MLR3 using trafo (transformations) within a learner branch/dependencies hitting “non-numeric argument” error when transposing

Question

I\'m trying to tune a few models using instances, so there is a branching step with several models. I have the pipe constructed and works without models that require trafos.