Neural-Machine-Translation
收藏Zenodo2025-05-13 更新2026-05-26 收录
下载链接：
https://zenodo.org/doi/10.5281/zenodo.15392601
下载链接
链接失效反馈官方服务：
资源简介：
The WMT 2014 English-German dataset is a cornerstone resource for researchers developing and evaluating machine translation (MT) systems. It's widely used in the annual WMT shared task, serving as a standard benchmark to compare different approaches and track progress in the field.
Key Features
Size: 4.5 million parallel sentence pairs, providing ample data for training and testing MT models
Origin: Comprises high-quality news articles from Europarl, News Commentary, and TED Talks, offering realistic and diverse text domains.
Preprocessing: Cleaned and normalized for consistency, ensuring model compatibility and training efficiency.
Task Diversity: Originally used for the WMT 2014 News Translation Task, but applicable to various MT research areas
提供机构：
Zenodo创建时间：
2025-05-13
