Struct LayerPrecision 

pub struct LayerPrecision {
    pub weight: Option<Operand>,
    pub activation: Option<Operand>,
}

2026-09-28: What one linear layer declares. None means that operand is not quantized (16-bit).

Fields

weight: Option<Operand>

2026-09-28: The weight’s declared format.

activation: Option<Operand>

2026-09-28: The input activations’ declared format.

Implementations

impl LayerPrecision

pub const UNQUANTIZED: LayerPrecision

2026-09-28: 16-bit weights and activations.

pub fn activation_is_fp4(&self) -> bool

2026-09-28: The activations are declared 4-bit float (W4A4 on an FP4 weight).

pub fn activation_is_fp8(&self) -> bool

2026-09-28: The activations are declared 8-bit float.

pub fn label(&self) -> String

2026-09-28: A short label, e.g. W4A4, W8A8, W4A16, W16A16.

Trait Implementations

impl Clone for LayerPrecision

fn clone(&self) -> LayerPrecision

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for LayerPrecision

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

Formats the value using the given formatter.

impl PartialEq for LayerPrecision

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

Tests for self and other values to be equal, and is used by ==.

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

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Copy for LayerPrecision

impl Eq for LayerPrecision

impl StructuralPartialEq for LayerPrecision