Datasets:
Tasks:
Question Answering
Modalities:
Text
Sub-tasks:
extractive-qa
Languages:
code
Size:
100K - 1M
License:
Refactor: Prefix subtokens
Browse files- codequeries.py +4 -1
codequeries.py
CHANGED
@@ -173,7 +173,10 @@ class Codequeries(datasets.GeneratorBasedBuilder):
|
|
173 |
]
|
174 |
features["example_type"] = datasets.Value("int8")
|
175 |
features["single_hop"] = datasets.Value("bool")
|
176 |
-
|
|
|
|
|
|
|
177 |
features["label_sequence"] = datasets.features.Sequence(datasets.Value("int8"))
|
178 |
if self.config.name == "twostep":
|
179 |
features["relevance_label"] = datasets.Value("int8")
|
|
|
173 |
]
|
174 |
features["example_type"] = datasets.Value("int8")
|
175 |
features["single_hop"] = datasets.Value("bool")
|
176 |
+
if self.config.name != "prefix":
|
177 |
+
features["subtokenized_input_sequence"] = datasets.features.Sequence(datasets.Value("string"))
|
178 |
+
else:
|
179 |
+
features["subtokenized_input_sequence"] = datasets.features.Sequence(datasets.Value("int32"))
|
180 |
features["label_sequence"] = datasets.features.Sequence(datasets.Value("int8"))
|
181 |
if self.config.name == "twostep":
|
182 |
features["relevance_label"] = datasets.Value("int8")
|