compute_standard_loss

compute_standard_loss(*, double_dqn: bool, gamma: float, loss_function: str, learner_model_container: ThreadSafeModel, target_model: DQNModel, apply_action_mask: Callable[[torch.Tensor], torch.Tensor], batch: LearnerBatch) -> LossResult

Compute the standard DQN loss.

Returns:

• LossResult – The computed loss and diagnostics.

Source