HPLT/ua-gec

UA-GEC是一个文档级别的乌克兰语法错误校正数据集，用于内部使用在HPLT项目中。数据集由源文档和目标文档组成，源文档包含有语法错误的文本，而目标文档是经过校正的语法正确的文本。该数据集适用于文档排名任务，模型需要能够偏好语法正确的文本版本。

UA-GEC is a document-level Ukrainian grammatical error correction dataset intended for internal use in the HPLT project. The dataset consists of source documents containing ungrammatical text and target documents with corrected grammatical text. It is suitable for a document ranking task where the model is required to prefer the grammatically correct text version.