ZurichNLP/mediomatix
收藏Hugging Face2026-03-21 更新2025-09-13 收录
下载链接：
https://hf-mirror.com/datasets/ZurichNLP/mediomatix
下载链接
链接失效反馈官方服务：
资源简介：
Mediomatix语料库是一个由UZH和PHGR准备的用于研究的平行语料库，包含了罗曼什语的五种方言：Sursilvan、Sutsilvan、Surmiran、Puter和Vallader。语料库中的数据以多对齐的形式展示，并且按照学校书籍的年级和卷数进行编码。数据集分为训练集、验证集、测试集和不包含rm-surmiran方言的数据集。
The Mediomatix corpus is a parallel corpus prepared by UZH and PHGR for research purposes, containing five Romansh idioms: Sursilvan, Sutsilvan, Surmiran, Puter, and Vallader. The data in the corpus is presented in a multi-aligned form and is encoded according to the school year and volume of the textbooks. The dataset is split into training, validation, test, and a split without the rm-surmiran idiom.
提供机构：
ZurichNLP
