Enum QuantizationMode 

pub enum QuantizationMode {
    Int8Symmetric,
    Int8Asymmetric,
    Int4,
    Fp16,
    Bf16,
}

Precision target for quantization.

Variants

Int8Symmetric

Symmetric INT8: scale = percentile(|x|) / 127.0; range [-127, 127].

Int8Asymmetric

Asymmetric INT8: separate zero-point; range [0, 255].

Int4

4-bit: scale = percentile(|x|) / 7.0; range [-7, 7] stored as i8.

Fp16

FP16 simulation: round to nearest f16 (5-bit exp, 10-bit mantissa).

Bf16

BFloat16 simulation: keep top 16 bits of the f32 representation.

Implementations

impl QuantizationMode

pub fn name(&self) -> &'static str

Human-readable name for statistics reporting.

pub fn bits_per_element(&self) -> f64

Nominal storage bits per element (for compression ratio).

Trait Implementations

impl Clone for QuantizationMode

fn clone(&self) -> QuantizationMode

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for QuantizationMode

impl Debug for QuantizationMode

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

Formats the value using the given formatter.

impl Eq for QuantizationMode

impl Hash for QuantizationMode

fn hash<__H: Hasher>(&self, state: &mut __H)

Feeds this value into the given Hasher.

fn hash_slice<H>(data: &[Self], state: &mut H)

where H: Hasher, Self: Sized,

Feeds a slice of this type into the given Hasher.

impl PartialEq for QuantizationMode

fn eq(&self, other: &QuantizationMode) -> 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 QuantizationMode