[V1] Enable Triton(ROCm) Attention backend for Nvidia GPUs (#14071)

Signed-off-by: Isotr0py <2037008807@qq.com>
Co-authored-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
This commit is contained in:
Isotr0py
2025-03-21 03:14:19 +00:00
committed by GitHub
co-authored by Woosuk Kwon
parent b15fd2be2a
commit f8a08cb90d
5 changed files with 23 additions and 16 deletions
+1
View File
@@ -29,6 +29,7 @@ def in_wsl() -> bool:
class _Backend(enum.Enum):
FLASH_ATTN = enum.auto()
FLASH_ATTN_VLLM_V1 = enum.auto()
TRITON_ATTN_VLLM_V1 = enum.auto()
XFORMERS = enum.auto()
ROCM_FLASH = enum.auto()
TORCH_SDPA = enum.auto()