aadelucia
commited on
Commit
•
a091e7d
0
Parent(s):
initial commit
Browse files- .gitignore +1 -0
- README.md +141 -0
.gitignore
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
tmp
|
README.md
ADDED
@@ -0,0 +1,141 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Dataset Card for Bernice Pre-train Data
|
2 |
+
|
3 |
+
## Table of Contents
|
4 |
+
- [Table of Contents](#table-of-contents)
|
5 |
+
- [Dataset Description](#dataset-description)
|
6 |
+
- [Dataset Summary](#dataset-summary)
|
7 |
+
- [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
|
8 |
+
- [Languages](#languages)
|
9 |
+
- [Dataset Structure](#dataset-structure)
|
10 |
+
- [Data Instances](#data-instances)
|
11 |
+
- [Data Fields](#data-fields)
|
12 |
+
- [Data Splits](#data-splits)
|
13 |
+
- [Dataset Creation](#dataset-creation)
|
14 |
+
- [Curation Rationale](#curation-rationale)
|
15 |
+
- [Source Data](#source-data)
|
16 |
+
- [Annotations](#annotations)
|
17 |
+
- [Personal and Sensitive Information](#personal-and-sensitive-information)
|
18 |
+
- [Considerations for Using the Data](#considerations-for-using-the-data)
|
19 |
+
- [Social Impact of Dataset](#social-impact-of-dataset)
|
20 |
+
- [Discussion of Biases](#discussion-of-biases)
|
21 |
+
- [Other Known Limitations](#other-known-limitations)
|
22 |
+
- [Additional Information](#additional-information)
|
23 |
+
- [Dataset Curators](#dataset-curators)
|
24 |
+
- [Licensing Information](#licensing-information)
|
25 |
+
- [Citation Information](#citation-information)
|
26 |
+
- [Contributions](#contributions)
|
27 |
+
|
28 |
+
## Dataset Description
|
29 |
+
|
30 |
+
- **Homepage:** N/A
|
31 |
+
- **Repository:** https://github.com/JHU-CLSP/Bernice-Twitter-encoder
|
32 |
+
- **Paper:** _Bernice: A Multilingual Pre-trained Encoder for Twitter_ at [EMNLP 2022](https://preview.aclanthology.org/emnlp-22-ingestion/2022.emnlp-main.415)
|
33 |
+
- **Leaderboard:** N/A
|
34 |
+
- **Point of Contact:** Alexandra DeLucia aadelucia (at) jhu.edu
|
35 |
+
|
36 |
+
### Dataset Summary
|
37 |
+
|
38 |
+
Tweet IDs for the 2.5 billion multilingual tweets used to train Bernice, a Twitter encoder.
|
39 |
+
The tweets are from the public 1% Twitter API stream from January 2016 to December 2021.
|
40 |
+
Twitter-provided language metadata is provided with the tweet ID. The data contains 66 unique languages, as identified by [ISO 639 language codes](https://www.wikiwand.com/en/List_of_ISO_639-1_codes), including `und` for undefined languages.
|
41 |
+
Tweets need to be re-gathered via the Twitter API. We suggest [Hydrator](https://github.com/DocNow/hydrator) or [tweepy](https://www.tweepy.org/).
|
42 |
+
|
43 |
+
|
44 |
+
### Supported Tasks and Leaderboards
|
45 |
+
|
46 |
+
N/A
|
47 |
+
|
48 |
+
### Languages
|
49 |
+
|
50 |
+
65 languages (ISO 639 codes shown below), plus an `und` (undefined) category.
|
51 |
+
All language identification provided by Twitter API.
|
52 |
+
|
53 |
+
| | | | | | | |
|
54 |
+
|----|-----|----|----|----|-----|----|
|
55 |
+
| en | ru | ht | zh | bn | ps | lt |
|
56 |
+
| es | bo | ur | ta | sr | ckb | km |
|
57 |
+
| pt | it | sv | ro | bg | si | dv |
|
58 |
+
| ja | th | ca | no | mr | hy | lo |
|
59 |
+
| ar | de | el | uk | ml | or | ug |
|
60 |
+
| in | hi | fi | cy | is | pa | |
|
61 |
+
| ko | pl | cs | ne | te | am | |
|
62 |
+
| tr | nl | iw | hu | gu | sd | |
|
63 |
+
| fr | fa | da | eu | kn | my | |
|
64 |
+
| tl | et | vi | sl | lv | ka | |
|
65 |
+
|
66 |
+
## Dataset Structure
|
67 |
+
|
68 |
+
### Data Instances
|
69 |
+
|
70 |
+
[More Information Needed]
|
71 |
+
|
72 |
+
### Data Fields
|
73 |
+
|
74 |
+
[More Information Needed]
|
75 |
+
|
76 |
+
### Data Splits
|
77 |
+
|
78 |
+
[More Information Needed]
|
79 |
+
|
80 |
+
|
81 |
+
## Dataset Creation
|
82 |
+
|
83 |
+
### Curation Rationale
|
84 |
+
|
85 |
+
Data was gathered to support the training of Bernice, a multilingual pre-trained Twitter encoder.
|
86 |
+
|
87 |
+
### Source Data
|
88 |
+
|
89 |
+
#### Initial Data Collection and Normalization
|
90 |
+
|
91 |
+
Data was gathered via the Twitter API public 1% stream from January 2016 through December 2021.
|
92 |
+
Tweets with less than three non-username or URL space-delimited words were removed.
|
93 |
+
All usernames and URLs were replaced with `@USER` and `HTTPURL`, respectively.
|
94 |
+
|
95 |
+
#### Who are the source language producers?
|
96 |
+
|
97 |
+
Data was produced by users on Twitter.
|
98 |
+
|
99 |
+
### Annotations
|
100 |
+
|
101 |
+
N/A
|
102 |
+
|
103 |
+
### Personal and Sensitive Information
|
104 |
+
|
105 |
+
As per Twitter guidelines, only tweet IDs and not full tweets are shared.
|
106 |
+
Tweets will only be accessible if user has not removed their account (or been banned), tweets were deleted or removed, or a user changed their account access to private.
|
107 |
+
|
108 |
+
## Considerations for Using the Data
|
109 |
+
|
110 |
+
### Social Impact of Dataset
|
111 |
+
|
112 |
+
[More Information Needed]
|
113 |
+
|
114 |
+
### Discussion of Biases
|
115 |
+
|
116 |
+
[More Information Needed]
|
117 |
+
|
118 |
+
### Other Known Limitations
|
119 |
+
|
120 |
+
[More Information Needed]
|
121 |
+
|
122 |
+
|
123 |
+
## Additional Information
|
124 |
+
|
125 |
+
### Dataset Curators
|
126 |
+
|
127 |
+
Dataset gathered and processed by Mark Dredze, Alexandra DeLucia, Shijie Wu, Aaron Mueller, Carlos Aguirre, and Philip Resnik.
|
128 |
+
|
129 |
+
### Licensing Information
|
130 |
+
|
131 |
+
[More Information Needed]
|
132 |
+
|
133 |
+
### Citation Information
|
134 |
+
|
135 |
+
Please cite the Bernice paper if you use this dataset:
|
136 |
+
|
137 |
+
> Alexandra DeLucia, Shijie Wu, Aaron Mueller, Carlos Aguirre, Philip Resnik, and Mark Dredze. 2022. Bernice: A Multilingual Pre-trained Encoder for Twitter. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 6191–6205, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics.
|
138 |
+
|
139 |
+
### Contributions
|
140 |
+
|
141 |
+
Dataset uploaded by [@AADeLucia](https://github.com/AADeLucia).
|