summaryrefslogtreecommitdiff
path: root/ot/lp/emd_wrap.pyx
diff options
context:
space:
mode:
authorRémi Flamary <remi.flamary@gmail.com>2020-01-30 08:04:36 +0100
committerRémi Flamary <remi.flamary@gmail.com>2020-01-30 08:04:36 +0100
commitf65073faa73b36280a19ff8b9c383e66f8bdbd2b (patch)
treee786f1ec744bbf28928ac72bb705f172b04a4562 /ot/lp/emd_wrap.pyx
parent9a9b3547837eac56349ce8df92bb5b0565daa2d6 (diff)
comlete documentation
Diffstat (limited to 'ot/lp/emd_wrap.pyx')
-rw-r--r--ot/lp/emd_wrap.pyx6
1 files changed, 6 insertions, 0 deletions
diff --git a/ot/lp/emd_wrap.pyx b/ot/lp/emd_wrap.pyx
index a4987f4..d345fd4 100644
--- a/ot/lp/emd_wrap.pyx
+++ b/ot/lp/emd_wrap.pyx
@@ -66,6 +66,12 @@ def emd_c(np.ndarray[double, ndim=1, mode="c"] a, np.ndarray[double, ndim=1, mod
.. warning::
Note that the M matrix needs to be a C-order :py.cls:`numpy.array`
+ .. warning::
+ The C++ solver discards all samples in the distributions with
+ zeros weights. This means that while the primal variable (transport
+ matrix) is exact, the solver only returns feasible dual potentials
+ on the samples with weights different from zero.
+
Parameters
----------
a : (ns,) numpy.ndarray, float64