parquet-converter commited on
Commit
81a3843
1 Parent(s): 75ec0e2

Update parquet files

Browse files
README.md DELETED
@@ -1,254 +0,0 @@
1
- ---
2
- annotations_creators:
3
- - no-annotation
4
- language_creators:
5
- - crowdsourced
6
- language:
7
- - en
8
- license:
9
- - cc-by-4.0
10
- multilinguality:
11
- - monolingual
12
- paperswithcode_id: null
13
- pretty_name: Reddit Webis-TLDR-17
14
- size_categories:
15
- - 1M<n<10M
16
- source_datasets:
17
- - original
18
- task_categories:
19
- - summarization
20
- task_ids: []
21
- train-eval-index:
22
- - config: default
23
- task: summarization
24
- task_id: summarization
25
- splits:
26
- train_split: train
27
- col_mapping:
28
- content: text
29
- summary: target
30
- metrics:
31
- - type: rouge
32
- name: Rouge
33
- tags:
34
- - reddit-posts-summarization
35
- dataset_info:
36
- features:
37
- - name: author
38
- dtype: string
39
- - name: body
40
- dtype: string
41
- - name: normalizedBody
42
- dtype: string
43
- - name: subreddit
44
- dtype: string
45
- - name: subreddit_id
46
- dtype: string
47
- - name: id
48
- dtype: string
49
- - name: content
50
- dtype: string
51
- - name: summary
52
- dtype: string
53
- splits:
54
- - name: train
55
- num_bytes: 18940542951
56
- num_examples: 3848330
57
- download_size: 3141854161
58
- dataset_size: 18940542951
59
- ---
60
-
61
- # Dataset Card for Reddit Webis-TLDR-17
62
-
63
- ## Table of Contents
64
- - [Dataset Description](#dataset-description)
65
- - [Dataset Summary](#dataset-summary)
66
- - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
67
- - [Languages](#languages)
68
- - [Dataset Structure](#dataset-structure)
69
- - [Data Instances](#data-instances)
70
- - [Data Fields](#data-fields)
71
- - [Data Splits](#data-splits)
72
- - [Dataset Creation](#dataset-creation)
73
- - [Curation Rationale](#curation-rationale)
74
- - [Source Data](#source-data)
75
- - [Annotations](#annotations)
76
- - [Personal and Sensitive Information](#personal-and-sensitive-information)
77
- - [Considerations for Using the Data](#considerations-for-using-the-data)
78
- - [Social Impact of Dataset](#social-impact-of-dataset)
79
- - [Discussion of Biases](#discussion-of-biases)
80
- - [Other Known Limitations](#other-known-limitations)
81
- - [Additional Information](#additional-information)
82
- - [Dataset Curators](#dataset-curators)
83
- - [Licensing Information](#licensing-information)
84
- - [Citation Information](#citation-information)
85
- - [Contributions](#contributions)
86
-
87
- ## Dataset Description
88
-
89
- - **Homepage:** [https://webis.de/data/webis-tldr-17.html](https://webis.de/data/webis-tldr-17.html)
90
- - **Repository:** [https://github.com/webis-de/webis-tldr-17-corpus](https://github.com/webis-de/webis-tldr-17-corpus)
91
- - **Paper:** [https://aclanthology.org/W17-4508]
92
- - **Point of Contact:** [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
93
- - **Size of downloaded dataset files:** 2996.31 MB
94
- - **Size of the generated dataset:** 18063.11 MB
95
- - **Total amount of disk used:** 21059.41 MB
96
-
97
- ### Dataset Summary
98
-
99
- This corpus contains preprocessed posts from the Reddit dataset (Webis-TLDR-17).
100
- The dataset consists of 3,848,330 posts with an average length of 270 words for content,
101
- and 28 words for the summary.
102
-
103
- Features includes strings: author, body, normalizedBody, content, summary, subreddit, subreddit_id.
104
- Content is used as document and summary is used as summary.
105
-
106
- ### Supported Tasks and Leaderboards
107
-
108
- Summarization (abstractive)
109
-
110
- Known ROUGE scores achieved for the Webis-TLDR-17:
111
-
112
- | Model | ROUGE-1 | ROUGE-2 | ROUGE-L | Paper/Source |
113
- |-------|-------|-------|-------|------:|
114
- | Transformer + Copy (Gehrmann et al., 2019) | 22 | 6 | 17 | Generating Summaries with Finetuned Language Models |
115
- | Unified VAE + PGN (Choi et al., 2019) | 19 | 4 | 15 | VAE-PGN based Abstractive Model in Multi-stage Architecture for Text Summarization |
116
-
117
- (Source: https://github.com/sebastianruder/NLP-progress/blob/master/english/summarization.md)
118
-
119
- ### Languages
120
-
121
- English
122
-
123
- ## Dataset Structure
124
-
125
- ### Data Instances
126
-
127
- #### default
128
-
129
- - **Size of downloaded dataset files:** 2996.31 MB
130
- - **Size of the generated dataset:** 18063.11 MB
131
- - **Total amount of disk used:** 21059.41 MB
132
-
133
- An example of 'train' looks as follows.
134
- ```
135
- {
136
- "author": "me",
137
- "body": "<>",
138
- "content": "input document.",
139
- "id": "1",
140
- "normalizedBody": "",
141
- "subreddit": "machinelearning",
142
- "subreddit_id": "2",
143
- "summary": "output summary."
144
- }
145
- ```
146
-
147
- ### Data Fields
148
-
149
- The data fields are the same among all splits.
150
-
151
- #### default
152
- - `author`: a `string` feature.
153
- - `body`: a `string` feature.
154
- - `normalizedBody`: a `string` feature.
155
- - `subreddit`: a `string` feature.
156
- - `subreddit_id`: a `string` feature.
157
- - `id`: a `string` feature.
158
- - `content`: a `string` feature.
159
- - `summary`: a `string` feature.
160
-
161
- ### Data Splits
162
-
163
- | name | train |
164
- |-------|------:|
165
- |default|3848330|
166
-
167
- This corpus does not contain a separate test set. Thus it is up to the users to divide the corpus into appropriate training, validation and test sets.
168
-
169
- ## Dataset Creation
170
-
171
- ### Curation Rationale
172
-
173
- In the scope of the task of absractive summarization the creators of the Webis-TLDR-17 propose mining social media for author-provided summaries and taking advantage of the common practice of appending a "TL;DR" to long posts. A large Reddit crawl was used to yield the Webis-TLDR-17 corpus. This dataset intends to complement the existing summarization corpora primarily from the news genre.
174
-
175
- ### Source Data
176
-
177
- Reddit subreddits posts (submissions & comments) containing "TL;DR" from 2006 to 2016. Multiple subreddits are included.
178
-
179
- #### Initial Data Collection and Normalization
180
-
181
- Initial data: a set of 286 million submissions and 1.6 billion comments posted to Reddit between 2006 and 2016.
182
- Then a five-step pipeline of consecutive filtering steps was applied.
183
-
184
- #### Who are the source language producers?
185
-
186
- The contents of the dataset are produced by human authors, bot-generated content was eliminated by filtering out all bot accounts with the help of an extensive list provided by the Reddit community, as well as manual inspection of cases where the user name contained the substring "bot."
187
-
188
- ### Annotations
189
-
190
- #### Annotation process
191
-
192
- [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
193
-
194
- #### Who are the annotators?
195
-
196
- [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
197
-
198
- ### Personal and Sensitive Information
199
-
200
- [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
201
-
202
- ## Considerations for Using the Data
203
-
204
- ### Social Impact of Dataset
205
-
206
- This dataset has been created to serve as a source of large-scale summarization training data. It is primarily geared towards the automatic abstractive summarization task, that can be considered one of the most challenging variants of automatic summarization. It also aims to tackle the lack of genre diversity in the summarization datasets (most are news-related).
207
-
208
- ### Discussion of Biases
209
-
210
- [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
211
-
212
- ### Other Known Limitations
213
-
214
- Reddit users write TL;DRs with various intentions, such as providing a “true” summary, asking questions or for help, or forming judgments and conclusions. As noted in the paper introducing the dataset, although the first kind of TL;DR posts are most important for training summarization models, yet, the latter allow for various alternative summarization-related tasks.
215
-
216
- Although filtering was performed abusive language maybe still be present.
217
-
218
- ## Additional Information
219
-
220
- ### Dataset Curators
221
-
222
- Michael Völske, Martin Potthast, Shahbaz Syed, Benno Stein
223
-
224
- ### Licensing Information
225
-
226
- [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
227
-
228
- ### Citation Information
229
-
230
- ```
231
-
232
- @inproceedings{volske-etal-2017-tl,
233
- title = "{TL};{DR}: Mining {R}eddit to Learn Automatic Summarization",
234
- author = {V{"o}lske, Michael and
235
- Potthast, Martin and
236
- Syed, Shahbaz and
237
- Stein, Benno},
238
- booktitle = "Proceedings of the Workshop on New Frontiers in Summarization",
239
- month = sep,
240
- year = "2017",
241
- address = "Copenhagen, Denmark",
242
- publisher = "Association for Computational Linguistics",
243
- url = "https://www.aclweb.org/anthology/W17-4508",
244
- doi = "10.18653/v1/W17-4508",
245
- pages = "59--63",
246
- abstract = "Recent advances in automatic text summarization have used deep neural networks to generate high-quality abstractive summaries, but the performance of these models strongly depends on large amounts of suitable training data. We propose a new method for mining social media for author-provided summaries, taking advantage of the common practice of appending a {``}TL;DR{''} to long posts. A case study using a large Reddit crawl yields the Webis-TLDR-17 dataset, complementing existing corpora primarily from the news genre. Our technique is likely applicable to other social media sites and general web crawls.",
247
- }
248
-
249
- ```
250
-
251
-
252
- ### Contributions
253
-
254
- Thanks to [@mariamabarham](https://github.com/mariamabarham), [@patrickvonplaten](https://github.com/patrickvonplaten), [@thomwolf](https://github.com/thomwolf) for adding this dataset.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
dataset_infos.json DELETED
@@ -1 +0,0 @@
1
- {"default": {"description": "\nThis corpus contains preprocessed posts from the Reddit dataset.\nThe dataset consists of 3,848,330 posts with an average length of 270 words for content,\nand 28 words for the summary.\n\nFeatures includes strings: author, body, normalizedBody, content, summary, subreddit, subreddit_id.\nContent is used as document and summary is used as summary.\n", "citation": "\n@inproceedings{volske-etal-2017-tl,\n title = \"{TL};{DR}: Mining {R}eddit to Learn Automatic Summarization\",\n author = {V{\"o}lske, Michael and\n Potthast, Martin and\n Syed, Shahbaz and\n Stein, Benno},\n booktitle = \"Proceedings of the Workshop on New Frontiers in Summarization\",\n month = sep,\n year = \"2017\",\n address = \"Copenhagen, Denmark\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://www.aclweb.org/anthology/W17-4508\",\n doi = \"10.18653/v1/W17-4508\",\n pages = \"59--63\",\n abstract = \"Recent advances in automatic text summarization have used deep neural networks to generate high-quality abstractive summaries, but the performance of these models strongly depends on large amounts of suitable training data. We propose a new method for mining social media for author-provided summaries, taking advantage of the common practice of appending a {``}TL;DR{''} to long posts. A case study using a large Reddit crawl yields the Webis-TLDR-17 dataset, complementing existing corpora primarily from the news genre. Our technique is likely applicable to other social media sites and general web crawls.\",\n}\n", "homepage": "https://github.com/webis-de/webis-tldr-17-corpus", "license": "", "features": {"author": {"dtype": "string", "id": null, "_type": "Value"}, "body": {"dtype": "string", "id": null, "_type": "Value"}, "normalizedBody": {"dtype": "string", "id": null, "_type": "Value"}, "subreddit": {"dtype": "string", "id": null, "_type": "Value"}, "subreddit_id": {"dtype": "string", "id": null, "_type": "Value"}, "id": {"dtype": "string", "id": null, "_type": "Value"}, "content": {"dtype": "string", "id": null, "_type": "Value"}, "summary": {"dtype": "string", "id": null, "_type": "Value"}}, "supervised_keys": null, "builder_name": "reddit", "config_name": "default", "version": {"version_str": "1.0.0", "description": null, "datasets_version_to_prepare": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 18940542951, "num_examples": 3848330, "dataset_name": "reddit"}}, "download_checksums": {"https://zenodo.org/record/1043504/files/corpus-webis-tldr-17.zip?download=1": {"num_bytes": 3141854161, "checksum": "c1a0f8c4374c7314d3c9ec50dd505303c536062d87037d4dca7035b89b36938a"}}, "download_size": 3141854161, "dataset_size": 18940542951, "size_in_bytes": 22082397112}}
 
 
default/partial/train/0000.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c4ce550dc47ca17268a19cc58aa1c7f76cacbedcba01d4e59d12f511709a19a9
3
+ size 320955109
default/partial/train/0001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:add95b75531a3bb14df2fdb7ae2a9f8193f3bc4afc01d689d59170d715662569
3
+ size 322267000
default/partial/train/0002.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8d7f566df9366bf62cbc446e5f31585c9f6ebb7606640cfe43e81c9bba908e98
3
+ size 320825346
default/partial/train/0003.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:18e7678e599ff9cc700d8c21167d23a353fe92e24a93f752d64ede2912aef9b5
3
+ size 320580976
default/partial/train/0004.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:768664c06d277c3105b41ffaeae08eadd4f906c1da9c7848bf1bcf1ab8834e62
3
+ size 320442406
default/partial/train/0005.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0524a8e578be1ce48dcd28d72aaae4477bd0a64a45669ebd7e0f144dad56cfe5
3
+ size 321098689
default/partial/train/0006.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:343e34cde36e59ce0aa3ad55bf98fac5d59dc39fdcfae375312317a939013373
3
+ size 321983919
default/partial/train/0007.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4d626c907f5c0abd37b356f6be3226bbffb90fca51b1e1ac24d1bca581cdef06
3
+ size 321267424
default/partial/train/0008.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a9d256a0ad567862c1ace3fd8fc9d9bb21d0dd600bf740fb1e406545dfc0b6c1
3
+ size 320484624
default/partial/train/0009.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:495db4cba3f531a03a380e2715143bcb791ebb6d98550c2aabefd0b4c1714c96
3
+ size 310980702
reddit.py DELETED
@@ -1,101 +0,0 @@
1
- # coding=utf-8
2
- # Copyright 2020 The TensorFlow Datasets Authors and the HuggingFace Datasets Authors.
3
- #
4
- # Licensed under the Apache License, Version 2.0 (the "License");
5
- # you may not use this file except in compliance with the License.
6
- # You may obtain a copy of the License at
7
- #
8
- # http://www.apache.org/licenses/LICENSE-2.0
9
- #
10
- # Unless required by applicable law or agreed to in writing, software
11
- # distributed under the License is distributed on an "AS IS" BASIS,
12
- # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13
- # See the License for the specific language governing permissions and
14
- # limitations under the License.
15
-
16
- # Lint as: python3
17
- """Reddit dataset using tldr as summaries."""
18
-
19
- import json
20
- import os
21
-
22
- import datasets
23
-
24
-
25
- _CITATION = """
26
- @inproceedings{volske-etal-2017-tl,
27
- title = {TL;DR: Mining {R}eddit to Learn Automatic Summarization},
28
- author = {V{\"o}lske, Michael and Potthast, Martin and Syed, Shahbaz and Stein, Benno},
29
- booktitle = {Proceedings of the Workshop on New Frontiers in Summarization},
30
- month = {sep},
31
- year = {2017},
32
- address = {Copenhagen, Denmark},
33
- publisher = {Association for Computational Linguistics},
34
- url = {https://www.aclweb.org/anthology/W17-4508},
35
- doi = {10.18653/v1/W17-4508},
36
- pages = {59--63},
37
- abstract = {Recent advances in automatic text summarization have used deep neural networks to generate high-quality abstractive summaries, but the performance of these models strongly depends on large amounts of suitable training data. We propose a new method for mining social media for author-provided summaries, taking advantage of the common practice of appending a {``}TL;DR{''} to long posts. A case study using a large Reddit crawl yields the Webis-TLDR-17 dataset, complementing existing corpora primarily from the news genre. Our technique is likely applicable to other social media sites and general web crawls.},
38
- }
39
- """
40
-
41
- _DESCRIPTION = """
42
- This corpus contains preprocessed posts from the Reddit dataset.
43
- The dataset consists of 3,848,330 posts with an average length of 270 words for content,
44
- and 28 words for the summary.
45
-
46
- Features includes strings: author, body, normalizedBody, content, summary, subreddit, subreddit_id.
47
- Content is used as document and summary is used as summary.
48
- """
49
-
50
- _URL = "https://zenodo.org/record/1043504/files/corpus-webis-tldr-17.zip?download=1"
51
-
52
- _DOCUMENT = "content"
53
- _SUMMARY = "summary"
54
- _ADDITIONAL_FEATURES = ["author", "body", "normalizedBody", "subreddit", "subreddit_id", "id"]
55
-
56
-
57
- class Reddit(datasets.GeneratorBasedBuilder):
58
- """Reddit Dataset."""
59
-
60
- VERSION = datasets.Version("1.0.0")
61
-
62
- def _info(self):
63
- return datasets.DatasetInfo(
64
- description=_DESCRIPTION,
65
- features=datasets.Features(
66
- {k: datasets.Value("string") for k in _ADDITIONAL_FEATURES + [_DOCUMENT, _SUMMARY]}
67
- ),
68
- supervised_keys=None,
69
- homepage="https://github.com/webis-de/webis-tldr-17-corpus",
70
- citation=_CITATION,
71
- )
72
-
73
- def _split_generators(self, dl_manager):
74
- """Returns SplitGenerators."""
75
- dl_path = dl_manager.download_and_extract(_URL)
76
- return [
77
- datasets.SplitGenerator(
78
- name=datasets.Split.TRAIN,
79
- gen_kwargs={"path": os.path.join(dl_path, "corpus-webis-tldr-17.json")},
80
- )
81
- ]
82
-
83
- def _generate_examples(self, path=None):
84
- """Yields examples."""
85
- with open(path, "rb") as f:
86
- for i, line in enumerate(f):
87
- # possible keys are:
88
- # author: string (nullable = true)
89
- # body: string (nullable = true)
90
- # normalizedBody: string (nullable = true)
91
- # content: string (nullable = true)
92
- # content_len: long (nullable = true)
93
- # summary: string (nullable = true)
94
- # summary_len: long (nullable = true)
95
- # id: string (nullable = true)
96
- # subreddit: string (nullable = true)
97
- # subreddit_id: string (nullable = true)
98
- # title: string (nullable = true)
99
- d = json.loads(line)
100
- if _SUMMARY in d and _DOCUMENT in d:
101
- yield i, {k: d.get(k, "") for k in _ADDITIONAL_FEATURES + [_DOCUMENT, _SUMMARY]}