Sampling#

Random Walks#

single-GPU#

cugraph.biased_random_walks(G[, ...])

Compute biased random walks for each nodes in 'start_vertices'.

cugraph.heterogeneous_neighbor_sample(G, ...)

Performs uniform/biased neighborhood sampling, which samples nodes from a graph based on the current node's neighbors, with a corresponding fan_out value at each hop.

cugraph.homogeneous_neighbor_sample(G, ...)

Performs uniform/biased neighborhood sampling, which samples nodes from a graph based on the current node's neighbors, with a corresponding fan_out value at each hop.

cugraph.uniform_neighbor_sample(G, ...[, ...])

Does neighborhood sampling, which samples nodes from a graph based on the current node's neighbors, with a corresponding fanout value at each hop.

multi-GPU#

cugraph.dask.sampling.biased_random_walks.biased_random_walks(...)

compute random walks under the biased sampling framework for each nodes in 'start_vertices' and returns a padded result along with the maximum path length.

cugraph.dask.sampling.random_walks.random_walks(...)

compute random walks for each nodes in 'start_vertices' and returns a padded result along with the maximum path length.

cugraph.dask.sampling.uniform_neighbor_sample.uniform_neighbor_sample(...)

Does neighborhood sampling, which samples nodes from a graph based on the current node's neighbors, with a corresponding fanout value at each hop.

cugraph.dask.sampling.uniform_random_walks.uniform_random_walks(...)

compute random walks under the uniform sampling framework for each nodes in 'start_vertices' and returns a padded result along with the maximum path length.

Node2Vec#

single-GPU#

multi-GPU#

cugraph.dask.sampling.node2vec_random_walks.node2vec_random_walks(...)

compute random walks under the node2vec sampling framework for each nodes in 'start_vertices' and returns a padded result along with the maximum path length.