Struct OnPolicyAlgorithm 

pub struct OnPolicyAlgorithm<A: Agent, S: Sampler, H: OnPolicyAlgorithmHooks<A = A, S = S>> {
    pub runtime: OnPolicyRuntime<A, S>,
    pub hooks: H,
}

Generic on-policy training loop combining a runtime with lifecycle hooks.

Fields

runtime: OnPolicyRuntime<A, S>

Coupled training runtime.

hooks: H

Lifecycle hooks.

Implementations

impl<A: Agent, S: Sampler, H: OnPolicyAlgorithmHooks<A = A, S = S>> OnPolicyAlgorithm<A, S, H>

pub fn new(runtime: OnPolicyRuntime<A, S>, hooks: H) -> Self

Creates an on-policy algorithm from its runtime and lifecycle hooks.

pub fn train(&mut self) -> Result<(), Error>

Runs training until a hook requests termination.

Errors

Returns an error if learning fails or a hook reports a deferred failure during end-of-training finalization.