Struct Dflash2Conv 

pub struct Dflash2Conv {
    pub base: CudaSlice<f32>,
    pub proj: GpuTensor,
}

One GroupedDynamicCausalConv module (reference model.py): a causal 2-tap depthwise conv over the BLOCK rows (block-local — row 0 zero-pads its missing predecessor; stateless across rounds), with per-position dynamic per-group coefficients projected from the module INPUT. prepare convolves the sublayer input with base_kernel[0] + dyn half 0; finish convolves the sublayer OUTPUT with base_kernel[1] + dyn half 1 (both dyn halves come from the SAME projection of the pre-conv input).

Fields

base: CudaSlice<f32>

base_kernel [2, k, hidden] flattened f32 (half-major: prepare then finish).

proj: GpuTensor

kernel_projection.weight [2kgroups, hidden] (row layout = view(2, k, groups)).