def run_m2s_model(env_name, config, trial):
env = create_env(env_name)
state_dim = env.observation_space.shape
action_dim = env.action_space.shape[0]
experiment = ExperimentDDPG(env_name, env, config)
agent = DDPGAerisM2SModelAgent(state_dim, action_dim, config)
experiment.run_forward_model(agent, trial)
env.close()
def run_metalearner_model(env_name, config, trial):
env = create_env(env_name)
state_dim = env.observation_space.shape
action_dim = env.action_space.shape[0]
experiment = ExperimentDDPG(env_name, env, config)
agent = DDPGAerisGatedMetacriticModelAgent(state_dim, action_dim, config)
experiment.run_metalearner_model(agent, trial)
env.close()
def run_dop_model(env_name, config, i):
env = create_env(env_name)
state_dim = env.observation_space.shape
action_dim = env.action_space.shape[0]
experiment = ExperimentDDPG(env_name, env, config)
agent = DDPGAerisDOPAgent(state_dim, action_dim, config, TYPE.continuous)
experiment.run_dop_model(agent, i)
env.close()
def run_baseline(config, i):
env = gym.make('HopperBulletEnv-v0')
state_dim = env.observation_space.shape[0]
action_dim = env.action_space.shape[0]
experiment = ExperimentDDPG('HopperBulletEnv-v0', env, config)
agent = DDPGBulletAgent(state_dim, action_dim, config)
experiment.run_baseline(agent, i)
env.close()
def run_metalearner_model(config, i):
env = gym.make('HopperBulletEnv-v0')
state_dim = env.observation_space.shape[0]
action_dim = env.action_space.shape[0]
experiment = ExperimentDDPG('HopperBulletEnv-v0', env, config)
agent = DDPGBulletGatedMetacriticModelAgent(state_dim, action_dim, config)
experiment.run_metalearner_model(agent, i)
env.close()
def run_forward_model(config, i):
env = gym.make('AntBulletEnv-v0')
state_dim = env.observation_space.shape[0]
action_dim = env.action_space.shape[0]
experiment = ExperimentDDPG('AntBulletEnv-v0', env, config)
agent = DDPGBulletForwardModelAgent(state_dim, action_dim, config)
experiment.run_forward_model(agent, i)
env.close()
def run_baseline(config, i):
env = gym.make('LunarLanderContinuous-v2')
state_dim = env.observation_space.shape[0]
action_dim = env.action_space.shape[0]
experiment = ExperimentDDPG('LunarLanderContinuous-v2', env, config)
agent = DDPGAgent(state_dim, action_dim, config)
experiment.run_baseline(agent, i)
env.close()
