mirror of
https://git.datalinker.icu/vllm-project/vllm.git
synced 2025-12-10 16:54:59 +08:00
[HOT-FIX] Add kv_sharing_target_layer_name argument to cutlass_mla backend (#19374)
Signed-off-by: Pavani Majety <pmajety@nvidia.com>
This commit is contained in:
parent
7d44c469fe
commit
8058c91108
@ -40,12 +40,13 @@ class CutlassMLAImpl(MLACommonImpl[MLACommonMetadata]):
|
|||||||
blocksparse_params: Optional[dict[str, Any]],
|
blocksparse_params: Optional[dict[str, Any]],
|
||||||
logits_soft_cap: Optional[float],
|
logits_soft_cap: Optional[float],
|
||||||
attn_type: str,
|
attn_type: str,
|
||||||
|
kv_sharing_target_layer_name: Optional[str],
|
||||||
# MLA Specific Arguments
|
# MLA Specific Arguments
|
||||||
**mla_args) -> None:
|
**mla_args) -> None:
|
||||||
super().__init__(num_heads, head_size, scale, num_kv_heads,
|
super().__init__(num_heads, head_size, scale, num_kv_heads,
|
||||||
alibi_slopes, sliding_window, kv_cache_dtype,
|
alibi_slopes, sliding_window, kv_cache_dtype,
|
||||||
blocksparse_params, logits_soft_cap, attn_type,
|
blocksparse_params, logits_soft_cap, attn_type,
|
||||||
**mla_args)
|
kv_sharing_target_layer_name, **mla_args)
|
||||||
|
|
||||||
unsupported_features = [
|
unsupported_features = [
|
||||||
alibi_slopes, sliding_window, blocksparse_params, logits_soft_cap
|
alibi_slopes, sliding_window, blocksparse_params, logits_soft_cap
|
||||||
|
|||||||
Loading…
x
Reference in New Issue
Block a user