dict output

jira-commentaries-mlm.py

@@ -65,4 +65,4 @@ class JiraComments(datasets.GeneratorBasedBuilder):
             sentences = f.read().split('\n')
 
         for i, sentence in enumerate(sentences):
-            yield i, sentence
+            yield i, {'sentence': sentence}