fixed bug in dataset script

prepare_data.py

@@ -23,8 +23,8 @@ def main():
 				ner_tags_ids = sentence['coarse_tags']
 				ner_tags_fine_ids = sentence['fine_tags']
 				
-				ner_tags = ['-'.join(['B', id2label[t]]) if id2label[t] != 'O' else id2label[t] for t in ner_tags_ids]
-				ner_tags_fine = ['-'.join(['B', id2label_fine[t]]) if id2label_fine[t] != 'O' else id2label[t] for t in ner_tags_fine_ids]
+				ner_tags = ['-'.join(['I', id2label[t]]) if id2label[t] != 'O' else id2label[t] for t in ner_tags_ids]
+				ner_tags_fine = ['-'.join(['I', id2label_fine[t]]) if id2label_fine[t] != 'O' else id2label[t] for t in ner_tags_fine_ids]
 				
 				for token, tag, tag_fine in zip(tokens, ner_tags, ner_tags_fine):
 					f.write(f'{token}\t{tag}\t{tag_fine}\n')

test.txt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7ee0b19161d128827bbd6d7bdc38c40c1b0be03d7a1bc0c82b10fb1748e07b08
+oid sha256:b1b5ec164a8800c9a70de28f38137780989db7f0f331a7d049bb85c0d3235f85
 size 13975793

train.txt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6a1631e7e3a1492daf0f5f1804787d49ebb09b33a8a679c55a919a65d62b90b6
+oid sha256:28a16e4eaa1c2caf717db49467ac91c222825ac231d8e662c053b913600cbe35
 size 49030821

validation.txt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2e5fa8c32b4b8f400890b19edc53c1ab2133f84870af405aaf6ea4a9152b4a1a
+oid sha256:b319ef76a7051b3447680d826c020de1a78f30dfc1285c1c769d815a0b731033
 size 7026135