With cluster bootstrap expanding like it has we should consider moving this into a class to be referenced by other parts of the code.
The ordering of the bootstrapping pipeline should be handled in init instead of main. The class should expose attributes for each of our variables in model state. For example -> model_id will probably turn into self.embedding_model and our new language model should live in something like self.language_model attributes.
Encapsulate all of cluster bootstrap into a class that can be used and referenced in other parts of the code.
Is your feature request related to a problem?
With cluster bootstrap expanding like it has we should consider moving this into a class to be referenced by other parts of the code.
The ordering of the bootstrapping pipeline should be handled in init instead of main. The class should expose attributes for each of our variables in model state. For example ->
model_id
will probably turn intoself.embedding_model
and our new language model should live in something likeself.language_model
attributes.Encapsulate all of cluster bootstrap into a class that can be used and referenced in other parts of the code.