Struct TrainingHistory 

pub struct TrainingHistory {
    pub epochs: Vec<EpochMetrics>,
}

Accumulated history of training metrics — returned after fit().

Access via model.history() on any iterative model.

Fields

epochs: Vec<EpochMetrics>

Per-epoch snapshots.

Implementations

impl TrainingHistory

pub fn new() -> Self

Create a new empty history.

pub fn push(&mut self, metrics: EpochMetrics)

Push a new epoch snapshot.

pub fn len(&self) -> usize

Number of recorded epochs.

pub fn is_empty(&self) -> bool

Whether the history is empty.

pub fn train_losses(&self) -> Vec<f64>

Training loss per epoch.

pub fn val_losses(&self) -> Vec<f64>

Validation loss per epoch (only epochs that have it).

pub fn train_metrics(&self) -> Vec<f64>

Training metric (accuracy / R²) per epoch.

pub fn val_metrics(&self) -> Vec<f64>

Validation metric per epoch.

pub fn grad_norms(&self) -> Vec<f64>

Gradient L2 norm per epoch.

pub fn learning_rates(&self) -> Vec<f64>

Learning rate per epoch.

pub fn epoch_times_ms(&self) -> Vec<u64>

Wall-clock milliseconds per epoch.

Trait Implementations

impl Clone for TrainingHistory

fn clone(&self) -> TrainingHistory

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for TrainingHistory

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

Formats the value using the given formatter.

impl Default for TrainingHistory

fn default() -> TrainingHistory

Returns the “default value” for a type.