Enum DflashQuantization 

pub enum DflashQuantization {
    Bf16,
    Fp8Weights,
}

2026-09-25: Drafter weight precision. from_weights builds the head as Bf16 and switches it to Fp8Weights unless METRALE_DFLASH_DRAFTER_FP8=0 or the fp8_gemm_n128_row_scaled kernels are missing.

Variants

Bf16

Fp8Weights

2026-09-25: Weight-only FP8: the q/k/v/o/gate/up/down weights get FP8 E4M3 copies with per-row f32 scales at load, and the LM head is FP8 (lm_head_shared_fp8). The KV cache stays bf16.

Trait Implementations

impl Clone for DflashQuantization

fn clone(&self) -> DflashQuantization

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for DflashQuantization

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

Formats the value using the given formatter.

impl PartialEq for DflashQuantization

fn eq(&self, other: &DflashQuantization) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Copy for DflashQuantization

impl Eq for DflashQuantization

impl StructuralPartialEq for DflashQuantization