def agent_worker(config, policy, learner_w_queue, global_episode, n_agent,
log_dir, training_on, replay_queue, update_step):
agent = Agent(config,
policy,
global_episode=global_episode,
n_agent=n_agent,
log_dir=log_dir)
agent.run(training_on, replay_queue, learner_w_queue, update_step)
def agent_worker(config, policy, learner_w_queue, global_episode, i, agent_type,
experiment_dir, training_on, replay_queue, update_step):
agent = Agent(config,
policy=policy,
global_episode=global_episode,
n_agent=i,
agent_type=agent_type,
log_dir=experiment_dir)
agent.run(training_on, replay_queue, learner_w_queue, update_step)
def main():
env = gym.make(FLAGS.env_name)
agent = Agent(num_actions=env.action_space.n, config=FLAGS)
if FLAGS.train: # Train mode
for _ in range(FLAGS.num_episodes):
terminal = False
observation = env.reset()
for _ in range(random.randint(1, FLAGS.no_op_steps)):
last_observation = observation
observation, _, _, _ = env.step(0) # Do nothing
state = agent.get_initial_state(observation, last_observation)
while not terminal:
last_observation = observation
action = agent.get_action(state)
observation, reward, terminal, _ = env.step(action)
# env.render()
processed_observation = preprocess(observation, last_observation)
state = agent.run(state, action, reward, terminal, processed_observation)
else: # Test mode
# env.monitor.start(ENV_NAME + '-test')
for _ in range(FLAGS.num_episodes_at_test):
terminal = False
observation = env.reset()
for _ in range(random.randint(1, FLAGS.no_op_steps)):
last_observation = observation
observation, _, _, _ = env.step(0) # Do nothing
state = agent.get_initial_state(observation, last_observation)
while not terminal:
last_observation = observation
action = agent.get_action_at_test(state)
observation, _, terminal, _ = env.step(action)
env.render()
processed_observation = preprocess(observation, last_observation)
state = np.append(state[1:, :, :], processed_observation, axis=0)