Replay Buffer Sample

Problem

Sample batch indices from a replay buffer of the given size, with replacement.

Use random.Random(seed) and call rng.randrange(buffer_size) once per sample so the output is reproducible. Return the list of integer indices.

Hint

randrange(n) returns an index in [0, n), which is exactly uniform sampling from the buffer.

case 1

in: [10,5,42]

expected: [1,0,4,3,3]

case 2

in: [5,3,0]

expected: [3,3,0]

case 3

in: [1,4,7]

expected: [0,0,0,0]

case 4

in: [100,6,123]

expected: [6,34,11,98,52,34]

Starter codePython

import random
def replay_sample_indices(buffer_size, batch_size, seed):
    # Your code here
    pass

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