multi_explanation

generate multiple explanations from a responsibility landscape <pixel_ranking>

Classes

MultiExplanation

Module Contents

class multi_explanation.MultiExplanation(maps, prediction_func, data, args, run_stats)

Bases: rex_xai.explanation.explanation.Explanation

explanations = []
explanation_confidences = []
__repr__()
Return type:

str

save(path, mask=None, clauses=None, multi_style=None)
Parameters:

  • path (str)

  • clauses (List[int] | None)

  • multi_style (str | None)

show(multi_style=None, clauses=None)
extract()
__dice(d1, d2)

calculates dice coefficient between two numpy arrays of the same dimensions

separate_by(dice_coefficient, reverse=True)
Parameters:

dice_coefficient (float)

__random_step_from(origin, width, height, step=5)
__random_location()