Struct QuantizationConfig 

pub struct QuantizationConfig {
    pub mode: QuantizationMode,
    pub bit_width: BitWidth,
    pub granularity: Granularity,
    pub eps: f32,
}

Configuration for quantization.

Fields

mode: QuantizationMode

Quantization mode (symmetric or asymmetric)

bit_width: BitWidth

Bit-width for quantization

granularity: Granularity

Granularity (per-tensor or per-channel)

eps: f32

Small epsilon to avoid division by zero

Implementations

impl QuantizationConfig

pub fn int8_symmetric() -> Self

Creates a default int8 symmetric per-tensor configuration.

pub fn int8_asymmetric() -> Self

Creates a default int8 asymmetric per-tensor configuration.

pub fn int4_per_channel() -> Self

Creates a default int4 symmetric per-channel configuration.

pub fn new( mode: QuantizationMode, bit_width: BitWidth, granularity: Granularity, ) -> Self

Creates a custom configuration.

Trait Implementations

impl Clone for QuantizationConfig

fn clone(&self) -> QuantizationConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for QuantizationConfig

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for QuantizationConfig

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for QuantizationConfig

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

where __S: Serializer,

Serialize this value into the given Serde serializer.