def test_trainer_increment_step(ppo_optimizer, dummy_config):
trainer_params = dummy_config
mock_optimizer = mock.Mock()
mock_optimizer.reward_signals = {}
ppo_optimizer.return_value = mock_optimizer
brain_params = BrainParameters(
brain_name="test_brain",
vector_observation_space_size=1,
camera_resolutions=[],
vector_action_space_size=[2],
vector_action_descriptions=[],
vector_action_space_type=0,
)
trainer = PPOTrainer(
brain_params.brain_name, 0, trainer_params, True, False, 0, "0"
)
policy_mock = mock.Mock(spec=NNPolicy)
policy_mock.get_current_step.return_value = 0
step_count = (
5
) # 10 hacked because this function is no longer called through trainer
policy_mock.increment_step = mock.Mock(return_value=step_count)
trainer.add_policy("testbehavior", policy_mock)
trainer._increment_step(5, "testbehavior")
policy_mock.increment_step.assert_called_with(5)
assert trainer.step == step_count
def test_trainer_increment_step(ppo_optimizer):
trainer_params = PPO_CONFIG
mock_optimizer = mock.Mock()
mock_optimizer.reward_signals = {}
ppo_optimizer.return_value = mock_optimizer
trainer = PPOTrainer("test_brain", 0, trainer_params, True, False, 0, "0")
policy_mock = mock.Mock(spec=NNPolicy)
policy_mock.get_current_step.return_value = 0
step_count = (
5 # 10 hacked because this function is no longer called through trainer
)
policy_mock.increment_step = mock.Mock(return_value=step_count)
trainer.add_policy("testbehavior", policy_mock)
trainer._increment_step(5, "testbehavior")
policy_mock.increment_step.assert_called_with(5)
assert trainer.step == step_count
def test_trainer_increment_step(ppo_optimizer, mock_create_model_saver):
trainer_params = PPO_CONFIG
mock_optimizer = mock.Mock()
mock_optimizer.reward_signals = {}
ppo_optimizer.return_value = mock_optimizer
trainer = PPOTrainer("test_brain", 0, trainer_params, True, False, 0, "0")
policy_mock = mock.Mock(spec=TFPolicy)
policy_mock.get_current_step.return_value = 0
step_count = (
5 # 10 hacked because this function is no longer called through trainer
)
policy_mock.increment_step = mock.Mock(return_value=step_count)
behavior_id = BehaviorIdentifiers.from_name_behavior_id(trainer.brain_name)
trainer.add_policy(behavior_id, policy_mock)
trainer._increment_step(5, trainer.brain_name)
policy_mock.increment_step.assert_called_with(5)
assert trainer.step == step_count
