Actually I'm gonna deleter this if it's ok: "The function solves the Wasserstein barycenter problem when the barycenter measure is constrained to be supported on k atoms" It's redundant with the first sentence. Le mer. 15 juil. 2020 à 16:01, Rémi Flamary <notifications@github.com> a écrit :

@rflamary commented on this pull request. ________________________________ In ot/lp/__init__.py: > @@ -443,18 +443,30 @@ def free_support_barycenter(measures_locations,

measures_weights, X_init, b=None

- we do not optimize over the weights - we do not do line search for the locations updates, we use i.e.

theta = 1 in [1] (Algorithm 2). This can be seen as a discrete implementation of the fixed-point algorithm of [2] proposed in the continuous setting.

+ Formally we solve the following problem: Could you please put the optimization problem after the first sentence

and before the discussion on the algorithm?

Also the itemize in the discussion does not work (it needs an empty line

maybe)

…