The dataset is currently empty. Upload or create new data files. Then, you will be able to explore them in the Dataset Viewer.

Subsampled version of https://huggingface.co/datasets/jvamvas/apertus-pretrain-romansh-backtranslated that has been processed as follows:

  • Subsampled to 30k samples
  • Balanced across the six Romansh varieties without duplication: if a variety has fewer than its equal share, all its samples are kept and the surplus budget is redistributed to the next-smallest variety
  • Samples with higher LID scores are prioritized
  • Formatted as German-to-Romansh translation instruction pairs in magpie prompt/completion format
  • Prompt: Übersetze den folgenden Text nach {variety}:\n{german_backtranslation}
  • Completion: the original Romansh text
Downloads last month
-