caozuobaandRoger Wang 9e19f8338b [Perf] add packed recurrent fast path for decode (#36596) ...

Signed-off-by: hdj <1293066020@qq.com>
Co-authored-by: Roger Wang <hey@rogerw.io>

2026-03-12 04:01:57 -07:00

..

__init__.py

[Perf] add packed recurrent fast path for decode (#36596)

2026-03-12 04:01:57 -07:00

chunk_delta_h.py

[PERF] Change GDN Attention State Layout from [N, HV, K, V] to [N, HV, V, K] (#33291)

2026-02-04 11:20:52 +00:00

chunk_o.py

[PERF] Change GDN Attention State Layout from [N, HV, K, V] to [N, HV, V, K] (#33291)

2026-02-04 11:20:52 +00:00

chunk_scaled_dot_kkt.py

[FLA] Introduce Kimi Delta Attention(KDA) to VLLM (#27654)

2025-10-28 22:56:28 +08:00

chunk.py

[Refactor] Deprecate head_first for chunk_gated_delta_rule (#34263)

2026-02-19 13:23:49 -05:00

cumsum.py

Update Optional[x] -> x | None and Union[x, y] to x | y (#26633)

2025-10-12 09:51:31 -07:00

fused_recurrent.py

[Perf] add packed recurrent fast path for decode (#36596)

2026-03-12 04:01:57 -07:00

fused_sigmoid_gating.py

[Kernel] Add fused_sigmoid_gating_delta_rule_update kernel for Qwen3 Next (#35777)

2026-03-08 23:41:01 -07:00

index.py

Convert formatting to use ruff instead of yapf + isort (#26247)

2025-10-05 07:06:22 -07:00

kda.py

[CustomOp] CustomOp FusedRMSNormGated (#35877)

2026-03-06 10:53:37 -08:00

l2norm.py

[Model] Add support for OLMo Hybrid (#32550)

2026-03-05 14:51:06 -05:00

layernorm_guard.py

[Model] Add support for OLMo Hybrid (#32550)

2026-03-05 14:51:06 -05:00

op.py

[Kernel] Accelerate solve_tril with TMA (#26746)

2025-10-20 05:39:02 +00:00

solve_tril.py

[Kernel] Accelerate solve_tril with TMA (#26746)

2025-10-20 05:39:02 +00:00

utils.py

replace with torch.cuda.device with with torch.accelerator.device_index (#36144)

2026-03-11 23:12:57 -07:00

wy_fast.py

Update Optional[x] -> x | None and Union[x, y] to x | y (#26633)

2025-10-12 09:51:31 -07:00