Struct QuantizedEmbedding 

pub struct QuantizedEmbedding {
    pub original_dim: usize,
    pub quantized_data: Vec<u8>,
    pub scale: f32,
    pub zero_point: f32,
}

A single embedding quantized to 8-bit precision via min-max scaling.

Fields

original_dim: usize

Original embedding dimensionality.

quantized_data: Vec<u8>

Quantized values as u8 in [0, 255].

scale: f32

Scale factor: (max - min) / 255.

zero_point: f32

Zero point (minimum value of original embedding).

Implementations

impl QuantizedEmbedding

pub fn quantize(embedding: &[f32]) -> Self

Quantize a floating-point embedding to 8-bit using min-max scaling.

Formula: v_q = round((v - min) / (max - min) * 255)

pub fn dequantize(&self) -> Vec<f32>

Dequantize back to f32.

Formula: v = v_q * scale + zero_point

pub fn approx_size_bytes(&self) -> usize

Approximate storage size in bytes (header overhead not included).

Trait Implementations

impl Clone for QuantizedEmbedding

fn clone(&self) -> QuantizedEmbedding

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for QuantizedEmbedding

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

Formats the value using the given formatter.