Enum Nvfp4Variant 

pub enum Nvfp4Variant {
    Standard,
    CompressedTensors,
    Fp8Dequanted,
    Bf16Raw,
}

2026-09-25: On-disk weight format, chosen by detect_nvfp4_variant.

Variants

Standard

2026-09-25: ModelOpt NVFP4: weight, weight_scale, weight_scale_2, optional input_scale (read by quantized).

CompressedTensors

2026-09-25: compressed-tensors NVFP4: weight_packed, weight_scale, weight_global_scale, optional input_global_scale (read by quantized_v2).

Fp8Dequanted

2026-09-25: Block-scaled FP8 E4M3 (weight plus its block scale). quantized_any turns such a weight into NVFP4 through a BF16 dequant (quantized_from_fp8); loaders that serve FP8 natively read it with load_fp8_block_scaled_as_fp8weight instead.

Bf16Raw

2026-09-25: No usable NVFP4 or FP8 metadata: the BF16 .weight is quantized to NVFP4 at load, and detection logs a warning.

Trait Implementations

impl Clone for Nvfp4Variant

fn clone(&self) -> Nvfp4Variant

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Nvfp4Variant

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

Formats the value using the given formatter.

impl PartialEq for Nvfp4Variant

fn eq(&self, other: &Nvfp4Variant) -> 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 Nvfp4Variant

impl Eq for Nvfp4Variant

impl StructuralPartialEq for Nvfp4Variant