sklearn wrappers without TF-DF?

[memeplex]

AFAICU the preferred way of using Yggdrasil is directly through YDF, not TF-DF.

Which is the way to integrate them with Sklearn worflows (say GridSearchCV), if possible at all?

I'm not sure but I assume that TF-DF models can be wrapped as regular Keras models in order to be used from Sklearn. But what about vanilla YDF models?

Thanks!