No total variation regulatization in unbalanced OT

[jolatechno]

My PhD research would gain from unbalanced OT with total variation regularization for the marginal relaxation term.

Most solvers in ot.unbalanced do not support total variation regularization for the marginal relaxation term, is there a reason not to support it ?

If it is reasonable to add support for total variation I would be happy to implement it and make a pull request !

[jolatechno]

I'm not sure but solving unbalanced OT with total variation regularization on the marginal seems trivially equivalent to solving balanced OT and removing from the transport plan transport with associated cost greater than the regularization term. If it is the case, is that the reason for not supporting total variation ?