Micka
Micka
As pointed by Tamas in his review, we should consider whether it is more efficient to pre-computing the norms before the distance computation kernel. https://github.com/rapidsai/cuvs/pull/179#pullrequestreview-2189437185 > We should ask the...
Currently the tests for ANN indexes are not checking that the distances computed are correct. https://github.com/rapidsai/cuvs/blob/branch-24.10/cpp/test/neighbors/ann_utils.cuh#L130-133
PR #1279 introduced a new API for sparse matrices. There is currently no bindings for it in Python, so any function in Python that accept sparse matrices as input have...
https://github.com/rapidsai/raft/pull/2481#discussion_r1828240971 Follow-up of Lanczos PR, to find a way to generate and validate test data without hardcoding
Follow-up of the Lanczos PR. https://github.com/rapidsai/raft/pull/2481#discussion_r1828183184
Related to #841. The IVF-PQ build metric is not properly initialized in the case of InnerProduct, so I am adding here correct initialization on the C layer, as well as...
This PR add support for half dtype for HNSW in C++, C and python, as well as some tests with it. I had to modify a bit the HNSW patch...
ANN algorithm with metric=`inner_product` and dtype=`int8` or `uint8` are not tested properly. In pytest the dataset generated is in (u)int8, then it gets normalized and transformed in float. So when...
This is a WIP. TODO: Add support for CAGRA+HNSW int8 issue: #1201