ted_multi

TEDMulti是一个大规模多语种数据集，源自TED演讲的文本记录，包含60种语言的平行语料。数据集的每个样本由演讲名称和多语种翻译文本构成，并已过滤缺失或不完整的翻译。该数据集包含训练集、验证集和测试集，可用于机器翻译等任务。TEDMulti is a large-scale multilingual dataset derived from the textual transcripts of TED Talks. It contains parallel corpora across 60 languages, with each sample consisting of talk titles and multilingual translated texts. Missing or incomplete translations have been filtered out from the dataset. This dataset includes training, validation, and test splits, and can be used for tasks such as machine translation.

搜集汇总

背景与挑战

背景概述

TEDMulti是一个基于TED演讲的大规模多语种平行语料数据集，覆盖60种语言，每个样本包含演讲名称和翻译文本，并已过滤不完整翻译。该数据集包含训练、验证和测试分割，适用于机器翻译等自然语言处理任务。