Datasets:

ArXiv:
License:

Can't read with Spark

#1
by egor-pakhomov - opened

This is a big dataset. zstd json on of the few format-compression pairs which can not be read with Spark

Hope that authors can convert to parquet for HF compatible.

ML Foundations org

We are working on parquet release! Please give us a few days.

ML Foundations org

We just released a parquet version of this dataset: https://huggingface.co/datasets/mlfoundations/dclm-baseline-1.0-parquet
There should be a 1:1 mapping between files in that repo and this one.

Please let us know if you have any more issues here!

mattjordan changed discussion status to closed

Thank you for your big effort!

Sign up or log in to comment