cugraph_pyg.sampler.BaseSampler#

class cugraph_pyg.sampler.BaseSampler( sampler: DistributedNeighborSampler, data: Tuple[FeatureStore, GraphStore], batch_size: int = 16, )[source]#

Adapt a distributed cuGraph sampler to PyG sampler interfaces.

Parameters:

samplerDistributedNeighborSampler: The distributed cuGraph neighbor sampler.
dataTuple[FeatureStore, GraphStore]: The PyG feature store and graph store to sample from.
batch_sizeint, optional: Number of input nodes or edges in each minibatch.

Methods

sample_from_edges
sample_from_nodes

__init__( sampler: DistributedNeighborSampler, data: Tuple[FeatureStore, GraphStore], batch_size: int = 16, )[source]#

Methods

`__init__`(sampler, data[, batch_size])
`sample_from_edges`(index[, neg_sampling])
`sample_from_nodes`(index, **kwargs)