[Python] multiprocessing hangs xgboost scoring

I am trying to score multiple datasets at the same time using multiprocessing. The following code hangs on run, yet when I run score on the base_model outside of the pool it executes instantly.

import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split

titanic = pd.read_csv('titanic.csv')
titanic['Age'].fillna(titanic['Age'].mean(), inplace=True) #fill all missing values with the average
#create indicator for cabin because there is areason it was missing
titanic['Cabin_ind'] = np.where(titanic['Cabin'].isnull(), 0, 1)
#convert sex to numeric
gender_num = {'male': 0, 'female': 1}
titanic['Sex'] = titanic['Sex'].map(gender_num)
#drop uncessary variables
titanic.drop(columns=['Cabin', 'Embarked', 'Name', 'Ticket'], inplace=True)
features = titanic.drop(columns=['Survived'])
labels = titanic['Survived']
X_train, X_test, y_train, y_test = train_test_split(features, labels, test_size=.2, random_state=42)

from xgboost import XGBClassifier

base_model = XGBClassifier(objective='binary:logistic', nthread=4, seed=27, scoring='auc')
base_model.fit(X_train, y_train)

import multiprocessing as mp

def my_func(X_test, y_test, base_model):
val = base_model.score(X_test, y_test)
print(val)

def main():
to_pass = [(X_test, y_test, base_model)]
pool = mp.Pool(1)
pool.starmap(my_func, to_pass)


if __name__ == "__main__":
main()

Continue reading...