mrc2dataset#
Functions
Converts the output of run_mrc.py (eval_predictions.json, not eval_predictions_processed.json) into a Dataset for use with the classifiers |
|
Converts the output of run_mrc.py (eval_predictions.json, not eval_predictions_processed.json) into a Dataset for use with the classifiers |
|
span_answer is a dictionary with two items - split into separate items for easier handling with pandas/datasets |
|
target_type_logits is an array of 5 numbers, with meanings determined by the TargetType enum. |