Open zhangchengxian-git opened 1 year ago
Yes, it only evaluated the coarse-grained setting I haven’t tested it on the fine-grained version. It does very well on Ontonotes which consist of 18 NER labels, I assume it might also work on the 47 labels, but it might depend on the training data size.
I got it! Thank you so much.
How do you get the "char_vocab.eng.genia.txt"
Sorry to bother you, but i have a question about GENIA in your code. Do you consider GENIA's categories as the following six categories?
If so, does your method perform well in fine-grained division (47 categories) in GENIA?