Open juliohm opened 1 year ago
A combination of this and https://github.com/ThummeTo/ForwardDiffChainRules.jl can work yes. And it can be used with Turing if you use ChainRules-compatible AD or ForwardDiff with the above package.
It can be used assuming we have BK implemented here which we don't yet.
Thank you @mohamed82008 for this package. Can it solve https://github.com/JuliaDiff/ForwardDiff.jl/issues/602?
I wonder if we can easily use the factorizations implemented here with Turing.jl models.