def run_dop_model(config, i):
env = gym.make('MountainCarContinuous-v0')
state_dim = env.observation_space.shape[0]
action_dim = env.action_space.shape[0]
experiment = ExperimentDDPG('MountainCarContinuous-v0', env, config)
agent = DDPGBulletDOPSimpleModelAgent(state_dim, action_dim, config)
experiment.run_dop_model(agent, i)
env.close()
def run_metalearner_rnd_model(config, i):
env = gym.make('MountainCarContinuous-v0')
state_dim = env.observation_space.shape[0]
action_dim = env.action_space.shape[0]
experiment = ExperimentDDPG('MountainCarContinuous-v0', env, config)
agent = DDPGBulletMetaCriticRNDModelAgent(state_dim, action_dim, config)
experiment.run_metalearner_rnd_model(agent, i)
env.close()
def run_qrnd_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 = DDPGBulletQRNDModelAgent(state_dim, action_dim, config)
experiment.run_qrnd_model(agent, i)
env.close()
def run_metalearner_model(config, i):
env = gym.make('HalfCheetahBulletEnv-v0')
state_dim = env.observation_space.shape[0]
action_dim = env.action_space.shape[0]
experiment = ExperimentDDPG('HalfCheetahBulletEnv-v0', env, config)
agent = DDPGBulletGatedMetacriticModelAgent(state_dim, action_dim, config)
experiment.run_metalearner_model(agent, i)
env.close()
def run_baseline(config, i):
env = gym.make('HalfCheetahBulletEnv-v0')
state_dim = env.observation_space.shape[0]
action_dim = env.action_space.shape[0]
experiment = ExperimentDDPG('HalfCheetahBulletEnv-v0', env, config)
agent = DDPGBulletAgent(state_dim, action_dim, config)
experiment.run_baseline(agent, i)
env.close()
def run_qrnd_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 = DDPGAerisQRNDModelAgent(state_dim, action_dim, config)
experiment.run_qrnd_model(agent, i)
env.close()
def run_forward_inverse_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 = DDPGAerisForwardInverseModelAgent(state_dim, action_dim, config)
experiment.run_forward_inverse_model(agent, trial)
env.close()
def run_baseline(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 = DDPGAerisAgent(state_dim, action_dim, config)
experiment.run_baseline(agent, trial)
env.close()
def run_metalearner_rnd_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 = DDPGAerisMetaCriticRNDModelAgent(state_dim, action_dim, config)
experiment.run_metalearner_rnd_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('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()