Commit
·
f6a8a5a
1
Parent(s):
724c8fd
Update gigaword card and info (#3775)
Browse files* Fix gigaword url
* Update card
* Revert url chang
* Regenerate info
Commit from https://github.com/huggingface/datasets/commit/ff3227aa40059ed101191a36aa7273c63633159d
- README.md +37 -29
- dataset_infos.json +1 -1
README.md
CHANGED
@@ -1,11 +1,27 @@
|
|
1 |
---
|
|
|
|
|
|
|
|
|
2 |
languages:
|
3 |
- en
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
4 |
paperswithcode_id: null
|
5 |
-
pretty_name:
|
6 |
---
|
7 |
|
8 |
-
# Dataset Card for
|
9 |
|
10 |
## Table of Contents
|
11 |
- [Dataset Description](#dataset-description)
|
@@ -33,10 +49,10 @@ pretty_name: gigaword
|
|
33 |
|
34 |
## Dataset Description
|
35 |
|
36 |
-
- **
|
37 |
-
- **
|
38 |
-
- **Paper:** [
|
39 |
-
- **Point of Contact:** [
|
40 |
- **Size of downloaded dataset files:** 551.61 MB
|
41 |
- **Size of the generated dataset:** 918.35 MB
|
42 |
- **Total amount of disk used:** 1469.96 MB
|
@@ -48,35 +64,23 @@ around 4 million articles. Use the 'org_data' provided by
|
|
48 |
https://github.com/microsoft/unilm/ which is identical to
|
49 |
https://github.com/harvardnlp/sent-summary but with better format.
|
50 |
|
51 |
-
There are two features:
|
52 |
-
- document: article.
|
53 |
-
- summary: headline.
|
54 |
-
|
55 |
### Supported Tasks and Leaderboards
|
56 |
|
57 |
-
[
|
58 |
|
59 |
### Languages
|
60 |
|
61 |
-
|
62 |
|
63 |
## Dataset Structure
|
64 |
|
65 |
-
We show detailed information for up to 5 configurations of the dataset.
|
66 |
-
|
67 |
### Data Instances
|
68 |
|
69 |
-
#### default
|
70 |
-
|
71 |
-
- **Size of downloaded dataset files:** 551.61 MB
|
72 |
-
- **Size of the generated dataset:** 918.35 MB
|
73 |
-
- **Total amount of disk used:** 1469.96 MB
|
74 |
-
|
75 |
An example of 'train' looks as follows.
|
76 |
```
|
77 |
{
|
78 |
-
|
79 |
-
|
80 |
}
|
81 |
```
|
82 |
|
@@ -84,7 +88,6 @@ An example of 'train' looks as follows.
|
|
84 |
|
85 |
The data fields are the same among all splits.
|
86 |
|
87 |
-
#### default
|
88 |
- `document`: a `string` feature.
|
89 |
- `summary`: a `string` feature.
|
90 |
|
@@ -104,17 +107,24 @@ The data fields are the same among all splits.
|
|
104 |
|
105 |
#### Initial Data Collection and Normalization
|
106 |
|
107 |
-
|
|
|
|
|
|
|
108 |
|
109 |
#### Who are the source language producers?
|
110 |
|
111 |
-
|
|
|
112 |
|
113 |
### Annotations
|
114 |
|
115 |
#### Annotation process
|
116 |
|
117 |
-
|
|
|
|
|
|
|
118 |
|
119 |
#### Who are the annotators?
|
120 |
|
@@ -150,8 +160,7 @@ The data fields are the same among all splits.
|
|
150 |
|
151 |
### Citation Information
|
152 |
|
153 |
-
```
|
154 |
-
|
155 |
@article{graff2003english,
|
156 |
title={English gigaword},
|
157 |
author={Graff, David and Kong, Junbo and Chen, Ke and Maeda, Kazuaki},
|
@@ -171,7 +180,6 @@ The data fields are the same among all splits.
|
|
171 |
author={Rush, Alexander M. and Chopra, Sumit and Weston, Jason},
|
172 |
year={2015}
|
173 |
}
|
174 |
-
|
175 |
```
|
176 |
|
177 |
|
|
|
1 |
---
|
2 |
+
annotations_creators:
|
3 |
+
- found
|
4 |
+
language_creators:
|
5 |
+
- found
|
6 |
languages:
|
7 |
- en
|
8 |
+
licenses:
|
9 |
+
- mit
|
10 |
+
multilinguality:
|
11 |
+
- monolingual
|
12 |
+
size_categories:
|
13 |
+
- 100K<n<1M
|
14 |
+
source_datasets:
|
15 |
+
- extended|gigaword_2003
|
16 |
+
task_categories:
|
17 |
+
- conditional-text-generation
|
18 |
+
task_ids:
|
19 |
+
- summarization
|
20 |
paperswithcode_id: null
|
21 |
+
pretty_name: Gigaword
|
22 |
---
|
23 |
|
24 |
+
# Dataset Card for Gigaword
|
25 |
|
26 |
## Table of Contents
|
27 |
- [Dataset Description](#dataset-description)
|
|
|
49 |
|
50 |
## Dataset Description
|
51 |
|
52 |
+
- **Repository:** [Gigaword repository](https://github.com/harvardnlp/sent-summary)
|
53 |
+
- **Leaderboard:** [Gigaword leaderboard](https://paperswithcode.com/sota/text-summarization-on-gigaword)
|
54 |
+
- **Paper:** [A Neural Attention Model for Abstractive Sentence Summarization](https://arxiv.org/abs/1509.00685)
|
55 |
+
- **Point of Contact:** [Alexander Rush](mailto:arush@cornell.edu)
|
56 |
- **Size of downloaded dataset files:** 551.61 MB
|
57 |
- **Size of the generated dataset:** 918.35 MB
|
58 |
- **Total amount of disk used:** 1469.96 MB
|
|
|
64 |
https://github.com/microsoft/unilm/ which is identical to
|
65 |
https://github.com/harvardnlp/sent-summary but with better format.
|
66 |
|
|
|
|
|
|
|
|
|
67 |
### Supported Tasks and Leaderboards
|
68 |
|
69 |
+
- `summarization`: This dataset can be used for Summarization, where given a dicument, the goal is to predict its summery. The model performance is evaluated using the [ROUGE](https://huggingface.co/metrics/rouge) metric. The leaderboard for this task is available [here](https://paperswithcode.com/sota/text-summarization-on-gigaword).
|
70 |
|
71 |
### Languages
|
72 |
|
73 |
+
English.
|
74 |
|
75 |
## Dataset Structure
|
76 |
|
|
|
|
|
77 |
### Data Instances
|
78 |
|
|
|
|
|
|
|
|
|
|
|
|
|
79 |
An example of 'train' looks as follows.
|
80 |
```
|
81 |
{
|
82 |
+
'document': "australia 's current account deficit shrunk by a record #.## billion dollars -lrb- #.## billion us -rrb- in the june quarter due to soaring commodity prices , figures released monday showed .",
|
83 |
+
'summary': 'australian current account deficit narrows sharply'
|
84 |
}
|
85 |
```
|
86 |
|
|
|
88 |
|
89 |
The data fields are the same among all splits.
|
90 |
|
|
|
91 |
- `document`: a `string` feature.
|
92 |
- `summary`: a `string` feature.
|
93 |
|
|
|
107 |
|
108 |
#### Initial Data Collection and Normalization
|
109 |
|
110 |
+
From the paper:
|
111 |
+
> For our training set, we pair the headline of each article with its first sentence to create an inputsummary pair. While the model could in theory be trained on any pair, Gigaword contains many spurious headline-article pairs. We therefore prune training based on the following heuristic filters: (1) Are there no non-stop-words in common? (2) Does the title contain a byline or other extraneous editing marks? (3) Does the title have a question mark or colon? After applying these filters, the training set consists of roughly J = 4 million title-article pairs. We apply a minimal preprocessing step using PTB tokenization, lower-casing, replacing all digit characters with #, and replacing of word types seen less than 5 times with UNK. We also remove all articles from the time-period of the DUC evaluation. release.
|
112 |
+
The complete input training vocabulary consists of 119 million word tokens and 110K unique word types with an average sentence size of 31.3 words. The headline vocabulary consists of 31 million tokens and 69K word types with the average title of length 8.3 words (note that this is significantly shorter than the DUC summaries). On average there are 4.6 overlapping word types between the headline and the input; although only 2.6 in the
|
113 |
+
first 75-characters of the input.
|
114 |
|
115 |
#### Who are the source language producers?
|
116 |
|
117 |
+
From the paper:
|
118 |
+
> For training data for both tasks, we utilize the annotated Gigaword data set (Graff et al., 2003; Napoles et al., 2012), which consists of standard Gigaword, preprocessed with Stanford CoreNLP tools (Manning et al., 2014).
|
119 |
|
120 |
### Annotations
|
121 |
|
122 |
#### Annotation process
|
123 |
|
124 |
+
Annotations are inherited from the annotatated Gigaword data set.
|
125 |
+
|
126 |
+
Additional information from the paper:
|
127 |
+
> Our model only uses annotations for tokenization and sentence separation, although several of the baselines use parsing and tagging as well.
|
128 |
|
129 |
#### Who are the annotators?
|
130 |
|
|
|
160 |
|
161 |
### Citation Information
|
162 |
|
163 |
+
```bibtex
|
|
|
164 |
@article{graff2003english,
|
165 |
title={English gigaword},
|
166 |
author={Graff, David and Kong, Junbo and Chen, Ke and Maeda, Kazuaki},
|
|
|
180 |
author={Rush, Alexander M. and Chopra, Sumit and Weston, Jason},
|
181 |
year={2015}
|
182 |
}
|
|
|
183 |
```
|
184 |
|
185 |
|
dataset_infos.json
CHANGED
@@ -1 +1 @@
|
|
1 |
-
{"default": {"description": "\nHeadline-generation on a corpus of article pairs from Gigaword consisting of\naround 4 million articles. Use the 'org_data' provided by\nhttps://github.com/microsoft/unilm/ which is identical to\nhttps://github.com/harvardnlp/sent-summary but with better format.\n\nThere are two features:\n - document: article.\n - summary: headline.\n\n", "citation": "\n@article{graff2003english,\n title={English gigaword},\n author={Graff, David and Kong, Junbo and Chen, Ke and Maeda, Kazuaki},\n journal={Linguistic Data Consortium, Philadelphia},\n volume={4},\n number={1},\n pages={34},\n year={2003}\n}\n\n@article{Rush_2015,\n title={A Neural Attention Model for Abstractive Sentence Summarization},\n url={http://dx.doi.org/10.18653/v1/D15-1044},\n DOI={10.18653/v1/d15-1044},\n journal={Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing},\n publisher={Association for Computational Linguistics},\n author={Rush, Alexander M. and Chopra, Sumit and Weston, Jason},\n year={2015}\n}\n", "homepage": "https://github.com/harvardnlp/sent-summary", "license": "", "features": {"document": {"dtype": "string", "id": null, "_type": "Value"}, "summary": {"dtype": "string", "id": null, "_type": "Value"}}, "supervised_keys": {"input": "document", "output": "summary"}, "builder_name": "gigaword", "config_name": "default", "version": {"version_str": "1.2.0", "description": null, "
|
|
|
1 |
+
{"default": {"description": "\nHeadline-generation on a corpus of article pairs from Gigaword consisting of\naround 4 million articles. Use the 'org_data' provided by\nhttps://github.com/microsoft/unilm/ which is identical to\nhttps://github.com/harvardnlp/sent-summary but with better format.\n\nThere are two features:\n - document: article.\n - summary: headline.\n\n", "citation": "\n@article{graff2003english,\n title={English gigaword},\n author={Graff, David and Kong, Junbo and Chen, Ke and Maeda, Kazuaki},\n journal={Linguistic Data Consortium, Philadelphia},\n volume={4},\n number={1},\n pages={34},\n year={2003}\n}\n\n@article{Rush_2015,\n title={A Neural Attention Model for Abstractive Sentence Summarization},\n url={http://dx.doi.org/10.18653/v1/D15-1044},\n DOI={10.18653/v1/d15-1044},\n journal={Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing},\n publisher={Association for Computational Linguistics},\n author={Rush, Alexander M. and Chopra, Sumit and Weston, Jason},\n year={2015}\n}\n", "homepage": "https://github.com/harvardnlp/sent-summary", "license": "", "features": {"document": {"dtype": "string", "id": null, "_type": "Value"}, "summary": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": {"input": "document", "output": "summary"}, "task_templates": null, "builder_name": "gigaword", "config_name": "default", "version": {"version_str": "1.2.0", "description": null, "major": 1, "minor": 2, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 915249388, "num_examples": 3803957, "dataset_name": "gigaword"}, "validation": {"name": "validation", "num_bytes": 45767096, "num_examples": 189651, "dataset_name": "gigaword"}, "test": {"name": "test", "num_bytes": 450782, "num_examples": 1951, "dataset_name": "gigaword"}}, "download_checksums": {"https://drive.google.com/uc?export=download&id=1USoQ8lJgN8kAWnUnRrupMGrPMLlDVqlV": {"num_bytes": 578402958, "checksum": "bc0c4a2e1aa19cf2123688b87bc2d778c0d8fc24a4090e3c10a27c5faa1b898b"}}, "download_size": 578402958, "post_processing_size": null, "dataset_size": 961467266, "size_in_bytes": 1539870224}}
|