Convert to Parquet format and remove loading script

#4
by elie-mada - opened

Dataset loading scripts are no longer supported by the datasets library (v3+).

This PR:

  • Converts all splits (train, test, validation) to Parquet format
  • Removes the ptb_text_only.py loading script

After this change, the dataset can be loaded with:

from datasets import load_dataset
ds = load_dataset('ptb-text-only/ptb_text_only')

No trust_remote_code=True needed.

Ready to merge
This branch is ready to get merged automatically.

Sign up or log in to comment