Enum Activation 

pub enum Activation {
    ReLU,
    Tanh,
    Sigmoid,
    Identity,
    Gelu,
    Swish,
}

Activation function types supported by the inference engine

Serialized as the enum variant name by default. Gelu and Swish use the custom serde names "GELU" and "Swish" for compatibility with the upstream JSON wire format produced by the training-side exporter.

Variants

ReLU

Rectified linear unit, max(x, 0). Most common hidden activation.

Tanh

Hyperbolic tangent, x.tanh(). Bounded output in (-1, 1).

Sigmoid

Logistic sigmoid, 1 / (1 + exp(-x)). Bounded output in (0, 1).

Identity

No-op pass-through; useful for explicit “linear” output heads.

Gelu

Approximate Gaussian error linear unit (tanh approximation, see Activation::apply). Serialized as "GELU".

Swish

x * sigmoid(x) (also known as SiLU). Serialized as "Swish".

Implementations

impl Activation

pub fn apply(&self, x: f32) -> f32

Apply activation function to a single value

pub fn apply_vec(&self, vec: &mut [f32])

Apply activation function to a vector in-place

Trait Implementations

impl Clone for Activation

fn clone(&self) -> Activation

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for Activation

impl Debug for Activation

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for Activation

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for Activation

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

Equality operator ==.

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

Inequality operator !=.

impl Serialize for Activation

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 Activation