You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Thanks for this great work which makes FG inference a lot easier!
In demo (ForneyLab.jl-master/demo/bayes_rule_2d.ipynb)
Slightly change of the calculation in factor graph makes it fail:
@rv y ~ GaussianMeanVariance(A*(x.^2) + b, Sigma_y) # Likelihood
The error message is:
MethodError: no method matching length(::Variable)
Could any one help? Thanks.
The text was updated successfully, but these errors were encountered:
Unfortunately, Variable type doesn't support broadcasting.
If you want to apply a quadratic function to a multivariate Gaussian random variable you can (1) use a Nonlinear node (see demo/variational_laplace_and_sampling.ipynb), (2) define your own node or Composite node and provide the required update rules for it (see demo/composite_nodes.ipynb).
It will be useful if you provide the entire snippet where you specify your model.
Thanks for this great work which makes FG inference a lot easier!
In demo (ForneyLab.jl-master/demo/bayes_rule_2d.ipynb)
Slightly change of the calculation in factor graph makes it fail:
@rv y ~ GaussianMeanVariance(A*(x.^2) + b, Sigma_y) # Likelihood
The error message is:
MethodError: no method matching length(::Variable)
Could any one help? Thanks.
The text was updated successfully, but these errors were encountered: