thomasjpfan

How to run benchmarks

Run pure python version

0. Set up a virtual env.
1. git clone -b ccp_prune_tree https://github.com/thomasjpfan/scikit-learn scikit-learn-ccp-python
2. cd scikit-learn-ccp-python
3. git checkout 25910e085cf7bb0a98ee33c050fa9233e247e523
4. Install scikit-learn
5. Go to directory with bench_prune_tree.py

thomasjpfan

#!/usr/bin/env python3

# pip install GitPython PyGithub

import argparse
import os
import sys

from github import Github
from git import Repo

thomasjpfan

from sklearn.tree import DecisionTreeRegressor
from sklearn.linear_model import Ridge
from sklearn.preprocessing import StandardScaler
from sklearn.pipeline import Pipeline
from sklearn.model_selection import GridSearchCV
from sklearn.datasets import load_boston
from sklearn.model_selection import train_test_split

boston = load_boston()
X_train, X_test, y_train, y_test = train_test_split(

thomasjpfan

thomasjpfan

For reference here are the time it takes to run the following:

OMP_NUM_THREADS=$i python benchmarks/bench_hist_gradient_boosting_higgsboson.py \
    --n-leaf-nodes 255 --n-trees 100

lightgbm==2.2.1
xgboost==0.90

thomasjpfan

thomasjpfan

thomasjpfan

{
  "TagRemovePreprocessor": { "remove_input_tags": ["to_remove"] },
  "SlidesExporter": { "reveal_theme": "simple" },
  "NbConvertApp": {
    "export_format": "slides"
  },
  "TemplateExporter": {
    "exclude_input_prompt": true,
    "exclude_output_prompt": true
  }

thomasjpfan

import sklearn
import numpy as np
import scipy
import csv
import argparse
from pathlib import Path

import openml
from openml.exceptions import OpenMLRunsExistError
from sklearn.experimental import enable_hist_gradient_boosting

thomasjpfan