Datasets:
Add norwegian subset of opensubtitles (~250M tokens)
#3
by KennethEnevoldsen - opened
Danish subset of opensubtitles is ~271 tokens
@KennethEnevoldsen Are we sure we want to add open subtitles ?
From my understanding it is closer to Common Crawl in a sense that source is not responsible for the licence and they refer to downstream licence
it is also not included in Danish Gigaword because of copyright concerns (https://huggingface.co/datasets/danish-foundation-models/danish-gigaword/blob/main/README.md)
V4ldeLund changed discussion status to closed
V4ldeLund changed discussion status to open
We fixed the opensubtitles concern in danish gigaword (which is why it is included in dynaword). I think you might be able to filter on license in the API (if we can't then we should not add it)