def _play(self, play_context: core.PlayContext,
callbacks: Union[List[core.AgentCallback], core.AgentCallback, None],
default_plots: Optional[bool]):
"""Plays episodes with the current policy according to play_context.
Hints:
o updates rewards in play_context
Args:
play_context: specifies the num of episodes to play
callbacks: list of callbacks called during the play of the episodes
default_plots: if set adds a set of default callbacks (plot.State, plot.Rewards, plot.Loss,...).
if None default callbacks are only added if the callbacks list is empty
Returns:
play_context containing the actions taken and the rewards received during training
"""
assert play_context, "play_context not set."
if callbacks is None:
callbacks = []
if not isinstance(callbacks, list):
assert isinstance(callbacks, core.AgentCallback), "callback not an AgentCallback or a list thereof."
callbacks = [callbacks]
callbacks = self._prepare_callbacks(callbacks, default_plots, [plot.Steps(), plot.Rewards()])
self._backend_agent.play(play_context=play_context, callbacks=callbacks)
return play_context
def train(self, train_context: core.TrainContext,
callbacks: Union[List[core.AgentCallback], core.AgentCallback,
None], default_plots: Optional[bool]):
"""Trains a new model using the gym environment passed during instantiation.
Args:
callbacks: list of callbacks called during the training and evaluation
train_context: training configuration to be used (num_iterations,num_episodes_per_iteration,...)
default_plots: if set adds a set of default callbacks (plot.State, plot.Rewards, plot.Loss,...).
if None default callbacks are only added if the callbacks list is empty
"""
assert train_context, "train_context not set."
if callbacks is None:
callbacks = []
if not isinstance(callbacks, list):
assert isinstance(
callbacks, core.AgentCallback
), "callback not a AgentCallback or a list thereof."
callbacks = [callbacks]
callbacks = self._prepare_callbacks(
callbacks, default_plots,
[plot.Loss(), plot.Steps(),
plot.Rewards()])
self._backend_agent.train(train_context=train_context,
callbacks=callbacks)
def play(self,
callbacks: Union[List[core.AgentCallback], core.AgentCallback, None] = None,
num_episodes: int = 1,
max_steps_per_episode: int = 1000,
play_context: core.PlayContext = None,
default_plots: bool = None):
"""Plays num_episodes with the current policy.
Args:
callbacks: list of callbacks called during each episode play
num_episodes: number of episodes to play
max_steps_per_episode: max steps per episode
play_context: play configuration to be used. If set override all other play context arguments
default_plots: if set addes a set of default callbacks (plot.State, plot.Rewards, ...)
Returns:
play_context containg the actions taken and the rewards received during training
"""
assert self._backend_agent._agent_context._is_policy_trained, "No trained policy available. Call train() first."
if play_context is None:
play_context = core.PlayContext()
play_context.max_steps_per_episode = max_steps_per_episode
play_context.num_episodes = num_episodes
callbacks = self._to_callback_list(callbacks=callbacks)
callbacks = self._add_plot_callbacks(callbacks, default_plots, [plot.Steps(), plot.Rewards()])
self._backend_agent.play(play_context=play_context, callbacks=callbacks)
return play_context
def test_train_multiple_subplots(self):
agent = agents.PpoAgent("CartPole-v0")
agent.train([
duration._SingleIteration(),
plot.State(),
plot.Rewards(),
plot.Loss(),
plot.Steps()
])
def test_train_plotsteps(self):
agent = agents.PpoAgent("CartPole-v0")
agent.train([duration._SingleIteration(), plot.Steps()])
