Repository navigation
fix(dbscan): GPU clustering metrics, and eps quantile in the estimator's metric - #227
Draft
Alexandr-Solovev wants to merge 3 commits into
Draft
Alexandr-Solovev wants to merge 3 commits into
Alexandr-Solovev wants to merge 3 commits into
Conversation
… on GPU The DBSCAN quality metrics passed the dpnp ground-truth labels to np.unique, which refuses the implicit conversion, so every sklearnex GPU DBSCAN case failed after the fit. Use the converted copy, as the other metrics do. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: Alexandr-Solovev <aleksandr.solovev@intel.com>
… a metrics experiment distances_quantile always used euclidean distances, so eps had the wrong scale for manhattan, chebyshev, cosine and minkowski. It now uses the metric (and p) from the estimator parameters; euclidean cases get the same value as before. configs/experiments/dbscan_metrics.json runs DBSCAN with the five metrics that scikit-learn-intelex runs in oneDAL on blobs, cifar and covtype, against scikit-learn's brute and auto algorithms. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: Alexandr-Solovev <aleksandr.solovev@intel.com>
david-cortes-intel
approved these changes
Nov 9, 2026
The experiment had at most 54 columns apart from cifar; the new blobs cover wide data with fewer rows. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: Alexandr-Solovev <aleksandr.solovev@intel.com>
david-cortes-intel
approved these changes
Nov 12, 2026
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Description
Two DBSCAN changes, one commit each:
get_subset_metrics_of_estimatorpassed the dpnp ground-truth labels tonp.unique, which refuses the implicit conversion, so every sklearnex GPU DBSCAN case failed after the fit. It now uses the convertedy_compat, as the other metrics already do.distances_quantilealways took euclidean distances, soepshad the wrong scale for manhattan, chebyshev, cosine and minkowski. It now usesmetric(andpfor minkowski) from the estimator parameters throughpairwise_distances. Euclidean cases get the same value as before.New
configs/experiments/dbscan_metrics.jsonruns DBSCAN with the five metrics that scikit-learn-intelex computes in oneDAL (uxlfoundation/scikit-learn-intelex#3483). It covers blobs, cifar and covtype, against scikit-learn'sbruteandautoalgorithms, sklearnex on CPU, and sklearnex on GPU with dpnp input. This config produced the CPU tables in uxlfoundation/oneDAL#4004.🤖 Generated with Claude Code