marcov/esnli_promptsource

该数据集包含多个字段，如前提（premise）、假设（hypothesis）、标签（label）等。标签字段是一个分类标签，包含蕴含（entailment）、中立（neutral）和矛盾（contradiction）三个类别。此外，数据集还包含多个解释字段（explanation_1, explanation_2, explanation_3）以及模板相关字段（template_name, template, rendered_input, rendered_output）。数据集分为训练集、验证集和测试集，分别包含5493570、98420和98240个样本。数据集的下载大小为772918906字节，总大小为4042717690.042371字节。

The dataset contains multiple fields such as premise, hypothesis, label, etc. The label field is a classification label with three categories: entailment, neutral, and contradiction. Additionally, the dataset includes multiple explanation fields (explanation_1, explanation_2, explanation_3) and template-related fields (template_name, template, rendered_input, rendered_output). The dataset is divided into training, validation, and test sets, containing 5493570, 98420, and 98240 samples respectively. The download size of the dataset is 772918906 bytes, and the total size is 4042717690.042371 bytes.