model_evaluation_speed/model_evaluation_benchmark.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import lightgbm as lgb\n",
    "import pandas as pd\n",
    "import catboost as cb\n",
    "import xgboost as xgb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import catboost.datasets"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Load data\n",
    "===="
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "epsilon_test = catboost.datasets.epsilon()[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test = epsilon_test.drop(0, axis=1).values\n",
    "xgb_test = xgb.DMatrix(X_test)\n",
    "catboost_pool = cb.Pool(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8000\n"
     ]
    }
   ],
   "source": [
    "epsilon_model = cb.CatBoost()\n",
    "epsilon_model.load_model('epsilon8k_64.bin')\n",
    "print(epsilon_model.tree_count_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "32 thread apply\n",
    "===="
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "bst = xgb.Booster({'nthread': 32})  # set 32 thread in openmp threadpool\n",
    "bst.load_model('XGBoost.model')  # load data\n",
    "lgb_model = lgb.Booster(model_file='LightGBM_model.txt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4.89 s ± 223 ms per loop (mean ± std. dev. of 5 runs, 1 loop each)\n"
     ]
    }
   ],
   "source": [
    "%timeit -r 5 bst.predict(xgb_test, ntree_limit=8000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "11.1 s ± 402 ms per loop (mean ± std. dev. of 5 runs, 1 loop each)\n"
     ]
    }
   ],
   "source": [
    "%timeit -r 5 lgb_model.predict(X_test, num_iteration=8000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "184 ms ± 806 µs per loop (mean ± std. dev. of 5 runs, 1 loop each)\n"
     ]
    }
   ],
   "source": [
    "%timeit -r 5 epsilon_model.predict(catboost_pool, thread_count=32, ntree_end=8000)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1 thread apply\n",
    "===="
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "bst = xgb.Booster({'nthread': 1})  # set 1 thread in openmp threadpool\n",
    "bst.load_model('XGBoost.model')  # load data\n",
    "lgb_model = lgb.Booster(model_file='LightGBM_model.txt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1min 12s, sys: 20 ms, total: 1min 12s\n",
      "Wall time: 1min 12s\n"
     ]
    }
   ],
   "source": [
    "%time _ = bst.predict(xgb_test, ntree_limit=8000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1min 53s, sys: 2.72 s, total: 1min 56s\n",
      "Wall time: 1min 56s\n"
     ]
    }
   ],
   "source": [
    "%time _ = lgb_model.predict(X_test, num_iteration=8000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1.96 s, sys: 24 ms, total: 1.98 s\n",
      "Wall time: 1.98 s\n"
     ]
    }
   ],
   "source": [
    "%time _ = epsilon_model.predict(catboost_pool, thread_count=1, ntree_end=8000)"
   ]
  },
  {
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}