Improving Spancat predictions #13430
Unanswered
Gitclop
asked this question in
Help: Model Advice
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
In the paper, that is referenced at the end of Spancat Blog-post, boundary distinctiveness is defined as the tokens that are immediatly in front/at the end of the span. Can I set the weight or bias for the boundary distinctiveness? To tell the modell to focus on learning those "external span indicators" more then the spans themselfs.
What kind of neural net is the spancat component made of? a cnn, lstm, or does it depend on the used base model, for example de_core_news_lg?
Beta Was this translation helpful? Give feedback.
All reactions