Add some documentation on the ontonotes/ww conversion

stanfordnlp · Nov 13, 2023 · 27d306f · 27d306f
1 parent b301532
commit 27d306f
Showing 1 changed file with 11 additions and 0 deletions.
diff --git a/stanza/utils/datasets/ner/ontonotes_multitag.py b/stanza/utils/datasets/ner/ontonotes_multitag.py
@@ -1,3 +1,14 @@
+"""
+Combines OntoNotes and WW into a single dataset with OntoNotes used for dev & test
+
+The resulting dataset has two layers saved in the multi_ner column.
+
+WW is kept as 9 classes, with the tag put in either the first or
+second layer depending on the flags.
+
+OntoNotes is converted to one column for 18 and one column for 9 classes.
+"""
+
 import argparse
 import json
 import os