def _setup_replay_memory(self, replay_memory):
replay_memory.sample_minibatch.side_effect = \
[[(0, _Transition(
np.array([0.1], np.float32),
0,
10,
np.array([0.2], np.float32),
0.01))],
[(1, _Transition(
np.array([0.3], np.float32),
1,
-10,
np.array([0.4], np.float32),
0.02))]]
def _setup_prioritized_replay_memory(self, replay_memory):
# Duplicated values can be returned.
replay_memory.sample_minibatch.return_value = \
[(3, _Transition(
np.array([0.1], np.float32),
0,
10,
np.array([0.2], np.float32),
2)),
(4, _Transition(
np.array([0.3], np.float32),
1,
11,
np.array([0.4], np.float32),
1)),
(3, _Transition(
np.array([0.1], np.float32),
0,
10,
np.array([0.2], np.float32),
2))]