sartifyllc/Swahili-NLi-Triplet
收藏Hugging Face2024-07-04 更新2024-06-29 收录
下载链接：
https://hf-mirror.com/datasets/sartifyllc/Swahili-NLi-Triplet
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含三个主要特征：anchor、positive和negative，均为字符串类型。数据集分为训练集、开发集和测试集，分别包含557850、6584和6609个样本。总下载大小为39607912字节，数据集总大小为113696173字节。这些特征表明数据集可能用于某种形式的对比学习或三元组损失训练，其中anchor、positive和negative分别代表锚点、正例和负例。
The dataset includes three main features: anchor, positive, and negative, all of string type. It is divided into training, development, and test sets, containing 557850, 6584, and 6609 samples respectively. The total download size is 39607912 bytes, and the total dataset size is 113696173 bytes. These features suggest that the dataset might be used for some form of contrastive learning or triplet loss training, where anchor, positive, and negative represent the anchor, positive example, and negative example respectively.
提供机构：
sartifyllc
