Enum CompressionStage 

pub enum CompressionStage {
    Prune(PruningConfig),
    QuantizeInt8,
    Clip {
        percentile: f32,
    },
}

A compression stage to apply in sequence.

Variants

Prune(PruningConfig)

Prune weights to target sparsity using the provided configuration.

QuantizeInt8

Quantize to INT8 (per-tensor, simulate INT8 precision loss while keeping f32 storage). Memory footprint is reported as original * 0.25 (theoretical INT8 size).

Clip

Apply magnitude-based weight clipping: zero out weights whose absolute value falls below the given percentile of absolute values in each tensor.

Fields

percentile: f32

Fraction in (0.0, 1.0] — e.g. 0.1 zeros the bottom 10% of weights.

Implementations

impl CompressionStage

pub fn name(&self) -> &'static str

Human-readable name for this stage.

Trait Implementations

impl Clone for CompressionStage

fn clone(&self) -> CompressionStage

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for CompressionStage

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

Formats the value using the given formatter.