Teacher Ensemble Soft Labels

Problem

Average the soft predictions of several teacher models into one target distribution.

prob_list is a list of probability vectors of equal length; return the elementwise mean rounded to 4 decimal places.

case 1

in: [[[0.8,0.2],[0.6,0.4]]]

expected: [0.7,0.3]

case 2

in: [[[1,0,0]]]

expected: [1,0,0]

case 3

in: [[[0.2,0.8],[0.5,0.5],[0.8,0.2]]]

expected: [0.5,0.5]

Starter codePython

def ensemble_soft_labels(prob_list):
    # Your code here
    pass

3 test cases · run your code against them in the editor