Expand description
A single Pre-LN residual layer wrapping one SSM block (Layer).
Structs§
- Layer
- A single Pre-LN block wrapper computing
M(RMSNorm(x))— the residual is not applied here. The enclosingLayersowns that decision (add the input back, suppress it on the first/last layer, or thread it through Multi-Gate streams), so no input clone / zero-add is wasted when no residual is wanted. - Layer
Record - The record type for the module.
- Layer
Record Item - The record item type for the module.