ott.solvers.linear.sinkhorn.SinkhornOutput

class ott.solvers.linear.sinkhorn.SinkhornOutput(f=None, g=None, errors=None, reg_ot_cost=None, ot_prob=None)

Implements the problems.Transport interface, for a Sinkhorn solution.

Methods

apply(inputs[, axis])	Apply the transport to a ndarray; axis=1 for its transpose.
cost_at_geom(other_geom)	Return reg-OT cost for matrix, evaluated at other cost matrix.
count(value, /)	Return number of occurrences of value.
index(value[, start, stop])	Return first index of value.
marginal(axis)	rtype Array
set(**kwargs)	Return a copy of self, with potential overwrites.
set_cost(ot_prob, lse_mode, use_danskin)	rtype SinkhornOutput
to_dual_potentials()	Return the entropic map estimator.
transport_cost_at_geom(other_geom)	Return bare transport cost of current solution at any geometry.

Attributes

a	rtype Array
b	rtype Array
converged	rtype bool
dual_cost	Return transport cost in dual form of current solution.
errors	Alias for field number 2
f	Alias for field number 0
g	Alias for field number 1
geom	rtype Geometry
linear	rtype bool
linear_output	rtype bool
matrix	Transport matrix if it can be instantiated.
n_iters	rtype int
ot_prob	Alias for field number 4
primal_cost	Return transport cost of current solution at geometry.
reg_ot_cost	Alias for field number 3
scalings	rtype Tuple[Array, Array]
transport_mass	Sum of transport matrix.

Parameters

• f (Optional[jax.Array]) –

• g (Optional[jax.Array]) –

• errors (Optional[jax.Array]) –

• reg_ot_cost (Optional[float]) –

• ot_prob (Optional[ott.problems.linear.linear_problem.LinearProblem]) –