error with demCmu35gplvmReconstruct.m about gpmat HOT 4 CLOSED

That might be difficult to trace, I've done an ack-grep in all files of GPmat and "indexAll" only appears in that line...
If I understand correctly, fgplvmOptimiseSequence is similar to fgplvmOptimisePoint but the points are correlated. In that context, indexAll might (?) denote which indexes of X form one (or more?) sequence(s).
Sorry I haven't been able to help more, but it seems to me it has to do with indexing elements of X (e.g. 1:size(model.X,1) for the whole set treated as sequence).

from gpmat.

Thank you. I try fgplvmOptimisePoint as suggested and the error is gone.

from gpmat.

Do you think a fix for this in the main repo is appropriate?

from gpmat.

Yes, I added it, thanks. But bear in mind that fgplvmOptimisePoint won't take sequential correlations into account, so the behaviour is not 100% the same.

from gpmat.