---
license: odc-by
language:
- ja
- en
---
This pre-training dataset was created for [shisa-base-7b-v1](https://huggingface.co/datasets/augmxnt/shisa-pretrain-en-ja-v1).

It is primarily composed of a DSIR sampling of [MADLAD-400](https://huggingface.co/datasets/allenai/MADLAD-400) JA/EN tokens in a 90%/10% ratio.