License clarification for reuse in a research artifact
#1
by NewArtY - opened
Hello,
The dataset card for codeparrot/codeparrot-clean-valid does not set a license field, and I would like to clarify the terms before publishing a research artifact alongside a paper.
What I used: 296 documents from revision 4db92d2ec0c1b4c41eeb439cfae16854511d9dcd, tokenized into 2000 windows of 256 tokens. Only aggregate statistics derived from them appear in the paper itself.
Two questions:
- Under what license is this dataset distributed?
- Would you permit redistributing a derived form - sequences of token IDs for that subset - as part of a public research artifact? If so, what attribution would you like me to use?
If there is no clear answer, I will simply not include the tokenized text in the artifact and will reference the dataset by path and revision instead, so nothing is required on your side.
Thank you.