Open albertpod opened 2 weeks ago
Can you try and see if it works as a workaround?
μ=PCA_block(in = [ x[i], w... ])
EDIT: ah, sorry, it won't work probably either
I think its a real issue, but should be fixable. We need to define iterate
, which probably should reuse some code from Base.broadcastable(ref::VariableRef)
(it may even call broadcastable
?). We may had some justification for not including iterate
on VariableRef
, but I cannot recall it.
I know the issue with splatting was addressed here, but I am having a trouble with it, when running the
infer
function with the provided PCA model (see below). The following error is encountered:The error occurs at the line: https://github.com/ReactiveBayes/GraphPPL.jl/blob/b3ae2a6917e898f3bfe653f538a191f61f7bde34/src/model_macro.jl#L542
The minimum code example to reproduce the error:
The issue appears to be related to the splatting of the
w
array in thePCA_block
function.