07_ensemble_learning_and_random_forests.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Chapter 7 – Ensemble Learning and Random Forests**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "_This notebook contains all the sample code and solutions to the exercises in chapter 7._"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<table align=\"left\">\n",
    "  <td>\n",
    "    <a target=\"_blank\" href=\"https://colab.research.google.com/github/ageron/handson-ml2/blob/master/07_ensemble_learning_and_random_forests.ipynb\"><img src=\"https://www.tensorflow.org/images/colab_logo_32px.png\" />Run in Google Colab</a>\n",
    "  </td>\n",
    "</table>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Setup"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, let's import a few common modules, ensure MatplotLib plots figures inline and prepare a function to save the figures. We also check that Python 3.5 or later is installed (although Python 2.x may work, it is deprecated so we strongly recommend you use Python 3 instead), as well as Scikit-Learn ≥0.20."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Python ≥3.5 is required\n",
    "import sys\n",
    "assert sys.version_info >= (3, 5)\n",
    "\n",
    "# Scikit-Learn ≥0.20 is required\n",
    "import sklearn\n",
    "assert sklearn.__version__ >= \"0.20\"\n",
    "\n",
    "# Common imports\n",
    "import numpy as np\n",
    "import os\n",
    "\n",
    "# to make this notebook's output stable across runs\n",
    "np.random.seed(42)\n",
    "\n",
    "# To plot pretty figures\n",
    "%matplotlib inline\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "mpl.rc('axes', labelsize=14)\n",
    "mpl.rc('xtick', labelsize=12)\n",
    "mpl.rc('ytick', labelsize=12)\n",
    "\n",
    "# Where to save the figures\n",
    "PROJECT_ROOT_DIR = \".\"\n",
    "CHAPTER_ID = \"ensembles\"\n",
    "IMAGES_PATH = os.path.join(PROJECT_ROOT_DIR, \"images\", CHAPTER_ID)\n",
    "os.makedirs(IMAGES_PATH, exist_ok=True)\n",
    "\n",
    "def save_fig(fig_id, tight_layout=True, fig_extension=\"png\", resolution=300):\n",
    "    path = os.path.join(IMAGES_PATH, fig_id + \".\" + fig_extension)\n",
    "    print(\"Saving figure\", fig_id)\n",
    "    if tight_layout:\n",
    "        plt.tight_layout()\n",
    "    plt.savefig(path, format=fig_extension, dpi=resolution)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Voting classifiers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving figure law_of_large_numbers_plot\n"
     ]
    },
    {
     "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(\"Number of coin tosses\")\n",
    "plt.ylabel(\"Heads ratio\")\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": 4,
   "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": [
    "**Note**: to be future-proof, we set `solver=\"lbfgs\"`, `n_estimators=100`, and `gamma=\"scale\"` since these will be the default values in upcoming Scikit-Learn versions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "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=\"lbfgs\", random_state=42)\n",
    "rnd_clf = RandomForestClassifier(n_estimators=100, random_state=42)\n",
    "svm_clf = SVC(gamma=\"scale\", random_state=42)\n",
    "\n",
    "voting_clf = VotingClassifier(\n",
    "    estimators=[('lr', log_clf), ('rf', rnd_clf), ('svc', svm_clf)],\n",
    "    voting='hard')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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='lbfgs', tol=0.0001,\n",
       "                                                 verbose=0, warm_start=False)),\n",
       "                             ('rf',\n",
       "                              RandomForestClassifier(bootstrap=True,\n",
       "                                                     class_weight=None,\n",
       "                                                     criterion='gini',\n",
       "                                                     m...\n",
       "                                                     n_estimators=100,\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='scale', 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": [
    "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 0.896\n",
      "SVC 0.896\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": [
    "Soft voting:"
   ]
  },
  {
   "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='lbfgs', tol=0.0001,\n",
       "                                                 verbose=0, warm_start=False)),\n",
       "                             ('rf',\n",
       "                              RandomForestClassifier(bootstrap=True,\n",
       "                                                     class_weight=None,\n",
       "                                                     criterion='gini',\n",
       "                                                     m...\n",
       "                                                     n_estimators=100,\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='scale', 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=\"lbfgs\", random_state=42)\n",
    "rnd_clf = RandomForestClassifier(n_estimators=100, random_state=42)\n",
    "svm_clf = SVC(gamma=\"scale\", 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 0.896\n",
      "SVC 0.896\n",
      "VotingClassifier 0.92\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": [
    "# Bagging ensembles"
   ]
  },
  {
   "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, 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, 2.45, -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": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving figure decision_tree_without_and_with_bagging_plot\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fix, axes = plt.subplots(ncols=2, figsize=(10,4), sharey=True)\n",
    "plt.sca(axes[0])\n",
    "plot_decision_boundary(tree_clf, X, y)\n",
    "plt.title(\"Decision Tree\", fontsize=14)\n",
    "plt.sca(axes[1])\n",
    "plot_decision_boundary(bag_clf, X, y)\n",
    "plt.title(\"Decision Trees with Bagging\", fontsize=14)\n",
    "plt.ylabel(\"\")\n",
    "save_fig(\"decision_tree_without_and_with_bagging_plot\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Random Forests"
   ]
  },
  {
   "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, 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, 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)  # almost identical predictions"
   ]
  },
  {
   "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, 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, 2.45, -1, 1.5], alpha=0.02, contour=False)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Out-of-Bag evaluation"
   ]
  },
  {
   "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, 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": [
    "## Feature importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import fetch_openml\n",
    "\n",
    "mnist = fetch_openml('mnist_784', version=1)\n",
    "mnist.target = mnist.target.astype(np.uint8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "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=100,\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=100, 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 = mpl.cm.hot,\n",
    "               interpolation=\"nearest\")\n",
    "    plt.axis(\"off\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving figure mnist_feature_importance_plot\n"
     ]
    },
    {
     "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(['Not important', 'Very important'])\n",
    "\n",
    "save_fig(\"mnist_feature_importance_plot\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# AdaBoost"
   ]
  },
  {
   "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": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving figure boosting_plot\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "m = len(X_train)\n",
    "\n",
    "fix, axes = plt.subplots(ncols=2, figsize=(10,4), sharey=True)\n",
    "for subplot, learning_rate in ((0, 1), (1, 0.5)):\n",
    "    sample_weights = np.ones(m)\n",
    "    plt.sca(axes[subplot])\n",
    "    for i in range(5):\n",
    "        svm_clf = SVC(kernel=\"rbf\", C=0.05, gamma=\"scale\", 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",
    "    if subplot == 0:\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",
    "    else:\n",
    "        plt.ylabel(\"\")\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": [
    "# Gradient Boosting"
   ]
  },
  {
   "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": [],
   "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)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving figure gradient_boosting_plot\n"
     ]
    },
    {
     "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": [
    "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=\"Training set\")\n",
    "plt.ylabel(\"$y$\", fontsize=16, rotation=0)\n",
    "plt.title(\"Residuals and tree predictions\", 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=\"Training set\")\n",
    "plt.ylabel(\"$y$\", fontsize=16, rotation=0)\n",
    "plt.title(\"Ensemble predictions\", 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=\"Residuals\")\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": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GradientBoostingRegressor(alpha=0.9, criterion='friedman_mse', init=None,\n",
       "                          learning_rate=1.0, 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": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "\n",
    "gbrt = GradientBoostingRegressor(max_depth=2, n_estimators=3, learning_rate=1.0, random_state=42)\n",
    "gbrt.fit(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "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": 43,
     "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": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving figure gbrt_learning_rate_plot\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fix, axes = plt.subplots(ncols=2, figsize=(10,4), sharey=True)\n",
    "\n",
    "plt.sca(axes[0])\n",
    "plot_predictions([gbrt], X, y, axes=[-0.5, 0.5, -0.1, 0.8], label=\"Ensemble predictions\")\n",
    "plt.title(\"learning_rate={}, n_estimators={}\".format(gbrt.learning_rate, gbrt.n_estimators), fontsize=14)\n",
    "plt.xlabel(\"$x_1$\", fontsize=16)\n",
    "plt.ylabel(\"$y$\", fontsize=16, rotation=0)\n",
    "\n",
    "plt.sca(axes[1])\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",
    "plt.xlabel(\"$x_1$\", fontsize=16)\n",
    "\n",
    "save_fig(\"gbrt_learning_rate_plot\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Gradient Boosting with Early stopping"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "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=56,\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": 45,
     "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) + 1\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": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "min_error = np.min(errors)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving figure early_stopping_gbrt_plot\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10, 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, \"Minimum\", ha=\"center\", fontsize=14)\n",
    "plt.axis([0, 120, 0, 0.01])\n",
    "plt.xlabel(\"Number of trees\")\n",
    "plt.ylabel(\"Error\", fontsize=16)\n",
    "plt.title(\"Validation error\", 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(\"Best model (%d trees)\" % bst_n_estimators, fontsize=14)\n",
    "plt.ylabel(\"$y$\", fontsize=16, rotation=0)\n",
    "plt.xlabel(\"$x_1$\", fontsize=16)\n",
    "\n",
    "save_fig(\"early_stopping_gbrt_plot\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "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  # early stopping"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "61\n"
     ]
    }
   ],
   "source": [
    "print(gbrt.n_estimators)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Minimum validation MSE: 0.002712853325235463\n"
     ]
    }
   ],
   "source": [
    "print(\"Minimum validation MSE:\", min_val_error)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Using XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "try:\n",
    "    import xgboost\n",
    "except ImportError as ex:\n",
    "    print(\"Error: the xgboost library is not installed.\")\n",
    "    xgboost = None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "Validation MSE: 0.0028512559726563943\n"
     ]
    }
   ],
   "source": [
    "if xgboost is not None:  # not shown in the book\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) # Not shown\n",
    "    print(\"Validation MSE:\", val_error)           # Not shown"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/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",
      "Validation MSE: 0.002528626115371327\n"
     ]
    }
   ],
   "source": [
    "if xgboost is not None:  # not shown in the book\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)  # Not shown\n",
    "    print(\"Validation MSE:\", val_error)            # Not shown"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:06] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:07] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:08] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:09] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "[15:53:10] WARNING: /home/conda/feedstock_root/build_artifacts/xgboost_1572314959925/work/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "4.42 ms ± 54.9 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": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "13.2 ms ± 24.4 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": {},
   "source": [
    "# My exercise solutions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. to 7."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "These were theoretical questions; I answered them and looked at the book for the answers."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 8."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Exercise: Load the MNIST data and split it into a training set, a validation set, and a test set (e.g., use 50,000 instances for training, 10,000 for validation, and 10,000 for testing). Then train various classifiers, such as a Random Forest classifier, an Extra-Trees classifier, and an SVM classifier. Next, try to combine them into an ensemble that outperforms each individual classifier on the validation set, using soft or hard voting. Once you have found one, try it on the test set. How much better does it perform compared to the individual classifiers?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import fetch_openml\n",
    "\n",
    "mnist = fetch_openml('mnist_784', version=1)\n",
    "mnist.target = mnist.target.astype(np.uint8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "# taken from https://datascience.stackexchange.com/questions/15135/train-test-validation-set-splitting-in-sklearn\n",
    "X_train, X_test, y_train, y_test = train_test_split(mnist[\"data\"], mnist.target, test_size=0.2, random_state=1)\n",
    "X_train, X_val, y_train, y_val = train_test_split(X_train, y_train, test_size=0.25, random_state=1) # 0.25 x 0.8 = 0.2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/mislav/anaconda3/envs/tf2/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Pipeline(memory=None,\n",
       "         steps=[('standardscaler',\n",
       "                 StandardScaler(copy=True, with_mean=True, with_std=True)),\n",
       "                ('svc',\n",
       "                 SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,\n",
       "                     decision_function_shape='ovr', degree=3, gamma='auto',\n",
       "                     kernel='rbf', max_iter=-1, probability=False,\n",
       "                     random_state=None, shrinking=True, tol=0.001,\n",
       "                     verbose=False))],\n",
       "         verbose=False)"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.ensemble import ExtraTreesClassifier\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.pipeline import make_pipeline\n",
    "\n",
    "random_forest_classifier = RandomForestClassifier(max_depth=2, random_state=0)\n",
    "random_forest_classifier.fit(X_train, y_train)\n",
    "\n",
    "extra_trees_classifier = ExtraTreesClassifier(n_estimators=100, random_state=0)\n",
    "extra_trees_classifier.fit(X_train, y_train)\n",
    "\n",
    "svm_classifier = make_pipeline(StandardScaler(), SVC(gamma='auto'))\n",
    "svm_classifier.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now let's evaluate each individual classifier on the validation set:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5578571428571428\n",
      "0.9673571428571428\n",
      "0.9599285714285715\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score\n",
    "\n",
    "preds_random_forest_classifier = random_forest_classifier.predict(X_val)\n",
    "preds_extra_trees_classifier = extra_trees_classifier.predict(X_val)\n",
    "preds_svm_classifier = svm_classifier.predict(X_val)\n",
    "\n",
    "print(accuracy_score(y_val, preds_random_forest_classifier))\n",
    "print(accuracy_score(y_val, preds_extra_trees_classifier))\n",
    "print(accuracy_score(y_val, preds_svm_classifier))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's look at each individual classifier on the test set:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5552142857142857\n",
      "0.9690714285714286\n",
      "0.9596428571428571\n"
     ]
    }
   ],
   "source": [
    "preds_random_forest_classifier = random_forest_classifier.predict(X_test)\n",
    "preds_extra_trees_classifier = extra_trees_classifier.predict(X_test)\n",
    "preds_svm_classifier = svm_classifier.predict(X_test)\n",
    "\n",
    "print(accuracy_score(y_test, preds_random_forest_classifier))\n",
    "print(accuracy_score(y_test, preds_extra_trees_classifier))\n",
    "print(accuracy_score(y_test, preds_svm_classifier))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's now construct an ensemble:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "VotingClassifier(estimators=[('random_forest_classifier',\n",
       "                              RandomForestClassifier(bootstrap=True,\n",
       "                                                     class_weight=None,\n",
       "                                                     criterion='gini',\n",
       "                                                     max_depth=2,\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...\n",
       "                                               StandardScaler(copy=True,\n",
       "                                                              with_mean=True,\n",
       "                                                              with_std=True)),\n",
       "                                              ('svc',\n",
       "                                               SVC(C=1.0, cache_size=200,\n",
       "                                                   class_weight=None, coef0=0.0,\n",
       "                                                   decision_function_shape='ovr',\n",
       "                                                   degree=3, gamma='auto',\n",
       "                                                   kernel='rbf', max_iter=-1,\n",
       "                                                   probability=False,\n",
       "                                                   random_state=None,\n",
       "                                                   shrinking=True, tol=0.001,\n",
       "                                                   verbose=False))],\n",
       "                                       verbose=False))],\n",
       "                 flatten_transform=True, n_jobs=None, voting='hard',\n",
       "                 weights=None)"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import VotingClassifier\n",
    "\n",
    "ensemble = VotingClassifier(estimators=[('random_forest_classifier', random_forest_classifier), ('extra_trees_classifier', extra_trees_classifier), ('svm_classifier', svm_classifier)], voting='hard')\n",
    "\n",
    "ensemble.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's see the performance of the ensemble on the validation and the test dataset:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9590714285714286\n",
      "0.9602142857142857\n"
     ]
    }
   ],
   "source": [
    "preds_ensemble = ensemble.predict(X_val)\n",
    "print(accuracy_score(y_val, preds_ensemble))\n",
    "\n",
    "preds_ensemble = ensemble.predict(X_test)\n",
    "print(accuracy_score(y_test, preds_ensemble))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Well, the ensemble doesn't beat the extra trees classifier, but it's close enough!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 9."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Exercise: _Run the individual classifiers from the previous exercise to make predictions on the validation set, and create a new training set with the resulting predictions: each training instance is a vector containing the set of predictions from all your classifiers for an image, and the target is the image's class. Train a classifier on this new training set._"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<zip object at 0x7fe763534280>\n"
     ]
    }
   ],
   "source": [
    "preds_random_forest_classifier = random_forest_classifier.predict(X_val)\n",
    "preds_extra_trees_classifier = extra_trees_classifier.predict(X_val)\n",
    "preds_svm_classifier = svm_classifier.predict(X_val)\n",
    "\n",
    "new_training_set = zip(preds_random_forest_classifier, preds_extra_trees_classifier, preds_svm_classifier)\n",
    "print(new_training_set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Pipeline(memory=None,\n",
       "         steps=[('standardscaler',\n",
       "                 StandardScaler(copy=True, with_mean=True, with_std=True)),\n",
       "                ('svc',\n",
       "                 SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,\n",
       "                     decision_function_shape='ovr', degree=3, gamma='auto',\n",
       "                     kernel='rbf', max_iter=-1, probability=False,\n",
       "                     random_state=None, shrinking=True, tol=0.001,\n",
       "                     verbose=False))],\n",
       "         verbose=False)"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_classifier = make_pipeline(StandardScaler(), SVC(gamma='auto'))\n",
    "new_classifier.fit(list(new_training_set), y_val)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Exercise: _Congratulations, you have just trained a blender, and together with the classifiers they form a stacking ensemble! Now let's evaluate the ensemble on the test set. For each image in the test set, make predictions with all your classifiers, then feed the predictions to the blender to get the ensemble's predictions. How does it compare to the voting classifier you trained earlier?_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "preds_random_forest_classifier = random_forest_classifier.predict(X_test)\n",
    "preds_extra_trees_classifier = extra_trees_classifier.predict(X_test)\n",
    "preds_svm_classifier = svm_classifier.predict(X_test)\n",
    "\n",
    "new_training_set = zip(preds_random_forest_classifier, preds_extra_trees_classifier, preds_svm_classifier)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9599285714285715\n"
     ]
    }
   ],
   "source": [
    "new_classifier_preds = new_classifier.predict(list(new_training_set))\n",
    "print(accuracy_score(y_test, list(new_classifier_preds)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This stack classifier does approximately as good as the ensemble."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Exercise solutions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. to 7."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "See Appendix A."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 8. Voting Classifier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Exercise: _Load the MNIST data and split it into a training set, a validation set, and a test set (e.g., use 50,000 instances for training, 10,000 for validation, and 10,000 for testing)._"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The MNIST dataset was loaded earlier."
   ]
  },
  {
   "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": [
    "Exercise: _Then train various classifiers, such as a Random Forest classifier, an Extra-Trees classifier, and an SVM._"
   ]
  },
  {
   "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=100, random_state=42)\n",
    "extra_trees_clf = ExtraTreesClassifier(n_estimators=100, random_state=42)\n",
    "svm_clf = LinearSVC(random_state=42)\n",
    "mlp_clf = MLPClassifier(random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training the 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=100, n_jobs=None,\n",
      "            oob_score=False, random_state=42, verbose=0, warm_start=False)\n",
      "Training the 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=100, n_jobs=None,\n",
      "           oob_score=False, random_state=42, verbose=0, warm_start=False)\n",
      "Training the LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,\n",
      "     intercept_scaling=1, loss='squared_hinge', max_iter=1000,\n",
      "     multi_class='ovr', penalty='l2', random_state=42, tol=0.0001,\n",
      "     verbose=0)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/ageron/miniconda3/envs/tf2b/lib/python3.7/site-packages/sklearn/svm/base.py:931: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training the 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(\"Training the\", estimator)\n",
    "    estimator.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.9692, 0.9715, 0.8641, 0.9603]"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[estimator.score(X_val, y_val) for estimator in estimators]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The linear SVM is far outperformed by the other classifiers. However, let's keep it for now since it may improve the voting classifier's performance."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Exercise: _Next, try to combine them into an ensemble that outperforms them all on the validation set, using a soft or hard voting classifier._"
   ]
  },
  {
   "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": [
      "/Users/ageron/miniconda3/envs/tf2b/lib/python3.7/site-packages/sklearn/svm/base.py:931: 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', 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",
       "   ...=True, solver='adam', tol=0.0001,\n",
       "       validation_fraction=0.1, verbose=False, warm_start=False))],\n",
       "         flatten_transform=None, n_jobs=None, voting='hard', 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": [
       "0.9704"
      ]
     },
     "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": [
       "[0.9692, 0.9715, 0.8641, 0.9603]"
      ]
     },
     "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": [
    "Let's remove the SVM to see if performance improves. It is possible to remove an estimator by setting it to `None` using `set_params()` like this:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "VotingClassifier(estimators=[('random_forest_clf', 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",
       "   ...=True, solver='adam', tol=0.0001,\n",
       "       validation_fraction=0.1, verbose=False, warm_start=False))],\n",
       "         flatten_transform=None, n_jobs=None, voting='hard', weights=None)"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "voting_clf.set_params(svm_clf=None)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This updated the list of estimators:"
   ]
  },
  {
   "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=100, n_jobs=None,\n",
       "              oob_score=False, random_state=42, verbose=0, 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=100, n_jobs=None,\n",
       "             oob_score=False, random_state=42, verbose=0, 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": [
    "However, it did not update the list of _trained_ estimators:"
   ]
  },
  {
   "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=100, n_jobs=None,\n",
       "             oob_score=False, random_state=42, verbose=0, 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=100, n_jobs=None,\n",
       "            oob_score=False, random_state=42, verbose=0, 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=1000,\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": [
    "So we can either fit the `VotingClassifier` again, or just remove the SVM from the list of trained estimators:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "del voting_clf.estimators_[2]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now let's evaluate the `VotingClassifier` again:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9732"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "voting_clf.score(X_val, y_val)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A bit better! The SVM was hurting performance. Now let's try using a soft voting classifier. We do not actually need to retrain the classifier, we can just set `voting` to `\"soft\"`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [],
   "source": [
    "voting_clf.voting = \"soft\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9672"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "voting_clf.score(X_val, y_val)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Nope, hard voting wins in this case."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "_Once you have found one, try it on the test set. How much better does it perform compared to the individual classifiers?_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9704"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "voting_clf.voting = \"hard\"\n",
    "voting_clf.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.9645, 0.9691, 0.9602]"
      ]
     },
     "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": [
    "The voting classifier only very slightly reduced the error rate of the best model in this case."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 9. Stacking Ensemble"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Exercise: _Run the individual classifiers from the previous exercise to make predictions on the validation set, and create a new training set with the resulting predictions: each training instance is a vector containing the set of predictions from all your classifiers for an image, and the target is the image's class. Train a classifier on this new training set._"
   ]
  },
  {
   "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_jobs=None,\n",
       "            oob_score=True, random_state=42, verbose=0, 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": [
       "0.9696"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rnd_forest_blender.oob_score_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You could fine-tune this blender or try other types of blenders (e.g., an `MLPClassifier`), then select the best one using cross-validation, as always."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Exercise: _Congratulations, you have just trained a blender, and together with the classifiers they form a stacking ensemble! Now let's evaluate the ensemble on the test set. For each image in the test set, make predictions with all your classifiers, then feed the predictions to the blender to get the ensemble's predictions. How does it compare to the voting classifier you trained earlier?_"
   ]
  },
  {
   "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": [
       "0.9669"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy_score(y_test, y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This stacking ensemble does not perform as well as the voting classifier we trained earlier, it's not quite as good as the best individual classifier."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.8"
  },
  "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": 4
}