Remove check of correct shape of predictions which becomes too complicated with the new output types supported.

Signed-off-by: abigailt <abigailt@il.ibm.com>
This commit is contained in:
abigailt 2024-05-02 18:56:31 +03:00
parent a4816878f9
commit 846de0f753
6 changed files with 27 additions and 25 deletions

View file

@ -227,14 +227,14 @@ def test_blackbox_classifier_predictions_multi_label_binary():
assert model.model_type is None
def test_blackbox_classifier_mismatch():
(x_train, y_train), (x_test, y_test) = dataset_utils.get_iris_dataset_np()
train = ArrayDataset(x_train, y_train)
test = ArrayDataset(x_test, y_test)
data = Data(train, test)
with pytest.raises(ValueError):
BlackboxClassifierPredictions(data, ModelOutputType.CLASSIFIER_SINGLE_OUTPUT_CLASS_PROBABILITIES)
# def test_blackbox_classifier_mismatch():
# (x_train, y_train), (x_test, y_test) = dataset_utils.get_iris_dataset_np()
#
# train = ArrayDataset(x_train, y_train)
# test = ArrayDataset(x_test, y_test)
# data = Data(train, test)
# with pytest.raises(ValueError):
# BlackboxClassifierPredictions(data, ModelOutputType.CLASSIFIER_SINGLE_OUTPUT_CLASS_PROBABILITIES)
def test_blackbox_classifier_no_test():