rapidsai/raft
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.
Provides header-only C++ templates with optional precompiled shared library to reduce compile times, alongside `pylibraft` Python bindings and `raft-dask` for multi-node GPU distribution. Built on mdspan multi-dimensional array views and RMM memory management, it integrates with cuBLAS, cuSolver, and the broader RAPIDS ecosystem while maintaining minimal dependencies for interoperability with CuPy, PyTorch, and other `__cuda_array_interface__`-compatible libraries.
988 stars. Actively maintained with 20 commits in the last 30 days.
Stars
988
Forks
227
Language
Cuda
License
Apache-2.0
Category
Last pushed
Mar 13, 2026
Commits (30d)
20
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/vector-db/rapidsai/raft"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.