Datasets:
Tasks:
Summarization
Modalities:
Text
Sub-tasks:
news-articles-summarization
Languages:
English
Size:
100K - 1M
ArXiv:
License:
Update files from the datasets library (from 1.16.0)
Browse filesRelease notes: https://github.com/huggingface/datasets/releases/tag/1.16.0
README.md
CHANGED
@@ -1,7 +1,12 @@
|
|
1 |
---
|
|
|
2 |
languages:
|
3 |
- en
|
4 |
paperswithcode_id: xsum
|
|
|
|
|
|
|
|
|
5 |
---
|
6 |
|
7 |
# Dataset Card for "xsum"
|
|
|
1 |
---
|
2 |
+
pretty_name: Extreme Summarization (XSum)
|
3 |
languages:
|
4 |
- en
|
5 |
paperswithcode_id: xsum
|
6 |
+
task_categories:
|
7 |
+
- conditional-text-generation
|
8 |
+
task_ids:
|
9 |
+
- summarization
|
10 |
---
|
11 |
|
12 |
# Dataset Card for "xsum"
|
xsum.py
CHANGED
@@ -95,7 +95,7 @@ class Xsum(datasets.GeneratorBasedBuilder):
|
|
95 |
"""Returns SplitGenerators."""
|
96 |
|
97 |
files_to_download = {"data": _URL_DATA, "splits": _URL_SPLITS}
|
98 |
-
downloaded_files = dl_manager.
|
99 |
|
100 |
return [
|
101 |
datasets.SplitGenerator(
|
@@ -103,7 +103,8 @@ class Xsum(datasets.GeneratorBasedBuilder):
|
|
103 |
gen_kwargs={
|
104 |
"split_path": downloaded_files["splits"],
|
105 |
"split_name": "train",
|
106 |
-
"data_dir":
|
|
|
107 |
},
|
108 |
),
|
109 |
datasets.SplitGenerator(
|
@@ -111,7 +112,8 @@ class Xsum(datasets.GeneratorBasedBuilder):
|
|
111 |
gen_kwargs={
|
112 |
"split_path": downloaded_files["splits"],
|
113 |
"split_name": "validation",
|
114 |
-
"data_dir":
|
|
|
115 |
},
|
116 |
),
|
117 |
datasets.SplitGenerator(
|
@@ -119,37 +121,50 @@ class Xsum(datasets.GeneratorBasedBuilder):
|
|
119 |
gen_kwargs={
|
120 |
"split_path": downloaded_files["splits"],
|
121 |
"split_name": "test",
|
122 |
-
"data_dir":
|
|
|
123 |
},
|
124 |
),
|
125 |
]
|
126 |
|
127 |
-
def _generate_examples(self, split_path, split_name, data_dir):
|
128 |
"""Yields examples."""
|
129 |
|
130 |
with open(split_path, "r", encoding="utf-8") as f:
|
131 |
split_ids = json.load(f)
|
132 |
-
|
133 |
-
|
134 |
-
|
135 |
-
|
136 |
-
|
137 |
-
|
138 |
-
|
139 |
-
|
140 |
-
|
141 |
-
|
142 |
-
|
143 |
-
|
144 |
-
|
145 |
-
|
146 |
-
|
147 |
-
|
148 |
-
|
149 |
-
|
150 |
-
|
151 |
-
|
152 |
-
|
153 |
-
|
154 |
-
|
155 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
95 |
"""Returns SplitGenerators."""
|
96 |
|
97 |
files_to_download = {"data": _URL_DATA, "splits": _URL_SPLITS}
|
98 |
+
downloaded_files = dl_manager.download(files_to_download)
|
99 |
|
100 |
return [
|
101 |
datasets.SplitGenerator(
|
|
|
103 |
gen_kwargs={
|
104 |
"split_path": downloaded_files["splits"],
|
105 |
"split_name": "train",
|
106 |
+
"data_dir": "bbc-summary-data",
|
107 |
+
"files": dl_manager.iter_archive(downloaded_files["data"]),
|
108 |
},
|
109 |
),
|
110 |
datasets.SplitGenerator(
|
|
|
112 |
gen_kwargs={
|
113 |
"split_path": downloaded_files["splits"],
|
114 |
"split_name": "validation",
|
115 |
+
"data_dir": "bbc-summary-data",
|
116 |
+
"files": dl_manager.iter_archive(downloaded_files["data"]),
|
117 |
},
|
118 |
),
|
119 |
datasets.SplitGenerator(
|
|
|
121 |
gen_kwargs={
|
122 |
"split_path": downloaded_files["splits"],
|
123 |
"split_name": "test",
|
124 |
+
"data_dir": "bbc-summary-data",
|
125 |
+
"files": dl_manager.iter_archive(downloaded_files["data"]),
|
126 |
},
|
127 |
),
|
128 |
]
|
129 |
|
130 |
+
def _generate_examples(self, split_path, split_name, data_dir, files):
|
131 |
"""Yields examples."""
|
132 |
|
133 |
with open(split_path, "r", encoding="utf-8") as f:
|
134 |
split_ids = json.load(f)
|
135 |
+
split_ids = {k: set(v) for k, v in split_ids.items()}
|
136 |
+
|
137 |
+
for path, f in files:
|
138 |
+
if not split_ids[split_name]:
|
139 |
+
break
|
140 |
+
elif path.startswith(data_dir) and path.endswith(".summary"):
|
141 |
+
i = os.path.basename(path).split(".")[0]
|
142 |
+
if i in split_ids[split_name]:
|
143 |
+
split_ids[split_name].remove(i)
|
144 |
+
text = "".join(
|
145 |
+
[
|
146 |
+
line.decode("utf-8")
|
147 |
+
for line in f.readlines()
|
148 |
+
if line.decode("utf-8") not in _REMOVE_LINES and line.strip()
|
149 |
+
]
|
150 |
+
)
|
151 |
+
# Each file follows below format:
|
152 |
+
# [SN]URL[SN]
|
153 |
+
# http://somelink
|
154 |
+
#
|
155 |
+
# [SN]TITLE[SN]
|
156 |
+
# some intro
|
157 |
+
#
|
158 |
+
# [SN]FIRST-SENTENCE[SN]
|
159 |
+
# some intro
|
160 |
+
#
|
161 |
+
# [SN]RESTBODY[SN]
|
162 |
+
# text line.
|
163 |
+
# another text line.
|
164 |
+
# "another text line."
|
165 |
+
|
166 |
+
# According to the following issue, FIRST-SENTENCE
|
167 |
+
# is the reference summary and TITLE is unused:
|
168 |
+
# https://github.com/EdinburghNLP/XSum/issues/22
|
169 |
+
segs = text.split("[SN]")
|
170 |
+
yield i, {_DOCUMENT: segs[8].strip(), _SUMMARY: segs[6].strip(), _ID: i}
|