ray.rllib.core.learner.learner.Learner.additional_update_for_module#

Learner.additional_update_for_module(*, module_id: str, config: AlgorithmConfig | None = None, timestep: int, **kwargs) None[source]#

Apply additional non-gradient based updates for a single module.

See additional_update for more details.

Parameters:
  • module_id – The id of the module to update.

  • config – The AlgorithmConfig specific to the given module_id.

  • timestep – The current global timestep (to be used with schedulers).

  • **kwargs – Keyword arguments to use for the additional update.

Returns:

A dictionary of results from the update