Closed
Description
Describe the bug
I am not sure if this counts as a bug or a regression (or whether it was discussed and purposefully done, but I can't see any documentation of it).
Previously (sklearn version 0.20.1) when running fit on a BaseSearchCV object with verbosity above 2, we would get output detailing the test scores for each fold. Under current sklearn (0.24.1), if we don't supply scorers through a dictionary, there is no output regarding fold scores. It can be seen that this is the case here
I assume intended logic was
if verbose > 2:
if isinstance(test_scores, dict):
# handle multiple scores
else:
# handle single score
...
Steps/Code to Reproduce
Run RandomizedSearchCV or GridSearchCV fit with verbose > 2 and some provided arg for scoring.
Example:
import numpy as np
from sklearn.model_selection import GridSearchCV
X = np.array([
[0.1, 0.2, 0.3],
[0.1, 0.2, 0.3],
])
y = np.array([1, 2])
from sklearn.linear_model import LinearRegression
cross_validator = GridSearchCV(
estimator=LinearRegression(),
param_grid={"fit_intercept": [True, False]},
scoring="neg_mean_squared_error",
verbose=3,
cv=2,
refit=False,
)
cross_validator.fit(X, y)Expected Results
Fitting 2 folds for each of 2 candidates, totalling 4 fits
[CV 1/2] END .............................fit_intercept=True; score=-1.0; total time= 0.0s
[CV 2/2] END .............................fit_intercept=True; score=-1.0; total time= 0.0s
[CV 1/2] END ............................fit_intercept=False; score=-1.0; total time= 0.0s
[CV 2/2] END ............................fit_intercept=False; score=-1.0; total time= 0.0s
Actual Results
Fitting 2 folds for each of 2 candidates, totalling 4 fits
[CV 1/2] END .............................fit_intercept=True; total time= 0.0s
[CV 2/2] END .............................fit_intercept=True; total time= 0.0s
[CV 1/2] END ............................fit_intercept=False; total time= 0.0s
[CV 2/2] END ............................fit_intercept=False; total time= 0.0s
Versions
Darwin-19.6.0-x86_64-i386-64bit
Python 3.7.4 (default, May 4 2020, 18:19:10)
[Clang 11.0.3 (clang-1103.0.32.59)]
NumPy 1.20.1
SciPy 1.6.1
Scikit-Learn 0.24.1