[mamba] unify KDA conv states into one cache to match 2-state SSM layout (#44539)

This commit is contained in:
Jiangyun Zhu
2026-06-04 20:38:05 +02:00
committed by GitHub
parent 41a4829f22
commit a55fccfc7c
3 changed files with 16 additions and 30 deletions
+3 -5
View File
@@ -600,7 +600,7 @@ class KimiLinearForCausalLM(
def get_mamba_state_dtype_from_config(
cls,
vllm_config: "VllmConfig",
) -> tuple[torch.dtype, torch.dtype, torch.dtype, torch.dtype]:
) -> tuple[torch.dtype, torch.dtype]:
return MambaStateDtypeCalculator.kda_state_dtype(
vllm_config.model_config.dtype, vllm_config.cache_config.mamba_cache_dtype
)
@@ -608,7 +608,7 @@ class KimiLinearForCausalLM(
@classmethod
def get_mamba_state_shape_from_config(
cls, vllm_config: "VllmConfig"
) -> tuple[tuple[int, ...], tuple[int, ...], tuple[int, ...], tuple[int, ...]]:
) -> tuple[tuple[int, ...], tuple[int, ...]]:
parallel_config = vllm_config.parallel_config
hf_config = vllm_config.model_config.hf_config
tp_size = parallel_config.tensor_parallel_size
@@ -628,9 +628,7 @@ class KimiLinearForCausalLM(
@classmethod
def get_mamba_state_copy_func(
cls,
) -> tuple[
MambaStateCopyFunc, MambaStateCopyFunc, MambaStateCopyFunc, MambaStateCopyFunc
]:
) -> tuple[MambaStateCopyFunc, MambaStateCopyFunc]:
return MambaStateCopyFuncCalculator.kda_state_copy_func()
def compute_logits(