Corey J. Nolet
Corey J. Nolet
We should always test what we expect to work, and not expect things to work without tests. It looks like we have been adding new index wrappers to cuVS bench...
This needs to be fixed. @singhmanas1 can you please add the error stack trace and minimal reproducer here?
While cuVS-bench is great for running exhaustive grid searches, it becomes tedious and time consuming having to rerun the full grid searches every single time, especially when a Pareto might...
We recently came across a situation where a user noticed one of our examples crashed. It was a legitimate bug, and we would have caught it if we had been...
I just noticed the CAGRA search functions don't really tell the user what filtering functions it accepts. We should do a better job of communicating this in the docs by...
Users have requested the ability to build an IVFPQ index add a bunch of vectors to it, then extract the encoded verctors from each list so they can be stored...
UMAP's has had a looming bug for awhile that affects the reproduciblity when the laplacian eigenmaps initialization method is used. The bug lies in the lanczos solver in RAFT and...
We used UCX in the mnmg knn algorithm for 2 main reasons: 1. The set of workers hosting index partitions and query partitions could be disjoint 2. The NCCL p2p...
There are several algorithms in cuML which need support for different distances. For instance, our NearestNeighbors algorithm should be able to support the distances that sklearn's NearestNeighbors supports. UMAP-contrib uses...