Spectral Clustering

cpp/CMakeLists.txt

@@ -349,6 +349,7 @@ if(NOT BUILD_CPU_ONLY)
     src/cluster/kmeans_transform_double.cu
     src/cluster/kmeans_transform_float.cu
     src/cluster/single_linkage_float.cu
+    src/cluster/spectral.cu
     src/core/bitset.cu
     src/core/omp_wrapper.cpp
     src/distance/detail/kernels/gram_matrix.cu

cpp/include/cuvs/cluster/spectral.hpp

@@ -0,0 +1,33 @@
+/*
+ * SPDX-FileCopyrightText: Copyright (c) 2025, NVIDIA CORPORATION.
+ * SPDX-License-Identifier: Apache-2.0
+ */
+#pragma once
+
+#include <raft/core/device_coo_matrix.hpp>
+#include <raft/core/device_mdspan.hpp>
+#include <raft/core/resources.hpp>
+#include <raft/random/rng_state.hpp>
+
+namespace cuvs::cluster::spectral {
+
+struct params {
+  int n_clusters;
+  int n_components;
+  int n_init;
+  int n_neighbors;
+  raft::random::RngState rng_state{0};
+};
+
+// TODO: int64_t nnz support (see https://github.com/rapidsai/cuvs/issues/1484)
+void fit_predict(raft::resources const& handle,
+                 params config,
+                 raft::device_coo_matrix_view<float, int, int, int> connectivity_graph,
+                 raft::device_vector_view<int, int> labels);
+
+void fit_predict(raft::resources const& handle,
+                 params config,
+                 raft::device_coo_matrix_view<double, int, int, int> connectivity_graph,
+                 raft::device_vector_view<int, int> labels);
+
+}  // namespace cuvs::cluster::spectral

cpp/include/cuvs/preprocessing/spectral_embedding.hpp

@@ -159,8 +159,22 @@ void transform(raft::resources const& handle,
                raft::device_coo_matrix_view<float, int, int, int> connectivity_graph,
                raft::device_matrix_view<float, int, raft::col_major> embedding);
 
+void transform(raft::resources const& handle,
+               params config,
+               raft::device_coo_matrix_view<double, int, int, int> connectivity_graph,
+               raft::device_matrix_view<double, int, raft::col_major> embedding);
+
 /**
  * @}
  */
 
 }  // namespace cuvs::preprocessing::spectral_embedding
+
+namespace cuvs::preprocessing::spectral_embedding::helpers {
+
+void create_connectivity_graph(raft::resources const& handle,
+                               params spectral_embedding_config,
+                               raft::device_matrix_view<float, int, raft::row_major> dataset,
+                               raft::device_coo_matrix<float, int, int, int>& connectivity_graph);
+
+}  // namespace cuvs::preprocessing::spectral_embedding::helpers

cpp/src/cluster/detail/spectral.cuh

@@ -0,0 +1,63 @@
+/*
+ * SPDX-FileCopyrightText: Copyright (c) 2025, NVIDIA CORPORATION.
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+#pragma once
+
+#include <cuvs/cluster/kmeans.hpp>
+#include <cuvs/cluster/spectral.hpp>
+#include <cuvs/preprocessing/spectral_embedding.hpp>
+#include <raft/core/device_mdarray.hpp>
+#include <raft/linalg/transpose.cuh>
+#include <raft/random/rng_state.hpp>
+
+namespace cuvs::cluster::spectral::detail {
+
+template <typename DataT>
+void fit_predict(raft::resources const& handle,
+                 params config,
+                 raft::device_coo_matrix_view<DataT, int, int, int> connectivity_graph,
+                 raft::device_vector_view<int, int> labels)
+{
+  int n_samples = connectivity_graph.structure_view().get_n_rows();
+  DataT inertia;
+  int n_iter;
+  auto embedding_col_major =
+    raft::make_device_matrix<DataT, int, raft::col_major>(handle, n_samples, config.n_components);
+  auto embedding_row_major =
+    raft::make_device_matrix<DataT, int, raft::row_major>(handle, n_samples, config.n_components);
+  cuvs::preprocessing::spectral_embedding::params spectral_embedding_config;
+  spectral_embedding_config.n_components   = config.n_components;
+  spectral_embedding_config.n_neighbors    = config.n_neighbors;
+  spectral_embedding_config.norm_laplacian = true;
+  spectral_embedding_config.drop_first     = false;
+  spectral_embedding_config.seed           = config.rng_state.seed;
+
+  cuvs::cluster::kmeans::params kmeans_config;
+  kmeans_config.n_clusters          = config.n_clusters;
+  kmeans_config.rng_state           = config.rng_state;
+  kmeans_config.n_init              = config.n_init;
+  kmeans_config.oversampling_factor = 0.0;
+
+  cuvs::preprocessing::spectral_embedding::transform(
+    handle, spectral_embedding_config, connectivity_graph, embedding_col_major.view());
+
+  raft::linalg::transpose(handle,
+                          embedding_col_major.data_handle(),
+                          embedding_row_major.data_handle(),
+                          n_samples,
+                          config.n_components,
+                          raft::resource::get_cuda_stream(handle));
+
+  cuvs::cluster::kmeans::fit_predict(handle,
+                                     kmeans_config,
+                                     embedding_row_major.view(),
+                                     std::nullopt,
+                                     std::nullopt,
+                                     labels,
+                                     raft::make_host_scalar_view(&inertia),
+                                     raft::make_host_scalar_view(&n_iter));
+}
+
+}  // namespace cuvs::cluster::spectral::detail

cpp/src/cluster/spectral.cu

@@ -0,0 +1,26 @@
+/*
+ * SPDX-FileCopyrightText: Copyright (c) 2025, NVIDIA CORPORATION.
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+#include "./detail/spectral.cuh"
+
+#include <cuvs/cluster/spectral.hpp>
+
+namespace cuvs::cluster::spectral {
+
+#define CUVS_INST_SPECTRAL(DataT)                                                         \
+  void fit_predict(raft::resources const& handle,                                         \
+                   params config,                                                         \
+                   raft::device_coo_matrix_view<DataT, int, int, int> connectivity_graph, \
+                   raft::device_vector_view<int, int> labels)                             \
+  {                                                                                       \
+    detail::fit_predict<DataT>(handle, config, connectivity_graph, labels);               \
+  }
+
+CUVS_INST_SPECTRAL(float);
+CUVS_INST_SPECTRAL(double);
+
+#undef CUVS_INST_SPECTRAL
+
+}  // namespace cuvs::cluster::spectral