ceteri

rllib rollout \
 tmp/ppo/cart/checkpoint_40/checkpoint-40 \
 - config "{\"env\": \"CartPole-v1\"}" \
 - run PPO \
 - steps 2000

ceteri

__________________________________________________________________________________________________
Layer (type)                    Output Shape         Param #     Connected to                     
==================================================================================================
observations (InputLayer)       [(None, 4)]          0                                            
__________________________________________________________________________________________________
fc_1 (Dense)                    (None, 256)          1280        observations[0][0]               
__________________________________________________________________________________________________
fc_value_1 (Dense)              (None, 256)          1280        observations[0][0]               
__________________________________________________________________________________________________
fc_2 (Dense)                    (None, 256)          65792       fc_1[0][0]                       

ceteri

N_ITER = 40
s = "{:3d} reward {:6.2f}/{:6.2f}/{:6.2f} len {:6.2f} saved {}"

for n in range(N_ITER):
  result = agent.train()
  file_name = agent.save(CHECKPOINT_ROOT)

  print(s.format(
    n + 1,
    result["episode_reward_min"],

ceteri

SELECT_ENV = "CartPole-v1"

config = ppo.DEFAULT_CONFIG.copy()
config["log_level"] = "WARN"

agent = ppo.PPOTrainer(config, env=SELECT_ENV)

ceteri

CHECKPOINT_ROOT = "tmp/ppo/cart"
shutil.rmtree(CHECKPOINT_ROOT, ignore_errors=True, onerror=None)

ray_results = os.getenv("HOME") + "/ray_results/"
shutil.rmtree(ray_results, ignore_errors=True, onerror=None)

ceteri

rllib rollout \
 tmp/ppo/froz/checkpoint_10/checkpoint-10 \
 - config "{\"env\": \"FrozenLake-v0\"}" \
 - run PPO \
 - steps 2000

ceteri

_____________________________________________________________________________
Layer (type) Output Shape Param # Connected to
=============================================================================
observations (InputLayer) [(None, 16)] 0 
_____________________________________________________________________________
fc_1 (Dense) (None, 256) 4352 observations[0][0] 
_____________________________________________________________________________
fc_value_1 (Dense) (None, 256) 4352 observations[0][0] 
_____________________________________________________________________________
fc_2 (Dense) (None, 256) 65792 fc_1[0][0] 

ceteri

N_ITER = 10
s = "{:3d} reward {:6.2f}/{:6.2f}/{:6.2f} len {:6.2f} saved {}"

for n in range(N_ITER):
  result = agent.train()
  file_name = agent.save(CHECKPOINT_ROOT)

  print(s.format(
    n + 1,
    result["episode_reward_min"],

ceteri

SELECT_ENV = "FrozenLake-v0"

config = ppo.DEFAULT_CONFIG.copy()
config["log_level"] = "WARN"

agent = ppo.PPOTrainer(config, env=SELECT_ENV)

ceteri

CHECKPOINT_ROOT = "tmp/ppo/froz"
shutil.rmtree(CHECKPOINT_ROOT, ignore_errors=True, onerror=None)

ray_results = os.getenv("HOME") + "/ray_results/"
shutil.rmtree(ray_results, ignore_errors=True, onerror=None)