Skip to content

Various speed-ups in the full chain routines#84

Merged
francois-drielsma merged 16 commits intoDeepLearnPhysics:developfrom
francois-drielsma:develop
May 2, 2025
Merged

Various speed-ups in the full chain routines#84
francois-drielsma merged 16 commits intoDeepLearnPhysics:developfrom
francois-drielsma:develop

Conversation

@francois-drielsma
Copy link
Member

Time savings in multiple locations:

  • Overhauled spine.utils.numba_local into spine.math
  • Vectorized form_clusters
  • Sped up cluster breaking in label adapter
  • In label adapter, moved from knn based adaptation to pure cdist (quicker)
  • Faster orphan assignment routine
  • Faster DBSCAN and connected components routine

Minor bug fixes:

  • Enable loading lite versions of objects
  • Add option to lexsort coordinates in the sparse3d parser
  • Corner case fix in the segmentation evaluation tool

@francois-drielsma francois-drielsma merged commit b02f465 into DeepLearnPhysics:develop May 2, 2025
1 check failed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant