StateCriticOutput

class maze.core.agent.state_critic_input_output.StateCriticOutput

Critic output holds the output of a critic for one full flat env step.

Individual CriticStepOutputs are stored in a list which can be (if needed) referenced to the corresponding ActorID.

property actor_ids: List[ActorID]

List of actor IDs for the individual sub-steps.

append(value: StateCriticStepOutput) → None

Set self[key] to value.

property detached_values: List[torch.Tensor]

List of detached values for the individual sub-steps

reshape(shape: Sequence[int]) → None

Reshape all the elements of the critic output to the given shape

property values: List[torch.Tensor]

List of values for the individual sub-steps