Continuous Action Target Entropy

Problem

Return the standard SAC target entropy for a continuous action space: the negative action dimensionality, -action_dim.

Hint

One nat of entropy per action dimension is the usual heuristic target.

case 1

in: [1]

expected: -1

case 2

in: [3]

expected: -3

case 3

in: [6]

expected: -6

case 4

in: [0]

expected: 0

Starter codePython

def continuous_target_entropy(action_dim):
    # Your code here
    pass

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