Struct QuantizationResult 

pub struct QuantizationResult {
    pub values: Vec<i8>,
    pub scales: Vec<f32>,
    pub shape: (usize, usize),
    pub group_size: usize,
}

Result of quantization operation.

Fields

values: Vec<i8>

Quantized ternary values as i8 (-1, 0, +1).

scales: Vec<f32>

Scale factors (one per group or per row).

shape: (usize, usize)

Original tensor shape.

group_size: usize

Group size used.

Implementations

impl QuantizationResult

pub fn to_packed(&self) -> Result<PackedTernary, TernaryError>

Convert to PackedTernary for efficient storage and matmul.

pub fn to_tensor(&self, device: &Device) -> Result<Tensor, QuantizationError>

Get values as f32 tensor with scales applied.

Trait Implementations

impl Clone for QuantizationResult

fn clone(&self) -> QuantizationResult

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for QuantizationResult

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

Formats the value using the given formatter.