Replace operation_weights in CategoryConstructor by a better mechanism #1678
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
When using the string "default" for create_func_*, one must now set the option
underlying_category
, from which the weight is determined. If create_func_* is a function, one can (optionally) set a weight by returning a pair of a string and an integer (instead of only a string).@mohamed-barakat This PR partially implements a suggestion you had: You suggested to pass an underlying category to CategoryConstructor from which the weights can be taken. I argued that this does not make sense in general. However, while working with the weights, I noticed that it indeed does make sense in a specific case, namely when using the string "default" for create_func_*. This PR now requires to set the option
underlying_category
when using the string "default". This sets the correct weights, which probably requires changes like recompiling code. Please tell me when I can merge, i.e. when you have to time to adjust things in CategoricalTowers.