Superlearn check weights about zepid

[pzivich]

Yes, calling SuperLearner.summary() returns the weights given to each learner.

If this is related to #167 then things aren't as straightforward. summary() is not called because of the volume of output. The cross-fit procedure means that there are multiple SL fit to each partition of the data (a min of 2 for single cross-fit). A number of different partitions are done by the cross-fitting procedure. As the recommendation is like 100 different partitions, the weights for each partition are not output (printed or saved).

I will consider how that could be added to things like SingleCrossfitTMLE

[miaow27]

Thank you @pzivich! able to know weights (at least average weights across each partition) for the model candidates would be very helpful!