Enum NumericImpact 

pub enum NumericImpact {
    TrainingLossNaN,
    GradientPoison,
    InferenceOutputRisk,
    LocalOnly,
}

How a numeric hazard can interfere with training or inference.

Variants

TrainingLossNaN

Hazard is on a loss sink (or reaches one): training can abort with NaN loss.

GradientPoison

Hazard lies on a trainable → loss path: gradients can be poisoned.

InferenceOutputRisk

Hazard reaches the entry return of a non-loss forward.

LocalOnly

Unstable value proven locally but not shown to reach loss or return.

Implementations

impl NumericImpact

pub fn label(self) -> &'static str

pub fn is_training_failure(self) -> bool

Trait Implementations

impl Clone for NumericImpact

fn clone(&self) -> NumericImpact

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for NumericImpact

impl Debug for NumericImpact

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

Formats the value using the given formatter.

impl Default for NumericImpact

fn default() -> NumericImpact

Returns the “default value” for a type.

impl Eq for NumericImpact

impl PartialEq for NumericImpact

fn eq(&self, other: &NumericImpact) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl Serialize for NumericImpact

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for NumericImpact