ibm

/

PowerLM-3b

like 11

Thanks for your paper.
As the paper said "training corpus is a mix of large-scale, medium-quality open-source datasets with permissive licenses."
Would you share details about training corpus, Thanks a lot.