$100 to try Ray on Anyscale — Start now.

ray.rllib.algorithms.algorithm_config.AlgorithmConfig.get_evaluation_config_object#

AlgorithmConfig.get_evaluation_config_object() → AlgorithmConfig | None[source]#

Creates a full AlgorithmConfig object from self.evaluation_config.

Returns:: A fully valid AlgorithmConfig object that can be used for the evaluation EnvRunnerGroup. If self is already an evaluation config object, return None.