MMTAfrica Test Set

该数据集是通过从每个平行源域中抽取相同数量的小样本句子构建的测试集，这些样本来自许多现有的测试集。它包含了来自众多领域的句子，使其成为评估多语种机器翻译任务的有用工具。尽管规模较小，但该数据集涵盖了广泛的领域。其主要任务是对多语种机器翻译进行评估。

This test dataset is constructed by extracting equal numbers of few-shot sentence samples from each parallel source domain, with all samples sourced from numerous existing test sets. It encompasses sentences spanning a diverse range of domains, making it a valuable tool for evaluating multilingual machine translation tasks. Despite its relatively small scale, this dataset covers a broad spectrum of fields. Its primary task is to evaluate multilingual machine translation.