遇见数据集

turkic-xwmt

收藏
OpenXLab2026-04-18 收录
官方服务:

资源简介:

To establish a comprehensive and challenging evaluation benchmark for Machine Translation in Turkic languages, we translate a test set originally introduced in WMT 2020 News Translation Task for English-Russian. The original dataset is profesionally translated and consists of sentences from news articles that are both English and Russian-centric. We adopt this evaluation set (X-WMT) and begin efforts to translate it into several Turkic languages. The current version of X-WMT includes covers 8 Turkic languages and 88 language directions with a minimum of 300 sentences per language direction.

提供机构:
OpenDataLab
创建时间:
2023-12-07
二维码
社区交流群
二维码
科研交流群
商业服务