Closed dirknbr closed 2 years ago
Ok I found the bug, my Y was a matrix but needs to be vector.
Error message was confusing
@dirknbr: We're working on trying to formulate error messages that'll be clearer to users. Do you have a suggestion for how we can rephrase the one you ran into (keeping in mind that multi_normal
has multiple signatures)?
Ill-typed arguments to '~' statement. No distribution 'multi_normal' was found with the correct signature.
P.S. If you want to vectorize, the variate can be an array of vectors (as can the location param, I believe).
I am getting this error
Ill-typed arguments to '~' statement. No distribution 'multi_normal' was found with the correct signature.Traceback (most recent call last):
So for some strange reason that distribution is not supported. I updated my pystan to latest version with pip. But the latest stan version shows it exists.
https://mc-stan.org/docs/2_28/functions-reference/multivariate-normal-distribution.html
Is maybe the error message misleading?