[ROCm] Split AITER unified attention into its own backend (#25507)

Signed-off-by: Gregory Shtrasberg <Gregory.Shtrasberg@amd.com>
This commit is contained in:
Gregory Shtrasberg
2025-10-06 18:49:23 -04:00
committed by GitHub
parent 2161efe978
commit f231e5bc21
8 changed files with 325 additions and 301 deletions

View File

@@ -1623,6 +1623,7 @@ class EngineArgs:
"TREE_ATTN",
"XFORMERS",
"ROCM_ATTN",
"ROCM_AITER_UNIFIED_ATTN",
]
if (
envs.is_set("VLLM_ATTENTION_BACKEND")