Update phrase_similarity.py

phrase_similarity.py

@@ -79,7 +79,8 @@ class PhraseSimilarity(datasets.GeneratorBasedBuilder):
                     "phrase2": datasets.Value("string"),
                     "sentence1": datasets.Value("string"),
                     "sentence2": datasets.Value("string"),
-                    "label": datasets.ClassLabel(num_classes=2, names=["negative", "positive"])
+                    "label": datasets.ClassLabel(num_classes=2, names=["negative", "positive"]),
+                    "idx": datasets.Value("int32")
                 }
             ),
             # No default supervised_keys (as we have to pass both question and context as input).
@@ -115,6 +116,7 @@ class PhraseSimilarity(datasets.GeneratorBasedBuilder):
                     "phrase2": example["phrase2"],
                     "sentence1": example["sentence1"],
                     "sentence2": example["sentence2"],
-                    "label": example["label"]
+                    "label": example["label"],
+                    "idx": example["idx"]
                 }
                 key += 1