Struct QuantizationMetadata 

pub struct QuantizationMetadata {
    pub scheme: QuantizationScheme,
    pub group_size: u32,
    pub scale_bits: u16,
    pub zero_point: i8,
    pub has_per_group_scales: bool,
}

Quantization metadata for HCT v2.

Contains information needed to dequantize INT4/INT8 weights.

Fields

scheme: QuantizationScheme

Quantization scheme used.

group_size: u32

Group size for group-wise quantization (0 = per-tensor).

scale_bits: u16

Global scale factor (f16 stored as u16 bits).

zero_point: i8

Global zero point (for asymmetric quantization).

has_per_group_scales: bool

Whether per-group scales are stored after compressed data.

Implementations

impl QuantizationMetadata

pub const SIZE: usize = 8

Size of quantization metadata in bytes.

pub fn to_bytes(&self) -> [u8; 8]

Serialize to bytes.

pub fn from_bytes(buf: &[u8; 8]) -> Result<Self>

Parse from bytes.

Trait Implementations

impl Clone for QuantizationMetadata

fn clone(&self) -> QuantizationMetadata

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for QuantizationMetadata

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

Formats the value using the given formatter.

impl Default for QuantizationMetadata

fn default() -> QuantizationMetadata

Returns the “default value” for a type.

impl PartialEq for QuantizationMetadata

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