Struct OptimizerMonitor 

pub struct OptimizerMonitor { /* private fields */ }

Main optimizer monitor that tracks various metrics.

Implementations

impl OptimizerMonitor

pub fn new(config: MonitoringConfig) -> Self

Create a new optimizer monitor.

pub fn with_defaults() -> Self

Create a monitor with default configuration.

pub fn before_step(&mut self)

Called before an optimizer step to start timing.

pub fn after_step( &mut self, learning_rate: f32, parameters: &[Tensor], loss: Option<f32>, ) -> Result<()>

Called after an optimizer step to record metrics.

pub fn update_loss(&mut self, loss: f32)

Update loss value (can be called independently of step).

pub fn get_metrics(&self) -> &OptimizerMetrics

Get current metrics.

pub fn should_log(&self) -> bool

Check if we should log detailed metrics this step.

pub fn get_summary_report(&self) -> String

Get a summary report of current optimizer status.

pub fn get_convergence_report(&self) -> String

Get convergence status report.

pub fn reset(&mut self)

Reset monitoring state.

Trait Implementations

impl Debug for OptimizerMonitor

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

Formats the value using the given formatter.