mirror of
https://git.datalinker.icu/vllm-project/vllm.git
synced 2026-04-21 01:47:03 +08:00
Add FLASH_SINK_ATTN to AttentionBackendEnum
Signed-off-by: yuantao <2422264527@qq.com>
This commit is contained in:
parent
b0e880632a
commit
e38739aefc
@ -42,6 +42,9 @@ class AttentionBackendEnum(enum.Enum, metaclass=_AttentionBackendEnumMeta):
|
|||||||
"""
|
"""
|
||||||
|
|
||||||
FLASH_ATTN = "vllm.v1.attention.backends.flash_attn.FlashAttentionBackend"
|
FLASH_ATTN = "vllm.v1.attention.backends.flash_attn.FlashAttentionBackend"
|
||||||
|
FLASH_SINK_ATTN = (
|
||||||
|
"vllm.v1.attention.backends.flash_sink_attn.FlashSinkAttentionBackend"
|
||||||
|
)
|
||||||
TRITON_ATTN = "vllm.v1.attention.backends.triton_attn.TritonAttentionBackend"
|
TRITON_ATTN = "vllm.v1.attention.backends.triton_attn.TritonAttentionBackend"
|
||||||
XFORMERS = "vllm.v1.attention.backends.xformers.XFormersAttentionBackend"
|
XFORMERS = "vllm.v1.attention.backends.xformers.XFormersAttentionBackend"
|
||||||
ROCM_ATTN = "vllm.v1.attention.backends.rocm_attn.RocmAttentionBackend"
|
ROCM_ATTN = "vllm.v1.attention.backends.rocm_attn.RocmAttentionBackend"
|
||||||
|
|||||||
Loading…
x
Reference in New Issue
Block a user