openrl.modules.common package¶

Submodules¶

openrl.modules.common.a2c_net module¶

class openrl.modules.common.a2c_net.A2CNet(env: Union[openrl.envs.vec_env.base_venv.BaseVecEnv, gymnasium.core.Env, str], cfg=None, device: Union[torch.device, str] = 'cpu', n_rollout_threads: int = 1, model_dict: Optional[Dict[str, Any]] = None, module_class: openrl.modules.base_module.BaseModule = <class 'openrl.modules.ppo_module.PPOModule'>)[source]¶: Bases: openrl.modules.common.ppo_net.PPONet

openrl.modules.common.base_net module¶

class openrl.modules.common.base_net.BaseNet[source]¶: Bases: abc.ABC

openrl.modules.common.bc_net module¶

class openrl.modules.common.bc_net.BCNet(env: openrl.envs.vec_env.wrappers.reward_wrapper.RewardWrapper, cfg=None, device: Union[torch.device, str] = 'cpu', n_rollout_threads: int = 1, model_dict: Optional[Dict[str, Any]] = None, module_class: abc.ABCMeta = <class 'openrl.modules.bc_module.BCModule'>)[source]¶: Bases: openrl.modules.common.ppo_net.PPONet

openrl.modules.common.ddpg_net module¶

class openrl.modules.common.ddpg_net.DDPGNet(env: Union[gymnasium.core.Env, str], cfg=None, device: Union[torch.device, str] = 'cpu', n_rollout_threads: int = 1, model_dict: Optional[Dict[str, Any]] = None)[source]¶

Bases: openrl.modules.common.base_net.BaseNet

act(observation: Union[numpy.ndarray, Dict[str, numpy.ndarray]], deterministic: bool) → Tuple[numpy.ndarray, Optional[Tuple[numpy.ndarray, ...]]][source]¶

load_policy(path: str) → None[source]¶

reset(env: Optional[gymnasium.core.Env] = None) → None[source]¶

openrl.modules.common.dqn_net module¶

class openrl.modules.common.dqn_net.DQNNet(env: Union[gymnasium.core.Env, str], cfg=None, device: Union[torch.device, str] = 'cpu', n_rollout_threads: int = 1, model_dict: Optional[Dict[str, Any]] = None)[source]¶

Bases: openrl.modules.common.base_net.BaseNet

act(observation: Union[numpy.ndarray, Dict[str, numpy.ndarray]]) → Tuple[numpy.ndarray, Optional[Tuple[numpy.ndarray, ...]]][source]¶

load_policy(path: str) → None[source]¶

reset(env: Optional[gymnasium.core.Env] = None) → None[source]¶

openrl.modules.common.gail_net module¶

class openrl.modules.common.gail_net.GAILNet(env: openrl.envs.vec_env.wrappers.reward_wrapper.RewardWrapper, cfg=None, device: Union[torch.device, str] = 'cpu', n_rollout_threads: int = 1, model_dict: Optional[Dict[str, Any]] = None, module_class: abc.ABCMeta = <class 'openrl.modules.gail_module.GAILModule'>)[source]¶: Bases: openrl.modules.common.ppo_net.PPONet

openrl.modules.common.mat_net module¶

class openrl.modules.common.mat_net.MATNet(env: Union[gymnasium.core.Env, str], cfg=None, device: Union[torch.device, str] = 'cpu', n_rollout_threads: int = 1, model_dict: Dict[str, Any] = {'model': <class 'openrl.modules.networks.MAT_network.MultiAgentTransformer'>})[source]¶: Bases: openrl.modules.common.ppo_net.PPONet

openrl.modules.common.ppo_net module¶

class openrl.modules.common.ppo_net.PPONet(env: Union[openrl.envs.vec_env.base_venv.BaseVecEnv, gymnasium.core.Env, str], cfg=None, device: Union[torch.device, str] = 'cpu', n_rollout_threads: int = 1, model_dict: Optional[Dict[str, Any]] = None, module_class: openrl.modules.base_module.BaseModule = <class 'openrl.modules.ppo_module.PPOModule'>)[source]¶

Bases: openrl.modules.common.base_net.BaseNet

act(observation: Union[numpy.ndarray, Dict[str, numpy.ndarray]], action_masks: Optional[numpy.ndarray] = None, deterministic: bool = False, episode_starts: Optional[numpy.ndarray] = None) → Tuple[numpy.ndarray, Optional[Tuple[numpy.ndarray, ...]]][source]¶

load_policy(path: str) → None[source]¶

reset(env: Optional[gymnasium.core.Env] = None) → None[source]¶

openrl.modules.common.ppo_net.reset_rnn_states(rnn_states, episode_starts, env_num, agent_num, rnn_layers, hidden_size)[source]¶

openrl.modules.common.sac_net module¶

class openrl.modules.common.sac_net.SACNet(env: Union[gymnasium.core.Env, str], cfg=None, device: Union[torch.device, str] = 'cpu', n_rollout_threads: int = 1, model_dict: Optional[Dict[str, Any]] = None, module_class: openrl.modules.base_module.BaseModule = <class 'openrl.modules.sac_module.SACModule'>)[source]¶

Bases: openrl.modules.common.base_net.BaseNet

act(observation: Union[numpy.ndarray, Dict[str, numpy.ndarray]], deterministic=True) → Tuple[numpy.ndarray, Optional[Tuple[numpy.ndarray, ...]]][source]¶

load_policy(path: str) → None[source]¶

reset(env: Optional[gymnasium.core.Env] = None) → None[source]¶

openrl.modules.common.vdn_net module¶

class openrl.modules.common.vdn_net.VDNNet(env: Union[gymnasium.core.Env, str], cfg=None, device: Union[torch.device, str] = 'cpu', n_rollout_threads: int = 1, model_dict: Optional[Dict[str, Any]] = None)[source]¶

Bases: openrl.modules.common.base_net.BaseNet

act(observation: Union[numpy.ndarray, Dict[str, numpy.ndarray]]) → Tuple[numpy.ndarray, Optional[Tuple[numpy.ndarray, ...]]][source]¶

load_policy(path: str) → None[source]¶

reset(env: Optional[gymnasium.core.Env] = None) → None[source]¶

Module contents¶

class openrl.modules.common.A2CNet(env: Union[openrl.envs.vec_env.base_venv.BaseVecEnv, gymnasium.core.Env, str], cfg=None, device: Union[torch.device, str] = 'cpu', n_rollout_threads: int = 1, model_dict: Optional[Dict[str, Any]] = None, module_class: openrl.modules.base_module.BaseModule = <class 'openrl.modules.ppo_module.PPOModule'>)[source]¶: Bases: openrl.modules.common.ppo_net.PPONet

class openrl.modules.common.BCNet(env: openrl.envs.vec_env.wrappers.reward_wrapper.RewardWrapper, cfg=None, device: Union[torch.device, str] = 'cpu', n_rollout_threads: int = 1, model_dict: Optional[Dict[str, Any]] = None, module_class: abc.ABCMeta = <class 'openrl.modules.bc_module.BCModule'>)[source]¶: Bases: openrl.modules.common.ppo_net.PPONet

class openrl.modules.common.BaseNet[source]¶: Bases: abc.ABC

class openrl.modules.common.DDPGNet(env: Union[gymnasium.core.Env, str], cfg=None, device: Union[torch.device, str] = 'cpu', n_rollout_threads: int = 1, model_dict: Optional[Dict[str, Any]] = None)[source]¶

Bases: openrl.modules.common.base_net.BaseNet

act(observation: Union[numpy.ndarray, Dict[str, numpy.ndarray]], deterministic: bool) → Tuple[numpy.ndarray, Optional[Tuple[numpy.ndarray, ...]]][source]¶

load_policy(path: str) → None[source]¶

reset(env: Optional[gymnasium.core.Env] = None) → None[source]¶

class openrl.modules.common.DQNNet(env: Union[gymnasium.core.Env, str], cfg=None, device: Union[torch.device, str] = 'cpu', n_rollout_threads: int = 1, model_dict: Optional[Dict[str, Any]] = None)[source]¶

Bases: openrl.modules.common.base_net.BaseNet

act(observation: Union[numpy.ndarray, Dict[str, numpy.ndarray]]) → Tuple[numpy.ndarray, Optional[Tuple[numpy.ndarray, ...]]][source]¶

load_policy(path: str) → None[source]¶

reset(env: Optional[gymnasium.core.Env] = None) → None[source]¶

class openrl.modules.common.GAILNet(env: openrl.envs.vec_env.wrappers.reward_wrapper.RewardWrapper, cfg=None, device: Union[torch.device, str] = 'cpu', n_rollout_threads: int = 1, model_dict: Optional[Dict[str, Any]] = None, module_class: abc.ABCMeta = <class 'openrl.modules.gail_module.GAILModule'>)[source]¶: Bases: openrl.modules.common.ppo_net.PPONet

class openrl.modules.common.MATNet(env: Union[gymnasium.core.Env, str], cfg=None, device: Union[torch.device, str] = 'cpu', n_rollout_threads: int = 1, model_dict: Dict[str, Any] = {'model': <class 'openrl.modules.networks.MAT_network.MultiAgentTransformer'>})[source]¶: Bases: openrl.modules.common.ppo_net.PPONet

class openrl.modules.common.PPONet(env: Union[openrl.envs.vec_env.base_venv.BaseVecEnv, gymnasium.core.Env, str], cfg=None, device: Union[torch.device, str] = 'cpu', n_rollout_threads: int = 1, model_dict: Optional[Dict[str, Any]] = None, module_class: openrl.modules.base_module.BaseModule = <class 'openrl.modules.ppo_module.PPOModule'>)[source]¶

Bases: openrl.modules.common.base_net.BaseNet

act(observation: Union[numpy.ndarray, Dict[str, numpy.ndarray]], action_masks: Optional[numpy.ndarray] = None, deterministic: bool = False, episode_starts: Optional[numpy.ndarray] = None) → Tuple[numpy.ndarray, Optional[Tuple[numpy.ndarray, ...]]][source]¶

load_policy(path: str) → None[source]¶

reset(env: Optional[gymnasium.core.Env] = None) → None[source]¶

class openrl.modules.common.SACNet(env: Union[gymnasium.core.Env, str], cfg=None, device: Union[torch.device, str] = 'cpu', n_rollout_threads: int = 1, model_dict: Optional[Dict[str, Any]] = None, module_class: openrl.modules.base_module.BaseModule = <class 'openrl.modules.sac_module.SACModule'>)[source]¶

Bases: openrl.modules.common.base_net.BaseNet

act(observation: Union[numpy.ndarray, Dict[str, numpy.ndarray]], deterministic=True) → Tuple[numpy.ndarray, Optional[Tuple[numpy.ndarray, ...]]][source]¶

load_policy(path: str) → None[source]¶

reset(env: Optional[gymnasium.core.Env] = None) → None[source]¶

class openrl.modules.common.VDNNet(env: Union[gymnasium.core.Env, str], cfg=None, device: Union[torch.device, str] = 'cpu', n_rollout_threads: int = 1, model_dict: Optional[Dict[str, Any]] = None)[source]¶

Bases: openrl.modules.common.base_net.BaseNet

act(observation: Union[numpy.ndarray, Dict[str, numpy.ndarray]]) → Tuple[numpy.ndarray, Optional[Tuple[numpy.ndarray, ...]]][source]¶

load_policy(path: str) → None[source]¶

reset(env: Optional[gymnasium.core.Env] = None) → None[source]¶