ray.rllib.core.learner.learner_group.LearnerGroup#

class ray.rllib.core.learner.learner_group.LearnerGroup(*, config: AlgorithmConfig = None, module_spec: SingleAgentRLModuleSpec | MultiAgentRLModuleSpec | None = None, max_queue_len: int = 20, learner_spec=None)[source]#

Coordinator of n (possibly remote) Learner workers.

Each Learner worker has a copy of the RLModule, the loss function(s), and one or more optimizers.

PublicAPI (alpha): This API is in alpha and may change before becoming stable.

Methods

`__init__`	Initializes a LearnerGroup instance.
`add_module`	Add a module to the Learners maintained by this LearnerGroup.
`additional_update`	Apply additional non-gradient based updates to the Learners.
`foreach_learner`	Calls the given function on each Learner L with the args: (L, **kwargs).
`get_state`	Get the states of this LearnerGroup.
`get_stats`	Returns the current stats for the input queue for this learner group.
`get_weights`	Get the weights of the MultiAgentRLModule maintained by each Learner.
`load_module_state`	Load the checkpoints of the modules being trained by this LearnerGroup.
`load_state`	Loads the state of the LearnerGroup.
`remove_module`	Remove a module from the Learners maintained by this LearnerGroup.
`save_state`	Saves the state of the LearnerGroup.
`set_state`	Sets the state of this LearnerGroup.
`set_weights`	Set the weights of the MultiAgentRLModule maintained by each Learner.
`shutdown`	Shuts down the LearnerGroup.
`update_from_batch`	Performs gradient based update(s) on the Learner(s), based on given batch.
`update_from_episodes`	Performs gradient based update(s) on the Learner(s), based on given episodes.

Attributes

`is_local`
`is_remote`