from argparse import Namespace
from gymnasium.spaces import Space
from xuance.common import List, Optional, MultiAgentBaseCallback
from xuance.environment import DummyVecMultiAgentEnv, SubprocVecMultiAgentEnv
from xuance.mindspore import Module
from xuance.mindspore.utils import NormalizeFunctions, InitializeFunctions, ActivationFunctions
from xuance.mindspore.policies import REGISTRY_Policy
from xuance.mindspore.agents import OffPolicyMARLAgents
[docs]
class ISAC_Agents(OffPolicyMARLAgents):
"""The implementation of Independent SAC agents.
Args:
config: the Namespace variable that provides hyperparameters and other settings.
envs: the vectorized environments.
callback: A user-defined callback function object to inject custom logic during training.
"""
def __init__(
self,
config: Namespace,
envs: Optional[DummyVecMultiAgentEnv | SubprocVecMultiAgentEnv] = None,
num_agents: Optional[int] = None,
agent_keys: Optional[List[str]] = None,
state_space: Optional[Space] = None,
observation_space: Optional[Space] = None,
action_space: Optional[Space] = None,
callback: Optional[MultiAgentBaseCallback] = None
):
super(ISAC_Agents, self).__init__(
config, envs, num_agents, agent_keys, state_space, observation_space, action_space, callback
)
# build policy, optimizers, schedulers
self.policy = self._build_policy() # build policy
self.memory = self._build_memory() # build memory
self.learner = self._build_learner(self.config, self.model_keys, self.agent_keys, self.policy, self.callback)
def _build_policy(self) -> Module:
"""
Build representation(s) and policy(ies) for agent(s)
Returns:
policy (torch.nn.Module): A dict of policies.
"""
normalize_fn = NormalizeFunctions[self.config.normalize] if hasattr(self.config, "normalize") else None
initializer = InitializeFunctions[self.config.initialize] if hasattr(self.config, "initialize") else None
activation = ActivationFunctions[self.config.activation]
agent = self.config.agent
# build representations
A_representation = self._build_representation(self.config.representation, self.observation_space, self.config)
C_representation = self._build_representation(self.config.representation, self.observation_space, self.config)
# build policies
if self.config.policy == "Gaussian_ISAC_Policy":
policy = REGISTRY_Policy["Gaussian_ISAC_Policy"](
action_space=self.action_space, n_agents=self.n_agents,
actor_representation=A_representation, critic_representation=C_representation,
actor_hidden_size=self.config.actor_hidden_size,
critic_hidden_size=self.config.critic_hidden_size,
normalize=normalize_fn, initialize=initializer, activation=activation,
activation_action=ActivationFunctions[self.config.activation_action],
use_parameter_sharing=self.use_parameter_sharing, model_keys=self.model_keys,
use_rnn=self.use_rnn, rnn=self.config.rnn if self.use_rnn else None)
self.continuous_control = True
elif self.config.policy == "Categorical_ISAC_Policy":
policy = REGISTRY_Policy["Categorical_ISAC_Policy"](
action_space=self.action_space, n_agents=self.n_agents,
actor_representation=A_representation, critic_representation=C_representation,
actor_hidden_size=self.config.actor_hidden_size,
critic_hidden_size=self.config.critic_hidden_size,
normalize=normalize_fn, initialize=initializer, activation=activation,
use_distributed_training=self.distributed_training,
use_parameter_sharing=self.use_parameter_sharing, model_keys=self.model_keys,
use_rnn=self.use_rnn, rnn=self.config.rnn if self.use_rnn else None)
self.continuous_control = False
else:
raise AttributeError(f"{agent} currently does not support the policy named {self.config.policy}.")
return policy
[docs]
def get_actions(self,
obs_dict: List[dict],
avail_actions_dict: Optional[List[dict]] = None,
rnn_hidden: Optional[dict] = None,
test_mode: Optional[bool] = False,
**kwargs):
"""
Returns actions for agents.
Parameters:
obs_dict (List[dict]): Observations for each agent in self.agent_keys.
avail_actions_dict (Optional[List[dict]]): Actions mask values, default is None.
rnn_hidden (Optional[dict]): The hidden variables of the RNN.
test_mode (Optional[bool]): True for testing without noises.
Returns:
rnn_hidden_state (dict): The new hidden states for RNN (if self.use_rnn=True).
actions_dict (dict): The output actions.
"""
batch_size = len(obs_dict)
obs_input, agents_id, avail_actions_input = self._build_inputs(obs_dict)
hidden_state, actions, _ = self.policy(observation=obs_input, agent_ids=agents_id, rnn_hidden=rnn_hidden)
if self.use_parameter_sharing:
key = self.model_keys[0]
if self.continuous_control:
actions[key] = actions[key].reshape([batch_size, self.n_agents, -1]).asnumpy()
else:
actions[key] = actions[key].reshape(batch_size, self.n_agents).asnumpy()
actions_dict = [{k: actions[key][e, i] for i, k in enumerate(self.agent_keys)} for e in range(batch_size)]
else:
for key in self.agent_keys:
if self.continuous_control:
actions[key] = actions[key].reshape([batch_size, -1]).asnumpy()
else:
actions[key] = actions[key].reshape(batch_size).asnumpy()
actions_dict = [{k: actions[k][i] for k in self.agent_keys} for i in range(batch_size)]
return {"hidden_state": hidden_state, "actions": actions_dict}