Struct QuantizedMatrix 

pub struct QuantizedMatrix {
    pub data: QuantizedData2D,
    pub shape: (usize, usize),
    pub data_type: QuantizedDataType,
}

A matrix with quantized values

Fields

data: QuantizedData2D

The quantized values can be stored in different formats

shape: (usize, usize)

The original shape of the matrix

data_type: QuantizedDataType

The data type used for quantization

Implementations

impl QuantizedMatrix

pub fn new_i8( data: Array2<i8>, shape: (usize, usize), data_type: QuantizedDataType, ) -> Self

Creates a new quantized matrix with int8 storage

pub fn new_f16(data: Array2<f16>, shape: (usize, usize)) -> Self

Creates a new f16 quantized matrix

pub fn new_bf16(data: Array2<bf16>, shape: (usize, usize)) -> Self

Creates a new bf16 quantized matrix

pub fn from_i8(data: Array2<i8>, shape: (usize, usize)) -> Self

Creates a standard Int8 quantized matrix (for backward compatibility)

pub fn get(&self, row: usize, col: usize) -> i8

👎Deprecated since 0.1.0:

Use get_i8 or get_f32 instead

pub fn shape(&self) -> (usize, usize)

Returns the shape of the matrix

pub fn nrows(&self) -> usize

Returns the number of rows in the matrix

pub fn ncols(&self) -> usize

Returns the number of columns in the matrix

pub fn get_i8(&self, row: usize, col: usize) -> i8

Get value at specified position as i8 (for int quantization)

pub fn get_f32(&self, row: usize, col: usize) -> f32

Get value at specified position as f32 (for all quantization types)

Trait Implementations

impl Clone for QuantizedMatrix

fn clone(&self) -> QuantizedMatrix

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for QuantizedMatrix

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

Formats the value using the given formatter.