Datasets:
Size:
10K - 100K
License:
update config
Browse files
jsick.py
CHANGED
@@ -50,6 +50,11 @@ class JSICKDataset(ds.GeneratorBasedBuilder):
|
|
50 |
version=VERSION,
|
51 |
description="fuga",
|
52 |
),
|
|
|
|
|
|
|
|
|
|
|
53 |
]
|
54 |
|
55 |
def _info(self) -> ds.DatasetInfo:
|
@@ -96,6 +101,22 @@ class JSICKDataset(ds.GeneratorBasedBuilder):
|
|
96 |
)
|
97 |
|
98 |
elif self.config.name == "stress":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
99 |
features = ds.Features(
|
100 |
{
|
101 |
"pair_ID": ds.Value("string"),
|
@@ -123,7 +144,7 @@ class JSICKDataset(ds.GeneratorBasedBuilder):
|
|
123 |
data_path = dl_manager.download_and_extract(_URLS[self.config.name])
|
124 |
df: pd.DataFrame = pd.read_table(data_path, sep="\t", header=0)
|
125 |
|
126 |
-
if self.config.name
|
127 |
df = df.rename(
|
128 |
columns={
|
129 |
"pair_ID": "id",
|
|
|
50 |
version=VERSION,
|
51 |
description="fuga",
|
52 |
),
|
53 |
+
ds.BuilderConfig(
|
54 |
+
name="stress_original",
|
55 |
+
version=VERSION,
|
56 |
+
description="fuga",
|
57 |
+
),
|
58 |
]
|
59 |
|
60 |
def _info(self) -> ds.DatasetInfo:
|
|
|
101 |
)
|
102 |
|
103 |
elif self.config.name == "stress":
|
104 |
+
features = ds.Features(
|
105 |
+
{
|
106 |
+
"id": ds.Value("string"),
|
107 |
+
"premise": ds.Value("string"),
|
108 |
+
"hypothesis": ds.Value("string"),
|
109 |
+
"label": labels,
|
110 |
+
"score": ds.Value("float32"),
|
111 |
+
"sentence_A_Ja_origin": ds.Value("string"),
|
112 |
+
"entailment_label_origin": labels,
|
113 |
+
"relatedness_score_Ja_origin": ds.Value("float32"),
|
114 |
+
"rephrase_type": ds.Value("string"),
|
115 |
+
"case_particles": ds.Value("string"),
|
116 |
+
}
|
117 |
+
)
|
118 |
+
|
119 |
+
elif self.config.name == "stress_original":
|
120 |
features = ds.Features(
|
121 |
{
|
122 |
"pair_ID": ds.Value("string"),
|
|
|
144 |
data_path = dl_manager.download_and_extract(_URLS[self.config.name])
|
145 |
df: pd.DataFrame = pd.read_table(data_path, sep="\t", header=0)
|
146 |
|
147 |
+
if self.config.name in ["base", "stress"]:
|
148 |
df = df.rename(
|
149 |
columns={
|
150 |
"pair_ID": "id",
|