Struct QuantParams 

pub struct QuantParams {
    pub scale: f32,
    pub zero_point: i8,
    pub anti_resonance: f32,
    pub bits: u8,
}

Quantization parameters for a tensor or channel

Fields

scale: f32

Scale factor (float = quant * scale + zero_point)

zero_point: i8

Zero point offset (for asymmetric quantization)

anti_resonance: f32

Anti-resonance offset from π

bits: u8

Quantization bits (7 for signed int8)

Implementations

impl QuantParams

pub fn symmetric(min_val: f32, max_val: f32) -> Self

Create symmetric quantization params (for weights)

Uses π-based anti-resonance to avoid boundary collapse.

pub fn asymmetric(min_val: f32, max_val: f32) -> Self

Create asymmetric quantization params (for activations)

Maps [min_val, max_val] to [-128, 127] with π-based calibration.

pub fn quantize(&self, value: f32) -> i8

Quantize a single f32 value to i8

pub fn dequantize(&self, quantized: i8) -> f32

Dequantize a single i8 value to f32

Trait Implementations

impl Clone for QuantParams

fn clone(&self) -> QuantParams

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for QuantParams

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

Formats the value using the given formatter.

impl Default for QuantParams

fn default() -> Self

Returns the “default value” for a type.

impl Copy for QuantParams