HGBC with categorical_crossentropy fails silently on binary classification #14858
Description
import numpy as np
from sklearn.experimental import enable_hist_gradient_boosting
from sklearn.ensemble import HistGradientBoostingClassifier
X = [[1, 0],
[1, 0],
[1, 0],
[0, 1],
[1, 1]]
y = [1, 1, 1, 0, 1]
gb = HistGradientBoostingClassifier(loss='categorical_crossentropy',
min_samples_leaf=1)
gb.fit(X, y)
print(gb.predict([[1, 0]]))
print(gb.predict([[0, 1]]))gives:
[0]
[0]
And binary_crossentropy works fine. categorical_crossentropy should either generalize or raise an error on binary classification.
Ping @NicolasHug @ogrisel