Datasets:

bible-nlp

/

biblenlp-corpus

like 25

File name was too long, resulting in an OS error when a path is created. The keyword argument name was shortened by passing it to an md5 hash function in the hashlib library. Note that according to the hashlib documentation, the md5 algorithm "may be missing or blocked if you are using a rare “FIPS compliant” build of Python."

Note that passing ['eng'] as a language to load_dataset results in 30 translations being loaded; I assumed this is because there is more than one english translation. Also, ['fra'] produced three translations and ['for'] produced one. However, if this should not happen, then maybe more than changing name to an md5 hash is necessary.