MCPcopy Create free account
hub / github.com/PythonOT/POT / OTResult

Class OTResult

ot/utils.py:1024–1315  ·  view source on GitHub ↗

Base class for OT results. Parameters ---------- potentials : tuple of array-like, shape (`n1`, `n2`) Dual potentials, i.e. Lagrange multipliers for the marginal constraints. This pair of arrays has the same shape, numerical type and properties as the input weig

Source from the content-addressed store, hash-verified

1022
1023
1024class OTResult:
1025 """Base class for OT results.
1026
1027 Parameters
1028 ----------
1029
1030 potentials : tuple of array-like, shape (`n1`, `n2`)
1031 Dual potentials, i.e. Lagrange multipliers for the marginal constraints.
1032 This pair of arrays has the same shape, numerical type
1033 and properties as the input weights "a" and "b".
1034 value : float, array-like
1035 Full transport cost, including possible regularization terms and
1036 quadratic term for Gromov Wasserstein solutions.
1037 value_linear : float, array-like
1038 The linear part of the transport cost, i.e. the product between the
1039 transport plan and the cost.
1040 value_quad : float, array-like
1041 The quadratic part of the transport cost for Gromov-Wasserstein
1042 solutions.
1043 plan : array-like, shape (`n1`, `n2`)
1044 Transport plan, encoded as a dense array.
1045 log : dict
1046 Dictionary containing potential information about the solver.
1047 backend : Backend
1048 Backend used to compute the results.
1049 sparse_plan : array-like, shape (`n1`, `n2`)
1050 Transport plan, encoded as a sparse array.
1051 lazy_plan : LazyTensor
1052 Transport plan, encoded as a symbolic POT or KeOps LazyTensor.
1053 status : int or str
1054 Status of the solver.
1055 batch_size : int
1056 Batch size used to compute the results/marginals for LazyTensor.
1057
1058 Attributes
1059 ----------
1060
1061 potentials : tuple of array-like, shape (`n1`, `n2`)
1062 Dual potentials, i.e. Lagrange multipliers for the marginal constraints.
1063 This pair of arrays has the same shape, numerical type
1064 and properties as the input weights "a" and "b".
1065 potential_a : array-like, shape (`n1`,)
1066 First dual potential, associated to the "source" measure "a".
1067 potential_b : array-like, shape (`n2`,)
1068 Second dual potential, associated to the "target" measure "b".
1069 value : float, array-like
1070 Full transport cost, including possible regularization terms and
1071 quadratic term for Gromov Wasserstein solutions.
1072 value_linear : float, array-like
1073 The linear part of the transport cost, i.e. the product between the
1074 transport plan and the cost.
1075 value_quad : float, array-like
1076 The quadratic part of the transport cost for Gromov-Wasserstein
1077 solutions.
1078 plan : array-like, shape (`n1`, `n2`)
1079 Transport plan, encoded as a dense array.
1080 sparse_plan : array-like, shape (`n1`, `n2`)
1081 Transport plan, encoded as a sparse array.

Callers 5

solveFunction · 0.85
solve_gromovFunction · 0.85
solve_sampleFunction · 0.85
solve_batchFunction · 0.85
solve_gromov_batchFunction · 0.85

Calls

no outgoing calls

Tested by

no test coverage detected