Bringing regularized optimal transport to lightspeed: a splitting method adapted for GPUs

Open in new window