RAFT contains fundamental widely-used algorithms and primitives for machine learning and information retrieval. The algorithms are CUDA-accelerated and form building blocks for more easily writing high performance applications.
CAGRA graph construction with IVF-PQ will support inner product with PR https://github.com/rapidsai/raft/pull/2260. The same needs to be done for NN Descent. This is not done currently because NN Descent does not support graph building with metrics other than L2Expanded.
CAGRA graph construction with IVF-PQ will support inner product with PR https://github.com/rapidsai/raft/pull/2260. The same needs to be done for NN Descent. This is not done currently because NN Descent does not support graph building with metrics other than L2Expanded.