Thanks for this great code. I am wondering how can I explicitly specify a kernal. In the readme file, you say there are three possible values for the kernel argument (i.e., CONST_KERNEL, DIAG_KERNEL (the default), and FULL_KERNEL) which specifies the kernel's precision-matrix Λ(m). Suppose I can precompute a Λ(m) and want to pass it when I call the addPairwiseEnergy() function, is there a way that I can do this?
Hi @lucasb-eyer,
Thanks for this great code. I am wondering how can I explicitly specify a kernal. In the readme file, you say there are three possible values for the kernel argument (i.e., CONST_KERNEL, DIAG_KERNEL (the default), and FULL_KERNEL) which specifies the kernel's precision-matrix Λ(m). Suppose I can precompute a Λ(m) and want to pass it when I call the addPairwiseEnergy() function, is there a way that I can do this?
Thanks.