moscot.base.solver.OTSolver.__call__

OTSolver.__call__(xy=None, x=None, y=None, tags=mappingproxy({}), device=None, **kwargs)

Solve an optimal transport problem.

Parameters:

• xy (Union[TaggedArray, ndarray[tuple[int, ...], dtype[floating]], Array, Tuple[Union[ndarray[tuple[int, ...], dtype[floating]], Array], Union[ndarray[tuple[int, ...], dtype[floating]], Array]], None]) – Data that defines the linear term.

• x (Union[TaggedArray, ndarray[tuple[int, ...], dtype[floating]], Array, None]) – Data of the first geometry that defines the quadratic term.

• y (Union[TaggedArray, ndarray[tuple[int, ...], dtype[floating]], Array, None]) – Data of the second geometry that defines the quadratic term.

• tags (Mapping[Literal['xy', 'x', 'y'], Tag]) – How to interpret the data in xy, x and y.

• device (Union[Literal['cpu', 'gpu', 'tpu'], str, None]) – Device to transfer the output to, see to().

• is_conditional – Whether the OT problem is conditional.

• kwargs (Any) – Keyword arguments for parent’s __call__().

Return type:

TypeVar(O, bound= BaseSolverOutput)

Returns:

: The optimal transport solution.