boun-tabilab/MultiNLI-TR

MultiNLI-TR 1.0是通过将英文MultiNLI 1.0语料库机器翻译成土耳其语而发布的数据集。它保留了原始的数据结构，包含三个主要字段：premise（提供NLI上下文的原句）、hypothesis（需要确定与premise关系的句子）和label（表示entailment、neutral或contradiction的标签）。数据集分为训练集、验证集和测试集，分别包含392702、4892和4923个例子。

MultiNLI-TR 1.0 is originally released by machine translating the English MultiNLI 1.0 corpora into Turkish. We kept the original data structure. The dataset contains three main fields: premise (the original sentence that provides the context for NLI), hypothesis (the sentence for which the relationship to the premise is to be determined), and label (0 for entailment, 1 for neutral, 2 for contradiction). The dataset is divided into training, validation, and test sets, containing 392702, 4892, and 4923 examples respectively.