Giters

ljynlp / MRN

Source code for ACL-IJCNLP 2021 findings paper: MRN: A Locally and Globally Mention-Based Reasoning Network for Document-Level Relation Extraction.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

why is the NT F1 exactly the same as F1 score running this code?

ZhichaoDuan opened this issue · comments

commented

I cloned this repo and try to replicate the results and I found that each time I evaluate the model performance after training, the NT f1 score and the f1 score are the exact same which is odd..

commented

There existed such a bug, but I thought it was fixed. I haven't encountered it since I updated the code.
Did you try the bert version?

commented

Yeah, I used the bert version but encountered this problem. As for the glove version, I haven't tested yet.