Datasets:
Review dataset script
#1
by
larrylawl
- opened
Dataset review request for larrylawl/opus
Description
Loads OPUS as HuggingFace dataset. OPUS is an open parallel corpus covering 700+ languages and 1100+ datasets.
Given a src and tgt language, this repository can load all available parallel corpus. To my knowledge, other OPUS datasets on HuggingFace loads a specific corpus
Files to review
opus.pyREADME.md
Cool ! This can be super convenient :)
Can you make this dataset public for reviewing ?
Hi Quentin, sorry for the late reply. I made the dataset public. However, I realised that the loading is slow when used at scale, thus this dataset may not be useful. The loading is slow because I used opustools under the hood, which downloads and parses xml files.