icemiliang / pyvot Goto Github PK

• add label to potential
• add curvature
• update arxiv
• curvature kp_ds replace for loop by matrix operations
• regularization flag replace int to string specifying regulairzation term
• use new results to demo potential
• add affine
• unit test affine

• readme reference links
• add comments following the Google style
• add try except
• add comments to vot saying it is outdated

The demos of VOTAP seem to be experimenting with 2D data. Does VOTAP support high-dimensional data?

Hi,

I was going through the paper "Regularized Wasserstein Means for Aligning Distributional Data" and wanted to know how can these losses be incorporated for distribution alignment problems on custom datasets.

Thank you