ValueError: min_samples_split must be at least 2 or in (0, 1], got 1

[avaranovich]

Description

Getting the following error, while trying to specify the min_samples_split parameter, and trying to fit (some featues of X are int, some are float. The type of y is float.)
ValueError: min_samples_split must be at least 2 or in (0, 1], got 1

Steps/Code to Reproduce

from sklearn.grid_search import GridSearchCV
from sklearn.ensemble import GradientBoostingRegressor
from sklearn.metrics import mean_squared_error
from time import time

tuned_params = {'n_estimators': [5000, 10000], 
                'max_depth': [2, 4], 
                'min_samples_split': [1, 2],
                'learning_rate': [0.001, 0.01]}

print("# Tuning hyper-parameters for mean_squared_error")

gscv = GridSearchCV(GradientBoostingRegressor(loss = 'ls', random_state=0), 
                    tuned_params, cv=5, scoring='mean_squared_error')

gscv.fit(X, y)

Versions

Linux-4.4.0-51-generic-x86_64-with-debian-stretch-sid
Python 3.5.2 |Anaconda 4.1.1 (64-bit)| (default, Jul 2 2016, 17:53:06)
[GCC 4.4.7 20120313 (Red Hat 4.4.7-1)]
NumPy 1.11.1
SciPy 0.17.1
Scikit-Learn 0.18.1

[jnothman]

you can't split on a minimum of 1 sample. You can split on a minimum of 1.0 x n_samples samples. That error message should say (0.0, 1.0] (fixed in 96cc49f). If you intended 1.0 you should have used that, not 1.