Closed JSzitas closed 2 years ago
It is quite obnoxious having to retrain many individual models (if using soothsayer for model combining/selection)
A quick sketch of an interface around this:
Take a mdl_df object, and add a new column for soothsayer by calling some function, i.e.
add_soothsayer( mdl_df, soothsayer_model, ... )
and potentially have it look for models it would be missing (and compute those) or have it take the fitted models (and combine them).
Resolved as of ecbc77a0022d3f921e5f82b83ea72c106ea88068
It is quite obnoxious having to retrain many individual models (if using soothsayer for model combining/selection)
A quick sketch of an interface around this:
Take a mdl_df object, and add a new column for soothsayer by calling some function, i.e.
add_soothsayer( mdl_df, soothsayer_model, ... )
and potentially have it look for models it would be missing (and compute those) or have it take the fitted models (and combine them).