setting to generate no split, so that it can do after loading the initial ds

Stab-Gurevych-Essays.py

@@ -277,7 +277,7 @@ class NewDataset(datasets.GeneratorBasedBuilder):
 
         # this dataset will return a "train" split only, allowing for
         # 5-fold cross-validation
-        train, test, validate = self.__range_generator(0.7, 0.2)
+        train, test, validate = self.__range_generator(1, 0)
         # essays = self._get_essay_list()
 
         if len(validate) > 0 and len(test) > 0: