flores_101

FLORES-101主要用于英语和低资源语言之间的机器翻译基准测试。它包含从英语维基百科提取的3001个句子，涵盖不同主题和领域，并由专业翻译人员翻译成101种语言。此数据集支持多语言机器翻译任务，并提供模型评估的排行榜。数据以ISO 639-3代码标识语言，并提供句子ID、句子内容、URL、领域、主题、是否包含图片和超链接等字段。FLORES-101使用CC-BY-SA 4.0协议授权。

FLORES-101 is primarily designed as a benchmark for machine translation between English and low-resource languages. It comprises 3001 sentences extracted from English Wikipedia, spanning various topics and domains, which were professionally translated into 101 languages. This dataset supports multilingual machine translation tasks and offers a leaderboard for model evaluation. Languages in the dataset are identified via ISO 639-3 codes, and it provides fields such as sentence ID, sentence content, URL, domain, topic, presence of images and hyperlinks, and more. FLORES-101 is licensed under the CC-BY-SA 4.0 license.

搜集汇总

背景与挑战

背景概述

FLORES-101是一个多语言机器翻译基准数据集，包含3001个英语维基百科句子的101种语言专业翻译，支持低资源语言翻译任务评估。数据集提供丰富的元信息，并采用CC-BY-SA 4.0协议开放使用。