Skip to content
Discussion options

You must be logged in to vote

Hi!

I think a lot of the confusion is coming from the fact that in spaCy v2.x, the tagger would assign both token.tag and token.pos, while it only assigns token.tag since v3.0. The migration guide specifies how to use the new attribute_ruler to obtain token.pos annotations.

The current release and documentation of Prodigy focuses on Prodigy v.10, which is compatible with spaCy 2. We're working on Prodigy v.11, with updated tagging recipes, to be compatible with spaCy v3, but this is not officially released yet.

Can an existing transform model be trained or only the sm/lg model can be trained?

Yep - you should be able to fine-tune an existing transformer-based pipeline as you would with …

Replies: 1 comment 5 replies

Comment options

You must be logged in to vote
5 replies
@adrianeboyd
Comment options

@qingyun1988
Comment options

@qingyun1988
Comment options

@svlandeg
Comment options

@qingyun1988
Comment options

Answer selected by svlandeg
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
feat / tagger Feature: Part-of-speech tagger
3 participants