[FEA] Integrate balanced KMeans
Open
@Nyrio is already working on this.
Since Aug 22, 2022.
feature request
- Dominant language
- Cuda
- Stars
- 1k
- Forks
- 251
- Avg merge
- 1d 8h
- Merged PRs (30d)
- 31
Description
The approximate nearest neighbors algorithms (IVF-Flat #652, IVF-PQ) use a balanced k-means implementation. In #652, balanced kmeans is separate implementation from #605. These two implementations should be merged to avoid code duplications.
One should ensure that the changes are benchmarked, not to loose performance for the approximate nearest neighbors methods.
Tasks
- move k-means benchmarks from cuML #795
- add balanced k-means benchmark #795
- define public API for balanced k-means #1113
- reduce duplication in primitive functions (various PRs)
About reducing the duplication of primitive functions, the following issues came up earlier:
- https://github.com/rapidsai/raft/pull/652#discussion_r915038999 copyRows
- https://github.com/rapidsai/raft/pull/652#discussion_r915047790 #912
- https://github.com/rapidsai/raft/pull/652#discussion_r915050104 #911
- https://github.com/rapidsai/raft/pull/652#discussion_r915052244 #909
- compare
minClusterAnddistanceComputeand_cuann_kmeans_predict[_core], - replace
_cuann_kmeans_update_centers[outdated] withreduce_cols_by_keyorreduce_rows_by_key,
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Assessment
This issue has not been assessed yet.