CoolFace
Datasetpublic

NEUDM/semeval-2015

上述数据集为ABSA(Aspect-Based Sentiment Analysis)领域数据集,基本形式为从句子中抽取:方面术语、方面类别(术语类别)、术语在上下文中情感极性以及针对该术语的观点词,不同数据集抽取不同的信息,这点在jsonl文件的“instruction”键中有分别提到,在此我将其改造为了生成任务,需要模型按照一定格式生成抽取结果。 补充:SemEval-2015数据集文件夹中有两个文件夹"laptop"和"restaurant",其实根据数据集文本的主要围绕主题区分的。抽取的元素方面,laptop和restaurant两文件夹中,数据的抽取元素也不同,laptop抽取的是方面类别和情感极性的二元组,restaurant抽取的是方面术语、方面类别和情感极性的三元组 以acos数据集中抽取的jsonl文件一条数据举例: { "task_type": "generation", "dataset": "acos", "input": ["the computer has difficulty… See the full description on the dataset page: https://huggingface.co/datasets/NEUDM/semeval-2015.

sourceHugging Faceupdated 3y agoView on Hugging Face
1likes51downloads
10 commits on main
c43c5063y ago

add researched SOTA result

xinxngxin
62ce52c3y ago

add some supplement note for unexpected folders

xinxngxin
5a589773y ago

Delete Generation/restaurant/a.txt

xinxngxin
e57b3433y ago

Upload 3 files

xinxngxin
10ad0773y ago

Delete Generation/laptop/a.txt

xinxngxin
80ecf873y ago

Upload 3 files

xinxngxin
e3474113y ago

Create restaurant/a.txt

xinxngxin
eeaf4e83y ago

Create Generation/laptop/a.txt

xinxngxin
cbbc22a3y ago

Create README.md

xinxngxin
6586e843y ago

initial commit

xinxngxin