From 57ea13bb41c73b7722b3e35cc35e268a76cb4895 Mon Sep 17 00:00:00 2001 From: Julien Jerphanion Date: Fri, 25 Nov 2022 15:09:44 +0100 Subject: [PATCH 1/4] DOC Add more details regarding the improved efficiency in 1.1 and 1.2 --- .../plot_release_highlights_1_2_0.py | 15 +++++++++------ 1 file changed, 9 insertions(+), 6 deletions(-) diff --git a/examples/release_highlights/plot_release_highlights_1_2_0.py b/examples/release_highlights/plot_release_highlights_1_2_0.py index be44e0465f13c..f89a46eb47191 100644 --- a/examples/release_highlights/plot_release_highlights_1_2_0.py +++ b/examples/release_highlights/plot_release_highlights_1_2_0.py @@ -120,10 +120,13 @@ # Improved efficiency of many estimators # -------------------------------------- # In version 1.1 the efficiency of many estimators relying on the computation of -# pairwise distances was greatly improved for float64 dense input. In version 1.2, -# the efficiency of these estimators was further improved for all combinations of -# float32/float64 and dense/sparse input (for all metrics except euclidean). It -# concerns essentially clustering, manifold learning and neighbor search algorithms. +# pairwise distances (essentially estimators related to clustering, manifold +# learning and neighbors search algorithms) was greatly improved for float64 +# dense input. Efficiency improvement especially were a reduced memory footprint +# a much better scalability on multi-core machines. +# In version 1.2, the efficiency of these estimators was further improved for all +# combinations of dense and CSR inputs on float32 and float64 datasets, except +# the CSR-dense and dense-CSR combinations for the Euclidean and Squared Euclidean +# Distance metrics. # A detailed list of the impacted estimators can be found in the -# :ref:`changelog `. The main benefits are a reduced memory footprint -# and a much better scalability on multi-core machines. +# :ref:`changelog `. From 39a2ff664112b49912559ea8bbf99949cd8042dc Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?J=C3=A9r=C3=A9mie=20du=20Boisberranger?= <34657725+jeremiedbb@users.noreply.github.com> Date: Fri, 25 Nov 2022 15:20:15 +0100 Subject: [PATCH 2/4] Update examples/release_highlights/plot_release_highlights_1_2_0.py --- examples/release_highlights/plot_release_highlights_1_2_0.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/examples/release_highlights/plot_release_highlights_1_2_0.py b/examples/release_highlights/plot_release_highlights_1_2_0.py index f89a46eb47191..bb4677288394b 100644 --- a/examples/release_highlights/plot_release_highlights_1_2_0.py +++ b/examples/release_highlights/plot_release_highlights_1_2_0.py @@ -123,7 +123,7 @@ # pairwise distances (essentially estimators related to clustering, manifold # learning and neighbors search algorithms) was greatly improved for float64 # dense input. Efficiency improvement especially were a reduced memory footprint -# a much better scalability on multi-core machines. +# and a much better scalability on multi-core machines. # In version 1.2, the efficiency of these estimators was further improved for all # combinations of dense and CSR inputs on float32 and float64 datasets, except # the CSR-dense and dense-CSR combinations for the Euclidean and Squared Euclidean From df457100a4db3eba8c4b33f73980d15bed32be81 Mon Sep 17 00:00:00 2001 From: Julien Jerphanion Date: Fri, 25 Nov 2022 15:23:10 +0100 Subject: [PATCH 3/4] DOC Prefer 'sparse' over 'CSR' not to use jargon MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Co-authored-by: Jérémie du Boisberranger --- .../release_highlights/plot_release_highlights_1_2_0.py | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/examples/release_highlights/plot_release_highlights_1_2_0.py b/examples/release_highlights/plot_release_highlights_1_2_0.py index bb4677288394b..0d94b77ef555f 100644 --- a/examples/release_highlights/plot_release_highlights_1_2_0.py +++ b/examples/release_highlights/plot_release_highlights_1_2_0.py @@ -125,8 +125,8 @@ # dense input. Efficiency improvement especially were a reduced memory footprint # and a much better scalability on multi-core machines. # In version 1.2, the efficiency of these estimators was further improved for all -# combinations of dense and CSR inputs on float32 and float64 datasets, except -# the CSR-dense and dense-CSR combinations for the Euclidean and Squared Euclidean -# Distance metrics. +# combinations of dense and sparse inputs on float32 and float64 datasets, except +# the sparse-dense and dense-sparse combinations for the Euclidean and Squared +# Euclidean Distance metrics. # A detailed list of the impacted estimators can be found in the # :ref:`changelog `. From 64cecc171c6aadfbcfba1cc34b8c32e4684adf49 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?J=C3=A9r=C3=A9mie=20du=20Boisberranger?= <34657725+jeremiedbb@users.noreply.github.com> Date: Fri, 25 Nov 2022 15:49:33 +0100 Subject: [PATCH 4/4] Update plot_release_highlights_1_2_0.py --- .../release_highlights/plot_release_highlights_1_2_0.py | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/examples/release_highlights/plot_release_highlights_1_2_0.py b/examples/release_highlights/plot_release_highlights_1_2_0.py index 0d94b77ef555f..32b1108caa920 100644 --- a/examples/release_highlights/plot_release_highlights_1_2_0.py +++ b/examples/release_highlights/plot_release_highlights_1_2_0.py @@ -77,10 +77,10 @@ from sklearn.metrics import PredictionErrorDisplay fig, axs = plt.subplots(nrows=1, ncols=2, figsize=(12, 5)) -PredictionErrorDisplay.from_estimator( +_ = PredictionErrorDisplay.from_estimator( hist_no_interact, X, y, kind="actual_vs_predicted", ax=axs[0] ) -PredictionErrorDisplay.from_estimator( +_ = PredictionErrorDisplay.from_estimator( hist_no_interact, X, y, kind="residual_vs_predicted", ax=axs[1] ) @@ -89,7 +89,7 @@ # results from :func:`~model_selection.learning_curve`. from sklearn.model_selection import LearningCurveDisplay -LearningCurveDisplay.from_estimator( +_ = LearningCurveDisplay.from_estimator( hist_no_interact, X, y, cv=5, n_jobs=2, train_sizes=np.linspace(0.1, 1, 5) )