Module qwen3_ssm 

2026-09-25: Qwen3SsmLayer, the Gated DeltaNet (linear-attention) layer, and the submodules its constructor, forward paths and tests live in.

Owner: model-layers (qwen3_ssm). Invariants: none beyond the types.

Re-exports

pub use gdn_flags::GdnFlags;

pub use gdn_flags::MAX_F16_TWIN_DFLASH_GAMMA;

pub use gdn_flags::MAX_F16_TWIN_K;

pub use gdn_flags::UNRESOLVED_DFLASH_GAMMA;

pub use gdn_flags::default_dflash_gamma;

pub use gdn_flags::gdn_fused_norm_enabled;

pub use gdn_flags::resolve_dflash_gamma;

pub use gdn_flags::ssm_batched_recurrent_enabled;

pub use gdn_flags::ssm_h_dtype_bits;

pub use gdn_flags::ssm_h_f16_pool_enabled;

pub use gdn_flags::ssm_h_fp16_enabled;

pub use gdn_flags::verify_exact_enabled;

Modules

gdn_flags

2026-09-25: GDN / SSM decode-path flags, resolved once per process from the serve command line or, failing that, the environment.

Structs

Qwen3SsmLayer

2026-09-25: The Gated DeltaNet layer. Its QKVZ projection comes in two layouts: interleaved per key-head group, which deinterleave_qkvz (or the fused w4a16_gemv_qkvz) turns into [Q | K | V | Z], and sequential (sequential_qkvz), already in that order.

Functions

carry_flush_kernel

2026-10-01: The fold of pending rows (gdn_carry_flush), or under the exact verify the model directory’s twin (gdn_exact_carry_flush), which every directory with the exact twins defines: one whose strided decode clamps the state after each update folds with the same clamp. Shared by the layer’s fold and the model’s (model-engine gdn_carry.rs).