src.model.reinforcement.RL_actorCritic module

class src.model.reinforcement.RL_actorCritic.RL_AC_manager(*args, **kwargs)

Bases: RLModelHandler

aggregate(agg, x)

calculate_confusion_matrix(labels: Tensor, exp_l: Tensor | None = None) → Dict[str, int]

calculate_reward(labels: Tensor, current_params: Tensor | None = None, previous_params: Tensor | None = None) → Tensor

discounted_sum(x: Tensor) → Tensor

evaluate_performances(*args, **kwargs) → None

execute_train(*args, **kwargs) → None

get_margin_values(*args, **kwargs) → Tensor

get_probabilities_values(*args, **kwargs) → Tensor

register(stat: str, value: Any, step: int | None = None) → None

reset() → None

step(observation: Tensor, labels: Tensor, game_over: bool) → int | None

train(*args, **kwargs) → None

update_memory() → None