vak.datasets.parametric_umap.parametric_umap.get_graph_elements#

vak.datasets.parametric_umap.parametric_umap.get_graph_elements(graph: coo_matrix, n_epochs: int) → tuple[coo_matrix, ndarray[Any, dtype[_ScalarType_co]], ndarray[Any, dtype[_ScalarType_co]], ndarray[Any, dtype[_ScalarType_co]], ndarray[Any, dtype[_ScalarType_co]], int][source]#

Get graph elements for Parametric UMAP Dataset.

Parameters:

  • graph (scipy.sparse.csr_matrix) – The graph returned by get_umap_graph().

  • n_epochs (int) – Number of epochs model will be trained

Returns:

  • graph (scipy.sparse._coo.coo_matrix) – The graph, now in COOrdinate format.

  • epochs_per_sample (int)

  • head (numpy.ndarray) – Graph rows.

  • tail (numpy.ndarray) – Graph columns.

  • weight (numpy.ndarray) – Graph data.

  • n_vertices (int) – Number of vertices in dataset.