ott.solvers.quadratic.gromov_wasserstein.GWOutput#
- class ott.solvers.quadratic.gromov_wasserstein.GWOutput(costs=None, linear_convergence=None, converged=False, errors=None, linear_state=None, geom=None, old_transport_mass=1.0)[source]#
Holds the output of the Gromov-Wasserstein solver.
- Parameters:
costs (Optional[Array]) – Holds the sequence of regularized GW costs seen through the outer loop of the solver.
linear_convergence (Optional[Array]) – Holds the sequence of bool convergence flags of the inner Sinkhorn iterations.
converged (bool) – Convergence flag for the outer GW iterations.
errors (Optional[Array]) – Holds sequence of vectors of errors of the Sinkhorn algorithm at each iteration.
linear_state (Optional[Union[SinkhornOutput, LRSinkhornOutput]]) – State used to solve and store solutions to the local linearization of GW.
geom (Optional[Geometry]) – The geometry underlying the local linearization.
old_transport_mass (float) – Holds total mass of transport at previous iteration.
Methods
apply
(inputs[, axis])Apply the transport to an array; axis=1 for its transpose.
count
(value, /)Return number of occurrences of value.
index
(value[, start, stop])Return first index of value.
set
(**kwargs)Return a copy of self, possibly with overwrites.
Attributes
Alias for field number 2
Alias for field number 0
Alias for field number 3
Alias for field number 5
Alias for field number 1
Alias for field number 4
Transport matrix.
Alias for field number 6
Return transport cost of current linear OT solution at geometry.
Regularized optimal transport cost of the linearization.