Struct QuantizedTensor 

pub struct QuantizedTensor {
    pub data: Vec<i8>,
    pub shape: Vec<usize>,
    pub scale: f32,
    pub zero_point: i8,
}

Quantized tensor representation: INT8 values + per-tensor scale + zero-point.

Fields

data: Vec<i8>

shape: Vec<usize>

scale: f32

zero_point: i8

Implementations

impl QuantizedTensor

pub fn from_tensor(tensor: &Tensor, mode: QuantMode) -> Self

Quantize an f32 tensor to INT8.

pub fn to_tensor(&self) -> Result<Tensor, ModelError>

Dequantize back to f32 tensor.

pub fn len(&self) -> usize

Number of elements.

pub fn is_empty(&self) -> bool

Whether empty.

pub fn compression_ratio(&self) -> f32

Compression ratio vs f32 (4x for INT8).

pub fn byte_size(&self) -> usize

Total bytes of quantized data (not including metadata).

Trait Implementations

impl Clone for QuantizedTensor

fn clone(&self) -> QuantizedTensor

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for QuantizedTensor

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

Formats the value using the given formatter.