Closed dpaetzel closed 8 months ago
(I'm referring to this line in the README based on which I'd expect just a <: Table(Continuous)
.)
This is not a bug, but your question is a good one. I think it is answered here. You can ignore the convention C
, which is always the default one in MLJ.
I see, thank you! The link clarified it for me. :slightly_smiling_face:
Hi! Thank you for creating and maintaining this project!
I'm rather new to Julia, so this may be a dumb question (sorry in that case).
I currently have some issues getting
MultitargetNeuralNetworkRegressor
to work properly using a custom layer and loss function. In the process of debugging that, I noticed, that even without my custom layer, thetarget_scitype
does not conform to the one given in the README:yields
Union{Table{<:AbstractVector{<:Continuous}}, AbstractMatrix{Continuous}}
.I'm a bit confused about the nesting of
Table
andAbstractVector
here; am I doing something wrong or is my mental model off?