Function main

rl2/cartpole/pg_tf.py:246–280

()

def main():
  env = gym.make('CartPole-v0')
  D = env.observation_space.shape[0]
  K = env.action_space.n
  pmodel = PolicyModel(D, K, [])
  vmodel = ValueModel(D, [10])
  init = tf.global_variables_initializer()
  session = tf.InteractiveSession()
  session.run(init)
  pmodel.set_session(session)
  vmodel.set_session(session)
  gamma = 0.99

  if 'monitor' in sys.argv:
    filename = os.path.basename(__file__).split('.')[0]
    monitor_dir = './' + filename + '_' + str(datetime.now())
    env = wrappers.Monitor(env, monitor_dir)

  N = 1000
  totalrewards = np.empty(N)
  costs = np.empty(N)
  for n in range(N):
    totalreward = play_one_mc(env, pmodel, vmodel, gamma)
    totalrewards[n] = totalreward
    if n % 100 == 0:
      print("episode:", n, "total reward:", totalreward, "avg reward (last 100):", totalrewards[max(0, n-100):(n+1)].mean())

  print("avg reward for last 100 episodes:", totalrewards[-100:].mean())
  print("total steps:", totalrewards.sum())

  plt.plot(totalrewards)
  plt.title("Rewards")
  plt.show()

  plot_running_avg(totalrewards)


if __name__ == '__main__':