bind_kda_weights

metrale_model_arch::glm5next_kda::binding

Function bind_kda_weights 

Source 
pub fn bind_kda_weights(
    gpu: &dyn GpuBackend,
    cfg: &Glm5NextKdaConfig,
    layer_idx: usize,
    src: &dyn KdaTensorSource,
) -> Result<(Glm5NextKdaWeights, KdaBindReport)>
Expand description

2026-09-25: Bind one KDA block: exact tensor set, dtypes, shapes and byte counts, or an error.

Weights are uploaded unchanged: BF16 stays BF16 and F32 stays F32.