From 733eb6b890d44cd258e65d7bf3274165d26bf6c0 Mon Sep 17 00:00:00 2001 From: "julia.roznova" Date: Thu, 12 Sep 2024 03:50:05 -0700 Subject: [PATCH] remove outdated tests from deselected_test.yaml --- deselected_tests.yaml | 487 ------------------------------------------ 1 file changed, 487 deletions(-) diff --git a/deselected_tests.yaml b/deselected_tests.yaml index 364e597393..514c6c671b 100755 --- a/deselected_tests.yaml +++ b/deselected_tests.yaml @@ -54,20 +54,7 @@ deselected_tests: - decomposition/tests/test_pca.py::test_pca_svd_solver_auto[1000-500-400-full] >=1.5 - decomposition/tests/test_pca.py::test_pca_svd_solver_auto[1000-500-0.5-full] >=1.5 - # test for KMeans FutureWarning is not removed from sklearn tests suit yet - - cluster/tests/test_k_means.py::test_change_n_init_future_warning[KMeans-10] ==1.4.dev0 - - # Non-critical, but there are significant numerical differences in doctest results - - pipeline.py::sklearn.pipeline.FeatureUnion - - ensemble/_forest.py::sklearn.ensemble._forest.RandomForestRegressor - - ensemble/_voting.py::sklearn.ensemble._voting.VotingRegressor - # Non-critical, but there are significant differences due to different implementations - - ensemble/tests/test_forest.py::test_importances[RandomForestClassifier-gini-float64] - - ensemble/tests/test_voting.py::test_set_estimator_drop - - linear_model/tests/test_common.py::test_balance_property[42-True-LinearRegression] - - linear_model/tests/test_logistic.py::test_logistic_regression_multinomial - - cluster/tests/test_k_means.py::test_k_means_fit_predict >=0.23,<0.24 - neighbors/tests/test_lof.py::test_lof_dtype_equivalence[0.5-True-brute] - neighbors/tests/test_lof.py::test_lof_dtype_equivalence[auto-True-brute] @@ -79,12 +66,6 @@ deselected_tests: - neighbors/tests/test_lof.py::test_lof_dtype_equivalence[auto-True-ball_tree] >=1.2 win32 - neighbors/tests/test_lof.py::test_lof_dtype_equivalence[auto-True-kd_tree] >=1.2 win32 - # TODO: fix for Linear Regression failed to converge on macOS only - - inspection/tests/test_partial_dependence.py::test_partial_dependence_easy_target[2-est0] >=0.23 darwin - - inspection/tests/test_partial_dependence.py::test_partial_dependence_easy_target[2-est1] >=0.23 darwin - - inspection/tests/test_partial_dependence.py::test_partial_dependence_easy_target[2-est2] >=0.23 darwin - - inspection/tests/test_partial_dependence.py::test_partial_dependence_easy_target[2-est3] >=0.23 darwin - # Sklearnex RandomForestClassifier RNG is different from scikit-learn and daal4py # resulting in different feature importances for small number of trees (10). # Issue dissappears with bigger number of trees (>=20) @@ -93,64 +74,23 @@ deselected_tests: - inspection/tests/test_permutation_importance.py::test_permutation_importance_correlated_feature_regression_pandas[1.0-1] - inspection/tests/test_permutation_importance.py::test_permutation_importance_correlated_feature_regression_pandas[1.0-2] - # Random forest classifier selects a different most-important feature - # Feature importances: - # scikit-learn-intelex [0. 0.00553064 0.71323666 0.2812327 ] - # scikit-learn [1.59232288e-04 2.65131818e-02 2.63581110e-01 7.09746476e-01] - - feature_selection/tests/test_from_model.py::test_prefit_get_feature_names_out - - # TODO: add support of subset invariance to SVM - - tests/test_common.py::test_estimators[SVC()-check_methods_subset_invariance] - - tests/test_common.py::test_estimators[NuSVC()-check_methods_subset_invariance] - - tests/test_common.py::test_estimators[SVR()-check_methods_subset_invariance] - - tests/test_common.py::test_estimators[NuSVR()-check_methods_subset_invariance] - - # SVR.fit fails when input is two samples of one class - - preprocessing/tests/test_data.py::test_cv_pipeline_precomputed - - # KDtree kNN rarely misses 0-distance points when kneighbors is used on same-fitting data - - manifold/tests/test_spectral_embedding.py::test_precomputed_nearest_neighbors_filtering - - # Cache directory is not accessible on some systems - - utils/tests/test_validation.py::test_check_memory - # oneDAL doesn't throw error if resulting coeffs are not finite - linear_model/tests/test_coordinate_descent.py::test_enet_nonfinite_params - svm/tests/test_svm.py::test_svc_nonfinite_params - # Different exception types in scikit-learn-intelex and scikit-learn - - utils/tests/test_validation.py::test_check_array_links_to_imputer_doc_only_for_X[asarray-X] - - utils/tests/test_validation.py::test_check_array_links_to_imputer_doc_only_for_X[csr_matrix-X] - # TODO: investigate copy failure of read-only buffer - linear_model/tests/test_coordinate_descent.py::test_read_only_buffer - # Warm starting issue on 1 iteration of LogReg - - linear_model/tests/test_logistic.py::test_warm_start[multinomial-True-True-lbfgs] - - linear_model/tests/test_logistic.py::test_warm_start[multinomial-True-True-newton-cg] - - linear_model/tests/test_logistic.py::test_warm_start[multinomial-False-True-newton-cg] - - linear_model/tests/test_logistic.py::test_warm_start[multinomial-False-True-lbfgs] - - # Difference of assigned dtype: int64 (scikit-learn) vs. int32 (oneDAL/scikit-learn-intelex) - - cluster/_dbscan.py::sklearn.cluster._dbscan.DBSCAN - - neighbors/_base.py::sklearn.neighbors._base.KNeighborsMixin.kneighbors - # Difference between scikit-learn and scikit-learn-intelex methods of kNN - neighbors/tests/test_neighbors.py::test_unsupervised_kneighbors[float64-euclidean-True-1000-5-100-1] - neighbors/tests/test_neighbors.py::test_unsupervised_kneighbors[float64-minkowski-True-1000-5-100-1] - neighbors/tests/test_neighbors.py::test_unsupervised_kneighbors[float64-l2-True-1000-5-100-1] - - neighbors/tests/test_neighbors.py::test_unsupervised_kneighbors[float64-l1-True-1000-5-100-1] - - neighbors/tests/test_neighbors.py::test_unsupervised_kneighbors[float64-cityblock-True-1000-5-100-1] - neighbors/tests/test_neighbors.py::test_neigh_predictions_algorithm_agnosticity[float64-KNeighborsClassifier-1-100-euclidean-1000-5-100] - neighbors/tests/test_neighbors.py::test_neigh_predictions_algorithm_agnosticity[float64-KNeighborsClassifier-1-100-minkowski-1000-5-100] - neighbors/tests/test_neighbors.py::test_neigh_predictions_algorithm_agnosticity[float64-KNeighborsRegressor-1-100-euclidean-1000-5-100] - neighbors/tests/test_neighbors.py::test_neigh_predictions_algorithm_agnosticity[float64-KNeighborsRegressor-1-100-minkowski-1000-5-100] - neighbors/tests/test_neighbors.py::test_neigh_predictions_algorithm_agnosticity[float64-KNeighborsClassifier-1-100-l2-1000-5-100] - - neighbors/tests/test_neighbors.py::test_neigh_predictions_algorithm_agnosticity[float64-KNeighborsClassifier-1-100-l1-1000-5-100] - neighbors/tests/test_neighbors.py::test_neigh_predictions_algorithm_agnosticity[float64-KNeighborsRegressor-1-100-l2-1000-5-100] - - neighbors/tests/test_neighbors.py::test_neigh_predictions_algorithm_agnosticity[float64-KNeighborsRegressor-1-100-l1-1000-5-100] - - neighbors/tests/test_neighbors.py::test_neigh_predictions_algorithm_agnosticity[float64-KNeighborsClassifier-1-100-cityblock-1000-5-100] - - neighbors/tests/test_neighbors.py::test_neigh_predictions_algorithm_agnosticity[float64-KNeighborsRegressor-1-100-cityblock-1000-5-100] - neighbors/tests/test_neighbors.py::test_KNeighborsClassifier_multioutput # Models with sparse data are different between oneAPI Data Analytics Library (oneDAL) and stock scikit-learn @@ -158,10 +98,6 @@ deselected_tests: - svm/tests/test_sparse.py::test_svc_iris - svm/tests/test_sparse.py::test_sparse_realdata - # Decision function is different, 1.83697605e-06 - - ensemble/tests/test_bagging.py::test_sparse_classification - - ensemble/tests/test_bagging.py::test_sparse_regression <0.23 - # Same results as in scikit-learn, but in a different order - svm/tests/test_svm.py::test_svc_ovr_tie_breaking[SVC] - svm/tests/test_svm.py::test_svc_ovr_tie_breaking[NuSVC] @@ -170,67 +106,24 @@ deselected_tests: - svm/tests/test_svm.py::test_svc_clone_with_callable_kernel - svm/tests/test_svm.py::test_precomputed - # scikit-learn expects an exception for sparse matrices with 64-bit integer indices, - # scikit-learn-intelex works correctly with 64-bit integer indices - - tests/test_common.py::test_estimators[NuSVC()-check_estimator_sparse_data] - - tests/test_common.py::test_estimators[NuSVC()-check_estimator_sparse_array] - - tests/test_common.py::test_estimators[NuSVC()-check_estimator_sparse_matrix] - - utils/tests/test_estimator_checks.py::test_xfail_ignored_in_check_estimator - # SVC._dual_coef_ is changing after fitting, but the result of prediction is still the same - svm/tests/test_svm.py::test_tweak_params # Bitwise comparison of SVR score using a print (diff = 2.220446049250313e-16) - svm/tests/test_svm.py::test_custom_kernel_not_array_input[SVR] - # This test needs a warning from scikit-learn. scikit-learn-intelex raises the same warning - - tests/test_common.py::test_estimators[SVC()-check_supervised_y_2d] - - tests/test_common.py::test_estimators[SVR()-check_supervised_y_2d] - - tests/test_common.py::test_estimators[NuSVC()-check_supervised_y_2d] - - tests/test_common.py::test_estimators[NuSVR()-check_supervised_y_2d] - - # Bitwise comparison of probabilities using a print. - - metrics/tests/test_classification.py >=0.22,<0.24 - - # Max absolute difference: 0.04 for rocauc, and 0.01 for precision_recall - - metrics/tests/test_ranking.py::test_roc_curve_hard >=0.23,<0.24 - # test_non_uniform_strategies fails due to differences in handling of vacuous clusters after update # See https://github.com/IntelPython/daal4py/issues/69 - cluster/tests/test_k_means.py::test_kmeans_relocated_clusters >=0.24 - # In scikit-learn, these algorithms are not included in this test. However, scikit-learn-intelex - # does and throws an error. This is due to the different structure of the transformer.__module__.split("."). - - tests/test_common.py::test_transformers_get_feature_names_out[KMeans()] >=1.0 - # oneAPI Data Analytics Library (oneDAL) does not check convergence for tol == 0.0 for ease of benchmarking - cluster/tests/test_k_means.py::test_kmeans_convergence >=0.23 - cluster/tests/test_k_means.py::test_kmeans_verbose >=0.23 - # The Newton-CG solver solution computed in float32 disagrees with that of float64 by a small - # margin above the test threshold, see https://github.com/scikit-learn/scikit-learn/pull/13645 - - linear_model/tests/test_logistic.py::test_dtype_match - - # Logistic Regression coeffs change due to fix for loss scaling - # (https://github.com/scikit-learn/scikit-learn/pull/26721) - - feature_selection/tests/test_from_model.py::test_importance_getter[estimator0-named_steps.logisticregression.coef_] - - inspection/_plot/tests/test_boundary_decision_display.py::test_class_of_interest_binary[predict_proba] - - linear_model/tests/test_sag.py::test_sag_pobj_matches_logistic_regression - - # This fails on certain platforms. While weighted data does not go through DAAL, - # unweighted does. Since convergence does not occur (comment in the test - # suggests that) and because coefficients are slightly different, - # it results in a prediction disagreement in 1 case. - - ensemble/tests/test_stacking.py::test_stacking_with_sample_weight[StackingClassifier] - # Insufficient accuracy of "coefs" and "intercept" in Elastic Net for multi-target problems # https://github.com/oneapi-src/oneDAL/issues/494 - linear_model/tests/test_coordinate_descent.py::test_enet_multitarget - # Insufficient accuracy of objective function in Elastic Net in case of warm_start - # https://github.com/oneapi-src/oneDAL/issues/495 - - linear_model/tests/test_coordinate_descent.py::test_warm_start_convergence_with_regularizer_decrement <0.24 - # oneAPI Data Analytics Library (oneDAL) doesn't support sample_weight (back to scikit-learn), # sufficient accuracy (similar to previous cases) - linear_model/tests/test_coordinate_descent.py::test_enet_sample_weight_consistency >=0.23 @@ -239,31 +132,19 @@ deselected_tests: # Looks like we need to align tree traversal. This problem will be fixed - ensemble/tests/test_forest.py::test_min_samples_leaf - # Different random number generation engine in oneDAL and scikit-learn - # The result is depend on random state, for random_state=777 in RandomForestClassifier the test is passed - - ensemble/tests/test_voting.py::test_majority_label_iris - # scikit-learn-intelex RF threads are used internally and are not explicitly specified - ensemble/tests/test_forest.py::test_backend_respected # scikit-learn-intelex does not support accessing trees through the result variable - ensemble/tests/test_forest.py::test_warm_start - - inspection/tests/test_partial_dependence.py::test_recursion_decision_tree_vs_forest_and_gbdt[0] >=0.23 # scikit-learn-intelex implementation builds different trees compared to scikit-learn # Comparison of tree forest will fail - ensemble/tests/test_forest.py::test_class_weights - ensemble/tests/test_forest.py::test_poisson_vs_mse - inspection/tests/test_permutation_importance.py::test_robustness_to_high_cardinality_noisy_feature >=0.23 - - tests/test_common.py::test_estimators[SVC()-check_sample_weights_invariance(kind=zeros)] <1.0 - - tests/test_common.py::test_estimators[SVR()-check_sample_weights_invariance(kind=zeros)] <1.0 - - tests/test_common.py::test_estimators[NuSVC()-check_sample_weights_invariance(kind=zeros)] <1.0 - - tests/test_common.py::test_estimators[NuSVR()-check_sample_weights_invariance(kind=zeros)] <1.0 - - tests/test_common.py::test_estimators[NuSVC()-check_class_weight_classifiers] <1.0 - - tests/test_multioutput.py::test_multi_output_classification # Linear Regression - minor mismatches in error/warning messages - - model_selection/tests/test_search.py::test_grid_search_pipeline_steps - linear_model/tests/test_base.py::test_linear_regression_pd_sparse_dataframe_warning # L1 Linear models with sklearn 1.1 + numpy > 1.25 - extra warnings from numpy lead to test fail @@ -274,53 +155,19 @@ deselected_tests: - linear_model/tests/test_coordinate_descent.py::test_assure_warning_when_normalize[deprecated-0-LassoCV] >=1.1,<1.2 - linear_model/tests/test_coordinate_descent.py::test_assure_warning_when_normalize[deprecated-0-ElasticNetCV] >=1.1,<1.2 - # Different results scikit-learn-intelex and scikit-learn linear regression with weights. Need to investigate. - - inspection/tests/test_permutation_importance.py::test_permutation_importance_sample_weight >=0.24 - # OOB scores in scikit-learn and oneDAL are different because of different random number generators - - ensemble/tests/test_forest.py::test_forest_classifier_oob[X1-y1-0.65-array-ExtraTreesClassifier] - - ensemble/tests/test_forest.py::test_forest_classifier_oob[True-X1-y1-0.65-array-ExtraTreesClassifier] >=1.3 - ensemble/tests/test_forest.py::test_forest_regressor_oob[True-X0-y0-0.7-array-ExtraTreesRegressor] >=1.3 - - ensemble/tests/test_forest.py::test_forest_regressor_oob[X0-y0-0.7-array-RandomForestRegressor] >=1.2 darwin - - ensemble/tests/test_forest.py::test_forest_regressor_oob[True-X0-y0-0.7-array-RandomForestRegressor] >=1.3 darwin - - ensemble/tests/test_forest.py::test_importances[ExtraTreesRegressor-squared_error-float64] >=0.23 darwin - ensemble/tests/test_forest.py::test_forest_regressor_oob[X0-y0-0.7-array-ExtraTreesRegressor] - ensemble/tests/test_forest.py::test_warm_start_oob - ensemble/tests/test_forest.py::test_distribution - # Need a warning from scikit-learn if some samples do not have OOB scores that can not be raised from oneDAL - - ensemble/tests/test_forest.py::test_forest_oob_warning - - # Different behavior when 1 class enters the input - - feature_selection/tests/test_rfe.py::test_rfe_cv_groups - # few-percent numerical differences in ExtraTreesRegressor, but 6 digits are checked - ensemble/tests/test_forest.py::test_memory_layout[float64-ExtraTreesRegressor] - ensemble/tests/test_forest.py::test_memory_layout[float32-ExtraTreesRegressor] - # The bugs are fixed in 2021.2 release - - ensemble/tests/test_stacking.py::test_stacking_cv_influence - # module name should starts with 'sklearn.' but we have 'daal4py.sklearn.' - - tests/test_common.py::test_check_n_features_in_after_fitting[LogisticRegression()] >=0.24,<1.0 - - tests/test_common.py::test_check_n_features_in_after_fitting[SVC()] >=0.24,<1.0 - - metrics/tests/test_score_objects.py::test_scoring_is_not_metric - utils/tests/test_estimator_checks.py::test_check_dataframe_column_names_consistency >=1.0 - # Stability issue with max absolute difference: 4.33846826e-08/1.17613697e-11. Remove in next release - - ensemble/tests/test_bagging.py::test_estimators_samples_deterministic - - # We use similar statements, but with different words - - ensemble/tests/test_gradient_boosting.py::test_gradient_boosting_with_init_pipeline >=1.0 - - # Data for the tests is generated by using SVC. It is not equal to stock scikit-learn - - metrics/tests/test_ranking.py::test_precision_recall_curve >=0.22,<0.24 - - # Some values in PCA.components_ (in the last component) aren't equal (0.6 on average - # for absolute error in this test) because of different implementations of PCA. - # The results are also not stable. - - decomposition/tests/test_incremental_pca.py::test_whitening - # The test fails because of changing of 'auto' strategy in PCA to improve performance. # 'randomized' PCA expected, but 'full' is given. - decomposition/tests/test_pca.py::test_pca_svd_solver_auto[data3-10-randomized] @@ -332,52 +179,20 @@ deselected_tests: # RandomForestRegressor sum(y_pred)!=sum(y_true) - ensemble/tests/test_forest.py::test_balance_property_random_forest[squared_error] >=1.0 - # This test fails on mac mini 8.1 with stock scikit-learn - - semi_supervised/tests/test_label_propagation.py - # This test fails because with patch config_context with new options, but the # test checks only the exact number of options that are used - tests/test_config.py::test_config_context - # Some scikit-learn-intelex docstrings differ from scikit-learn. - - tests/test_docstrings.py >=1.0.2 # Accuracy of scikit-learn-intelex and scikit-learn may differ due to different approaches - - manifold/tests/test_t_sne.py::test_preserve_trustworthiness_approximately_with_precomputed_distances - manifold/tests/test_t_sne.py::test_bh_match_exact - manifold/tests/test_t_sne.py::test_uniform_grid[barnes_hut] - - manifold/tests/test_t_sne.py::test_sparse_precomputed_distance - - manifold/tests/test_t_sne.py::test_tsne_different_square_distances >=0.24 - - # Temporary deselected up to 2021.6 release. Need to fix - - ensemble/tests/test_bagging.py::test_classification - - # Failure related to incompatibility of older sklearn versions with updated dependencies - - ensemble/_hist_gradient_boosting/tests/test_compare_lightgbm.py::test_same_predictions_multiclass_classification >=0.24,<1.0 - - ensemble/tests/test_gradient_boosting.py::test_gradient_boosting_with_init_pipeline >=0.24,<1.0 - - utils/tests/test_validation.py::test_check_array_pandas_dtype_casting >=1.0,<1.2 - - utils/tests/test_validation.py::test_check_sparse_pandas_sp_format <1.2 - - # Failure due to non-uniformity in the MT2203 engine causing - # bad Random Forest fits for small datasets with large n_estimators - # Had been solved by using MT19937, but oneDAL forces use of MT2203 - - tests/test_multioutput.py::test_classifier_chain_tuple_order # oneDAL decision forest trains individual trees differently than # sklearn. Attempts to compare individual sklearn trees to oneDAL # trees will fail, especially since two different RNGs are used. - ensemble/tests/test_forest.py::test_estimators_samples - # Tests migrated from gpu deselection set starting from sklearn 1.4 for unknowm reason(s) - - 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_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 - - tests/test_multioutput.py::test_classifier_chain_fit_and_predict_with_sparse_data >=1.4 - # Deselected tests for incremental algorithms # Need to rework getting policy to correctly obtain it for method without data (finalize_fit) # and avoid keeping it in class attribute, also need to investigate how to implement @@ -388,329 +203,43 @@ deselected_tests: - tests/test_common.py::test_estimators[IncrementalLinearRegression()-check_estimators_pickle(readonly_memmap=True)] - tests/test_common.py::test_estimators[IncrementalPCA()-check_estimators_pickle] - tests/test_common.py::test_estimators[IncrementalPCA()-check_estimators_pickle(readonly_memmap=True)] - - tests/test_common.py::test_estimators[IncrementalRidge()-check_estimators_pickle] - - tests/test_common.py::test_estimators[IncrementalRidge()-check_estimators_pickle(readonly_memmap=True)] # There are not enough data to run onedal backend - tests/test_common.py::test_estimators[IncrementalLinearRegression()-check_fit2d_1sample] - tests/test_common.py::test_estimators[IncrementalRidge()-check_fit2d_1sample] - # Deselection of LogisticRegression tests over accuracy comparisons with sample_weights - # and without. Because scikit-learn-intelex does not support sample_weights, it's doing - # a fallback to scikit-learn in one case and not in the other, and needs to be investigated. - - model_selection/tests/test_classification_threshold.py::test_fit_and_score_over_thresholds_sample_weight >=1.5 - - model_selection/tests/test_classification_threshold.py::test_tuned_threshold_classifier_cv_zeros_sample_weights_equivalence >=1.5 - # Deselections for 2025.0 - - ensemble/tests/test_forest.py::test_importances[ExtraTreesRegressor-squared_error-float64] - cluster/tests/test_k_means.py::test_kmeans_elkan_results - - # -------------------------------------------------------- - # No need to test daal4py patching reduced_tests: - cluster/tests/test_affinity_propagation.py - - cluster/tests/test_bicluster.py - - cluster/tests/test_birch.py - - cluster/tests/test_mean_shift.py - - cluster/tests/test_optics.py - - - compose/tests/test_column_transformer.py - - - decomposition/tests/test_dict_learning.py - - decomposition/tests/test_factor_analysis.py - - decomposition/tests/test_nmf.py - - decomposition/tests/test_online_lda.py - - - ensemble/tests/test_gradient_boosting.py - - ensemble/tests/test_gradient_boosting_loss_functions.py - - ensemble/tests/test_iforest.py - - - feature_selection/tests/test_chi2.py - - feature_selection/tests/test_feature_select.py - - feature_selection/tests/test_mutual_info.py - - feature_selection/tests/test_sequential.py - - feature_selection/tests/test_from_model.py - - - manifold/tests/test_isomap.py - manifold/tests/test_locally_linear.py - - manifold/tests/test_spectral_embedding.py - - - model_selection/tests/test_successive_halving.py - - - neighbors/tests/test_ball_tree.py - - neighbors/tests/test_kd_tree.py - - neighbors/tests/test_quad_tree.py - - - tests/test_kernel_approximation.py - - tests/test_docstring_parameters.py - - tests/test_dummy.py - - tests/test_random_projection.py - - tests/test_naive_bayes.py - - - utils/tests/test_arpack.py - - utils/tests/test_cython_blas.py - - utils/tests/test_encode.py - - utils/tests/test_estimator_html_repr.py - - utils/tests/test_extmath.py - - utils/tests/test_fast_dict.py - - utils/tests/test_mocking.py - - utils/tests/test_murmurhash.py - - utils/tests/test_sparsefuncs.py - - utils/tests/test_utils.py - - - _loss/ - - cross_decomposition/ - - datasets/ - - ensemble/_hist_gradient_boosting/ - - experimental/ - - feature_extraction/ - - gaussian_process/ - - impute/ - - inspection/ - - neural_network/ - - preprocessing/ - public: - - tests/test_common.py::test_estimators - # Failed in stock scikit-learn - - metrics/tests/test_common.py::test_not_symmetric_metric[precision_recall_curve] - - metrics/tests/test_common.py::test_binary_sample_weight_invariance[precision_recall_curve] # Fails from numpy 2.0 and sklearn 1.4+ - neighbors/tests/test_neighbors.py::test_KNeighborsClassifier_raise_on_all_zero_weights - - # -------------------------------------------------------- # The following tests currently fail with GPU offloading gpu: - # Segfaults - - ensemble/tests/test_weight_boosting.py - # Fails - cluster/tests/test_dbscan.py::test_weighted_dbscan - cluster/tests/test_k_means.py::test_kmeans_elkan_results[42-1e-100-sparse-normal] - - cluster/tests/test_k_means.py::test_kmeans_elkan_results[42-1e-100-sparse-blobs] - model_selection/tests/test_search.py::test_unsupervised_grid_search - - - ensemble/tests/test_bagging.py::test_gridsearch - - ensemble/tests/test_bagging.py::test_estimators_samples - - ensemble/tests/test_common.py::test_ensemble_heterogeneous_estimators_behavior - - ensemble/tests/test_voting.py::test_parallel_fit - ensemble/tests/test_voting.py::test_sample_weight - - - feature_selection/tests/test_rfe.py::test_number_of_subsets_of_features - - - manifold/tests/test_t_sne.py::test_preserve_trustworthiness_approximately - - manifold/tests/test_t_sne.py::test_uniform_grid - - manifold/tests/test_t_sne.py::test_tsne_different_square_distances - - - metrics/tests/test_classification.py::test_precision_recall_f1_score_binary - - metrics/tests/test_classification.py::test_precision_recall_fscore_support_errors - - metrics/tests/test_classification.py::test_confusion_matrix_binary - - metrics/tests/test_classification.py::test_multilabel_confusion_matrix_binary - - metrics/tests/test_ranking.py::test_roc_curve - - metrics/tests/test_ranking.py::test_roc_returns_consistency - - metrics/tests/test_ranking.py::test_roc_curve_confidence - - metrics/tests/test_ranking.py::test_precision_recall_curve - - metrics/tests/test_ranking.py::test_score_scale_invariance - - metrics/tests/test_ranking.py::test_partial_roc_auc_score - metrics/tests/test_score_objects.py::test_average_precision_pos_label - - - mixture/tests/test_bayesian_mixture.py::test_bayesian_mixture_weights_prior_initialisation - - mixture/tests/test_bayesian_mixture.py::test_bayesian_mixture_mean_prior_initialisation - - mixture/tests/test_bayesian_mixture.py::test_bayesian_mixture_precisions_prior_initialisation - - mixture/tests/test_bayesian_mixture.py::test_bayesian_mixture_weights - - mixture/tests/test_bayesian_mixture.py::test_monotonic_likelihood - - mixture/tests/test_bayesian_mixture.py::test_compare_covar_type - - mixture/tests/test_bayesian_mixture.py::test_check_covariance_precision - - mixture/tests/test_bayesian_mixture.py::test_invariant_translation - - mixture/tests/test_bayesian_mixture.py::test_bayesian_mixture_fit_predict - - mixture/tests/test_bayesian_mixture.py::test_bayesian_mixture_fit_predict_n_init - - mixture/tests/test_bayesian_mixture.py::test_bayesian_mixture_predict_predict_proba - - mixture/tests/test_gaussian_mixture.py::test_check_weights - - mixture/tests/test_gaussian_mixture.py::test_check_means - - mixture/tests/test_gaussian_mixture.py::test_check_precisions - - mixture/tests/test_gaussian_mixture.py::test_gaussian_mixture_estimate_log_prob_resp - - mixture/tests/test_gaussian_mixture.py::test_gaussian_mixture_predict_predict_proba - - mixture/tests/test_gaussian_mixture.py::test_gaussian_mixture_fit_predict - - mixture/tests/test_gaussian_mixture.py::test_gaussian_mixture_fit_predict_n_init - - mixture/tests/test_gaussian_mixture.py::test_gaussian_mixture_fit - - mixture/tests/test_gaussian_mixture.py::test_gaussian_mixture_fit_best_params - - mixture/tests/test_gaussian_mixture.py::test_gaussian_mixture_fit_convergence_warning - - mixture/tests/test_gaussian_mixture.py::test_multiple_init - - mixture/tests/test_gaussian_mixture.py::test_gaussian_mixture_n_parameters - - mixture/tests/test_gaussian_mixture.py::test_bic_1d_1component - - mixture/tests/test_gaussian_mixture.py::test_gaussian_mixture_aic_bic - - mixture/tests/test_gaussian_mixture.py::test_gaussian_mixture_verbose - - mixture/tests/test_gaussian_mixture.py::test_warm_start - - mixture/tests/test_gaussian_mixture.py::test_convergence_detected_with_warm_start - - mixture/tests/test_gaussian_mixture.py::test_score - - mixture/tests/test_gaussian_mixture.py::test_score_samples - - mixture/tests/test_gaussian_mixture.py::test_monotonic_likelihood - - mixture/tests/test_gaussian_mixture.py::test_regularisation - - mixture/tests/test_gaussian_mixture.py::test_property - - mixture/tests/test_gaussian_mixture.py::test_sample - - mixture/tests/test_gaussian_mixture.py::test_init - - mixture/tests/test_mixture.py::test_gaussian_mixture_n_iter - - - model_selection/tests/test_search.py::test_grid_search_one_grid_point - model_selection/tests/test_search.py::test_search_default_iid - - model_selection/tests/test_search.py::test_random_search_cv_results_multimetric - - model_selection/tests/test_search.py::test_predict_proba_disabled - - - model_selection/tests/test_validation.py::test_cross_val_predict_sparse_prediction - - model_selection/tests/test_validation.py::test_fit_and_score_verbosity - - neighbors/tests/test_neighbors.py::test_unsupervised_kneighbors - - - neighbors/tests/test_neighbors.py::test_kneighbors_classifier - - neighbors/tests/test_neighbors.py::test_KNeighborsClassifier_multioutput - neighbors/tests/test_neighbors.py::test_neighbors_metrics - - - neighbors/tests/test_neighbors.py::test_neighbors_iris - - - semi_supervised/tests/test_self_training.py::test_early_stopping - - - svm/tests/test_sparse.py::test_svc - - svm/tests/test_sparse.py::test_svc_with_custom_kernel - - svm/tests/test_sparse.py::test_svc_iris - - svm/tests/test_sparse.py::test_error - - svm/tests/test_sparse.py::test_sample_weights - - svm/tests/test_sparse.py::test_sparse_realdata - - svm/tests/test_svm.py::test_precomputed - - svm/tests/test_svm.py::test_tweak_params - svm/tests/test_svm.py::test_svm_classifier_sided_sample_weight[estimator0] - svm/tests/test_svm.py::test_svm_equivalence_sample_weight_C - svm/tests/test_svm.py::test_negative_weights_svc_leave_two_labels[partial-mask-label-1-SVC] - svm/tests/test_svm.py::test_negative_weights_svc_leave_two_labels[partial-mask-label-2-SVC] - - svm/tests/test_svm.py::test_svc_clone_with_callable_kernel - - svm/tests/test_svm.py::test_custom_kernel_not_array_input[SVR] - - - tests/test_common.py::test_estimators[BayesianGaussianMixture()-check_estimators_dtypes] - - tests/test_common.py::test_estimators[BayesianGaussianMixture()-check_fit_score_takes_y] - - tests/test_common.py::test_estimators[BayesianGaussianMixture()-check_estimators_fit_returns_self] - - tests/test_common.py::test_estimators[BayesianGaussianMixture()-check_estimators_fit_returns_self(readonly_memmap=True)] - - tests/test_common.py::test_estimators[BayesianGaussianMixture()-check_dtype_object] - - tests/test_common.py::test_estimators[BayesianGaussianMixture()-check_pipeline_consistency] - - tests/test_common.py::test_estimators[BayesianGaussianMixture()-check_estimators_nan_inf] - - tests/test_common.py::test_estimators[BayesianGaussianMixture()-check_estimators_overwrite_params] - - tests/test_common.py::test_estimators[BayesianGaussianMixture()-check_estimators_pickle] - - tests/test_common.py::test_estimators[BayesianGaussianMixture()-check_estimators_pickle(readonly_memmap=True)] - - tests/test_common.py::test_estimators[BayesianGaussianMixture()-check_methods_sample_order_invariance] - - tests/test_common.py::test_estimators[BayesianGaussianMixture()-check_methods_subset_invariance] - - tests/test_common.py::test_estimators[BayesianGaussianMixture()-check_fit2d_1feature] - - tests/test_common.py::test_estimators[BayesianGaussianMixture()-check_dict_unchanged] - - tests/test_common.py::test_estimators[BayesianGaussianMixture()-check_dont_overwrite_parameters] - - tests/test_common.py::test_estimators[BayesianGaussianMixture()-check_fit_idempotent] - - tests/test_common.py::test_estimators[BayesianGaussianMixture()-check_n_features_in] - - tests/test_common.py::test_estimators[BayesianGaussianMixture()-check_fit2d_predict1d] # sparse input is not implemented for DBSCAN. - - tests/test_common.py::test_estimators[DBSCAN()-check_estimator_sparse_data] - - tests/test_common.py::test_estimators[GaussianMixture()-check_estimators_dtypes] - - tests/test_common.py::test_estimators[GaussianMixture()-check_fit_score_takes_y] - - tests/test_common.py::test_estimators[GaussianMixture()-check_estimators_fit_returns_self] - - tests/test_common.py::test_estimators[GaussianMixture()-check_estimators_fit_returns_self(readonly_memmap=True)] - - tests/test_common.py::test_estimators[GaussianMixture()-check_dtype_object] - - tests/test_common.py::test_estimators[GaussianMixture()-check_pipeline_consistency] - - tests/test_common.py::test_estimators[GaussianMixture()-check_estimators_nan_inf] - - tests/test_common.py::test_estimators[GaussianMixture()-check_estimators_overwrite_params] - - tests/test_common.py::test_estimators[GaussianMixture()-check_estimators_pickle] - - tests/test_common.py::test_estimators[GaussianMixture()-check_estimators_pickle(readonly_memmap=True)] - - tests/test_common.py::test_estimators[GaussianMixture()-check_methods_sample_order_invariance] - - tests/test_common.py::test_estimators[GaussianMixture()-check_methods_subset_invariance] - - tests/test_common.py::test_estimators[GaussianMixture()-check_fit2d_1feature] - - tests/test_common.py::test_estimators[GaussianMixture()-check_dict_unchanged] - - tests/test_common.py::test_estimators[GaussianMixture()-check_dont_overwrite_parameters] - - tests/test_common.py::test_estimators[GaussianMixture()-check_fit_idempotent] - - tests/test_common.py::test_estimators[GaussianMixture()-check_n_features_in] - - tests/test_common.py::test_estimators[GaussianMixture()-check_fit2d_predict1d] - tests/test_common.py::test_estimators[RandomForestClassifier()-check_class_weight_classifiers] - - tests/test_common.py::test_estimators[SVC()-check_sample_weights_pandas_series] - tests/test_common.py::test_estimators[SVC()-check_sample_weights_not_an_array] - - tests/test_common.py::test_estimators[SVC()-check_sample_weights_shape] - - tests/test_common.py::test_estimators[SVC()-check_pipeline_consistency] - - tests/test_common.py::test_estimators[SVC()-check_estimators_nan_inf] - - tests/test_common.py::test_estimators[SVC()-check_estimators_pickle] - tests/test_common.py::test_estimators[SVC()-check_classifier_data_not_an_array] - - tests/test_common.py::test_estimators[SVC()-check_classifiers_classes] - - tests/test_common.py::test_estimators[SVC()-check_classifiers_train] - - tests/test_common.py::test_estimators[SVC()-check_class_weight_classifiers] - - tests/test_common.py::test_estimators[SVC()-check_fit2d_1feature] - - tests/test_common.py::test_estimators[SVC()-check_dict_unchanged] - - tests/test_common.py::test_estimators[SVC()-check_fit_idempotent] - - tests/test_common.py::test_estimators[SVC()-check_n_features_in] - - tests/test_common.py::test_estimators[SelfTrainingClassifier(base_estimator=LogisticRegression(C=1))-check_classifiers_classes] - - tests/test_common.py::test_estimators[SelfTrainingClassifier(base_estimator=LogisticRegression(C=1))-check_decision_proba_consistency] - - tests/test_common.py::test_estimators[StackingClassifier(estimators=[('est1',LogisticRegression(C=0.1)),('est2',LogisticRegression(C=1))])-check_sample_weights_invariance(kind=ones)] - - tests/test_common.py::test_estimators[StackingClassifier(estimators=[('est1',LogisticRegression(C=0.1)),('est2',LogisticRegression(C=1))])-check_sample_weights_invariance(kind=zeros)] - - tests/test_common.py::test_estimators[TSNE()-check_fit_idempotent] - - tests/test_common.py::test_estimators[TSNE()-check_n_features_in] - - tests/test_common.py::test_search_cv[RandomizedSearchCV(estimator=LogisticRegression(),param_distributions={'C':[0.1,1.0]})-check_classifiers_classes] - - tests/test_common.py::test_search_cv[RandomizedSearchCV(estimator=LogisticRegression(),param_distributions={'C':[0.1,1.0]})-check_decision_proba_consistency] - - tests/test_common.py::test_check_n_features_in_after_fitting[TSNE()] - - - tests/test_multiclass.py::test_ovr_fit_predict_sparse - - tests/test_multiclass.py::test_ovr_binary - - tests/test_multiclass.py::test_ovr_fit_predict_svc - - tests/test_multiclass.py::test_ovr_multilabel_predict_proba - - tests/test_multiclass.py::test_ovr_multilabel_decision_function - - tests/test_multiclass.py::test_ovr_single_label_decision_function - - tests/test_multiclass.py::test_ovr_coef_ - - tests/test_multiclass.py::test_ovr_deprecated_coef_intercept - - tests/test_multiclass.py::test_pairwise_cross_val_score - - tests/test_multioutput.py::test_multiclass_multioutput_estimator_predict_proba - - tests/test_multioutput.py::test_classifier_chain_fit_and_predict_with_sparse_data - - # Very slow execution due to SVC - - model_selection/tests/test_validation.py::test_validation_curve_cv_splits_consistency - - model_selection/tests/test_search.py::test_grid_search_cv_results - - model_selection/tests/test_search.py::test_random_search_cv_results - - # Segmentation faults on GPU - - tests/test_common.py::test_search_cv - - manifold/tests/test_t_sne.py::test_n_iter_without_progress - # Other device issues - - tests/test_metaestimators.py::test_meta_estimators_delegate_data_validation[StackingClassifier] - - tests/test_multiclass.py::test_ovr_always_present - - tests/test_multiclass.py::test_support_missing_values[OneVsRestClassifier] - - tests/test_multiclass.py::test_support_missing_values[OneVsOneClassifier] - - tests/test_multioutput.py::test_multi_output_delegate_predict_proba - - tests/test_multioutput.py::test_classifier_chain_vs_independent_models - - tests/test_multioutput.py::test_base_chain_fit_and_predict - - tests/test_multioutput.py::test_base_chain_crossval_fit_and_predict - - tests/test_multioutput.py::test_multi_output_classes_[estimator1] - - tests/test_multioutput.py::test_multi_output_classes_[estimator2] - - tests/test_multioutput.py::test_support_missing_values[MultiOutputClassifier-LogisticRegression] - - tests/test_multioutput.py::test_classifier_chain_tuple_order[list] - - tests/test_multioutput.py::test_classifier_chain_tuple_order[array] - tests/test_multioutput.py::test_classifier_chain_tuple_order[tuple] - - tests/test_pipeline.py::test_pipeline_methods_anova - - tests/test_pipeline.py::test_score_samples_on_pipeline_without_score_samples - - tests/test_pipeline.py::test_pipeline_methods_preprocessing_svm - - tests/test_pipeline.py::test_pipeline_transform - - tests/test_pipeline.py::test_feature_union_weights - - tests/test_pipeline.py::test_classes_property - - tests/test_pipeline.py::test_set_feature_union_passthrough - - tests/test_pipeline.py::test_pipeline_missing_values_leniency - - tests/test_pipeline.py::test_pipeline_set_output_integration - - tests/test_pipeline.py::test_feature_union_set_output - - tests/test_parallel.py::test_dispatch_config_parallel[1] - - tests/test_parallel.py::test_dispatch_config_parallel[2] # KD Tree (not implemented for GPU) - neighbors/tests/test_neighbors.py::test_neigh_predictions_algorithm_agnosticity[float64-KNeighborsClassifier-50-500-l2-1000-5-100] - neighbors/tests/test_neighbors.py::test_neigh_predictions_algorithm_agnosticity[float64-KNeighborsClassifier-100-1000-l2-1000-5-100] - neighbors/tests/test_neighbors.py::test_neigh_predictions_algorithm_agnosticity[float64-KNeighborsRegressor-50-500-l2-1000-5-100] - neighbors/tests/test_neighbors.py::test_neigh_predictions_algorithm_agnosticity[float64-KNeighborsRegressor-100-1000-l2-1000-5-100] - # failing due to numeric/code error - - linear_model/tests/test_common.py::test_balance_property[42-False-LogisticRegressionCV] - - sklearn/manifold/tests/test_t_sne.py::test_n_iter_without_progress - - model_selection/tests/test_search.py::test_searchcv_raise_warning_with_non_finite_score[RandomizedSearchCV-specialized_params1-False] - - model_selection/tests/test_search.py::test_searchcv_raise_warning_with_non_finite_score[RandomizedSearchCV-specialized_params1-True] - - tests/test_calibration.py::test_calibrated_classifier_cv_double_sample_weights_equivalence - - tests/test_calibration.py::test_calibrated_classifier_cv_zeros_sample_weights_equivalence - - tests/test_common.py::test_estimators[FeatureAgglomeration()-check_parameters_default_constructible] - - neighbors/tests/test_lof.py::test_novelty_true_common_tests[LocalOutlierFactor(novelty=True)-check_methods_subset_invariance] - - tests/test_common.py::test_transformers_get_feature_names_out[StackingRegressor(estimators=[('est1',Ridge(alpha=0.1)),('est2',Ridge(alpha=1))])] - - tests/test_common.py::test_transformers_get_feature_names_out[VotingRegressor(estimators=[('est1',Ridge(alpha=0.1)),('est2',Ridge(alpha=1))])] - - tests/test_common.py::test_f_contiguous_array_estimator[TSNE] - - manifold/tests/test_t_sne.py::test_tsne_works_with_pandas_output # GPU Forest algorithm implementation does not follow certain Scikit-learn standards - ensemble/tests/test_forest.py::test_max_leaf_nodes_max_depth @@ -720,24 +249,8 @@ gpu: - ensemble/tests/test_forest.py::test_max_samples_boundary_regressors # numerical issues in GPU Forest algorithms which require further investigation - - ensemble/tests/test_forest.py::test_forest_classifier_oob[X0-y0-0.9-array-ExtraTreesClassifier] - - ensemble/tests/test_forest.py::test_forest_classifier_oob[X0-y0-0.9-array-RandomForestClassifier] - - ensemble/tests/test_forest.py::test_forest_classifier_oob[X1-y1-0.65-array-RandomForestClassifier] - - ensemble/tests/test_forest.py::test_forest_classifier_oob[X2-y2-0.65-array-ExtraTreesClassifier] - - ensemble/tests/test_forest.py::test_forest_classifier_oob[X2-y2-0.65-array-RandomForestClassifier] - - ensemble/tests/test_forest.py::test_forest_regressor_oob[X0-y0-0.7-array-RandomForestRegressor] - - ensemble/tests/test_stacking.py::test_stacking_regressor_drop_estimator - ensemble/tests/test_voting.py::test_predict_on_toy_problem[42] - tests/test_common.py::test_estimators[ExtraTreesClassifier()-check_class_weight_classifiers] - - tests/test_common.py::test_estimators[ExtraTreesRegressor()-check_sample_weights_invariance(kind=zeros)] - - tests/test_common.py::test_estimators[RandomForestRegressor()-check_regressor_data_not_an_array] - - ensemble/tests/test_forest.py::test_max_samples_boundary_classifiers[ExtraTreesClassifier] - - tests/test_common.py::test_estimators[ExtraTreesClassifier()-check_classifier_data_not_an_array] - - tests/test_common.py::test_estimators[ExtraTreesClassifier()-check_classifiers_train] - - tests/test_common.py::test_estimators[ExtraTreesClassifier()-check_classifiers_train(readonly_memmap=True)] - - tests/test_common.py::test_estimators[ExtraTreesClassifier()-check_fit_idempotent] - - tests/test_common.py::test_estimators[ExtraTreesRegressor()-check_fit_idempotent] - - tests/test_common.py::test_estimators[ExtraTreesRegressor()-check_regressor_data_not_an_array] # GPU implementation of Extra Trees doesn't support sample_weights # comparisons to GPU with sample weights will use different algorithms