DistributedReinforcementLearning.Trainer
— TypeTrainer(;policy, sealer=deepcopy)
A wrapper around AbstractPolicy
, the sealer
is used to create an immutable object from the inner parameters of policy
when received a FetchParamMsg
.
DistributedReinforcementLearning.Worker
— TypeWorker(()->ex::Experiment)
DistributedReinforcementLearning.actor
— Methodactor(f;sz=DEFAULT_MAILBOX_SIZE)
Create a task to handle messages one-by-one by calling f(msg)
. A mailbox (RemoteChannel
) is returned.
DistributedReinforcementLearning.self
— Methodself()
Get the mailbox in current task.