Parallel and monolingual data for training machine translation systems translating English texts into Vietnamese, with a focus on news domain. The data was crawled from high-quality bilingual or multilingual websites of news and one-speaker educational talks on various topics, mostly technology, entertainment, and design (hereby referred to as TED-like talks). The dataset also includes noisy movie subtitles from the OpenSubtitle dataset.
Dataloader name:
vlsp2020_mt_envi/vlsp2020_mt_envi.py
DataCatalogue: http://seacrowd.github.io/seacrowd-catalogue/card.html?vlsp2020_mt_envi