pub struct NemotronSsmWeights {
pub in_proj: QuantizedWeight,
pub out_proj: QuantizedWeight,
pub conv1d_weight: DenseWeight,
pub conv1d_bias: DenseWeight,
pub a_log: DenseWeight,
pub d_param: DenseWeight,
pub dt_bias: DenseWeight,
pub ssm_norm: DenseWeight,
}Expand description
2026-09-25: Nemotron-H Mamba-2 mixer weights.
in_proj produces [z(d_inner), x(d_inner), B(n_groups*state), C(n_groups*state), dt(num_heads)].
Fields§
§in_proj: QuantizedWeight2026-09-25: NVFP4, or QuantizedWeight::null() when the layer is not NVFP4.
out_proj: QuantizedWeight2026-09-25: NVFP4, or QuantizedWeight::null() when the layer is not NVFP4.
conv1d_weight: DenseWeight§conv1d_bias: DenseWeight2026-09-25: F32, widened from the BF16 checkpoint tensor at load.
a_log: DenseWeight2026-09-25: F32, widened from the BF16 checkpoint tensor at load.
d_param: DenseWeight2026-09-25: The D skip connection; F32, widened from BF16 at load.
dt_bias: DenseWeight2026-09-25: F32, widened from the BF16 checkpoint tensor at load.
ssm_norm: DenseWeightAuto Trait Implementations§
impl Freeze for NemotronSsmWeights
impl RefUnwindSafe for NemotronSsmWeights
impl Send for NemotronSsmWeights
impl Sync for NemotronSsmWeights
impl Unpin for NemotronSsmWeights
impl UnwindSafe for NemotronSsmWeights
Blanket Implementations§
Source§impl<T> BorrowMut<T> for Twhere
T: ?Sized,
impl<T> BorrowMut<T> for Twhere
T: ?Sized,
Source§fn borrow_mut(&mut self) -> &mut T
fn borrow_mut(&mut self) -> &mut T
Mutably borrows from an owned value. Read more