HistGradientBoostingClassifier fails for early stopping on the test set with warm start #16661
Description
Describe the bug
I found a combination of input arguments to HistGradientBoostingClassifier which causes it to fail for two repeated calls to fit when warm_start is turned on and it uses the training set for early stopping.
Steps/Code to Reproduce
import sklearn.ensemble
import sklearn.datasets
from sklearn.experimental import enable_hist_gradient_boosting
X, y = sklearn.datasets.load_wine(return_X_y=True)
gb = sklearn.ensemble.HistGradientBoostingClassifier(
max_iter=1000,
scoring='loss',
warm_start=True,
n_iter_no_change=1,
validation_fraction=None,
)
gb.fit(X, y)
gb.fit(X, y)Expected Results
No error is thrown.
Actual Results
Traceback (most recent call last):
File "/home/feurerm/sync_dir/projects/automl_competition_2015/auto-sklearn/test.py", line 16, in <module>
gb.fit(X, y)
File "/home/feurerm/miniconda/3-4.5.4/envs/autosklearn/lib/python3.7/site-packages/sklearn/ensemble/_hist_gradient_boosting/gradient_boosting.py", line 356, in fit
raw_predictions_val, y_val
UnboundLocalError: local variable 'raw_predictions_val' referenced before assignment
Versions
System:
python: 3.7.1 (default, Dec 14 2018, 19:28:38) [GCC 7.3.0]
executable: /home/feurerm/miniconda/3-4.5.4/envs/autosklearn/bin/python3.7
machine: Linux-4.15.0-88-generic-x86_64-with-debian-buster-sid
Python dependencies:
pip: 10.0.1
setuptools: 39.2.0
sklearn: 0.22.2
numpy: 1.14.5
scipy: 1.2.0
Cython: 0.28.4
pandas: 0.25.0
matplotlib: None
joblib: 0.12.1
Built with OpenMP: True