cugraph_pyg.loader.dask_node_loader.DaskNeighborLoader#
- class cugraph_pyg.loader.dask_node_loader.DaskNeighborLoader(data: Union[DaskGraphStore, Tuple[DaskGraphStore, DaskGraphStore]], input_nodes: Union[Tensor, None, str, Tuple[str, Optional[Tensor]], int] = None, batch_size: Optional[int] = None, **kwargs)[source]#
Duck-typed version of the PyG NeighborLoader interface that uses Dask to sample nodes using the uniform neighbor sampling algorithm.
- Attributes:
- batch_size
- __init__(data: Union[DaskGraphStore, Tuple[DaskGraphStore, DaskGraphStore]], input_nodes: Union[Tensor, None, str, Tuple[str, Optional[Tensor]], int] = None, batch_size: Optional[int] = None, **kwargs)[source]#
Constructs a new DaskNeighborLoader object.
- Parameters:
- data: DaskGraphStore or (DaskGraphStore, DaskGraphStore)
The DaskGraphStore or stores where the graph/feature data is held.
- batch_size: int (required)
The number of input nodes in each batch.
- input_nodes: Union[InputNodes, int] (required)
The input nodes associated with this sampler.
- **kwargs: kwargs
Keyword arguments to pass through for sampling. i.e. “shuffle”, “fanout” See BulkSampleLoader.
Methods
__init__(data[, input_nodes, batch_size])Constructs a new DaskNeighborLoader object.
Attributes
batch_size