keras-nlp
keras-nlp copied to clipboard
Consider an `early_stopping=True` flag for `generate()`
Currently our generate function always ceases when all sequences contain a new endoftext token. We might instead want to make this configurable, so that generation can continue past end of document markers.
I am not sure if we need this, is it a valid/common use case to generate passing the end token?
This is now configurable via stop_token_ids, closing.