quantized_auto

metrale_model_layers::weight_map::nvfp4_detect

Function quantized_auto 

Source 
pub fn quantized_auto(
    store: &WeightStore,
    prefix: &str,
    gpu: &dyn GpuBackend,
    variant: Nvfp4Variant,
) -> Result<QuantizedWeight>
Expand description

2026-09-25: Load a pre-quantized NVFP4 weight by variant: quantized for Standard, quantized_v2 for CompressedTensors. Panics for Fp8Dequanted and Bf16Raw, which need quantized_any.