ParaCrawl

该数据集是一个用于评估跨多种语言的翻译模型的能力的分布外数据集。它成功地被转换成了多种语言，为评估翻译模型的鲁棒性提供了一个基准。该数据集的任务是机器翻译。

This dataset is an out-of-distribution (OOD) benchmark dataset intended to evaluate the capabilities of translation models across diverse languages. It has been successfully adapted to multiple languages, providing a reliable benchmark for assessing the robustness of translation models. The core task of this dataset is machine translation.