augmxnt/shisa-pretrain-en-ja-v1 download history

augmxnt/shisa-pretrain-en-ja-v1 is a dataset on the Hugging Face Hub. In the last 30 days it was downloaded 53 times (15 in the last 7 days), and 1,788 times in total. It ranks #251,917 among datasets by monthly downloads.

This pre-training dataset was created for shisa-base-7b-v1. It is primarily composed of a DSIR sampling of MADLAD-400 JA/EN tokens in a 90%/10% ratio.

Models trained on shisa-pretrain-en-ja-v1

6 models list it as training data.

Open augmxnt/shisa-pretrain-en-ja-v1 on Hugging Face

Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.