Some problems with the dataset. about ncrfpp

[zhenlanghuo]

@caoxu915683474 you need transform the BIO tags to BIOES tags by yourself.

[caoxu915683474]

@zhenlanghuo Thanks a lot , I forget to do this transform work by myself. Haha!

[jiesutd]

Thanks @zhenlanghuo !

[udion]

It seems like the code is written such that even if I pass the original CoNLL-2003 dataset, where the first column is the words and the last column is the NER tags, and inbetween columns is some more tags, it will only use the first columns for words and last column for tags, so I don't really have to convert the CoNLL-2003 format to the format given by you in sample_dataset, right?

i.e the original CoNLL-2003 is like

EU NNP I-NP I-ORG
rejects VBZ I-VP O
German JJ I-NP I-MISC
call NN I-NP O
to TO I-VP O
boycott VB I-VP O
British JJ I-NP I-MISC
lamb NN I-NP O
. . O O

but whats given in sample_dataset is

EU S-ORG
rejects O
German S-MISC
call O
to O
boycott O
British S-MISC
lamb O
.  O

both should work right?

[udion]

@caoxu915683474 were you able to resolve the issue? were you able to get BIOES dataset?