Struct ValidationSettings

pub struct ValidationSettings {
    pub enabled: bool,
    pub validation_split: f64,
    pub batch_size: usize,
    pub num_workers: usize,
}

Configuration for validation during training

Validation helps monitor model performance on unseen data during training, which is crucial for detecting overfitting and ensuring good generalization.

Examples

use scirs2_neural::training::ValidationSettings;

let validation = ValidationSettings {
    enabled: true,
    validation_split: 0.2,  // Use 20% of data for validation
    batch_size: 64,
    num_workers: 2,
};

Fields

enabled: bool

Whether to enable validation during training

When disabled, no validation will be performed even if this struct is provided.

validation_split: f64

Fraction of training data to use for validation (0.0 to 1.0)

For example, 0.2 means 20% of the data will be used for validation and 80% for training. The data is split before training begins.

batch_size: usize

Batch size for validation

Can be larger than training batch size since no gradients are computed. Larger validation batches are more memory efficient and faster.

num_workers: usize

Number of parallel workers for validation data loading

Similar to training workers, but for validation data. Setting to 0 uses the main thread.

Trait Implementations

impl Clone for ValidationSettings

fn clone(&self) -> ValidationSettings

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ValidationSettings

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

Formats the value using the given formatter.

impl Default for ValidationSettings

fn default() -> Self

Returns the “default value” for a type.