ferret.Tokenf1_PlausibilityEvaluation#

class ferret.Tokenf1_PlausibilityEvaluation(model, tokenizer, task_name)[source]#
__init__(model, tokenizer, task_name)#

Methods

__init__(model, tokenizer, task_name)

aggregate_score(score, total, **aggregation_args)

compute_evaluation(explanation_with_rationale)

Evaluate an explanation on the Token-f1 score Plausibility metric.

Attributes

LOWER_IS_BETTER

MAX_VALUE

METRIC_FAMILY

MIN_VALUE

NAME

SHORT_NAME

tokenizer