Enum FusedOp 

pub enum FusedOp {
    ScaleBias {
        scale: f64,
        bias: f64,
    },
    ScaleReluBias {
        scale: f64,
        bias: f64,
    },
    ClampNormalize {
        min: f64,
        max: f64,
    },
    Identity,
    Passthrough(TensorOp),
}

A fused (compound) operation produced by TensorOpFusion::fuse.

Variants

ScaleBias

Scale followed immediately by Bias: y = x * scale + bias.

Fields

scale: f64

bias: f64

ScaleReluBias

Scale, then ReLU, then Bias: y = max(0, x * scale) + bias.

Fields

scale: f64

bias: f64

ClampNormalize

Clamp followed by L2-normalisation.

Fields

min: f64

max: f64

Identity

No-op — used when the input sequence was empty.

Passthrough(TensorOp)

A single operation that could not be fused with its neighbours.

Trait Implementations

impl Clone for FusedOp

fn clone(&self) -> FusedOp

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for FusedOp

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

Formats the value using the given formatter.

impl PartialEq for FusedOp

fn eq(&self, other: &FusedOp) -> 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 StructuralPartialEq for FusedOp