Closed
Description
Description
Once transformer fit is done here and again here during fit_transform. I don't think this is intended as fitting can be costly and there is no reason to repeat it for the same data.
Steps/Code to Reproduce
from sklearn.base import TransformerMixin, BaseEstimator
from sklearn.compose import TransformedTargetRegressor
from sklearn.tree import DecisionTreeRegressor
from sklearn import datasets
class DummyTransformer(BaseEstimator, TransformerMixin):
def __init__(self):
pass
def fit(self, X):
print('in fit')
return self
def transform(self, X):
print('in transform')
return X
def inverse_transform(self, X):
print('in inverse_transform')
return X
estimator = TransformedTargetRegressor(
regressor=DecisionTreeRegressor(),
transformer=DummyTransformer(),
check_inverse = False
)
X, y = datasets.load_linnerud(return_X_y=True)
estimator.fit(X, y)returns:
in fit
in fit
in transform
Expected Results
in fit
in transform
Metadata
Metadata
Assignees
Labels
No labels