contrastive_unsupervised_rl

curl_agent

class xuance.torch.agents.contrastive_unsupervised_rl.curl_agent.CURL_Agent(config: Namespace, envs: DummyVecEnv | SubprocVecEnv, callback: BaseCallback | None = None)

Bases: OffPolicyAgent

class xuance.torch.agents.contrastive_unsupervised_rl.curl_agent.CURL_Policy(*args: Any, **kwargs: Any)

Bases: Module

copy_target()

forward(x: torch.Tensor)

target(x: torch.Tensor)

drq_agent

class xuance.torch.agents.contrastive_unsupervised_rl.drq_agent.DrQ_Agent(config: Namespace, envs: DummyVecEnv | SubprocVecEnv)

Bases: OffPolicyAgent

The implementation of Deep Q-Networks (DQN) agent.

Parameters:

• config – the Namespace variable that provides hyper-parameters and other settings.

• envs – the vectorized environments.

spr_agent

class xuance.torch.agents.contrastive_unsupervised_rl.spr_agent.SPR_Agent(config: Namespace, envs: DummyVecEnv | SubprocVecEnv, callback: BaseCallback | None = None)

Bases: OffPolicyAgent

class xuance.torch.agents.contrastive_unsupervised_rl.spr_agent.SPR_Policy(*args: Any, **kwargs: Any)

Bases: Module

copy_target()

forward(x: torch.Tensor)

target(x: torch.Tensor)