DivEMT

该数据集包含了对FLORES数据集中一部分维基文本的mBART-50翻译后编辑的单个集合，这些翻译覆盖了六种类型学上多样化的目标语言。此外，该数据集还用于评估无监督的机器翻译质量评估（WQE）指标。该任务是对一组固定示例进行跨语言比较。

This dataset contains a single set of post-edited mBART-50 translations of a subset of Wikipedia texts from the FLORES dataset. These translations cover six typologically diverse target languages. Additionally, this dataset is utilized to evaluate unsupervised machine translation quality estimation (WQE) metrics. The task involves cross-lingual comparison of a fixed set of examples.