From 5db17521f689b2def31ed708b6c68c9f00d3ab61 Mon Sep 17 00:00:00 2001 From: David Cortes Date: Wed, 29 Oct 2025 17:02:07 +0100 Subject: [PATCH 1/3] deselect cases that fail due to expected small numeric differences --- deselected_tests.yaml | 2 ++ 1 file changed, 2 insertions(+) diff --git a/deselected_tests.yaml b/deselected_tests.yaml index 54eb99632f..38f2f3a074 100755 --- a/deselected_tests.yaml +++ b/deselected_tests.yaml @@ -374,6 +374,8 @@ deselected_tests: - ensemble/tests/test_voting.py::test_sample_weight >=1.0,<1.1 - tests/test_multioutput.py::test_multiclass_multioutput_estimator_predict_proba - model_selection/tests/test_search.py::test_search_cv_sample_weight_equivalence[estimator0] + - tests/test_calibration.py::test_calibrated_classifier_cv_double_sample_weights_equivalence < 1.1 + - tests/test_common.py::test_estimators[StackingClassifier(estimators=[('est1',LogisticRegression(C=0.1)) < 1.1 # Scikit-learn does not constraint multinomial logistic intercepts to sum to zero. # Softmax function is invariant to additions by a constant, so even though the numbers From 3133e7e8dbf771932cc799223840827069bf882c Mon Sep 17 00:00:00 2001 From: David Cortes Date: Thu, 30 Oct 2025 08:22:06 +0100 Subject: [PATCH 2/3] deduplicate --- deselected_tests.yaml | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/deselected_tests.yaml b/deselected_tests.yaml index 38f2f3a074..54d2fc729c 100755 --- a/deselected_tests.yaml +++ b/deselected_tests.yaml @@ -295,7 +295,7 @@ deselected_tests: - ensemble/tests/test_bagging.py::test_estimators_samples >=1.4 - ensemble/tests/test_voting.py::test_sample_weight >=1.4 - svm/tests/test_svm.py::test_auto_weight >=1.4 - - tests/test_calibration.py::test_calibrated_classifier_cv_double_sample_weights_equivalence >=1.4 + - tests/test_calibration.py::test_calibrated_classifier_cv_double_sample_weights_equivalence >=1 - tests/test_calibration.py::test_calibrated_classifier_cv_zeros_sample_weights_equivalence >=1.4 - tests/test_common.py::test_estimators[LogisticRegression()-check_sample_weights_invariance(kind=ones)] >=1.4 - tests/test_common.py::test_estimators[LogisticRegression()-check_sample_weights_invariance(kind=zeros)] >=1.4 @@ -374,7 +374,6 @@ deselected_tests: - ensemble/tests/test_voting.py::test_sample_weight >=1.0,<1.1 - tests/test_multioutput.py::test_multiclass_multioutput_estimator_predict_proba - model_selection/tests/test_search.py::test_search_cv_sample_weight_equivalence[estimator0] - - tests/test_calibration.py::test_calibrated_classifier_cv_double_sample_weights_equivalence < 1.1 - tests/test_common.py::test_estimators[StackingClassifier(estimators=[('est1',LogisticRegression(C=0.1)) < 1.1 # Scikit-learn does not constraint multinomial logistic intercepts to sum to zero. From dcee7e6851d6d76fe3d06e7e47b6a9a221c96005 Mon Sep 17 00:00:00 2001 From: ethanglaser <42726565+ethanglaser@users.noreply.github.com> Date: Mon, 3 Nov 2025 20:28:11 -0500 Subject: [PATCH 3/3] Update deselected_tests.yaml --- deselected_tests.yaml | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/deselected_tests.yaml b/deselected_tests.yaml index 54d2fc729c..ceb1e3d63a 100755 --- a/deselected_tests.yaml +++ b/deselected_tests.yaml @@ -374,7 +374,8 @@ deselected_tests: - ensemble/tests/test_voting.py::test_sample_weight >=1.0,<1.1 - tests/test_multioutput.py::test_multiclass_multioutput_estimator_predict_proba - model_selection/tests/test_search.py::test_search_cv_sample_weight_equivalence[estimator0] - - tests/test_common.py::test_estimators[StackingClassifier(estimators=[('est1',LogisticRegression(C=0.1)) < 1.1 + - tests/test_common.py::test_estimators[StackingClassifier(estimators=[('est1',LogisticRegression(C=0.1)),('est2',LogisticRegression(C=1))])-check_sample_weights_invariance(kind=ones)] <1.1 + - tests/test_common.py::test_estimators[StackingClassifier(estimators=[('est1',LogisticRegression(C=0.1)),('est2',LogisticRegression(C=1))])-check_sample_weights_invariance(kind=zeros)] <1.1 # Scikit-learn does not constraint multinomial logistic intercepts to sum to zero. # Softmax function is invariant to additions by a constant, so even though the numbers