Update fs.py
Browse files
fs.py
CHANGED
@@ -62,10 +62,10 @@ _GOV_REPORT_CITATION = r"""
|
|
62 |
|
63 |
|
64 |
class FSConfig(datasets.BuilderConfig):
|
65 |
-
"""BuilderConfig for
|
66 |
|
67 |
def __init__(self, features, data_url, citation, url, **kwargs):
|
68 |
-
"""BuilderConfig for
|
69 |
Args:
|
70 |
features: `list[string]`, list of the features that will appear in the
|
71 |
feature dict. Should not include "label".
|
@@ -84,7 +84,7 @@ class FSConfig(datasets.BuilderConfig):
|
|
84 |
self.url = url
|
85 |
|
86 |
|
87 |
-
class
|
88 |
"""The SCROLLS benchmark."""
|
89 |
|
90 |
features = ["id", "pid", "input", "output"]
|
@@ -94,7 +94,7 @@ class Scrolls(datasets.GeneratorBasedBuilder):
|
|
94 |
name="summ_screen_fd",
|
95 |
description=_SUMM_SCREEN_DESCRIPTION,
|
96 |
features=features,
|
97 |
-
data_url="https://
|
98 |
citation=_SUMM_SCREEN_CITATION,
|
99 |
url="https://github.com/mingdachen/SummScreen",
|
100 |
),
|
@@ -102,7 +102,7 @@ class Scrolls(datasets.GeneratorBasedBuilder):
|
|
102 |
name="gov_report",
|
103 |
description=_GOV_REPORT_CITATION,
|
104 |
features=features,
|
105 |
-
data_url="https://
|
106 |
citation=_GOV_REPORT_DESCRIPTION,
|
107 |
url="https://gov-report-data.github.io/",
|
108 |
),
|
@@ -163,9 +163,9 @@ def _get_task_name_from_data_url(data_url):
|
|
163 |
return data_url.split("/")[-1].split(".")[0]
|
164 |
|
165 |
|
166 |
-
if __name__ == '__main__':
|
167 |
-
builder = Scrolls("scrolls", "summ_screen_fd")
|
168 |
-
builder.download_and_prepare()
|
169 |
-
dataset = builder.as_dataset("validation")
|
170 |
-
x = 5
|
171 |
|
|
|
62 |
|
63 |
|
64 |
class FSConfig(datasets.BuilderConfig):
|
65 |
+
"""BuilderConfig for FS."""
|
66 |
|
67 |
def __init__(self, features, data_url, citation, url, **kwargs):
|
68 |
+
"""BuilderConfig for FS.
|
69 |
Args:
|
70 |
features: `list[string]`, list of the features that will appear in the
|
71 |
feature dict. Should not include "label".
|
|
|
84 |
self.url = url
|
85 |
|
86 |
|
87 |
+
class Fs(datasets.GeneratorBasedBuilder):
|
88 |
"""The SCROLLS benchmark."""
|
89 |
|
90 |
features = ["id", "pid", "input", "output"]
|
|
|
94 |
name="summ_screen_fd",
|
95 |
description=_SUMM_SCREEN_DESCRIPTION,
|
96 |
features=features,
|
97 |
+
data_url="https://huggingface.co/datasets/tau/fs/resolve/main/data/summ_screen_fd.zip",
|
98 |
citation=_SUMM_SCREEN_CITATION,
|
99 |
url="https://github.com/mingdachen/SummScreen",
|
100 |
),
|
|
|
102 |
name="gov_report",
|
103 |
description=_GOV_REPORT_CITATION,
|
104 |
features=features,
|
105 |
+
data_url="https://huggingface.co/datasets/tau/fs/resolve/main/data/gov_report.zip",
|
106 |
citation=_GOV_REPORT_DESCRIPTION,
|
107 |
url="https://gov-report-data.github.io/",
|
108 |
),
|
|
|
163 |
return data_url.split("/")[-1].split(".")[0]
|
164 |
|
165 |
|
166 |
+
#if __name__ == '__main__':
|
167 |
+
# builder = Scrolls("scrolls", "summ_screen_fd")
|
168 |
+
# builder.download_and_prepare()
|
169 |
+
# dataset = builder.as_dataset("validation")
|
170 |
+
# x = 5
|
171 |
|