Datasets:

EleutherAI

/

pile

like 391

As reported by @shaoyuta, a UnidoceDecodeError is raised when loading the "enron_emails" subset:

datasets.load_dataset("the_pile","enron_emails")

raises:

UnicodeDecodeError: 'utf-8' codec can't decode byte 0xbf in position 500: invalid start byte