def run_ddpg_all_together():
train.train_ddpg_loop(
run_name="test_run_all_together_02",
seed=2019,
env_normalize_observations=True,
env_normalize_returns=True,
critic_l2_reg=0.0,
ops_together=True,
roller_normalize_observations=False,
roller_normalize_returns=False
)
def run_ddpg_all_together_better_normalization_with_reg():
train.train_ddpg_loop(
run_name="test_run_all_together_better_normalization_with_reg_03",
seed=2019,
env_normalize_observations=False,
env_normalize_returns=False,
critic_l2_reg=0.01,
ops_together=True,
roller_normalize_observations=True,
roller_normalize_returns=True
)
def run_ddpg_simple():
train.train_ddpg_loop(
run_name="test_run_simple_03",
# run_name="test_run_all_together_better_norm",
# run_name="test_run_all_together_better_norm_with_reg",
seed=2020,
env_normalize_observations=True,
env_normalize_returns=True,
critic_l2_reg=0.0,
ops_together=False,
roller_normalize_observations=False,
roller_normalize_returns=False
)