Datasets:
annotations_creators:
- shibing624
language_creators:
- shibing624
language:
- zh
license:
- cc-by-4.0
multilinguality:
- zh
size_categories:
- 100K<n<20M
source_datasets:
- https://www.biendata.xyz/competition/sohu_2021/data/
task_categories:
- text-classification
- sentence-similarity
task_ids:
- natural-language-inference
- semantic-similarity-scoring
- text-scoring
paperswithcode_id: sts
pretty_name: Sentence Text Similarity SOHU2021
Dataset Card for sts-sohu2021
Dataset Description
- Repository: Chinese NLI dataset
- Leaderboard: NLI_zh leaderboard (located on the homepage)
- Size of downloaded dataset files: 218 MB
- Total amount of disk used: 218 MB
Dataset Summary
2021搜狐校园文本匹配算法大赛数据集
分为 A 和 B 两个文件,A 和 B 文件匹配标准不一样。其中 A 和 B 文件又分为“短短文本匹配”、“短长文本匹配”和“长长文本匹配”。 A 文件匹配标准较为宽泛,两段文字是同一个话题便视为匹配,B 文件匹配标准较为严格,两段文字须是同一个事件才视为匹配。
数据类型:
type | 数据类型 |
---|---|
dda | 短短匹配 A 类 |
ddb | 短短匹配 B 类 |
dca | 短长匹配 A 类 |
dcb | 短长匹配 B 类 |
cca | 长长匹配 A 类 |
ccb | 长长匹配 B 类 |
Supported Tasks and Leaderboards
Supported Tasks: 支持中文文本匹配任务,文本相似度计算等相关任务。
中文匹配任务的结果目前在顶会paper上出现较少,我罗列一个我自己训练的结果:
Leaderboard: NLI_zh leaderboard
Languages
数据集均是简体中文文本。
Dataset Structure
Data Instances
An example of 'train' looks as follows.
# A 类 短短 样本示例
{
"sentence1": "小艺的故事让爱回家2021年2月16日大年初五19:30带上你最亲爱的人与团团君相约《小艺的故事》直播间!",
"sentence2": "香港代购了不起啊,宋点卷竟然在直播间“炫富”起来",
"label": 0
}
# B 类 短短 样本示例
{
"sentence1": "让很多网友好奇的是,张柏芝在一小时后也在社交平台发文:“给大家拜年啦。”还有网友猜测:谢霆锋的经纪人发文,张柏芝也发文,并且配图,似乎都在证实,谢霆锋依旧和王菲在一起,而张柏芝也有了新的恋人,并且生了孩子,两人也找到了各自的归宿,有了自己的幸福生活,让传言不攻自破。",
"sentence2": "陈晓东谈旧爱张柏芝,一个口误暴露她的秘密,难怪谢霆锋会离开她",
"label": 0
}
label: 0表示不匹配,1表示匹配。
Data Fields
The data fields are the same among all splits.
sentence1
: astring
feature.sentence2
: astring
feature.label
: a classification label, with possible values includingsimilarity
(1),dissimilarity
(0).
Data Splits
> wc -l *.jsonl
11690 cca.jsonl
11690 ccb.jsonl
11592 dca.jsonl
11593 dcb.jsonl
11512 dda.jsonl
11501 ddb.jsonl
69578 total
Curation Rationale
作为中文NLI(natural langauge inference)数据集,这里把这个数据集上传到huggingface的datasets,方便大家使用。
Who are the source language producers?
数据集的版权归原作者所有,使用各数据集时请尊重原数据集的版权。
Who are the annotators?
原作者。
Social Impact of Dataset
This dataset was developed as a benchmark for evaluating representational systems for text, especially including those induced by representation learning methods, in the task of predicting truth conditions in a given context.
Systems that are successful at such a task may be more successful in modeling semantic representations.
Licensing Information
用于学术研究。
Contributions
shibing624 upload this dataset.