Struct QuantizedVector 

pub struct QuantizedVector {
    pub data: Vec<u8>,
    pub min: f32,
    pub max: f32,
}

A quantized vector using 8-bit scalar quantization.

Each f32 value is mapped to a u8 (0-255) using min/max scaling. The original value can be reconstructed as: value = (data[i] / 255.0) * (max - min) + min

Fields

data: Vec<u8>

Quantized data (1 byte per dimension instead of 4).

min: f32

Minimum value in the original vector.

max: f32

Maximum value in the original vector.

Implementations

impl QuantizedVector

pub fn from_f32(vector: &[f32]) -> Self

Creates a new quantized vector from f32 data.

Arguments

• vector - The original f32 vector to quantize

Panics

Panics if the vector is empty.

pub fn to_f32(&self) -> Vec<f32>

Reconstructs the original f32 vector from quantized data.

Note: This is a lossy operation. The reconstructed values are approximations.

pub fn dimension(&self) -> usize

Returns the dimension of the vector.

pub fn memory_size(&self) -> usize

Returns the memory size in bytes.

pub fn to_bytes(&self) -> Vec<u8>

Serializes the quantized vector to bytes.

pub fn from_bytes(bytes: &[u8]) -> Result<Self>

Deserializes a quantized vector from bytes.

Errors

Returns an error if the bytes are invalid.

Trait Implementations

impl Clone for QuantizedVector

fn clone(&self) -> QuantizedVector

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for QuantizedVector

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

Formats the value using the given formatter.