT9DLX9PMKgSjqvY3MPYVI2rQ1v41S/upqzMi+azSHT8U0rHN0u8wVa/l1hSx7JMTEtGUEwUVebBJ7fiLZlaLj0WjHB5/xH+6v/7PCPhZ2DUk0iSZZxOO2eEfnbu/A5BB6PVFoGGlQvZ+uQWbI7Z8RDnCzc7WbgeE83ce021K3UNFhMNytQkIkSDMsrobrizSJ+1X8JsMJ2y1Ewx3fO/3kV/+vWTcWw+atWCD2NsuC0TC8syMAxjXuT8Zguf7wDJ5BCxWDugEIm40A0R2a5x4dpyREHANhsH6nGIhKK8/dIB2lo9BO1RRO8IRUVufum5B6mrTrVZKYqcUnxKDiKqXlz1e3NUobK5FuvWDHPg1MYJylEDnQN88OODdHQUo5cOgGAhCEy5ezTQa6c+vYizIDQcItA/QndfBV1iDKF6CEWR2LZjHdvvXTutA+zNQNqVO41ISGKH+xEk2aR28dju3Vz4EpKc64sRDsioNnOCDOpcF7m1i2J8/+ShG66ANB4zVY+6Ls+Nm4j5Vu+6EbjeYJovWF87n5ho6HEb41bGBRgfG2zY7dUoin1O96JpOm+/fZzOziQ9AQ+UDwBgn8RM8vh7GiO+DZRVNuMpChDVHMglT/PUE5/IzOkZHkWWuIOtfi+6ejfn9h+ip9vBiBgHVxhBEFCmimOWSTIS5b2/fIOrXR5G4i7cFV0IokhRSQFiLExkWGfQX06woh9BNnB73dy7ZwfldbfWyBByXbnTiIckfqHocaoWj821c+FLKA6DaEDOLAajARnZZk2QQJ3rAnflzqGMdOrNVECaTlkqGx8FkveHUcY1Hz6MseG2TCxAmDc5v9kgTcwTRRe6Xk4y2Y8sJzBlN8c77qIv0sCehzdRWTpzQtV4JBNJLMvEMGREp4Vqk3numXszSQWkEofx8rLjMdk58Wico68e5cJRHx3DDqzKtPxfGEFbRGdrbjApz5MVJ+NJhrv9hEcEEgggJxE9MWrqytm95z4KizyT3tf1tsDMFmlX7jQiIYnCEo1oWMpZrM+FL1G7OJbjmXCzE4DZYDbPXSndxn/94uPIZltGPWrAfzehSH1mMT4Vwft3PvfrXGt00due69at2A1KK2++otKtbHO7g5uJWxMXIJe0bbcvJJHoIR5vQxAaqK//7Kzupbm5gxdfPMiVJicBh4UwajK5ftVitkziW5SMJ+nrW8zRK2uQFnRyz7ZVPHjPlgnnZYs7mIZJy6GLnHvrNa52FJAclZOVVZl77l9HWWVqJ7qkKpZD1DY0jVggiiQO0RhIeRwhWNjdDjwOG/HBKJGIjGkzQEkiO2DNzrWs2bEacRLp6evdlZ0tsl2504iHJCSJnEXYXPgSK3cM5Xgm3Cr505lgNs89rSz1/d+rIhGViMYL6fOvmeCjMRXJ+w8/9yW6G130t+fOvardwFt5c/dBbnWb20cZt2Vi4XDUsmHD/7rp100T88JhEZ9PJBRajFroQ7MUHCX38LV/98CU7qfTIRaNc+StU/T3yQTRRneOQL2OCkgalmXRcrqZI6+dpaXNQ6woiVA7jKyI3LNlNV/9ageK3DPtGIHeIYK+OJGEDLY4SCZCXGT3x+/nrpULp+2X/bDsEE+H7LYmSTEziUd2peNmO3dfD2b73K+1r6J2ycKMelRhFRQy1n41FcHb1/s7KDYzJ6EDiIVvfQXrRkFxGBm98mxn3xu1MLqDibhVcQFySduKUojTWYGmBRFFx6ySCr8/yIsvvsuVK16C7hFE7wiFRS4+s/cBFtZVTTjfNE0uHLrIQG+M3qGiTGXDNkllI42hzkFO7DtE8yWVYUVOeUhIULO4isee2Ykry0MiLSmbjCZofPU4LUf9dA47M5tSzgIn7gI3jmSAoD+ekuZwxkC0kG0qe359N67CqRdrH6Yd4qkwvq2p96qT7kYXomzlVDo+Kt/72T53s3QrR86MJkoqOKvASW771VQk7+He30WxWTm8A7h9uAfjkR0XYCw2fFQ+H5Ph9ny3bhB8vgMZ/fF8WuhpYl4yGQJSJtgJQ8HrTvDlvR+b83Uty+LS6Sbee/U0TW0e4l4NoaYHWZHYsWU1tdVl0w8yBQKDAQ68cIAr561RRY5uRNmi8v9n772C4zqzPM/fd016eE8YekoiRZESJYoSKVJSSaKoonyZLtM9bWZ6urdntD07D7uzETMb9bCxGzGxsxsduzExHTHdE6Xd7q7oLlVJKqlUMiWpRIoiRStaeO9dAkh/zbcPiUwkgASQABKOyN+DBEqgOtkAACAASURBVCauy7zIe875zjn/s6WU77x6nJKSwnn3l1LSdaud8YG9RMZNLM0GdwgUQWFhHh69iL37dizrGtea+cqaDh7zTws0Egz1unjj9EMbJljKKllo8F4tuprd9MzIngDYy5yZlbraV1QRnfb6fAZ7o6Tgc0yRqW1IZSlN2+FwBNuWWJYD4bBwuXX++LvPUpvGBgx2DfLZL85w56bOiKYnFaNqd1Ty0KF70x7fiMT45tfnuX1uaNrAUk+em2+9+Bjbd9fO2kdKSc+VZm7/+gYtrXmECw1ETQ+qpnDfoXsojtp8OBZhNDa12KRqKsVVxYz15y0YVKxnFiprGu13JgONHHGy1eS9GiQCxVSMqOAnpw8v61mcsA2pdiHx+lzH3Uh2IRdYZEgm+uOJxrxUnLpBxFpeSUXTrVa++OArbt+uwdjageKKsWVLKT989UlKF3D6FyIWifHp339CS4OTflOiVA3icjo4deowBx7YnZEix2BbP5fePUck8hJqfhQ0C6fuoKqmFKfLQVfT+nuoJkp/ElO5E8z1dhcqa/IP6ggl7oymOqldzbMd1qVea7rXM/l9NpjZh5AoZdJdFgeP+WfvsECDt9NtEwrMfGArC17zYvshMvlsLFPBV2DM2mZseHmZwKU+7HMp+I3FYmzDcpq2I5Eo585dZ3DQQVCJxp10IdD12WY8MBbgo599QnNDASOOIKJ0FLfXxYsvH2P37ro5z3HpF2dp/WaEtr5CxPYOVF3wwCP38fjTD6GlOQ/AwO1OGj+4RP2tGsy6DoQ7RklVMU+99gSDn92g+9IA0ZiGXhwGBfIK8yisKERRBGPrb52BoqoI3fXe5FTuJGlsw0JlTeODDpRJuzCQUt7T25ydMsuFSpVWo4RspsObKGVKldVOZaEm75mr+BBfyV/omhfreGfy2RhhdVb2RKKlPc9iWIpt2Eh2IRdYZEgm+uOVlS9MGhQbsHA6gzh1k/rh9HWvmRKcCGFbAstWUXSb/EI3//pPXs6KDF8sEsOMmRimD8UVRHMovPzyUfbetz3jY0QCYWxb4nJNEPCXomgCT2kRg13x61uPJUFzqQZ9/XHmykapSAmaFv+/J6XEJ9U5XcjJnc9pTg1qZrIaGZGZfQg9bW7cPmvO8qWFGrwPHJ0tvdzV5FnwvSy2H2Kxn41/yIE1efukTVK9az01ci+F9EbUuTayO3cZi7MN8UzFYhrIpZRcvdrA++9for7FGy9TrR1F1VSeeHgfFaWF07YFiAQj2JaNYeiIAguHS+O733t6QRW+aCCEaWrgkCg63H9oF8dPzu7FSCU2EULaEtvSEA4Lb6GbV/75txFC0D0exrR0nJ5RAqNlOD0uorqX/pb4vuux5ON/+dWFtH0PVz9eQmWAjA89k1JMc1AnUuxCJg7ufE5zIqhJx2qsZM90eAfaPLh85pzlSws1ee89OjsY6W3yLlIha/q+6VjMZzM+pGNb8WoEaceDp7thkB2sjG3IBRYZkkkqO2FEbtz4e1wuP4FAMVeb9xHWty75vBNjARqvtzI45EXmj4Gw0TQ1K0GFbdk0XWlgfAzGowKKAwgETmfmK7WRQJi2q82MDHt4/Pn/DDV95BU6+LO/+F7albT1xkxn37LjwcDM0qdsBEfZdppXk5nlQqEJlXBIRcr02ycaQGMpCjOOutfWpSqUqtvJAMk0QEzGSkKQvB/r4R4sh3TG71XHzdtrcCl3HYuxDfFyqcU1kF++fIdf//orbt6uxtzajnAZVFWW8AevPUVFihBIIqiwTIv6Kw2MjWoEpQXO0IK9eFJK+uo7GRsMMTxaCIUjSCQO1/y9GJHxIH3X2xkczEMW+EHYKKqCEAJ/cy+B/gmGhgs5+vr/hVIyzj0P7+bR5x9d8D2vNemcLdsG5GxHdbnBUSaO6Xpdrf7J6cN013unZWPCEyqxiILustPuk2jyjqtCDWI7ionWvbTuVKESmSuJhmUoSbug6hLdKe+KQXawMrZh/Xt+64RMU9lxQ+Hm0qXr3LxTC3uaKF/CqAbLsrh89jpf/vYOTR15GGUBRHEEp1Pn2ScOLfPdTMrJ/uJLGm7rjGoa1HWiaLBzdy11GcyWkFLS/HU9lz+4QVN7PtHiMKJmBM2hcuTxBzZEUAEbp1l8rZlZLhSLqNh2fPVmrgAsVWEmldVW/lqImh3hZADx2dtT6mpSwoXJDNZy+y1y3L0sxjYsRYlqdHQcy1KwUVAckuJSH//2X7yKMmNxSQhBZ2Mnv3v7PA31HsY8ErG1A1UV7Nu3m/KK9PMEwuNBLr99jvrLQXpDXmRVN0KzKSwp4P4H96Tdx7ZsOs7eov63TTR35GGU+RHFETSnxgOH7+P2331O68UJegJeZGUPQrfwFfq495H0vR3rjY2+Cr1ajPa60GY0W8ciKpYhEEJMc7xTAzC79Ejaadyrrf41H6mZq6/envKJbFsQnlC59nFZMrudYzobw/tbB2Sayg4EQjQ1dTEy6kF6JxBCzptdGB86g7/vA8zYMJqjhMLK58kvPcZn757l2tddNHaUIre3o2iSXbu28P2XniRvGcpSkVCECx+c58ZXw3SMurDL4815eXkeXnrpKLt3zW7Om4m/b5QLvzhDww2VIXRkdbzZe0tNGS+/9iSF88jJ3i2kc45NA9bBaA5gaobEm//5Rbp7diTlYBMst7SnsDQuCzs2rM9bppWOdR3M2aBNJoEtpsraUkvafvrT/4GhwTKkw0BoNmffLEBV1bsiLZ5j8SynzGk+Eg3hmtbB/fcLgpEjtKOj6/qsoAKg5UYLv3v7S27d2oJR14niilFQnMcrrxynprYizRkgGopw5m8/pKXBwYBUUKr60B0aj544yIOP7UNRZsu/+jsHufGL87TcdjCqKUnFqKqdVdxbU0HPrxtpac+LLzZtGUHTVe4/up/9T+xfF7OLVpJ0jrFlCBQ1/er9WqAMfbWgJOxyyC+NEglolG8LzVumlY51+/y0QXXKlH/GS9smZvThbaQG65UkF1hkyEKpbNu2OX/+Bh9+eJ2G9jxipUFE9RC6Q+PkkYNpjzk+dIbBjjdRNB+KowzLDjLY8SYAI0N+YjEHOG1UXfLA/dv5watPL/n6pZQ0XW3kq/cTcrJRRPUImqZw+PA+nn7q0IJZBjNmcu3Di9w820PbgAerYhDhjOF2uzj5wqPs3bczKyVaG4FU5/iN0w/RdN2HtMG0Yah3snxAAZd79ZY0Ej0aed4OdtSUE43+GTdvHKKiopvHHv4HOoZOEojuBjZ+ac98LLfB25pcfFOUqZ4LacNR37NIKTENBZCoWhTVGWPPw+Noupa1tPh6WrXLsTDLKXOai9SGcChA00Z49PAX0LWbCHvT7jM66MeyBCYqisMir8jNn/93r6POMRsCIDIewojFiEYLUApH0Fwqz7/2BDv2zG7wNiIxGt7/msavRuhMLkoZuPM8PHJsP6GLrVw728eQUKGmG6FKyurKOP7qMXwFviV/FhuJVOfxJ6cP0349D2mDZSuM9k6uWCjgWEW7kOrsFng72FnTwc1vHsGTH+DBh8+zY/tNmodOMhHddVeU9szHchq8bXPKtxFKvO/CMhS+X/h88ndmTAEBbp+F7raS/SLZ+Fw3kl3IBRaLYL5U9jvv/I6vv+6ksaMEub0DRZNs31rJn7zyNMUF6Vfw/X0foGg+dH2y+U4txAD6299l3H+Q8Qkf0jMBAvK8y1MXuvzxJb4508Sd5grk9jaEblFZWcLrr5+gLBM52dsdXHz3Ek1NHgI+C1Hbg6oJ9h+4h2eeO7ygLvpyWc/TkYd6XdRs76TQ243TFSAa8TEwtJ3RkXIOpmlSXtHr2BVie+lHqGoE07RxNUcYmygmZnopy7uYDCwWS2ofwszXV5KZ972r2U17vRdVt6nZMTUZPbWMajkN3kd9z1JQMlXyNdzvQNPBiEJBiYEZieEf9QACzRXBirlIKxezDDbTytbdwlLLnOYi0RCuKD5CoQHCYS+2iLJvSwuXRmcHFtFIjP72AcbHXOAOIhWJw6nPG1RIKRlo6SUUgJABOKIIIfCksTVm1ODSf/2QjnqFnoiOUh2Xk7330L2UGpLOtxpoH/RiVQwgnDGcHhePfvsRtt23bcUXm9brKvFor4v88hhOZZSy0lbcrgDhiI/urj1sP7x6gUVqf8Y9pR+hqFFUDaJhL4YZD/i25F2kPrprScd3uKxZjdqZqDgth3T3vLfZQ2+Lh6od05/9qdex1Abv7/pOkVcyXTnQ3x8PFFUVvJNlwv4BJ0hw+cxZylbLZSPZhVxgkSX6+oaJRJzgslB1yeEHdvIHLz0970PVjA2jpOg527aNfyhKJNTN5dtPTD6kDTweNw8+sDSHMMFwzxDRqA4OE9Uh2b5zCz/6wcm0KfVUQmNBLvzyS+5cCdMXcSIr+xCaTXFpAS+/doKqqszlEpfDajc2LyaQyfN28NyJfyBmejFML7oWxKG9z7mLv8df/aor43Nmq/fA6RwlHCmd5u4aphe3a2hRx0kltQ8hlZXOfMy876nN1JmWYCVmjIQmVI76nk2+nhhmmHpPZwZQ0gZzMoMRCYSxzISjNpkWVwWadneXd+RYfWKxIcJhN3193UyEdGzdRKg2Be4wx3fcn9xOSknj1UbOvH+VphYf4aJovNdNV3nk4X1zHn980M+lt76k4bpkyPYgt8bLbStqKihN04sRHQthhCKEwqUo+aOoToWDB3cTvdbHxSYfwXwLUduNqinsPLiHR547hJ6Fwa2ZsNqNzYsJZB59+gI7S39D1PRhmB50LYR/6Ap/+Fc92GTerJyt1WqXc5RgZLrNNkwP3mXYhnSSspmoOC2HdPc80UydaflVYkZFOKCCjAcPQHKYYer9VHQ5K3iSNtleU7pryAUWK4AQUFyQt+BKjeYowbKDoBZi2TY9rb1Eo0EmInnYtd2oquDB/bt56dRjuJaREQiOBwn4gwQCTnCFAUlBvnfBoGKsf5TfvfkRjXeKGfPFEFXjOBwax588xCOPpq+/vVtYTCBTXvINMdOLafoQgDm5ClRe8g2QvmEyHdnKvESjRehaMHkdALoWJBotmmevKdIFVV3Nbrpa3NOyBLA+pYRnEg0reHwWoQl1WjYiFFCp2RWadk9nBlBff1yCwxljdMiNEVNBnWpSdLqdGBEtZ1xyZJ3hYUEgMMh4IB/pDSGEJN8jqajcwdYHpxqgv3zvHN+c64j34m1rR+g25ZXFvPb6k2kHm1qGya3fXuXm5x209XsxywcRrihOl5MTzz/CvQ/MLmeVUjLWPUQ0LIkYCjhiYEsGv2qns2UL9tZ2hMOioLSAJ147Sknl0iS7NwqLCWS25F2cDCriz2LD9BGNetD7PkrbvDwX2XLSI9EidG26XdO1EJEMbMNSswTrlcSMikhQjc9XmbQNkYA2S/Gpakcorfxw7tGfnlxgsYYUVj7PYMebGIARc2DLAKpqc7N7J7pD5aVTRzj84H1LPr5t29w8e5NLn9ymqTOfWMkQwhvB4XSw//6dC+4/2jeCETMIhtyIqh48Pgd/+Ccvb4jm7MWWTqUbANfT5sbpttPOXEjF7RrDML1pMgRjLCawyBaDEw+zo+rnOPUmqrcUYNsqHnc/Lb2ZNZOmC6oSDvhiG7VXi8T962pxEwmqNF2fXtMdmohnFfxDOoWls4fhJZiZNYpFFSxDBWQ8gJiMpYWioGoKcx8pR46l09OzjeLiyzhdQaLCprRAo6zYQVHti9O26+/sJxxxIl0GikOyd992Xn71ybSLWv2NXVx6+wJN9V7GvDaithtFFezZt5MnX3gUZxpp2ch4kNtvn6fxcoi+UD6yNh68uBQVGXZjO0wUp82WnVV86wff2hA9dostnUo3AG6gzTOthn4u0mUITMsTnz69BvRMPMx9VW9RUdGEqphUFl8lZuRxu/e1BffNRpZgtUm9d70tHmJBlfbrU/5LeGIq2zw+5CB/UpRkJukyRmZU4Ck0Z03mzpELLNaU/MmaXH/fB0RjvVimxteXD9Nu5OEojFFemtkKczosy+LjNz+k6UaI9uF8qIsrd2zbsYXXXn6CvLz508RSSsb6R4hGNCzFBEWiqAo+3/InSa8Gqc7x1TOFGJH4l7+93pt26NlMZ7p3cgDczOnQ6QhHCijV2mdlCMKRqfklq94jIgEpcLmDDA7UMNC3he62IiaCnuR57yYS96+3zY1QUtLUM+ZsWPb8js/Me/HDQ4/jc3dy80rKwEjFRtoK4YCGGZ2SVNwIq3Q51j9jYwEGBmrp6HBQt/s6ea4gmruKorrv4S19PLldcDxIJBQjFPaBM4IQUFJakNa5v/nJZW5/3kRDaxl2XQeKwyK/OI/nXz1GZU35rO1t26bj7G3qP2mgpTOfWEkQsSWCrms88Pg+5JUOukIOcEVAQH5J/oYIKmC6g3zrTDGxSduQGHoG04OMmQ51/+QAuExq6BMZAiPFNmhqKDllGla/R0RKicc9zshIFQMD1USjPvrbihgLeu+6Z1jqvetv80zZhgQpNsKexzakuw8JKdprKYMTFUViGWLT24ZcYLHG6J6HuXkrxpWvh+j0u5AVAwhPiKKiQirKlr7aHRoLMTY0xth4MaJwAtUhePLJB3ni6IEF9x3rH+X8L87ScF0whBu5owVFg7ptlRtSLtCIqLgnZUMlmQ09CwZVQiEVaU3NMoD08wxMZTujQ6NEox5ilgeHGsLpVDCV7cAYsDo9IonV9p11nbQEDxCz4sZs+55u/uV//yYoP8d7/0+ydr71iqrG1blmBhVLobQqQuedIqSEaMQNio0QAt1rUbEttOYNojnuHizL4uzZa3zyyW0aO/MxymJcrz+M2+Pkz350Cm9p3IGxbZvrZ29w4ZPbNHXkEysbQngi6A4HO3bUpD32QFMPoZAL22mguGy27t7C6e99K20561jXENff+orWWw5GNA1q4otSldsr2bu1iu7PmiflZEcQvjCarlGzJ/151zuxiIp7cgZDBC3phM7XnxEJqkQnbUOqUzlzpkFRVYQr159lZ83viEY9mJYHTQ1RWdWJUTnV67XSPSKpq+35dV00Bw9i4qV2dw8//De/BWMMW/knovf/+6ycbz2jqDaWnfI3vwwbkfhcLWtqkroQEofX3vS2YdMHFgmt8Pj01FIqK1/IqrrHfAz2DvPu333Indv5jPssRHUvmq7yxJEHePbEoWU1hYYmgpiGjWGqoBkoimDLAo3WpmFy/ePLXP+ii7b+KYUPt9vFc6ceZd/9aycnuxpD1VIzG9hTz5xYRE3ObRjqdSQzHgnyvB3U7hvlX/34fwQkuKtx1P0AvXQsmalor/fSmzK5WndZHDzmz9q1w9Rqe/DyfwBHBUpKEGhb+fEJ2GlYThlYJiwnW7OU+15YGkuqOVlm/D4KmHNK+Hz81a8uc/mXZ2m51ENDWwViVysVO0s4+eOTabdfrwo1ORbPatqGaDTGm2++z82bMboCnuRsiB27tvD7Lz1F/uTsIiNm8MGbH9J4M0bnuC+5Xe2OKl55+XjaTHQ0FCEWjhGLukA3QEBZRUnaoKLjy9vUf3yLhsYyzJouhCuG2+fhkSf2E77UytWf9TKMjqzuQWiSstoynnj1KHlrWB67GjKcqZmNmbYhvzQKwGivM5nxgLi0a2XJdcrKOnj9+3+N1F3Y7lqMymexS48knxXd9V76UyZXO1xW2obopZL63HFf/g/gKJs+cEnxocQGZ+23nBKwTFjOs3Kp9zy/1MDf70SI+D1UNYkZFWhOiWUuzrdZzPN8s9mFTR1YpGqFOxwV2HYgOehoNYKLjuYuwkGDQNCLqO7Fm+fgL/7oZUqKCxbeeQ5Mw+TKJ5e4+rtOWvuLMCu6Ea4YHq+X8vL0pVVSSrrvdPL1OxdT5GTjzeP79u/muVNHVlxOdiFWQ1I2NbMRTKm9NA2mKQWlZh58zkbqSn9DR/tOKDgA9jiYU79PZCp6JkurEsyUbl2M873gto6S+HWoKffbHo+/nobllIHNReo1ttd7cTjj+efUgCqTbE027ruqSExDxKMLe2rYnarZdDV5shqcrrZCTY6VYbVtQ1/fMMPDfkZGqhAVfeguhVdOHuGxh6ZLy44OjDI27GdktBJROoDmFDz97CM88shsBSgpJW1XGrn26+s0t+YRLhxFlATRNI3qrekH5vXfbCc44cR0mCgek/KaMu4pKaHzFw20p8wucrkn5WT3rryc7EKshmM2LbORYhssg2lqQYnvfp6zKakG1db+ENJXg2IGiU4GFTD1rBiYLK1KMFN9KFOnNJPtbEcxih2AlCnx2IFppVkJllMCNhep19hd70WbHDqXGkxl8qzMxj1XlKlgwozG/z8xrKNokt6m7JaFbTa7sKkDi4RWuK7Hv2Tq5Jetr+/9RRmPWMwgFjMwDA1Uc3HZNRH/j0DgcjuWFVR0NnTy5TvnaZxszlMmg4P9+3dx6tSRtMpSUkou/PIM9Rf6aekpRNZ1IzQrLif76nGqtpSlOdPdj1BIrni7PBaPPDMMwNn3pn8eZXkXiZleYpYvniFQi7CBWN8H6KWPZXy+xZRKLbSto/J5Yh3/LzaAkp8Mdhx1CzfoQdz5DwdUjKgy7fyLccBTrzE1qEo3C2OlKSw1CAVUDj8zvK4b0HOsH7JlGxaLEPEAWNUUttdWzbMdoICqqdTVVs76vWVanP+Hz2i+Nk7HUB6ythuhSUqrijn12nGKSmbbGcswMaMGhqGCHnd05cA41887GdYForYHRRPsPLCLw889gu5cHTnZ9YZQQNUltilweiwOPBNf7b/43lSvynQ1KAF6ATYsWg0KMndKM9nOqHwWd8ffT9oGH9iBeMBT99KC15GYV5HaOwCLywylXmNqQDUzmFoNEo3akYDGgWcG13UT+kZjUwcW8RT39JUbRfERm6NkZCZSSu7caeWdd85zp8lNwBdAVI6jqSrbqtOvCK0EwfEgX77zJTcvBegNuZCVvSiaTVFxPq+/dpya6tnNeQnC4yEGWvoYHS2EwglUp82BB/dw8tTjG1pONrWEJhZVksGe051+oFtpVYSm675pmQojnuGe1xluvL2d4ZFKImEP/+u/+9eTr9rUVDfyZ//Pct/F0kgENLG+D+LlT44SHHWvoZc+ljbb0V7vZbjfmSx7Ss0orIQTnig5i0WVaWVly2lkT9xvazIjIe2p+5eYhH72vTJU3eaN0w+t+VDFHOub5dqGtWZi0M9o9xAjw2VQMozqFDxydD+HTxxMm2EYbOzm1tsXaan3MuadQKkcRAiBM2YzMOZD2d2Hw6vz3I+fpaRqY8vJppbRmFFBZNIN0ueYhj3U7SQyoU7LVCRWuOfC5Rzl/LmniIZ9RMIu/tO/ewOQeNwjtE0cXrPyF7v0CGHiAY4SG8R2FBOte2laaVaCmaVPqRmFlXLAb50pJjSmTSspg6WXDE2bnG3HMxJyspQtOQmdeFCo6JKfnF67e3M3sakDC4ejFNsOJFejAGw7gMOx8NA3y7J5661PuHRpkJb+PGRNN0KzKS8t4I9f/RZbM1jpj0Zi2LYCytI7TdtvtXH27a+ory8hXBZGVA3HZ02ceJDHH9uPukBwYEQNbNPGsgSoNqqqsOeerRs6qIDpJTSzpjc3zVZG+qtfXZ42dTlRow9TDmo6Rv3lFBcPMjJazpYtfQBoWoDOzh28cfqhZG9FOKAmJU8RoOsy62U4qeilj6XNmKTLdvS2uYmGV+9+J0rOUhvpYXmN7OkChR8eenxZDfOJv5vw2ANEQybRmAbOKPnlQU7+uHXJ15pj/bMc27AUDMPEtiW2BMTctsCImViWjKubibmn3ptRA2nbWFIB1UbTFbbfUzt7RoVtc/OX52i9MEhrXz6ytgdFs8grzuO+2goC3/TE7ZMAp8ex4YMKYN7yoXQqPqXVUSxjqhTK3+9MZizmIhItwrJUvHkBEJLyLf3oWgDbcjLa4OInpw8neysiAXVK9lSApttZL8VJxS49kjZrshKlT4slFlHRnXJW5mWpJUNzBQkJRaeZZHqezdYzsVg2dWBRWflCsm5WUXzYdgDTDFBXt7Def3//MK2tPQwMlEPJEJoTnjy8n9eeObKgUx4Khvns3S+5cnGUrrE85PZmFM2irHT2QKOFuH3+NmN+nbAiEb4wJaX5/MHvn6Igf2E52dbLjVx6/zpNbUVEioYQeSE0h4OiovxFX8d6Zimr06oa761IkHBGVX26MR8Y2k5ZeTe6FkUiJ6duBxkYfoCJoAvdaeP2WbP6K6q2hfm7S1/yxumH+OGhx5MN0wmy1TidY+kkgjB/9yghf5iJkAu8QUJj1Wt9aTlWmOXYhsVgWRa/+90Vfvvbepq7SzCqOhHuKG63j4I8z7Ttvjlzna9/20BTZwlGRRfCHcXp9pBfMPWsty2bpi9v8s0n9bR0FBMr70F4Ijicbnz5vlnnDw6OMdrcy9BgKbJkBNVps/uBnZQGTdrOjNA9lo/c3oLQLPJL7y67AEur1VdUG8tQJrOiU2VBij4VEPZMPIxTj6DrUcCDrgVwagGa/UeBuBOvOyUun5kMWCBeEpRQFBrtdfGXh44nMwdA1hqncyyPzdYzsVg2dWCRqJWNK3/043CUUlf3ekY1tDIpMSNAAd2hcnj/7gWDipY77Xz8y7PcqS8iVBJBVA+j6yonHj/I08cfXNT1SykxYya2rSAU4ioi26sWDCoiwQjnfvYZjd8Y9IRcyOpuFE1SuaWUl187QfEy+jw2MqpuJ8uenK6pQMCySZYEvXH6oWkr3v4RL7duHaOu5gZu1xDRaBEd/mNMBOvix3Hbs5qgjaiSzFSkzl9IDT6W0zi9WHSXRXBMn7WSv9xsSqLkaVq2ZhLJ3GVp6w4pkRnOWF0NhZocK89ybEOmDA35+dnPPuTGDQdDiga1XSiqYPeeWn704gncrnipxsToBB/9/SfcuakxyNR2W3dt4aUXn8A1Ebzg0wAAIABJREFUuV3IH+D8339K402VAdsBtT0oKlTvrOLkS8fweGevsFqmhW1JpFQQKiiqwLzRzeXGckIlUUT1CJqusvfx+zlw/IGsvfeNRqK/AEB3ScDCjAruf3I4GZz85PTh5He/lwP4Ow2qaxooKerDtpw0+48yEd2VPKbutmZlA8yoSAYVqfMXEr0Iq5k9yEZPRToSClvTMjXA5Q/LsI340Lm7jc1mFzZ1YAFxA7Ja8rIAF89cY3hYJ6RIRF6IkpI8/vmPX6B4kXJ9Y0N+vvj5Ge5clwxaCmzpRAgoLV04KOhr7GK0b5i+/u2wswmHW+HkC4+x/4Hda67wsZbU7AgvWDqTboBazS4B7OdO9/5Z+6bLOnQ1eWYdJ9EwnSDROJ3Ouc+29O7BY/6s91OUVkVor/cmMzYJnG6bgF/j8GQz/HpG2jYjHQNMjJpETBU8wWRj7Vzk0uB3DyttGy5fvsPgYJjh8ULEzn5cXp0//M4z7Jkxi6L+cgOjg0GGRqsRu1pweHRe/85TbN8Rz5wlntkdV5vxDwQYHK5F7G7C4VE5+dpxtu+unXVuKSU9V1q4/evrtLQWEy4eQHiDCAvC405CIm6f8kp8PPvjZ9ZUTnY9kE7+tbfJO+37PvO7/5eHjpO/axsm26gfOjz7mGkyD4ljpvYYpAY18w1dy7bzuhI9FUVVEbrrvehOOc0u6G6Loop4zXG6TMBGZ7PZhU0fWKw2Ml5EixACRYX9921bVFAhpeTaZ1e5/GkjTR2FmFV9CFcUl8vJyZOPcPDAngWPYZkWUk52dSjg8ujcv3/Xpg4qZpI60yK1yXilJmWnzrS4eqaQWDTuvA71upLnHuh2Ul49u+Ejm9eUrQnhf/Wry3P2OMxU1lqPdFxtYnxwL1EHWE4LXDFQFApK8wmPzd2jsdza21zt7ubBtuPj4YVQQIHikrxZQQWAZVjYUoAiEAoUFHvZsTPNdqYJxI8lFIGv0JM2qIgGwtz4xy9ovmrSF3Eht8QVo0qqitkSsRn1q6BKhAp199Zu+qAildR5FmZUpJ3UnW1Sg5rLH049OxOlUr0tHpBQtXP6szab15St51IiaNpsZUSbzS7kAosNRmA0QMOlevp6izALx1E8Bjt3VPP6608lU+dzYVsWd85c5+onzbR0lWHXtSI0E483b9MGFamOdFezm/b6+MPNiAk8efEVFV+hOee07vmyB+kc9EwwIiq6057llLfXe6nZlT4DkikLZTsylb1dTgCi6va817DYY8+1/UB3+u/DQpkd27Ko//wbTOMVVJeN0C10p4Py2jI0XSU8Nvd3Zbm1t7na3RwJjJjB1x9d4sqZTloHCpF1rQjFIq9g+t+CaZjc+vgKt77opLW/ELm1DRQLb37650Lf1Rb8naP0DdTBrmZ0t8LDJw4g7/TS1mTSHwWqehCCTR1UJJy53mYP3ZN2wYwpIMDts/AUmvNO6p4vg5DOScwE2xCzng/9bR4Es1f6F/vMWOh6M30ureTgu8Uce75tV7s0abPZhVxgscpIOVlXPo/yx3zYth3PNAgFoUh0XeHo0QcWDCr8vSOc+/kXNNxwMKIr8TpdTbB9VzWnX3xi0wYWqY50qkN99r2yjMp15nOiZ/ZjJFgpJahMyFZmYzFzN2ZSsyM8b9nVYo+dbvtrZ4smywemZ3gyCXykLZFS4vH4GR8vhUgMb76Pwfb4d+xurYvNsbpM9enNtgU9LT18+ouzNNzxMe61obYbVVXYf2A3z558NLndYGsfF986R2O9hzGPTA423bN/JydOzS6/gXgjeDytIRAqOB0qo79ppLmjEKN8BFEcRXfoHHrmIHsOLZwBv1tJOHOpDt21j8uQkJxdMR/zOdGp/RiprOWzZb7rnSn/Oh/LcYIXCjwWc+y5tr38m7JZ2R1Y3Or/ZuuZWCy5wGKVCAZCfPrOWRrrLXoCOlTGV4TyF2i0zhbXf3uFwW6T0YgPUdmH2+vg5deeTJtSz5EdMnHi02UQYlEF3zIb2LJV0rQculrc9E4qXYUCKjLFf1rp0rJoWMGRJuuzmOzOiy/+Nb0xUGv7eez0EXYd2LXwTjlyLIBpWnz++SXOnm2juacIu6YDRbHw+qZU4c5/eIGBHpUxCUrROL4CD9/7/jNUVU2Xu7350SWGehXGLYFSNIY33823v/8UlWnkzm3Lpv3MLRo/baa5oxhZ2w6KBRGb3u5KjMIxFF+UsrpynvzOcdxe96xj5MgOS125V7SlLUimsh7KalJLyiIBNRlbf7/oeap2hFb0emxzdtYHFrf6vx7Lj9YTucBiFbj+9W2++M1VGpoLiZSGEFvCaLrG08ce5MhDexd/wCU8WyzDREoFVIGiCfbcuzUXVKwSi3Xy5+pNWAzLyShkDTn1pyolCHXq57lKy3LkuJvp6Ojj5z//jJu33Ix5BNR2omiC/fds57unjya3s0wbWybU/gSPPHLfrKAC4mV78f6L+Hb3H9qdNqgY7x3hxj99SfMtByOqBrXdKCqU15RSMhCgEwWhSTSHxsPfeigXVKwS8zn56RqmF5M5mIu1LqtJbeAGYNI2KIpEUZi3vCzHxmBNAgshxAng/5w8fwz4V1LKr9Jsdwj4CGhIefkfpZT/x6pc6Bx0dw/w1luf09Kaz4THj/AEUBUdT5pypNGhMS58fpn2tmIihWMoeWGqtpTwB999lqKC2bri8+Ef8PPFW1/Q0eZlTA9C3gRCUXG75y6DMmMm33x0iY76IO2DXmRFD2Dj9S2txnMzstxG7nXh5K8BNTunVLYufFyCZ1IFZL5J5gmunS2aNbTPiCq8cfohgGmBWmIIoe6ypjXB59h4bHTbsBDvvXeGzk6BX4JS7Mfnc/EHr3+LXdu2zLtf5pWq6TdsfP8iAx2SkaiGqOvF6XXw6ImDhC620tpRyLhzAuGdQCgaDrdjcW9qE5NYeU9t5IbMV9vX2slfC2Y2cF/7uCwppxteQE63t9mTnOmRimXNDtISAwgdLiutqleOlWPVAwshRCHwFnBaSnlOCPEk8LYQYruUcqb3VQr8XEr5L1b7OtMRixl88ME5vvyym9ZBD3ZlP8Jhkudz88PTT1CaZrCcaRiTNdsqQpM4nBqvnjq6qKDCMi0ufXyRa7/roKUvD6uyH+GM4XK7eOHUo1RWpJ+G2n2nk4vvXqCxwceEz4rX32qC/Q/s5rHHDyz5c9gMqNpUg3FwTEd3xntj5mvkzhZzNVi7fWbGTc8JZxtYlMOdbSnbpRINK8lAJEGYeEDR1exGSYlNjJjAstRVnf0xF8utvd3Mtbsb2TZkimmaSKklMxHPHDuwYFAxLxlmr6cy1qBogsoCH52/aKB9wItVMYBwxnB6XDx2+lEKSjbnHKNM0N0WIb+W/I6GxjR0p5zWyA0rFxikez7YNiBnn3OuhuelONzr5blkmwJvgTHr9YlhndFeF6P9TozwpGpXTMG2IDKhcuts8ZoOFtxsdmEtMhYngXop5TkAKeVnQohe4FvAuzO2LQWeE0Kcn/z3x8D/LqWcWLWrTeHixVtcvdpMa0ctcmczmlNy7NBeXnvmCA5dn7W9f3icz947R2enlwk9AJ4AQmg4HbO3nY/W6600XmqkpbUOe2cLqtPigf27ef7UEVzO9KtLkUCYK++dp73Zx7gzglI4QUGRj9e/+zSVaVLqm5W5HOmDx0aT2YhslCYthplZkETAkJCa7WpxYxkKqmZTszOclKTtanZz5GS84bwnZeBeIkNw7WwRAb+WzLgkSM28ZNrvsJIBSGIGRnjG6/rk0ELLVPClGJdYRMGyBVJOBXtGVElrgFaa5dbebvLa3Q1rG5bO9AyDETU4/+HX9HVZ9I65oKIfBDhd05/zZszk5keX6Os06fF7oWIABLgmt5NSIoTAMkyaPr7KQLtB96gPKvqRUhLtH6O/bQf2zlZUl8XOg3t45LlD6Iu0TXcz6Zy5ooooOw6OJb+nc0mnrhSpz4dEwJDoSQCmSc+mTu72FJpJx3pgcuBeanbg1tliQn5tVqlVIvOymOfSSjrBii7TZjUSU8+NsJrMfsQiCrYd/36lBoOpE9JXi81mF1YssBBCOIB0U1V+DTTPeK0Z2JFm27eAv5NSSiFEMfB/Az8FXk1zvj8F/hSgrq58GVc+N+FwNL7qI+JTrsvK8vi9U0/M2s4yLS58fplzn7bQ0u3DrPQjXFGcTiennztMRVnRos4bi0aRtkAiEKokv9DNK6/MX2tpxkyktLFsDUUPoztVTn378VxQMYNURzp1xT91fkRXi3tVA4uZzCyl6m1z4ysxCAXUaa8npHLnIhsNzQmW2nDtH3JgGvHSKIgHAD889Pis4CZdMHf1TCHt9V6MmGC4f8rRUlUoKY8xNqxPm5A+1OvK+jTxHMtnNW3DatiFbNF6q5Uv3v2axsY8AvmJDLPCwYN72L9/d3K7njsdXHnnIk2NXiZ8VlwxSlO478Au9j64J6k2NdjQxc23L9LSEFeWEpMTu7eUF5M3EKZPUVA0ia/Yy+Onj6zV2163pHPiYWp+BMQd+bUa6JaujCqd9Gx/mye5ij8XRlhFc8plS9bC0pzg8SEn0gbbipdGARhRwU9OH552vKodoVnXeOtMMaExje56L2ZMIRKMv1dFtSksjxEOaFRsCyV7VtajItfdxooFFlLKGDDraSWE+J8Ba8bLk5N9Zh0jnPLziBDi3wKdQgh36u8mf//XwF8DHDp0z6qEpHNJtN66Us+Vc7dpbqnB3tGK6rTYd+92Xv32UbyexTXF9bX3cetcPZ3dRdjlAwjFRtfnv22RYISrH16iv9fBmDTAN44QAkduNWpe5uqFWMhhXy4roeDkdNvJ0qDEFO/VXsVPzWqEJlRCE1PGLfGzosUbuecLbhL9Fqn7J1SbdSeYad7SUj83M2Zy88OLjPQLBoIaVPaDEGgLfOdyZM5q2oa1sAtLYXRglC/fO0dzQwmB/FFEwQSFJXl85ztPU1E5VeoaGBrj2q/O0dJQzIRvHFE4QX6xj1PfOUFFyqJReDRA/TvnaW8oYsIzEVeWKvJy+LF9DH/eQGtXMbKsD4SN5sj9bS/EXL0Q3StsG+Zr7l4qutsiEtCmTfE2ogJPwfKUCBdDalYjPDE96Amn2IaF5n3cOlvM+OD0bF7CNtizHxvAxlv934isxROlC3hmxmt1wD9lsK8KRJgpTr/OCAcj2LZATiptFJV4+fF3Z77l+YmEIpx/7zzXLwzR6XchK/oRukV+gZfTLxxLu4+UkuaLDVz54DpNrflESiKI6hE0XeXwo/uorlnfK3YblYUCg2wNpVsMB45ODdLravLwd5e+XNOSrsS5v/64JFmiBWTUF5HotwgHVRAgLeIVJFl0E6WU9N7u4Mq7l2hq8hDIs+KzXlSFXQd2Urtn9gTjHFnnrrcNcxENR7FtiWXpCKeN06Pz3e9+i/KK4mnbxSIxbFti2jrCaeFwa7zwnROUz8hEG+Eoth3PWAunhe5S2ZmXR/0/NNE97kJW9CF0C2+RlyOnctmKlSATWdeFyoZWork7URLV2+RNruKvZUnXXx46Tn+bB7dvemATWaCRG+KZFqGQrChcCduQY/GsRWDxNvCfhBD7pZTXhRCHgXuBj4QQJcA7wB9KKRuFEL8HfCCl9AshNOB/A96UySlzGwNFSR85z8e5d87SeG2Ejv5CxPYONIfgsUf38+STD6Fp6Z2x7tsdfPPRJerv1GDUdaC4Y1RVl/HyayeIBaL8s8f+DXcuN/HF+FuzanY3Mtla8e9qdtPTNjujZMTEsgKD1ZobcTeQLGFqmT4FPRRUwQZNl5iWSBoO04yvUIUDKqq+9MeCv3uIa786R+PtCoKlA4i8IHkleZx4/QlKqtKLI+TIOnetbbBtm7NnrzM4aNDv90LFIAjQ53iWAwhlekY8GozQ8MV1BvtdhJQIOMMIoczKpsVCEdq+uMnQgJOAEgVXCNuwGGoI0DVcgtgatyf7Ht/PA8f3o6prL3qwUmRrZsNcakRmTFnWtOrc6nnm/OT0YXpbPNOyRGYs7lupk30TVoptiNsFDYdrZhI0x0qz6oGFlHJMCPFd4G+EEJJ4qvuFSQNRC2wFErIUbuATIcSk7gGfA/9hta95IcaHzuDv+wAzNozmKAHuWfYxw4EwpqUhHKDo8OCDu3jmmUfm3ScSDCNtOdlXYeErcPPP/vg0V8/e5N///n/k8ecf5s7lpmVf23ojWyv+M5uCExhRZd5J0YvljdMPcfVMEZYZfygaMUHTdR8oUFgSW7Jkqqrb60LRCWYHewmVqmBQnZaxmEniXqbezwsflyRbXN0+C/+QA3uyHtftsTCiClXbwst6n9FgBGlJTCu+Euz0OTj5B8/iXaUBljnuTtswNHSGtrZfMDTUwchIHmHnboy6AIoKO3du4eC+nQseQ0pJ28UGriUy0cURRM0Iqq7y4JG9FE2qOEkp6bncxO1f36S5NY9IURRRPYqqq1S43Ai/E3QbocM9h/fw4FMHV/rtrznZWvGfS43IiIq08yZgaTMnfnL6MLfOFGOb8SeeGVNov54HCuSXxJYkm+pwWYTGtDlVo1abdCpVZkzBiAjyS+cu1R3tdfHQc9Mnnl/7uIzwhEp+aQyY3quhOWwqtk0N28uxeqxJcaWU8lNglpcspewEalL+/bfA367ipS2a8aEzDHa8iaL5UBxlRCNjKLH3UNTDyEIvQtgoSsYi5Egp6bjdxkh/kIGRAmThMCBxzqH+lCDoD9B2tYXBwXxkwSgoNqqqIIRg2z21/OON/8Ktiw388r/+ZpnvOMdyGep1oagkg5iRAQeqFu8VSMzLSGVmKZVlw9iwPk0SF+Dg0dF5syOrKSU7M9hrvuHDtOJZh9TGa7mElHXhpBEJBVQOPzOcLPXKJoK5e6hyrBx3k20YGjpDS8vf0N8fZmwsH8Ud5siDF3D3HOTYiR+xd/dWAGLRGHcu1jMy7CCkxMARRSCSme72Sw1888Flbt+uxtrehnCalNeU8vyrT1CYInHee6WF+l9f4c7taqytbQiXScmWEg7u3UbPJ/W0jnigZASQOBawJznWhtFeF6pKMogZG3CiaBLLEMlJ1QkylZ4tqoyy48GxebMjqylnOjPYa7+RB4BlKPj7p2ZyLcU25JfGKyEjAY3ylIbtHKtLrmtrmfj7PkDRfKhqPsMDo4wMhjGli7o9V7luenC5XTz/5OGMjhUYC/DlL89y80qYvogHWdWD0G3Kyop45OH0E7pty+L2F9e5+tsmWjvzMSpGEcVRnC6d408eAqCoLKdLngmqbqcd3racEptssJxSqmyViS33OFKCroGRqIFNYMczS5kEN7rLmnZ/Ek3pOaWnHOuRvr73sW0Xpmlj204My0axLE4+OMbW3VvjPXHXWzjzq0s0NucRKjQRdZ0omsJDh/ZSXBwPGoL+ALatYAsFRZcUlPn4/h+9MCvwDY8FsG2BjYLikHgL3ewSDm78rIP+qBtZ043QbIoqith1cNdafCQbloVkTteC5ZZRZatMbNnHSf0IZ9iGjAIbZfZgPTMqclmKNSQXWGTI2FiApqZOhke8yLwxEDaKEJixYRRHGf3dQ4z7IwQjDvBEyHOHObh/J6+8cBR3moncM4mEInz4335Dc4ODAVugVPXjcGg8/dTDPHp475x9GpffO0/DhS4a28qRO9pQdMnO3TV8+/QxvL67e7JztqnZEV6TCdmKAtZkr0Bs0lmG+TMJmTr62SoTW+5xhIi/R4iXLyWwLRbMNjjdNgG/RtW26ZMtlqOalSPHShOLDSFEPjD1vYlZGthjANRfbuDc+19z60415tZ4JqK0opBXX3uS8vLiOY4KqqZmlE0zJ8I03S5kULFRqobRHRqHnj3EPYfuyWXjFkk6mVNY+QnZQomXYUmbaSpOcznNi3Hys1UmtuzjiLgNtC1wpdgGy1o4eNLdFkZUJEueEiw2OMqRXXKBxQJYlsWZM1f55JM7NHXlYZQNIoqjOJw6p44eQrO/wbKDGLEYtq2AInE6DByuUn7w7aczPk94IowRM4jF8lEK/OhOlddfPcG992ydd7+JoXGihgOcFqoO9+7byiuvZX7eHMtnuSVGidKecEClals4o7KelVCSmsnMSd6Jxnan256mOpUJbp8VV3YKqDzyzPCirvfA0dEVKXfKkWMlcThKicWm90s5VBO0+Bwj/6Af01KxRDzDUFiex5/+y1ez5/RLSTTmQikJoLtUTnznONU7q7Nz7BwLko3yokRpz8xZDHOxEipS6UgNYLrrvcnGdt1tLXrCtdtnJQf2HXhmqocik2vee3RkmrpVjvVBLrBYgLfe+pQrV/po6ipGbmtH0eCe3TX80YtPke/zMD70fLzHQgkDDpx6FKdmMK5mVv4E8b6KrsZOQgEIGgKcEYQQeDzzazj7+0aYGJnA789HeidASNwL7HM3kq3egaUe525dNU8NXnpTJnlnIhG7VFazDySBZVr0N3QxMaERU2KgGiAyWxXOkWMuystPMjT0XxBCYqHjdARx6iZ6YZqFHwEOhz7rby4SCDPUNoDf7558xttp/y5jwQijrf34/W7wBJDY2JYgZErQoyAErk1oG7LVO7CU49zNK+apAUz/5CRvyEwidqmsZh9IjuWRCywWYHBwlGjUCU4T1QHHHrqXH7wwpfaQXxqfKTHQ+9/weAKMTuRxrW0vO/am74mYyWj/KF+89QX1NwRDuJB17Si6pLZuC1vmmJJtxgyufXiJm2d6aBvMw6oYQDgNPF43Bw8uX5Fqo5DtwXKrFSCUVkXoanYzNjx9YKGq2xk50G+cfmhaBiGB021TUrFyMv6pakypk7PfOP3QrM9uZoBgRBXCxPsk5mO1g7SB5l4u//IrmurdjLkV2NaGogu27d2Jy7v5HLEc2aGzs4+33urHP36IbbuvkVc8TDDmxm8/waFtC880klLS+vUdrn5wk+b2PKLFQcSWMJpD48HD+6Zt132xkTsf3KS5NZ9IcRiqhxAIouNumFwMq9hWRVF50Uq+5XVFtvoHEqxWkFBUFaG32cPEDNug6DIjB3ouWVxrhRVXM52cDbMDBCMqkCwsC3s3B2p3G7nAImPiq0SFebMj5vzSY7R1DtLaaNM57kbZ2sWODI7Y29rLZz/7lNu3K4iU9yN8QTxuFy+88Cj379uZdmXKjBn89m8+oPWOpDvoQqntQdUEDxzYwzMnH91U07VXoxxoJViu8zzU68LhtGfJtq5kJgHiiiOqFu+180yeOzx5PTOZ+R5Tg8DU+zPQ7eSHhx6ftf9q9E+0fl3Pjd9c4dadKsy6DoTTwFvo5YlXj1JRV7Gi585x93L9ehPvvHOGW3cqiVUXcrP+CHn5Hn740nHu2RkftBiNxBjqHSIw4QRXkJkTvS7/8iyNX/fS3FWK3NaO0Gwqa8t4/tXj5Bf6ktvdefc8rRd6aOooQW7tQGgWhFxIbER+CKfXzWPfPkzdvXWbKgO3WiVB2Wa5zvNcsrgzA5VsI21QNIlEJLMXEi1tcDfzPaYGgYn709viATlbsjfXO7ExyAUWa8hQ9xBmTBKLOVA8UXwFbv7iz7+Dax4pwNB4iEggRChYipLnR3MoPHfqCAcfvHcVrzzHfGQ7k7IcllpaNDMrEgqoIOMrUkBSnWmhDESCud73XNPAVyM4HOkYIBLRsBRQnDYFVfmc/pPTcw4My/YqaI67k46OPmIxBUOqKC6TwhIf/9OffRdd05BS0ni1kbPvX6Wp1UeoIIKoGUbVVQ49NPUMH+0eJhJxIV0xFF1y38GdfOv00VnBwXj3MJGwE+k0UXSbfFTGumoRu1rxFLp55c9fQd9Ei03rnfX0DFlKaVHi+lP7KiKTtkFKkRxKB2Q8mC7d+55rEvh6DAzX0z1dL+QCiywwMRYgHIwQCvuQzvDCO6QSF8wHBLpDmzeoABgfGCUWkUQMBRwGQhEUpWiZz8ehEw9wIfarxV1fjkWzGpmUmdKrEC83mhkwLCWQeeP0Q1z6rAQjJrCs+DkSqh1CBY/HWnQD9rpHgNvnnncK8UZdBc2xNgghQIDX50LX4qb2zDtnuXGhi8bOyQyDblNeWcyrrz1JaWkhAMGRCYxwjGAwH1xhEFBYUjBbXtYfIBaMEgzmIV1hQCIMDXQDhMThcuSCinXGSj9DFiOLuxSn99aZYlQ1PrgvsciUsA2FFVHCAY2Di2zA3ujk7MJscoHFMrAsi0tnrnH2kwZauvLjilGeKE6nk/t212X1XJFgmEvvnuf2RT9dY/nImg6Ew6KouJCKypKsnivH+ifddO6uJk9WMiKJUisp1Vm1udKKBzVXzxQmh/nFokqynCknAZsjx9z0dfQTDLnjmQiHzf4DOzn94nGEEFimxZ3Pr3Hz01ZaewowygcR7ihOl5OtO7Ykj2FbFm2/u0H9py20dsW3U1wRpKUwGtRgWzuqLti2d9vavdEca8JKy+ImSq1iEYFtTZfADwfifRK3zhQnh/mZUZEsZ9rMK/ibjVxgsUSMmMEvf/oB9TdjdE14oLYbRYPdu6r53osnyFtghoRt24wPjxGNqkg1BmLuQTv+vhG++P8+oeF2AeM+C1Hdh66rPPr4AY4dPzjvKmuOHEslIYObYKjXgTfP4uAxP19/XJLs8ZCQzNBslOyFZZiExoJEoxroMWbWuMPsFHci/b8UScUcOaYhBEIIKipKEEJgRGKc/emHtNyy6Qm5obYLRRPsvKeOp04/llT7M6MGV376Ce23LLoDXqjtQggJQQ/SEUUUhHJ9QjlWnPzS6X0co73OZKbi6sdluBMqUWjJQOduWsHPptzu3UgusFgiI4N+xkbG8PvLEaWD6C6Fbz97mMce3rfgvsO9w3zxizM03FAZEg7Y0YqqwrbtW9JuP9DWRzRkEAh6EVt68eQ5+MM/eZnCwrxsv60NRTakSddTP0SmrIUkK8Qr9hID/GJRJemKR8IqX0+qRG2E7EV/Yxftl+GcAAAc2UlEQVSX375Ac4MXv9tEqe1EqApb90yfGTMzxZ2QVVxJScUcm5Px/lGCowH8/ipEeR+aS+XJU4+y9+DuadsFB8cIj07g91ciygZQVEl+yMuovxilrpvqXVt48ntPbvrFpmxIk27E2vm1lGRNnNeMCiJoyVLdq5MqUYnsxXr+/DJlLeR2NxK5T2GZxOtoBaoqqKteeIXo9vlbfP2bb6hvKsOo7ka4ong8br797SPsvW97BucDh1Pf9EEFZEeadCMqS62Vs+72WWyZHOCX2nS9nOzFagZJUkquvXeehi+7aOoqQtZ1oegWvuI8nnjlccpry7N+zhybC9u2GR8PYhgqUjEQKZmwSCiCETOJxRzxWSnpEIACqqZQUV0294nibXkIIXBY+uTPUFpduumDCsiONOlGrJ1fK4ddc9jJIXWJxuvUzAVMZS8y+fxyMys2NrnAYpVpvtbMxLgTw2GiuA1qayv40Y9O4sw12d01rJSzvJrZlUybw5fLagZJZsxgoLGLsXEvti+M6rTZfv82jr50FEVVFtzf4bIIBzTMqJhm9HLGLgdAX98wb731KddvKAwJLT4TRYW6ugrqLzdw9tdXaWrJJ1w0ivAF0XSdqi3pZxXluDtZKYd5tbIri2kOXw4bKaORsAvANNuwme1CLrCYB8MwMQwDw3CAYk77XSwSw7JsLFuAssjpMwIEAqFAbV35gkFFLBTFtgVSsdk8SuTrj0wd+5VwlhNKTbrTnva6023T1ezO6iyIuYKH9VratFgE8fp2oUD1ruqMggqAvcfitbO9Td7k6lyOHABffHGFTz65Tn1LKWZND8IZw+1289qpw/RfbuH99/ro9HuR1d0ITVKxpYRXXz1BSUlcCcqIxLCt+LRshD3neczJ7SxLgLAAiWUqIKw0XUI5VoPFOPUr4TD/5PRhbnxWgu6c/hcQDSt013uzOgti79GROd/rZiVhFyBnGxLkAos5aGzs4O23z3K73suYdwKlahxVVakuL+ar317k3KdNtHSXYFR2Idz/f3t3HhzHfR14/Pu658Z98ABPESIlUiR1URJFkaJoybIjWrJkq6SopE2tK3HsKjlOUpVU/kil4pQ3iXe3Uuvd2uxu1utNqqKtTVyb0JLiWL4oUjJBW+JN8xAIgAeIkwSIm8BgZvq3f2CGHOIcco7uAd6niiXMoAf95qeeef36d3SUcLiEqrQbF+XC6PAoh9/9BWcPD9ExVIFZcx7xJahdVJnT/ajMuDlsarab4iXiVk7jyqR4SO+VSZ9vEQzfelKUy14Wt+fDpFY7SV/pBLw95lrlnzGGY8ca6e+PEAuNY4fHuad+GV965TNc6+jhTGcvV3vqYGkHvqCwc9fDbHvi/hsrQZ3df5xT+y5ysauW+PJWJDROKFJKafnNz68TT3Dhg1/RuO8iFzpqidddRsJRnJjNVRPDquvDsm2ql1a72BILk9tDpvo6Q/iC5sY4/5TRoSCBMmdKbNnElcn3XKpXJjXXIsUfvpm7ct3D4vZ8GM0Nt9LCYpJEIsHbb+/n0KFuzneVYVZ2YPkS1FaX8+ozj3HyZ0c5ezow0dWdXLnj3ntW8srzO4mEpx7Y6RzHIRGLT1yVmqOXo/3sJT5+5yMaG2sYrRlFlvfiD9g8sf0htu14IJdvWanbln4iP9NN7mBqMXaioYroqMWlxpJbelkyKQ6yLeycuINxTLKXcearwimThy1cH/DhDxoilfFbkrWXx1yrwjBmorSW5LyHzevXEAz4ST6NIGAJ/oDNvevvQkQYvNrPR/+wj6YzIa75bFgxkU/WbljNrt2PE0ze0+h67yAn/+EDWk6H6LHtGytBmeEwJhTFqhrWeULKM1In0jPd5A6mFmNnGqqJjdpTelgyPTEvdHGnuWF2WlhM0tnZS0tLG11dS2BRN76g4alHN/HFZ7dx+INj9PVcp7e/ClnbTajEz2+8+ixr75p+Nad0Pe09HNhzgMYzAXrtKFLXgWUJixdVTbt948Ez9F8LMGo5SNko1bVlvP5vnqO8Ire9IgudWyss5VrqhD1dLGrxu88/nPer+bO14eQehuioRaQ0wSjcUiTks9fHGENX42WOvnOY5uYyhkqHsCoGsW0fZdUzL4IwOaHNliiVul2tR5sZ7I3SN1SDrL1CsNTPC689w7KVty4C0nn8PEM9UXoHapC1Xdi2wddRx1jZMHYwyj2P3Mujzz6S8ZA+lZn5MoE4ddKeEivQ6kyztd/k3oXYqE2oNI5JW54WvHtirrlhdlpYTGKMwZjkak+WEAjYbHtwPbZl4TgOJjk+GwsqKiNzFhXGGA796GNOHrhEc3slzrIOJBijpDTC85/bxvp7V0/7mngsjnEEEbBsuHvdSi0q8mA+zBmAmyfs6UZh2qFDuTZbG04396OQ4uNxjnz/5zQfHuDStRLMsk4sn0NZTRk7v7iDWp08q/LIcRwcY5iuf8wYB8QCsRALyqtLpxQVMHHPFYMBC7DAL4JES5DKIXwBm3UPrtWiIg/myxCW1El7irmN1ZmyMVv7TZ73oeYXLSzyrP9KPxdOXaCruwZT3YcdTrBhwxpefHEnAf/U5h/uHeKj7zfQfFq4ErNgWTsiUF1V7kL0yktmWqnJ9s89rGeh6rnQSe/5Ljo6V8LKS9gh2LR9Mw/svB/L0pMxlSfGcPbQJxz80UmaL9YwtqgDCV3HHwgTiWRW7MejMc79+AhNDVdo7anErL6ASILxkRCJmi4kPILtDxIuDef5zSivS1+ZaPLzsTFdflgVlhYWeeY4TrIXxAIbfAGbbds2TSkqjDGc/fAkJ95voqW1gtiSXiQ8RjAY5Olnt/Dgw+tdegcqJd/DpmabnDzbSk0AR/bXMDrpd/5QZquVuT0pOp8cx0xc7UXAhkDYx4ZH12tRoXLKcQzGCBN3cjGcbDjFQHOI7njgxkpQdStq+cIXdhEOBwEwTnJzmbqeU09TO6ff/pjmxnKGy+OwsgMBzHCYRHgMiUSpXlrNzpd3aGHhsnwPmZprYvJM++lpD1K1NEp7Ywkm7VQvkGFecHtCtCpeWljkUXtzGwfe+YiW5gqGynqRkiFs209JydREMHxtiOZDjXR2VBOrHMAqibL6rmW89PKnMr7CpfIr3yfZs01O/r9HDt7yXKoYmFwQ+EMJHtzRf8trs9lvtiYXY7GoxSiZFz2z/a305wthvoy5VrnT09PPnj37aW4JcDXhIEuuAjDSN0b3tcWw9jyBkM3u559g46a7ERHisThn9x7nzIE2LnZXYFZcQqwEkbKbx3bL3uP0dvoZwmBVDOEHYhdWIava8IVtHvvso6x9aO3EsFzlqnyfZM81MTl9/+nFQO3y6I3nA6HELcui5mK/2Zj8XRqLCgZfxkXPXH8v/flCcHv/XqOFxe3IcKHw6GiUg+82cPrQAJcHQpil3YgvQVVVGV/4wk6qprlrthNPJFcWsRAb/AEfT+56SIsKNa3JxUDHxTDh0qlDpdw2uRhL7x1JLxIyKQ7uqLAzuVvdX6/SqRTHcdi37zD795+jua2c+NJ+JBQlGAyxua6GofMDYAliCbV1FWzavBaAqxe7OPzPv6CpMcJAxCAr27Fs4d5N9Tz13Nabfz/uYLBTUzCodHxcjYcQSyirLWXdw+vceuvKw6YrBq5cjDDuseFQk79L0wuiO7n5qNvfzW7v32u0sMiAcRyOHjjJoQPNNLdW4yy/jGUlZlxetvlYE22NHbS1r4K1LfhD8OSTD7Fj+wPYOslO5UEw7HB92CYWtW77hL2QCjW0yhhD59lWjv3LEZrO1zC+uAPxj2P7w9h+byVZVXza2q5w+PAZWluXEl/cjRWJcd+61bz+0i6O7T3KqfOD077uV+99zJV2mwEHrKoBSsrDfO6VXSxdocvEqvzwhxNc7/dNuaLupavpemI+v2hhMQdjDD/83vu0NobptZP3rrBhzZo6Xntx17SviY/HMVgggthQWVPCUzsfKmzgakF5YHsfMP2wKTe5MX9jbHiUo3sOcO7oGB3Xg5ilnYjPoby2nJ0v78A/x53ulZpLLBbHGDDGRnyGUNjPS5/dRjgUnPV1iXgCx/gR22DZwsNPbNSiQuXVfduvee6O0Dp/Y37TwmIO8fE4Xb0OvVEfsqqTcEmAVz//FBvuWZWz8a2DPQN8/M4vaLtUwqBvBEqGELFv3CBJqWLlxt3Kz75/jI6mPjp6FkP9RXwh4aFPbWHDVp20rbxFuDWHxMbGOffeEa60GjoGA7CkGxzDtWslmLo2sBIEQpoXVHFz+27lKr+0sJiWueUnY6yJK0w+4cmtm7hvmntP3IlEPMGpfcc4uf8SFzvLiC+5MjFONxTk2c8+xuLF1TnZj/K21FX9tpYwlxpvfrHafocV9aN5H87k9qToXItHxzGODbYgPmHp3YvZuO0+t8NSatZ5el0nL3DmByc431LOSGUMVrQjgDMcwqkYQHxQs7yG7c9vL1i4yj2pq/qdLRHa0/KC5TfU1V/P+1AmnZCs7pQWFmmuXRvgJz/5JW3tpYwEhyB0HQGstKtKc3VStDVd5pNDzbRersJZ0ollOfinuV8FwKWTLTR9dI4L51fh3H0eO5hg/X31/NrnniA0R5e6mj9SV/UnX9mfbVhTLouBYl9SNsUYw8XD52g/c5XWzkrM4k5EHALa86cKwBhD88kWGo+3cqmtErMkefwF/MRjcc787Cg9nYauwQAsuQIC/uBEbhju7qPpx0e42FLDSEUfUj6EL2Ez3lWLtewK/pCfrbsfpX5zva4EtUCkrupPvrI/27CmXBYDOiRJ3SktLIB4PMGHHx5l375ztLSXEVvcj9RECQWDrAj4iPbP3UwjgyMcfPcgp48M0zkSujGuu6q6nM/vfnLa10SvR8EIRgTxQXl1hJde/lSu354qgELPJZgvxUCuDHb3cWRPA02nhB6CmBVtiM9Qu2IRW57e4nZ4ar4zhvf/6QMunDJ0j4cwyzsQn0PtkioeWVfPj7/9Ls1NJQyVJmBlG7ZP2HD/3dyzsR6A+FgM4xgSjh8JONiWUD1UQTd+xCds2nEfd99/t8tvUt2uQs8l0GJAeYEWFsChQ6c5cOAU51qWY+ovYAcd7r9nDffVVHPkw2Y6+0KYxVcACMww8bPh7QOcPz1ER185svoy/qDFrqe2sG3bZuwMx3VbeiWqaOViLsGJhiqiozePlVjU4vUtT8yLG9XlUyIW59D/+4DWJpsrMQtreQ+BYIDHn3+UNRvX6BVelVMDA8N8+OExOjojjIUHITBGPOrQei5K50gp1so2AkEfn352K/UrlnDwrZ/S3FjDUHkfUjlEeXUpz728kyXLFmW8T50bVJxyMZfgTEM1sdGbK9nFosLvb9mpE52VZ2lhAQwNXcdxLLAEyyfUVJUQ7h7mvf1RrvnsiZWgfLBu7XIevv+eaf9G9PoY8YQP8RssP2zduoEd2x8o8DtRxSw6ahEpvXmDoFFgxdrreZ3onG+FmL+RiCeIx+LE4yHEH8fyWWzd/Qj1m+pztg+lEokEBw4cZ+/eT2huKyO2aAipjhII+FnuCxKNBxCfg+UTdj39MA8/soGrF7owCYdEIoCEEgQjfl547RlqFlW6/XZUkYiN2oRK4zceG3zUrR0p6onOOn9jftPC4hYCGEb6hzlzqZL+yDBS3U9paZhXXtzJ+rUrM/5LPnv2tfIHrvRz8cQFurrLMVW9iCT0qpSadwrb03KzZ8Ke4/On1O1qaDjB/v2/orFlOc6a81gBh7vXLueNF57ix3/7Hq1p2/p8Mx9/lnVrD9r49SiXfvkJPT0BRu0xCIxhHEPfYBAqr4EYLL3/kZpHtKdlftPCYpLUoh3xeBAJ9REK+/nSa8+ycnlu1hqPx+L86mfHOPXzy1zoLiGx5AoSHCcUCfH0M4/lZB+quKSu6seiFqNpz/tDiRlfo5QqrP7+YRIJG2OB5Ye6FVW8+fpuzB3e3d0YQ8exFs6+d4rzF8oYTVsJKjESJrG0C/HB4pWLqd+svW8LTeqqfiwqmLRTtYDmBeVxrhQWIuIHfg/4FvAbxph/nGE7Ab4JvAokgKPAV40xUwct5tFsV5+MMbR+0spAzxg9fZWYqokblc10henQ2wdoOdZDS1s1rLmE7YONm9fxmece1/tWLFCpq/qvb3li2nkaambGcWg91sxIv6FvxIcs7UUExNJ5FcWomHKD3+fDGEPT8SaG+mP0DpZBbQ/IxJyI+HiM1uPNDPQHGbWi4I8iYt3omW4/3ETjj47zydnlJNZcRAIxrGiAxPUgUjVEIBxk626dJ7RQpa7q//6WndPO01DKq9zqsfhtJjoHfjnHdv8W2A08aIwZFZG/A/4D8Dt5ji8jIwMjNLzdwOlj1+kai2CWtSN+h9pFVTxw/7rpX9M3zHjMDwEHyw8PPLSO5z63o8CRq1ybb/eCKAYDnb0c+f5Bmk7Z9IgfVrdi2bB49RLq6uvcDk/dmaLJDfFYnHe/+6+cOZHgaiKMWXEZ8RmW1NVSaoT3/tO7NDeXMlKeQFZdxvJZ3PfgOsorSwG43jtEImHjWIL4HYIWSOdKxpZ0Ewj7+fQbT7NoeeYTvJX36FwCtRC5UlgYY/47gIi8MMemvw78T2NMaoTIfwH2kuPkUVoaIRz2EwnZxCREyBfFKoGYFaa0xMbvm76ZTn54gsvN1+jqXQxrLhIIW3xq1xYe37pxxvkSkYpSAqEh/AEftj9ETW1FLt+Kckku5hIUujgp9BK5uWSM4cyPDtF5IU7PaAhZ3U0gEuTx5x/hrg136RXeIuW13JCurCxCKOQjHLRxrDDO8Bjt3XGuDtbAmosEw34+/ZlH2HDPaj787g9pb40wHIxiVQxRtaiCX/vikyxaWnPj7wXLwvhDNsGwkLCCRPzjEBHiVohwqU0oMvWzqYpLLuYSuFGcFHqZXDW/yJ2OD53zD4sEgOnu4rLTGDOe3GY/8DezdHc3Am8aY/YmH5cBg0ClMWZg0rZfAb6SfLgJOJWL95FHtUCP20HMoRhihOKI04MxbtwA0ejNxyYIEoVgEE6fdS+uOXmwLacohhjvNcaUFXqnhcwNRZgXoDiOHY0xdzwYZ1HmBg+247SKIc6sckPeeiySCeLxLP+MMDF+NiW15tqU7gBjzHeA7wCIyGFjzCNZ7juvNMbcKYY4NcbcKYY4iyVGN/ZbyNxQbHkBiiNOjTF3iiFOjTF3iiHObHOD19ewawNWpT1eBQwD/e6Eo5RSygM0NyillAd5qrAQkRoRaRCR1Mznt4AvJ7vOAb4O7DH5Gr+llFLKczQ3KKVUcfDafSwiwGogNaP574G1wMciEgfOkNnkvO/kJ7yc0hhzpxji1Bhzpxji1BhzKxe5oVjebzHEqTHmTjHEqTHmTjHEmVWMeZu8rZRSSimllFo4PDUUSimllFJKKVWctLBQSimllFJKZU0LC6WUUkoppVTWir6wEBG/iPyhiMRE5LVZttsiItdE5Jdp//7Ag3GKiPw7EWkUkTMi8n9EZOptN/MT41MiclRETorIYRGZdq35QrdlJnG52W63Gafnj0O32zLDGF1tx2QMXxWRE8n/1ydF5M0ZtvuaiHwiIqdE5AcissRLMSZXXBqZ1JZ/VagY86UYckMx5IXk/jU35DdGL3yfaW7IXZwLOzcYY4r6H/Am8AfAz4HXZtnus8D/KoI4vwQcAcLJx38H/HUB4qsEeoFtyce7gG4g4mZbZhqXW+12B3F6/jj0QFtmEqPb7WgDfwWUJh8vB0aB5ZO22wW0AouSj/8M+IHHYrwX+KlbbenmceT2seT1vJDcl+aG/Mfo6vdZpsei5oaMYlzwucGVhs9TQ+2f44v5DeAS8FHy318AZR6M8z3gK2mPHwR6CxDXrwMHJz13HHjBzbbMNC632u0O4vT8ceh2W2YYoyfaMS2eEDAArJr0/P8A/jLtcRUTd4mu8FCM25k42fkFcAj4r8ASt9qykMeRV44lr+aF5L40N+Q/RtePwbRYNDfkNtYFlxuKYiiUiAQmdcWk/gXmfvUNe4C7jDFbgeeANUyshe61OOuBlrTHLUC1iFTMsH1OYgTWT9pvat/10/yZvLdlmsntMVNceW23DGQaZyHb7k653ZaZ8Fo7/mfge8aY1knP39KWxpg+Jr7A7ypcaDfMFONhYJkxZhsTV9HGgR+KiBQ4vttWDLmhGPLCbHGiuSEb8ykvgOaGO7HgcoPXbpA3LWPMODDtmM7b+BujaT9fS465uywi4fTfZbmPrOMEBEikPY4n/5uTInCmGEXkjyftN7XvKfstRFumh5ZhXHlttwxkFGeB2+5Oud2Wc/JSO4rInzPRlfzydL8mw89VPs0WozEmmvbziIj8ETDIxA3omgoW5B0ohtxQDHkBNDfkyXzKC6C54bYs1NzgmYPBBTYwBkTn2rDA2oBVaY9XAcNAf4H3m9p3WwavzWdbZhqXW+020/5TMczVfl48Dt1uyzvhSjsmJ7JtBF5OnphNdktbikgEqCGzz1VOZBDjlJcwkRsG8xqYd+lncvZ9p/avuWFu8ykvgOaGjC3k3DBvC4vkbPYGEVmXfPyaiFQmf/YB3wLeMsY4XooTeAv4clo3+deBPSY54C2P3gHuF5HNybgeY6IL/Kcut+VMcb3vkXa7rTi9eBx66Bickdc+zyJiicjfACuBV1JfyiJii8heEXkyuelbwBtpQwW+BjQYY656JUYR+byI1CV/FuCbwH5jTHe+Y3SD146lTGLE3c+k5oY8x+jFYzAZi5eOw2l57fOsuaFIhkLdoQiwGkj9TwsDe0XEAQzwAfCnLsWWbnKcf89EN9PHIhIHzgC/k+8gjDEDIvIK8LciYpjoktttjOkXkZW41JYzxYVH2u0O4vTiceiptpyB19pxN/BVJsagHpCbQ07/nIkxvdUAxph9IvLfgA9EJAZ0ADMuK+pGjEy03x4R8QMOcJKJCZDzldeOpel45jOpuaEgMXrxGAQPteUsvNaWCz43iIuFplJKKaWUUmqemLdDoZRSSimllFKFo4WFUkoppZRSKmtaWCillFJKKaWypoWFUkoppZRSKmtaWCillFJKKaWypoWFUkoppZRSKmtaWCillFJKKaWypoWFUkoppZRSKmtaWCillFJKKaWypoWFUkoppZRSKmtaWCillFJKKaWypoWFUjkgImERaRORVhEJTvrdd0UkISKvuRWfUkqpwtPcoBYaLSyUygFjzCjwDWAl8GbqeRH5FvBbwNeNMf/oUnhKKaVcoLlBLTRijHE7BqXmBRGxgRPAYqAe+DLwbeAbxphvuhmbUkopd2huUAuJFhZK5ZCIPA/8C7AXeBr4a2PM77oblVJKKTdpblALhQ6FUiqHjDE/AI4CzwDfA35v8jYi8jUR+VhExkRkf4FDVEopVWCaG9RC4XM7AKXmExF5FXgw+XDITN8l2An8e+BRYFuhYlNKKeUOzQ1qodDCQqkcEZHPAG8B3wdiwG+KyLeNMWfTtzPG7Eluv6rwUSqllCokzQ1qIdGhUErlgIhsBfYADcAbwJ8ADvAtN+NSSinlHs0NaqHRwkKpLInIBuBfgXPAS8aYqDGmBfjfwIsist3VAJVSShWc5ga1EGlhoVQWkl3WPwEGgOeMMYNpv/4mMAr8RzdiU0op5Q7NDWqh0jkWSmXBGNPKxI2PpvtdJxApbERKKaXcprlBLVRaWChVYCLiY+Kz5wMsEQkBjjFm3N3IlFJKuUVzg5oPtLBQqvD+BPhG2uNR4ANglyvRKKWU8gLNDaro6Z23lVJKKaWUUlnTydtKKaWUUkqprGlhoZRSSimllMqaFhZKKaWUUkqprGlhoZRSSimllMqaFhZKKaWUUkqprGlhoZRSSimllMqaFhZKKaWUUkqprP1/6GC4oINYeI4AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 792x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "m = len(X_train)\n",
    "\n",
    "plt.figure(figsize=(11, 4))\n",
    "for subplot, learning_rate in ((121, 1), (122, 0.5)):\n",
    "    sample_weights = np.ones(m)\n",
    "    plt.subplot(subplot)\n",
    "    if subplot == 121:\n",
    "        plt.text(-0.7, -0.65, \"1\", fontsize=14)\n",
    "        plt.text(-0.6, -0.10, \"2\", fontsize=14)\n",
    "        plt.text(-0.5,  0.10, \"3\", fontsize=14)\n",
    "        plt.text(-0.4,  0.55, \"4\", fontsize=14)\n",
    "        plt.text(-0.3,  0.90, \"5\", fontsize=14)        \n",
    "    for i in range(5):\n",
    "        svm_clf = SVC(kernel=\"rbf\", C=0.05, gamma='auto', random_state=42)\n",
    "        svm_clf.fit(X_train, y_train, sample_weight=sample_weights)\n",
    "        y_pred = svm_clf.predict(X_train)\n",
    "        sample_weights[y_pred != y_train] *= (1 + learning_rate)\n",
    "        plot_decision_boundary(svm_clf, X, y, alpha=0.2)\n",
    "        plt.title(\"learning_rate = {}\".format(learning_rate), fontsize=16)\n",
    "\n",
    "save_fig(\"boosting_plot\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['base_estimator_',\n",
       " 'classes_',\n",
       " 'estimator_errors_',\n",
       " 'estimator_weights_',\n",
       " 'estimators_',\n",
       " 'feature_importances_',\n",
       " 'n_classes_']"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list(m for m in dir(ada_clf) if not m.startswith(\"_\") and m.endswith(\"_\"))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 그래디언트 부스팅"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(42)\n",
    "X = np.random.rand(100, 1) - 0.5\n",
    "y = 3*X[:, 0]**2 + 0.05 * np.random.randn(100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeRegressor(criterion='mse', max_depth=2, max_features=None,\n",
       "                      max_leaf_nodes=None, min_impurity_decrease=0.0,\n",
       "                      min_impurity_split=None, min_samples_leaf=1,\n",
       "                      min_samples_split=2, min_weight_fraction_leaf=0.0,\n",
       "                      presort=False, random_state=42, splitter='best')"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.tree import DecisionTreeRegressor\n",
    "\n",
    "tree_reg1 = DecisionTreeRegressor(max_depth=2, random_state=42)\n",
    "tree_reg1.fit(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeRegressor(criterion='mse', max_depth=2, max_features=None,\n",
       "                      max_leaf_nodes=None, min_impurity_decrease=0.0,\n",
       "                      min_impurity_split=None, min_samples_leaf=1,\n",
       "                      min_samples_split=2, min_weight_fraction_leaf=0.0,\n",
       "                      presort=False, random_state=42, splitter='best')"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y2 = y - tree_reg1.predict(X)\n",
    "tree_reg2 = DecisionTreeRegressor(max_depth=2, random_state=42)\n",
    "tree_reg2.fit(X, y2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeRegressor(criterion='mse', max_depth=2, max_features=None,\n",
       "                      max_leaf_nodes=None, min_impurity_decrease=0.0,\n",
       "                      min_impurity_split=None, min_samples_leaf=1,\n",
       "                      min_samples_split=2, min_weight_fraction_leaf=0.0,\n",
       "                      presort=False, random_state=42, splitter='best')"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y3 = y2 - tree_reg2.predict(X)\n",
    "tree_reg3 = DecisionTreeRegressor(max_depth=2, random_state=42)\n",
    "tree_reg3.fit(X, y3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_new = np.array([[0.8]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = sum(tree.predict(X_new) for tree in (tree_reg1, tree_reg2, tree_reg3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.75026781])"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 792x792 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot_predictions(regressors, X, y, axes, label=None, style=\"r-\", data_style=\"b.\", data_label=None):\n",
    "    x1 = np.linspace(axes[0], axes[1], 500)\n",
    "    y_pred = sum(regressor.predict(x1.reshape(-1, 1)) for regressor in regressors)\n",
    "    plt.plot(X[:, 0], y, data_style, label=data_label)\n",
    "    plt.plot(x1, y_pred, style, linewidth=2, label=label)\n",
    "    if label or data_label:\n",
    "        plt.legend(loc=\"upper center\", fontsize=16)\n",
    "    plt.axis(axes)\n",
    "\n",
    "plt.figure(figsize=(11,11))\n",
    "\n",
    "plt.subplot(321)\n",
    "plot_predictions([tree_reg1], X, y, axes=[-0.5, 0.5, -0.1, 0.8], label=\"$h_1(x_1)$\", style=\"g-\", data_label=\"훈련 세트\")\n",
    "plt.ylabel(\"$y$\", fontsize=16, rotation=0)\n",
    "plt.title(\"잔여 오차와 트리의 예측\", fontsize=16)\n",
    "\n",
    "plt.subplot(322)\n",
    "plot_predictions([tree_reg1], X, y, axes=[-0.5, 0.5, -0.1, 0.8], label=\"$h(x_1) = h_1(x_1)$\", data_label=\"훈련 세트\")\n",
    "plt.ylabel(\"$y$\", fontsize=16, rotation=0)\n",
    "plt.title(\"앙상블의 예측\", fontsize=16)\n",
    "\n",
    "plt.subplot(323)\n",
    "plot_predictions([tree_reg2], X, y2, axes=[-0.5, 0.5, -0.5, 0.5], label=\"$h_2(x_1)$\", style=\"g-\", data_style=\"k+\", data_label=\"잔여 오차\")\n",
    "plt.ylabel(\"$y - h_1(x_1)$\", fontsize=16)\n",
    "\n",
    "plt.subplot(324)\n",
    "plot_predictions([tree_reg1, tree_reg2], X, y, axes=[-0.5, 0.5, -0.1, 0.8], label=\"$h(x_1) = h_1(x_1) + h_2(x_1)$\")\n",
    "plt.ylabel(\"$y$\", fontsize=16, rotation=0)\n",
    "\n",
    "plt.subplot(325)\n",
    "plot_predictions([tree_reg3], X, y3, axes=[-0.5, 0.5, -0.5, 0.5], label=\"$h_3(x_1)$\", style=\"g-\", data_style=\"k+\")\n",
    "plt.ylabel(\"$y - h_1(x_1) - h_2(x_1)$\", fontsize=16)\n",
    "plt.xlabel(\"$x_1$\", fontsize=16)\n",
    "\n",
    "plt.subplot(326)\n",
    "plot_predictions([tree_reg1, tree_reg2, tree_reg3], X, y, axes=[-0.5, 0.5, -0.1, 0.8], label=\"$h(x_1) = h_1(x_1) + h_2(x_1) + h_3(x_1)$\")\n",
    "plt.xlabel(\"$x_1$\", fontsize=16)\n",
    "plt.ylabel(\"$y$\", fontsize=16, rotation=0)\n",
    "\n",
    "save_fig(\"gradient_boosting_plot\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GradientBoostingRegressor(alpha=0.9, criterion='friedman_mse', init=None,\n",
       "                          learning_rate=0.1, loss='ls', max_depth=2,\n",
       "                          max_features=None, max_leaf_nodes=None,\n",
       "                          min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                          min_samples_leaf=1, min_samples_split=2,\n",
       "                          min_weight_fraction_leaf=0.0, n_estimators=3,\n",
       "                          n_iter_no_change=None, presort='auto',\n",
       "                          random_state=42, subsample=1.0, tol=0.0001,\n",
       "                          validation_fraction=0.1, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "\n",
    "gbrt = GradientBoostingRegressor(max_depth=2, n_estimators=3, learning_rate=0.1, random_state=42)\n",
    "gbrt.fit(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GradientBoostingRegressor(alpha=0.9, criterion='friedman_mse', init=None,\n",
       "                          learning_rate=0.1, loss='ls', max_depth=2,\n",
       "                          max_features=None, max_leaf_nodes=None,\n",
       "                          min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                          min_samples_leaf=1, min_samples_split=2,\n",
       "                          min_weight_fraction_leaf=0.0, n_estimators=200,\n",
       "                          n_iter_no_change=None, presort='auto',\n",
       "                          random_state=42, subsample=1.0, tol=0.0001,\n",
       "                          validation_fraction=0.1, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gbrt_slow = GradientBoostingRegressor(max_depth=2, n_estimators=200, learning_rate=0.1, random_state=42)\n",
    "gbrt_slow.fit(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 792x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(11,4))\n",
    "\n",
    "plt.subplot(121)\n",
    "plot_predictions([gbrt], X, y, axes=[-0.5, 0.5, -0.1, 0.8], label=\"앙상블의 예측\")\n",
    "plt.title(\"learning_rate={}, n_estimators={}\".format(gbrt.learning_rate, gbrt.n_estimators), fontsize=14)\n",
    "\n",
    "plt.subplot(122)\n",
    "plot_predictions([gbrt_slow], X, y, axes=[-0.5, 0.5, -0.1, 0.8])\n",
    "plt.title(\"learning_rate={}, n_estimators={}\".format(gbrt_slow.learning_rate, gbrt_slow.n_estimators), fontsize=14)\n",
    "\n",
    "save_fig(\"gbrt_learning_rate_plot\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 조기 종료를 사용한 그래디언트 부스팅"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GradientBoostingRegressor(alpha=0.9, criterion='friedman_mse', init=None,\n",
       "                          learning_rate=0.1, loss='ls', max_depth=2,\n",
       "                          max_features=None, max_leaf_nodes=None,\n",
       "                          min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                          min_samples_leaf=1, min_samples_split=2,\n",
       "                          min_weight_fraction_leaf=0.0, n_estimators=55,\n",
       "                          n_iter_no_change=None, presort='auto',\n",
       "                          random_state=42, subsample=1.0, tol=0.0001,\n",
       "                          validation_fraction=0.1, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy as np\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import mean_squared_error\n",
    "\n",
    "X_train, X_val, y_train, y_val = train_test_split(X, y, random_state=49)\n",
    "\n",
    "gbrt = GradientBoostingRegressor(max_depth=2, n_estimators=120, random_state=42)\n",
    "gbrt.fit(X_train, y_train)\n",
    "\n",
    "errors = [mean_squared_error(y_val, y_pred)\n",
    "          for y_pred in gbrt.staged_predict(X_val)]\n",
    "bst_n_estimators = np.argmin(errors)\n",
    "\n",
    "gbrt_best = GradientBoostingRegressor(max_depth=2,n_estimators=bst_n_estimators, random_state=42)\n",
    "gbrt_best.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "min_error = np.min(errors)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 792x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(11, 4))\n",
    "\n",
    "plt.subplot(121)\n",
    "plt.plot(errors, \"b.-\")\n",
    "plt.plot([bst_n_estimators, bst_n_estimators], [0, min_error], \"k--\")\n",
    "plt.plot([0, 120], [min_error, min_error], \"k--\")\n",
    "plt.plot(bst_n_estimators, min_error, \"ko\")\n",
    "plt.text(bst_n_estimators, min_error*1.2, \"최소\", ha=\"center\", fontsize=14)\n",
    "plt.axis([0, 120, 0, 0.01])\n",
    "plt.xlabel(\"트리 개수\")\n",
    "plt.title(\"검증 오차\", fontsize=14)\n",
    "\n",
    "plt.subplot(122)\n",
    "plot_predictions([gbrt_best], X, y, axes=[-0.5, 0.5, -0.1, 0.8])\n",
    "plt.title(\"최적 모델 (트리 %d개)\" % bst_n_estimators, fontsize=14)\n",
    "\n",
    "save_fig(\"early_stopping_gbrt_plot\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "gbrt = GradientBoostingRegressor(max_depth=2, warm_start=True, random_state=42)\n",
    "\n",
    "min_val_error = float(\"inf\")\n",
    "error_going_up = 0\n",
    "for n_estimators in range(1, 120):\n",
    "    gbrt.n_estimators = n_estimators\n",
    "    gbrt.fit(X_train, y_train)\n",
    "    y_pred = gbrt.predict(X_val)\n",
    "    val_error = mean_squared_error(y_val, y_pred)\n",
    "    if val_error < min_val_error:\n",
    "        min_val_error = val_error\n",
    "        error_going_up = 0\n",
    "    else:\n",
    "        error_going_up += 1\n",
    "        if error_going_up == 5:\n",
    "            break  # 조기 종료"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "61\n"
     ]
    }
   ],
   "source": [
    "print(gbrt.n_estimators)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "최소 검증 MSE: 0.002712853325235463\n"
     ]
    }
   ],
   "source": [
    "print(\"최소 검증 MSE:\", min_val_error)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "try:\n",
    "    import xgboost\n",
    "except ImportError as ex:\n",
    "    print(\"에러: xgboost 라이브러리가 설치되지 않았습니다.\")\n",
    "    xgboost = None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[11:34:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "검증 MSE: 0.0028512559726563943\n"
     ]
    }
   ],
   "source": [
    "if xgboost is not None:  # 책에는 없음\n",
    "    xgb_reg = xgboost.XGBRegressor(random_state=42)\n",
    "    xgb_reg.fit(X_train, y_train)\n",
    "    y_pred = xgb_reg.predict(X_val)\n",
    "    val_error = mean_squared_error(y_val, y_pred)\n",
    "    print(\"검증 MSE:\", val_error)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[0]\tvalidation_0-rmse:0.286719\n",
      "Will train until validation_0-rmse hasn't improved in 2 rounds.\n",
      "[1]\tvalidation_0-rmse:0.258221\n",
      "[2]\tvalidation_0-rmse:0.232634\n",
      "[3]\tvalidation_0-rmse:0.210526\n",
      "[4]\tvalidation_0-rmse:0.190232\n",
      "[5]\tvalidation_0-rmse:0.172196\n",
      "[6]\tvalidation_0-rmse:0.156394\n",
      "[7]\tvalidation_0-rmse:0.142241\n",
      "[8]\tvalidation_0-rmse:0.129789\n",
      "[9]\tvalidation_0-rmse:0.118752\n",
      "[10]\tvalidation_0-rmse:0.108388\n",
      "[11]\tvalidation_0-rmse:0.100155\n",
      "[12]\tvalidation_0-rmse:0.09208\n",
      "[13]\tvalidation_0-rmse:0.084791\n",
      "[14]\tvalidation_0-rmse:0.078699\n",
      "[15]\tvalidation_0-rmse:0.073248\n",
      "[16]\tvalidation_0-rmse:0.069391\n",
      "[17]\tvalidation_0-rmse:0.066277\n",
      "[18]\tvalidation_0-rmse:0.063458\n",
      "[19]\tvalidation_0-rmse:0.060326\n",
      "[20]\tvalidation_0-rmse:0.0578\n",
      "[21]\tvalidation_0-rmse:0.055643\n",
      "[22]\tvalidation_0-rmse:0.053943\n",
      "[23]\tvalidation_0-rmse:0.053138\n",
      "[24]\tvalidation_0-rmse:0.052415\n",
      "[25]\tvalidation_0-rmse:0.051821\n",
      "[26]\tvalidation_0-rmse:0.051226\n",
      "[27]\tvalidation_0-rmse:0.051135\n",
      "[28]\tvalidation_0-rmse:0.05091\n",
      "[29]\tvalidation_0-rmse:0.050893\n",
      "[30]\tvalidation_0-rmse:0.050725\n",
      "[31]\tvalidation_0-rmse:0.050471\n",
      "[32]\tvalidation_0-rmse:0.050285\n",
      "[33]\tvalidation_0-rmse:0.050492\n",
      "[34]\tvalidation_0-rmse:0.050348\n",
      "Stopping. Best iteration:\n",
      "[32]\tvalidation_0-rmse:0.050285\n",
      "\n",
      "검증 MSE: 0.002528626115371327\n"
     ]
    }
   ],
   "source": [
    "if xgboost is not None:  # 책에는 없음\n",
    "    xgb_reg.fit(X_train, y_train,\n",
    "                eval_set=[(X_val, y_val)], early_stopping_rounds=2)\n",
    "    y_pred = xgb_reg.predict(X_val)\n",
    "    val_error = mean_squared_error(y_val, y_pred)\n",
    "    print(\"검증 MSE:\", val_error)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[11:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "7.93 ms ± 14.2 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit xgboost.XGBRegressor().fit(X_train, y_train) if xgboost is not None else None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "9.94 ms ± 5.51 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit GradientBoostingRegressor().fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# 연습문제 해답"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. to 7.\n",
    "\n",
    "부록 A를 참조."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 8. 투표 기반 분류기\n",
    "\n",
    "*문제: MNIST 데이터를 불러들여 훈련 세트, 검증 세트, 테스트 세트로 나눕니다(예를 들면 훈련에 40,000개 샘플, 검증에 10,000개 샘플, 테스트에 10,000개 샘플).*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "# from sklearn.datasets import fetch_mldata\n",
    "# mnist = fetch_mldata('MNIST original')\n",
    "from sklearn.datasets import fetch_openml\n",
    "mnist = fetch_openml('mnist_784', version=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_val, X_test, y_train_val, y_test = train_test_split(\n",
    "    mnist.data, mnist.target, test_size=10000, random_state=42)\n",
    "X_train, X_val, y_train, y_val = train_test_split(\n",
    "    X_train_val, y_train_val, test_size=10000, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*문제: 그런 다음 랜덤 포레스트 분류기, 엑스트라 트리 분류기, SVM 같은 여러 종류의 분류기를 훈련시킵니다.*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier, ExtraTreesClassifier\n",
    "from sklearn.svm import LinearSVC\n",
    "from sklearn.neural_network import MLPClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "random_forest_clf = RandomForestClassifier(n_estimators=10, random_state=42)\n",
    "extra_trees_clf = ExtraTreesClassifier(n_estimators=10, random_state=42)\n",
    "svm_clf = LinearSVC(max_iter=10000, random_state=42)\n",
    "mlp_clf = MLPClassifier(random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "훈련 예측기:  RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
      "                       max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
      "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
      "                       min_samples_leaf=1, min_samples_split=2,\n",
      "                       min_weight_fraction_leaf=0.0, n_estimators=10,\n",
      "                       n_jobs=None, oob_score=False, random_state=42, verbose=0,\n",
      "                       warm_start=False)\n",
      "훈련 예측기:  ExtraTreesClassifier(bootstrap=False, class_weight=None, criterion='gini',\n",
      "                     max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
      "                     min_impurity_decrease=0.0, min_impurity_split=None,\n",
      "                     min_samples_leaf=1, min_samples_split=2,\n",
      "                     min_weight_fraction_leaf=0.0, n_estimators=10, n_jobs=None,\n",
      "                     oob_score=False, random_state=42, verbose=0,\n",
      "                     warm_start=False)\n",
      "훈련 예측기:  LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,\n",
      "          intercept_scaling=1, loss='squared_hinge', max_iter=10000,\n",
      "          multi_class='ovr', penalty='l2', random_state=42, tol=0.0001,\n",
      "          verbose=0)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/haesun/anaconda3/envs/handson-ml/lib/python3.7/site-packages/sklearn/svm/base.py:929: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "훈련 예측기:  MLPClassifier(activation='relu', alpha=0.0001, batch_size='auto', beta_1=0.9,\n",
      "              beta_2=0.999, early_stopping=False, epsilon=1e-08,\n",
      "              hidden_layer_sizes=(100,), learning_rate='constant',\n",
      "              learning_rate_init=0.001, max_iter=200, momentum=0.9,\n",
      "              n_iter_no_change=10, nesterovs_momentum=True, power_t=0.5,\n",
      "              random_state=42, shuffle=True, solver='adam', tol=0.0001,\n",
      "              validation_fraction=0.1, verbose=False, warm_start=False)\n"
     ]
    }
   ],
   "source": [
    "estimators = [random_forest_clf, extra_trees_clf, svm_clf, mlp_clf]\n",
    "for estimator in estimators:\n",
    "    print(\"훈련 예측기: \", estimator)\n",
    "    estimator.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.9469, 0.9492, 0.8768, 0.9621]"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[estimator.score(X_val, y_val) for estimator in estimators]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "선형 SVM이 다른 분류기보다 성능이 많이 떨어집니다. 그러나 투표 기반 분류기의 성능을 향상시킬 수 있으므로 그대로 두겠습니다."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*문제: 그리고 검증 세트에서 개개의 분류기보다 더 높은 성능을 내도록 이들을 간접 또는 직접 투표 분류기를 사용하는 앙상블로 연결해보세요.*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import VotingClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "named_estimators = [\n",
    "    (\"random_forest_clf\", random_forest_clf),\n",
    "    (\"extra_trees_clf\", extra_trees_clf),\n",
    "    (\"svm_clf\", svm_clf),\n",
    "    (\"mlp_clf\", mlp_clf),\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "voting_clf = VotingClassifier(named_estimators)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/haesun/anaconda3/envs/handson-ml/lib/python3.7/site-packages/sklearn/svm/base.py:929: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "VotingClassifier(estimators=[('random_forest_clf',\n",
       "                              RandomForestClassifier(bootstrap=True,\n",
       "                                                     class_weight=None,\n",
       "                                                     criterion='gini',\n",
       "                                                     max_depth=None,\n",
       "                                                     max_features='auto',\n",
       "                                                     max_leaf_nodes=None,\n",
       "                                                     min_impurity_decrease=0.0,\n",
       "                                                     min_impurity_split=None,\n",
       "                                                     min_samples_leaf=1,\n",
       "                                                     min_samples_split=2,\n",
       "                                                     min_weight_fraction_leaf=0.0,\n",
       "                                                     n_estimators=10,\n",
       "                                                     n_jobs=None,\n",
       "                                                     oob_score=False,\n",
       "                                                     random_sta...\n",
       "                                            beta_2=0.999, early_stopping=False,\n",
       "                                            epsilon=1e-08,\n",
       "                                            hidden_layer_sizes=(100,),\n",
       "                                            learning_rate='constant',\n",
       "                                            learning_rate_init=0.001,\n",
       "                                            max_iter=200, momentum=0.9,\n",
       "                                            n_iter_no_change=10,\n",
       "                                            nesterovs_momentum=True,\n",
       "                                            power_t=0.5, random_state=42,\n",
       "                                            shuffle=True, solver='adam',\n",
       "                                            tol=0.0001, validation_fraction=0.1,\n",
       "                                            verbose=False, warm_start=False))],\n",
       "                 flatten_transform=True, n_jobs=None, voting='hard',\n",
       "                 weights=None)"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "voting_clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9611"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "voting_clf.score(X_val, y_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.0, 0.0, 0.0, 0.0]"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[estimator.score(X_val, y_val) for estimator in voting_clf.estimators_]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "SVM 모델을 제거해서 성능이 향상되는지 확인해 보죠. 다음과 같이 `set_params()`를 사용하여 `None`으로 지정하면 특정 예측기를  제외시킬 수 있습니다:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "VotingClassifier(estimators=[('random_forest_clf',\n",
       "                              RandomForestClassifier(bootstrap=True,\n",
       "                                                     class_weight=None,\n",
       "                                                     criterion='gini',\n",
       "                                                     max_depth=None,\n",
       "                                                     max_features='auto',\n",
       "                                                     max_leaf_nodes=None,\n",
       "                                                     min_impurity_decrease=0.0,\n",
       "                                                     min_impurity_split=None,\n",
       "                                                     min_samples_leaf=1,\n",
       "                                                     min_samples_split=2,\n",
       "                                                     min_weight_fraction_leaf=0.0,\n",
       "                                                     n_estimators=10,\n",
       "                                                     n_jobs=None,\n",
       "                                                     oob_score=False,\n",
       "                                                     random_sta...\n",
       "                                            beta_2=0.999, early_stopping=False,\n",
       "                                            epsilon=1e-08,\n",
       "                                            hidden_layer_sizes=(100,),\n",
       "                                            learning_rate='constant',\n",
       "                                            learning_rate_init=0.001,\n",
       "                                            max_iter=200, momentum=0.9,\n",
       "                                            n_iter_no_change=10,\n",
       "                                            nesterovs_momentum=True,\n",
       "                                            power_t=0.5, random_state=42,\n",
       "                                            shuffle=True, solver='adam',\n",
       "                                            tol=0.0001, validation_fraction=0.1,\n",
       "                                            verbose=False, warm_start=False))],\n",
       "                 flatten_transform=True, n_jobs=None, voting='hard',\n",
       "                 weights=None)"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "voting_clf.set_params(svm_clf=None)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "예측기 목록이 업데이트되었습니다:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('random_forest_clf',\n",
       "  RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
       "                         max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
       "                         min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                         min_samples_leaf=1, min_samples_split=2,\n",
       "                         min_weight_fraction_leaf=0.0, n_estimators=10,\n",
       "                         n_jobs=None, oob_score=False, random_state=42, verbose=0,\n",
       "                         warm_start=False)),\n",
       " ('extra_trees_clf',\n",
       "  ExtraTreesClassifier(bootstrap=False, class_weight=None, criterion='gini',\n",
       "                       max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
       "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                       min_samples_leaf=1, min_samples_split=2,\n",
       "                       min_weight_fraction_leaf=0.0, n_estimators=10, n_jobs=None,\n",
       "                       oob_score=False, random_state=42, verbose=0,\n",
       "                       warm_start=False)),\n",
       " ('svm_clf', None),\n",
       " ('mlp_clf',\n",
       "  MLPClassifier(activation='relu', alpha=0.0001, batch_size='auto', beta_1=0.9,\n",
       "                beta_2=0.999, early_stopping=False, epsilon=1e-08,\n",
       "                hidden_layer_sizes=(100,), learning_rate='constant',\n",
       "                learning_rate_init=0.001, max_iter=200, momentum=0.9,\n",
       "                n_iter_no_change=10, nesterovs_momentum=True, power_t=0.5,\n",
       "                random_state=42, shuffle=True, solver='adam', tol=0.0001,\n",
       "                validation_fraction=0.1, verbose=False, warm_start=False))]"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "voting_clf.estimators"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "하지만 훈련된 예측기 목록은 업데이트되지 않습니다:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
       "                        max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
       "                        min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                        min_samples_leaf=1, min_samples_split=2,\n",
       "                        min_weight_fraction_leaf=0.0, n_estimators=10,\n",
       "                        n_jobs=None, oob_score=False, random_state=42, verbose=0,\n",
       "                        warm_start=False),\n",
       " ExtraTreesClassifier(bootstrap=False, class_weight=None, criterion='gini',\n",
       "                      max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
       "                      min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                      min_samples_leaf=1, min_samples_split=2,\n",
       "                      min_weight_fraction_leaf=0.0, n_estimators=10, n_jobs=None,\n",
       "                      oob_score=False, random_state=42, verbose=0,\n",
       "                      warm_start=False),\n",
       " LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,\n",
       "           intercept_scaling=1, loss='squared_hinge', max_iter=10000,\n",
       "           multi_class='ovr', penalty='l2', random_state=42, tol=0.0001,\n",
       "           verbose=0),\n",
       " MLPClassifier(activation='relu', alpha=0.0001, batch_size='auto', beta_1=0.9,\n",
       "               beta_2=0.999, early_stopping=False, epsilon=1e-08,\n",
       "               hidden_layer_sizes=(100,), learning_rate='constant',\n",
       "               learning_rate_init=0.001, max_iter=200, momentum=0.9,\n",
       "               n_iter_no_change=10, nesterovs_momentum=True, power_t=0.5,\n",
       "               random_state=42, shuffle=True, solver='adam', tol=0.0001,\n",
       "               validation_fraction=0.1, verbose=False, warm_start=False)]"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "voting_clf.estimators_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`VotingClassifier`를 다시 훈련시키거나 그냥 훈련된 예측기 목록에서 SVM 모델을 제거할 수 있습니다:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "del voting_clf.estimators_[2]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`VotingClassifier`를 다시 평가해 보죠:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9639"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "voting_clf.score(X_val, y_val)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "훨씬 나아졌네요! SVM 모델이 성능을 저하시켰습니다. 이제 간접 투표 분류기를 사용해 보죠. 분류기를 다시 훈련시킬 필요는 없고 `voting`을 `\"soft\"`로 지정하면 됩니다:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [],
   "source": [
    "voting_clf.voting = \"soft\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9694"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "voting_clf.score(X_val, y_val)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "성능이 많이 나아졌고 개개의 분류기보다 훨씬 좋습니다."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*문제: 앙상블을 얻고 나면 테스트 세트로 확인해보세요. 개개의 분류기와 비교해서 성능이 얼마나 향상되나요?*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.972"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "voting_clf.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.0, 0.0, 0.0]"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[estimator.score(X_test, y_test) for estimator in voting_clf.estimators_]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "투표 기반 분류기는 에러율을 제일 좋은 모델(`MLPClassifier`)의 4.9%에서 3.5%로 줄였습니다. 약 28%나 오류가 적습니다. 나쁘지 않네요!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 9. 스태킹 앙상블"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*문제: 이전 연습문제의 각 분류기를 실행해서 검증 세트에서 예측을 만들고 그 결과로 새로운 훈련 세트를 만들어보세요. 각 훈련 샘플은 하나의 이미지에 대한 전체 분류기의 예측을 담은 벡터고 타깃은 이미지의 클래스입니다. 새로운 훈련 세트에 분류기 하나를 훈련시켜 보세요.*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_val_predictions = np.empty((len(X_val), len(estimators)), dtype=np.float32)\n",
    "\n",
    "for index, estimator in enumerate(estimators):\n",
    "    X_val_predictions[:, index] = estimator.predict(X_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[5., 5., 5., 5.],\n",
       "       [8., 8., 8., 8.],\n",
       "       [2., 2., 2., 2.],\n",
       "       ...,\n",
       "       [7., 7., 7., 7.],\n",
       "       [6., 6., 6., 6.],\n",
       "       [7., 7., 7., 7.]], dtype=float32)"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_val_predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
       "                       max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
       "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                       min_samples_leaf=1, min_samples_split=2,\n",
       "                       min_weight_fraction_leaf=0.0, n_estimators=200,\n",
       "                       n_jobs=None, oob_score=True, random_state=42, verbose=0,\n",
       "                       warm_start=False)"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rnd_forest_blender = RandomForestClassifier(n_estimators=200, oob_score=True, random_state=42)\n",
    "rnd_forest_blender.fit(X_val_predictions, y_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9588"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rnd_forest_blender.oob_score_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "이 블렌더를 세밀하게 튜닝하거나 다른 종류의 블렌더(예를 들어, MLPClassifier)를 시도해 볼 수 있습니다. 그런 늘 하던대로 다음 교차 검증을 사용해 가장 좋은 것을 선택합니다."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*문제: 축하합니다. 방금 블렌더를 훈련시켰습니다. 그리고 이 분류기를 모아서 스태킹 앙상블을 구성했습니다. 이제 테스트 세트에 앙상블을 평가해보세요. 테스트 세트의 각 이미지에 대해 모든 분류기로 예측을 만들고 앙상블의 예측 결과를 만들기 위해 블렌더에 그 예측을 주입합니다. 앞서 만든 투표 분류기와 비교하면 어떤가요?*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test_predictions = np.empty((len(X_test), len(estimators)), dtype=np.float32)\n",
    "\n",
    "for index, estimator in enumerate(estimators):\n",
    "    X_test_predictions[:, index] = estimator.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = rnd_forest_blender.predict(X_test_predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import accuracy_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9619"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy_score(y_test, y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "이 스태킹 앙상블은 앞서 만든 간접 투표 분류기만큼 성능을 내지는 못합니다. 하지만 개개의 분류기보다는 당연히 좋습니다."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  },
  "nav_menu": {
   "height": "252px",
   "width": "333px"
  },
  "toc": {
   "navigate_menu": true,
   "number_sections": true,
   "sideBar": true,
   "threshold": 6,
   "toc_cell": false,
   "toc_section_display": "block",
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}