def test_plas_impl(
observation_shape,
action_size,
actor_learning_rate,
critic_learning_rate,
imitator_learning_rate,
actor_optim_factory,
critic_optim_factory,
imitator_optim_factory,
encoder_factory,
q_func_factory,
gamma,
tau,
n_critics,
bootstrap,
share_encoder,
target_reduction_type,
lam,
beta,
scaler,
action_scaler,
augmentation,
):
impl = PLASImpl(
observation_shape,
action_size,
actor_learning_rate,
critic_learning_rate,
imitator_learning_rate,
actor_optim_factory,
critic_optim_factory,
imitator_optim_factory,
encoder_factory,
encoder_factory,
encoder_factory,
create_q_func_factory(q_func_factory),
gamma,
tau,
n_critics,
bootstrap,
share_encoder,
target_reduction_type,
lam,
beta,
use_gpu=False,
scaler=scaler,
action_scaler=action_scaler,
augmentation=augmentation,
)
torch_impl_tester(
impl, discrete=False, deterministic_best_action=q_func_factory != "iqn"
)
def test_plas_impl(
observation_shape,
action_size,
actor_learning_rate,
critic_learning_rate,
imitator_learning_rate,
actor_optim_factory,
critic_optim_factory,
imitator_optim_factory,
encoder_factory,
q_func_factory,
gamma,
tau,
n_critics,
target_reduction_type,
lam,
beta,
scaler,
action_scaler,
reward_scaler,
):
impl = PLASImpl(
observation_shape=observation_shape,
action_size=action_size,
actor_learning_rate=actor_learning_rate,
critic_learning_rate=critic_learning_rate,
imitator_learning_rate=imitator_learning_rate,
actor_optim_factory=actor_optim_factory,
critic_optim_factory=critic_optim_factory,
imitator_optim_factory=imitator_optim_factory,
actor_encoder_factory=encoder_factory,
critic_encoder_factory=encoder_factory,
imitator_encoder_factory=encoder_factory,
q_func_factory=create_q_func_factory(q_func_factory),
gamma=gamma,
tau=tau,
n_critics=n_critics,
target_reduction_type=target_reduction_type,
lam=lam,
beta=beta,
use_gpu=None,
scaler=scaler,
action_scaler=action_scaler,
reward_scaler=reward_scaler,
)
torch_impl_tester(impl,
discrete=False,
deterministic_best_action=q_func_factory != "iqn")