to_huggingface_dataset#

scikitplot.corpus.to_huggingface_dataset(documents)[source]#

Convert documents to a HuggingFace Dataset.

Parameters:
documentsSequence[CorpusDocument]

Source documents.

Returns:
datasets.Dataset or dict[str, list]

HuggingFace Dataset. Falls back to a column-dict if datasets is not installed.

Parameters:

documents (Sequence[Any])

Return type:

Any

Notes

User note: Directly usable for fine-tuning or upload:

ds = to_huggingface_dataset(docs)
ds.push_to_hub("my-org/my-corpus")