Replies: 1 comment 1 reply
-
REL's architecture already uses the same three components you mention. It's already possible to use the API with and without a given mention. It shouldn't be too hard to also make a call for the last step only, this doesn't have a very high priority here though... What's the use case, easing research or do you have a "real" application in mind? |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hey there!
Love the linker!
I'd like to include (parts of) REL in an entity linking framework I've been working on for a while - if the devs are fine with it, of course. I'm working on a framework dividing up linkers into 3 main steps (additionally to a bunch of interactions through merging/splitting capabilities): mention detection, candidate generation and entity disambiguation.
At an initial glance, it seems a tad "difficult" to separate those steps in REL despite your wonderful Youtube video showing that it's basically "multiple awesome components" executed after one another. Could you perchance give me some pointers how to potentially achieve my task or add API functionality for calling on different sub-tasks?
Thanks again for the great work & wishing a great week!
Best regards,
kmdn
Beta Was this translation helpful? Give feedback.
All reactions