Skip to content

Handling all identical vectors in UMAP smooth_knn_dist_kernel #6904

Merged
rapids-bot[bot] merged 8 commits intorapidsai:branch-25.08from
jinsolp:smooth-zero-dist-handle
Jul 18, 2025
Merged

Handling all identical vectors in UMAP smooth_knn_dist_kernel #6904
rapids-bot[bot] merged 8 commits intorapidsai:branch-25.08from
jinsolp:smooth-zero-dist-handle

Conversation

@jinsolp
Copy link
Copy Markdown
Contributor

@jinsolp jinsolp commented Jun 12, 2025 •
edited by csadorf
Loading

smoot_knn_dist_kernel just fails if all neighbors have zero distance (i.e. are identical vectors).

Instead of failing, set rhos and sigmas to specific values in such cases.

Closes #7017

@jinsolp jinsolp self-assigned this Jun 12, 2025
@jinsolp jinsolp requested a review from a team as a code owner June 12, 2025 19:50
@jinsolp jinsolp added the improvement Improvement / enhancement to an existing function label Jun 12, 2025
@jinsolp jinsolp requested review from lowener and viclafargue June 12, 2025 19:50
@jinsolp jinsolp added the non-breaking Non-breaking change label Jun 12, 2025
@jinsolp jinsolp mentioned this pull request Jul 7, 2025
Closed
viclafargue
viclafargue approved these changes Jul 8, 2025
View reviewed changes
Copy link
Copy Markdown
Contributor

@viclafargue viclafargue left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM!

@csadorf
Copy link
Copy Markdown
Contributor

csadorf commented Jul 10, 2025

Looks like this change actually fixed one of the umap tests? Or is it now flaky?

@jinsolp
Copy link
Copy Markdown
Contributor Author

jinsolp commented Jul 10, 2025

I'll try rerunning after other tests are done

@csadorf csadorf requested a review from a team as a code owner July 10, 2025 23:06
@csadorf csadorf requested review from bdice and jcrist July 10, 2025 23:06
@github-actions github-actions Bot added the Cython / Python Cython or Python issue label Jul 10, 2025
@jinsolp jinsolp mentioned this pull request Jul 16, 2025
Closed
cjnolet
cjnolet reviewed Jul 17, 2025
View reviewed changes
Comment thread python/cuml/cuml_accel_tests/upstream/umap/xfail-list.yaml
@csadorf csadorf linked an issue Jul 17, 2025 that may be closed by this pull request
[BUG] UMAP fails when the KNN graph has no non-zero neighbors due to duplicates in the input #7017
Closed
@jinsolp jinsolp mentioned this pull request Jul 17, 2025
Closed
@cjnolet cjnolet removed request for bdice and jcrist July 17, 2025 21:24
@cjnolet cjnolet removed the request for review from lowener July 17, 2025 21:24
@cjnolet
Copy link
Copy Markdown
Member

cjnolet commented Jul 17, 2025

/merge

1 similar comment
@cjnolet
Copy link
Copy Markdown
Member

cjnolet commented Jul 18, 2025

/merge

@rapids-bot rapids-bot Bot merged commit a009b7f into rapidsai:branch-25.08 Jul 18, 2025
192 of 199 checks passed
@csadorf csadorf mentioned this pull request Jul 18, 2025
Closed
@jinsolp jinsolp deleted the smooth-zero-dist-handle branch July 18, 2025 17:56
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@cjnolet cjnolet cjnolet left review comments

@csadorf csadorf csadorf approved these changes

@viclafargue viclafargue viclafargue approved these changes

Assignees

@jinsolp jinsolp

Labels

CUDA/C++ Cython / Python Cython or Python issue improvement Improvement / enhancement to an existing function non-breaking Non-breaking change

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

[BUG] UMAP fails when the KNN graph has no non-zero neighbors due to duplicates in the input

5 participants

@jinsolp @csadorf @cjnolet @viclafargue @GPUtester