primeqa.boolqa.processors.dataset.mrc2dataset.create_dataset_from_run_mrc_output#
- primeqa.boolqa.processors.dataset.mrc2dataset.create_dataset_from_run_mrc_output(mrcfn: str, unpack: bool) datasets.arrow_dataset.Dataset#
Converts the output of run_mrc.py (eval_predictions.json, not eval_predictions_processed.json) into a Dataset for use with the classifiers
- Parameters
mrcfn (str) – path to the eval_predictions.json file produced by run_mrc.py
- Returns
a dataset containing all of the fields in mrcfn as features, for the top-ranked answer
- Return type