[Attention]: Use KVCacheSpec for AttentionMetadataBuilder type hints (#50148)

Signed-off-by: Martin Hickey <martin.hickey@ie.ibm.com>
This commit is contained in:
Martin Hickey
2026-07-31 11:20:18 +00:00
committed by GitHub
parent 92643d68f5
commit b2fb83e7ff
14 changed files with 34 additions and 33 deletions
+1 -1
View File
@@ -169,7 +169,7 @@ def init_attn_backend(
# Check cudagraph support for the attention backend
cg_support = builder.get_cudagraph_support(
vllm_config,
cast(AttentionSpec, group.kv_cache_spec),
group.kv_cache_spec,
)
if cg_support.value < min_cg_support.value:
min_cg_support = cg_support