Struct EnhancedAdaptiveLRController 

pub struct EnhancedAdaptiveLRController<A: Float + Send + Sync> { /* private fields */ }

Enhanced adaptive learning rate controller with multiple adaptation mechanisms

Implementations

impl<A: Float + Default + Clone + Sum + Send + Sync> EnhancedAdaptiveLRController<A>

pub fn new(config: AdaptiveLRConfig<A>) -> Result<Self>

Create a new enhanced adaptive learning rate controller

pub fn record_memory_usage_mb(&mut self, usage_mb: f64)

Report the memory this optimizer’s workload is currently using, so the resource signal has a real figure to work from (E4).

pub fn record_energy_sample(&mut self, joules: f64)

Report measured energy consumption for the most recent step.

pub fn record_throughput(&mut self, samples_per_second: A)

Report the observed sample throughput so the resource signal can compare it against the configured requirement.

pub fn add_source_task(&mut self, task: TaskData<A>)

Register a previously solved task so the meta-optimizer can transfer its learning-rate schedule (it contributes nothing while none are known).

pub fn update_learning_rate( &mut self, gradients: &Array1<A>, loss: A, metrics: &HashMap<String, A>, step: usize, ) -> Result<A>

Update learning rate based on multiple adaptation signals

pub fn previous_lr(&self) -> Option<A>

Learning rate before the most recent adaptation, if there was one.

pub fn last_decision(&self) -> Option<&AdaptationDecision<A>>

The most recent adaptation decision, including its rationale.

pub fn voting_history(&self) -> &VecDeque<SignalVote<A>>

Recorded signal votes, newest last.

pub fn get_current_lr(&self) -> A

Get current learning rate

pub fn get_adaptation_statistics(&self) -> AdaptationStatistics<A>

Get adaptation statistics

pub fn evaluate_adaptation_effectiveness(&mut self, performance_improvement: A)

Evaluate adaptation effectiveness retrospectively

pub fn meta_arm_rewards(&self) -> &HashMap<usize, A>

Measured mean reward per meta-learning arm.

pub fn meta_arm_counts(&self) -> &HashMap<usize, usize>

Times each meta-learning arm has been played.

pub fn signal_reliability(&self) -> &HashMap<AdaptationSignalType, A>

Measured reliability of each adaptation signal.

pub fn reset(&mut self)

Reset controller state

Trait Implementations

impl<A: Clone + Float + Send + Sync> Clone for EnhancedAdaptiveLRController<A>

fn clone(&self) -> EnhancedAdaptiveLRController<A>

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<A: Debug + Float + Send + Sync> Debug for EnhancedAdaptiveLRController<A>

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.