Struct DenseFfnWeightsBf16 

pub struct DenseFfnWeightsBf16 {
    pub gate_proj: DenseWeight,
    pub up_proj: DenseWeight,
    pub down_proj: DenseWeight,
}

BF16 dense MLP weights — alternative to NVFP4 for precision-sensitive models (Gemma-4-31B). Each is [N, K] row-major BF16. When installed on a DenseFfnLayer via set_bf16_weights, the forward paths dispatch to dense_gemv_bf16 / dense_gemm_bf16 instead of the w4a16 NVFP4 kernels. Costs ~3.4 GB extra GPU memory on Gemma-4-31B (3 × hidden×intermediate × 2 bytes) vs NVFP4’s 0.5 bytes/weight.

Fields

gate_proj: DenseWeight

up_proj: DenseWeight

down_proj: DenseWeight