Datasets:
Update DiaBLa.py
Browse files
DiaBLa.py
CHANGED
@@ -27,11 +27,10 @@ English-French parallel dataset for the evaluation of \
|
|
27 |
Machine Translation (MT) for informal, written bilingual dialogue.
|
28 |
'''
|
29 |
|
30 |
-
|
31 |
-
|
32 |
-
|
33 |
-
|
34 |
-
#}
|
35 |
|
36 |
|
37 |
class DiablaConfig(datasets.BuilderConfig):
|
@@ -85,8 +84,8 @@ class Diabla(datasets.GeneratorBasedBuilder):
|
|
85 |
],
|
86 |
)
|
87 |
|
88 |
-
|
89 |
-
|
90 |
|
91 |
# return [
|
92 |
# datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": downloaded_files["train"]}),
|
|
|
27 |
Machine Translation (MT) for informal, written bilingual dialogue.
|
28 |
'''
|
29 |
|
30 |
+
_URLS = {
|
31 |
+
'dialogues': 'dialogues.json',
|
32 |
+
'users': 'users.json'
|
33 |
+
}
|
|
|
34 |
|
35 |
|
36 |
class DiablaConfig(datasets.BuilderConfig):
|
|
|
84 |
],
|
85 |
)
|
86 |
|
87 |
+
def _split_generators(self, dl_manager):
|
88 |
+
downloaded_files = dl_manager.download_and_extract(_URLS)
|
89 |
|
90 |
# return [
|
91 |
# datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": downloaded_files["train"]}),
|