add offload_folder
Browse files- README.md +165 -0
- super_tweet_eval.py +2 -1
README.md
ADDED
@@ -0,0 +1,165 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# SuperTweetEval
|
2 |
+
|
3 |
+
# Dataset Card for "super_glue"
|
4 |
+
|
5 |
+
## Table of Contents
|
6 |
+
- [Dataset Description](#dataset-description)
|
7 |
+
- [Dataset Summary](#dataset-summary)
|
8 |
+
- [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
|
9 |
+
- [Languages](#languages)
|
10 |
+
- [Dataset Structure](#dataset-structure)
|
11 |
+
- [Data Instances](#data-instances)
|
12 |
+
- [Data Fields](#data-fields)
|
13 |
+
- [Data Splits](#data-splits)
|
14 |
+
- [Dataset Creation](#dataset-creation)
|
15 |
+
- [Curation Rationale](#curation-rationale)
|
16 |
+
- [Source Data](#source-data)
|
17 |
+
- [Annotations](#annotations)
|
18 |
+
- [Personal and Sensitive Information](#personal-and-sensitive-information)
|
19 |
+
- [Considerations for Using the Data](#considerations-for-using-the-data)
|
20 |
+
- [Social Impact of Dataset](#social-impact-of-dataset)
|
21 |
+
- [Discussion of Biases](#discussion-of-biases)
|
22 |
+
- [Other Known Limitations](#other-known-limitations)
|
23 |
+
- [Additional Information](#additional-information)
|
24 |
+
- [Dataset Curators](#dataset-curators)
|
25 |
+
- [Licensing Information](#licensing-information)
|
26 |
+
- [Citation Information](#citation-information)
|
27 |
+
- [Contributions](#contributions)
|
28 |
+
|
29 |
+
## Dataset Description
|
30 |
+
|
31 |
+
- **Homepage:** [https://github.com/google-research-datasets/boolean-questions](https://github.com/google-research-datasets/boolean-questions)
|
32 |
+
- **Repository:** [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
|
33 |
+
- **Paper:** [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
|
34 |
+
- **Point of Contact:** [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
|
35 |
+
- **Size of downloaded dataset files:** 58.36 MB
|
36 |
+
- **Size of the generated dataset:** 249.57 MB
|
37 |
+
- **Total amount of disk used:** 307.94 MB
|
38 |
+
|
39 |
+
### Dataset Summary
|
40 |
+
|
41 |
+
SuperGLUE (https://super.gluebenchmark.com/) is a new benchmark styled after
|
42 |
+
GLUE with a new set of more difficult language understanding tasks, improved
|
43 |
+
resources, and a new public leaderboard.
|
44 |
+
|
45 |
+
BoolQ (Boolean Questions, Clark et al., 2019a) is a QA task where each example consists of a short
|
46 |
+
passage and a yes/no question about the passage. The questions are provided anonymously and
|
47 |
+
unsolicited by users of the Google search engine, and afterwards paired with a paragraph from a
|
48 |
+
Wikipedia article containing the answer. Following the original work, we evaluate with accuracy.
|
49 |
+
|
50 |
+
### Supported Tasks and Leaderboards
|
51 |
+
|
52 |
+
[More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
|
53 |
+
|
54 |
+
### Languages
|
55 |
+
|
56 |
+
[More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
|
57 |
+
|
58 |
+
## Dataset Structure
|
59 |
+
### Data Fields
|
60 |
+
|
61 |
+
The data fields are the same among all splits.
|
62 |
+
|
63 |
+
#### tweet_topic
|
64 |
+
- `text`: a `string` feature.
|
65 |
+
- `label_list`: a list of `string` feature.
|
66 |
+
- `id`: a `string` feature.
|
67 |
+
- `date`: a `string` feature.
|
68 |
+
|
69 |
+
#### tweet_ner7
|
70 |
+
- `text`: a `string` feature.
|
71 |
+
- `label_list`: a list of `string` feature.
|
72 |
+
- `id`: a `string` feature.
|
73 |
+
- `date`: a `string` feature.
|
74 |
+
|
75 |
+
#### axg
|
76 |
+
- `premise`: a `string` feature.
|
77 |
+
- `hypothesis`: a `string` feature.
|
78 |
+
- `idx`: a `int32` feature.
|
79 |
+
- `label`: a classification label, with possible values including `entailment` (0), `not_entailment` (1).
|
80 |
+
|
81 |
+
#### boolq
|
82 |
+
- `question`: a `string` feature.
|
83 |
+
- `passage`: a `string` feature.
|
84 |
+
- `idx`: a `int32` feature.
|
85 |
+
- `label`: a classification label, with possible values including `False` (0), `True` (1).
|
86 |
+
|
87 |
+
#### cb
|
88 |
+
- `premise`: a `string` feature.
|
89 |
+
- `hypothesis`: a `string` feature.
|
90 |
+
- `idx`: a `int32` feature.
|
91 |
+
- `label`: a classification label, with possible values including `entailment` (0), `contradiction` (1), `neutral` (2).
|
92 |
+
|
93 |
+
#### copa
|
94 |
+
- `premise`: a `string` feature.
|
95 |
+
- `choice1`: a `string` feature.
|
96 |
+
- `choice2`: a `string` feature.
|
97 |
+
- `question`: a `string` feature.
|
98 |
+
- `idx`: a `int32` feature.
|
99 |
+
- `label`: a classification label, with possible values including `choice1` (0), `choice2` (1).
|
100 |
+
|
101 |
+
### Data Splits
|
102 |
+
|
103 |
+
|
104 |
+
## Citation Information
|
105 |
+
- TweetTopic
|
106 |
+
```
|
107 |
+
@inproceedings{antypas-etal-2022-twitter,
|
108 |
+
title = "{T}witter Topic Classification",
|
109 |
+
author = "Antypas, Dimosthenis and
|
110 |
+
Ushio, Asahi and
|
111 |
+
Camacho-Collados, Jose and
|
112 |
+
Silva, Vitor and
|
113 |
+
Neves, Leonardo and
|
114 |
+
Barbieri, Francesco",
|
115 |
+
booktitle = "Proceedings of the 29th International Conference on Computational Linguistics",
|
116 |
+
month = oct,
|
117 |
+
year = "2022",
|
118 |
+
address = "Gyeongju, Republic of Korea",
|
119 |
+
publisher = "International Committee on Computational Linguistics",
|
120 |
+
url = "https://aclanthology.org/2022.coling-1.299",
|
121 |
+
pages = "3386--3400",
|
122 |
+
abstract = "Social media platforms host discussions about a wide variety of topics that arise everyday. Making sense of all the content and organising it into categories is an arduous task. A common way to deal with this issue is relying on topic modeling, but topics discovered using this technique are difficult to interpret and can differ from corpus to corpus. In this paper, we present a new task based on tweet topic classification and release two associated datasets. Given a wide range of topics covering the most important discussion points in social media, we provide training and testing data from recent time periods that can be used to evaluate tweet classification models. Moreover, we perform a quantitative evaluation and analysis of current general- and domain-specific language models on the task, which provide more insights on the challenges and nature of the task.",
|
123 |
+
}
|
124 |
+
```
|
125 |
+
|
126 |
+
- TweetNER7
|
127 |
+
```
|
128 |
+
@inproceedings{ushio-etal-2022-named,
|
129 |
+
title = "Named Entity Recognition in {T}witter: A Dataset and Analysis on Short-Term Temporal Shifts",
|
130 |
+
author = "Ushio, Asahi and
|
131 |
+
Barbieri, Francesco and
|
132 |
+
Sousa, Vitor and
|
133 |
+
Neves, Leonardo and
|
134 |
+
Camacho-Collados, Jose",
|
135 |
+
booktitle = "Proceedings of the 2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)",
|
136 |
+
month = nov,
|
137 |
+
year = "2022",
|
138 |
+
address = "Online only",
|
139 |
+
publisher = "Association for Computational Linguistics",
|
140 |
+
url = "https://aclanthology.org/2022.aacl-main.25",
|
141 |
+
pages = "309--319",
|
142 |
+
abstract = "Recent progress in language model pre-training has led to important improvements in Named Entity Recognition (NER). Nonetheless, this progress has been mainly tested in well-formatted documents such as news, Wikipedia, or scientific articles. In social media the landscape is different, in which it adds another layer of complexity due to its noisy and dynamic nature. In this paper, we focus on NER in Twitter, one of the largest social media platforms, and construct a new NER dataset, TweetNER7, which contains seven entity types annotated over 11,382 tweets from September 2019 to August 2021. The dataset was constructed by carefully distributing the tweets over time and taking representative trends as a basis. Along with the dataset, we provide a set of language model baselines and perform an analysis on the language model performance on the task, especially analyzing the impact of different time periods. In particular, we focus on three important temporal aspects in our analysis: short-term degradation of NER models over time, strategies to fine-tune a language model over different periods, and self-labeling as an alternative to lack of recently-labeled data. TweetNER7 is released publicly (https://huggingface.co/datasets/tner/tweetner7) along with the models fine-tuned on it (NER models have been integrated into TweetNLP and can be found at https://github.com/asahi417/tner/tree/master/examples/tweetner7{\_}paper).",
|
143 |
+
}
|
144 |
+
```
|
145 |
+
- TweetQA
|
146 |
+
```
|
147 |
+
@inproceedings{xiong2019tweetqa,
|
148 |
+
title={TweetQA: A Social Media Focused Question Answering Dataset},
|
149 |
+
author={Xiong, Wenhan and Wu, Jiawei and Wang, Hong and Kulkarni, Vivek and Yu, Mo and Guo, Xiaoxiao and Chang, Shiyu and Wang, William Yang},
|
150 |
+
booktitle={Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics},
|
151 |
+
year={2019}
|
152 |
+
}
|
153 |
+
```
|
154 |
+
|
155 |
+
- TweetIntimacy
|
156 |
+
```
|
157 |
+
@misc{pei2023semeval,
|
158 |
+
title={SemEval 2023 Task 9: Multilingual Tweet Intimacy Analysis},
|
159 |
+
author={Jiaxin Pei and Vítor Silva and Maarten Bos and Yozon Liu and Leonardo Neves and David Jurgens and Francesco Barbieri},
|
160 |
+
year={2023},
|
161 |
+
eprint={2210.01108},
|
162 |
+
archivePrefix={arXiv},
|
163 |
+
primaryClass={cs.CL}
|
164 |
+
}
|
165 |
+
```
|
super_tweet_eval.py
CHANGED
@@ -2,7 +2,7 @@
|
|
2 |
import json
|
3 |
import datasets
|
4 |
|
5 |
-
_VERSION = "0.0.
|
6 |
_SUPER_TWEET_EVAL_CITATION = """TBA"""
|
7 |
_SUPER_TWEET_EVAL_DESCRIPTION = """TBA"""
|
8 |
_TWEET_TOPIC_DESCRIPTION = """
|
@@ -160,6 +160,7 @@ class SuperTweetEval(datasets.GeneratorBasedBuilder):
|
|
160 |
'B-corporation', 'B-creative_work', 'B-event', 'B-group', 'B-location', 'B-person', 'B-product',
|
161 |
'I-corporation', 'I-creative_work', 'I-event', 'I-group', 'I-location', 'I-person', 'I-product', 'O']
|
162 |
features["label_sequence"] = datasets.Sequence(datasets.features.ClassLabel(names=names))
|
|
|
163 |
if self.config.name in ["tweet_intimacy", "tweet_similarity"]:
|
164 |
features["label_float"] = datasets.Value("float32")
|
165 |
|
|
|
2 |
import json
|
3 |
import datasets
|
4 |
|
5 |
+
_VERSION = "0.0.3"
|
6 |
_SUPER_TWEET_EVAL_CITATION = """TBA"""
|
7 |
_SUPER_TWEET_EVAL_DESCRIPTION = """TBA"""
|
8 |
_TWEET_TOPIC_DESCRIPTION = """
|
|
|
160 |
'B-corporation', 'B-creative_work', 'B-event', 'B-group', 'B-location', 'B-person', 'B-product',
|
161 |
'I-corporation', 'I-creative_work', 'I-event', 'I-group', 'I-location', 'I-person', 'I-product', 'O']
|
162 |
features["label_sequence"] = datasets.Sequence(datasets.features.ClassLabel(names=names))
|
163 |
+
features["text_tokenized"] = datasets.Sequence(datasets.Value("string"))
|
164 |
if self.config.name in ["tweet_intimacy", "tweet_similarity"]:
|
165 |
features["label_float"] = datasets.Value("float32")
|
166 |
|