biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[V0 deprecation] Remove _VLLM_V1 suffixes from attention backend names (#25489)

Browse Source 
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
Signed-off-by: Matthew Bonanni <mbonanni001@gmail.com>
This commit is contained in:
Matthew Bonanni
2025-09-25 13:37:50 -04:00
committed by GitHub
parent 71b25b0d48
commit 3468f17ebe
42 changed files with 131 additions and 174 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
tests/kernels/utils.py
View File

@@ -524,14 +524,14 @@ def make_backend(backend_name: str) -> AttentionBackend:
* Backend instance
'''
if backend_name in (STR_XFORMERS_ATTN_VAL, "XFORMERS_VLLM_V1"):
if backend_name == STR_XFORMERS_ATTN_VAL:
from vllm.v1.attention.backends.xformers import (
XFormersAttentionBackend)
return XFormersAttentionBackend()
if backend_name in (STR_FLASH_ATTN_VAL, "FLASH_ATTN_VLLM_V1"):
if backend_name == STR_FLASH_ATTN_VAL:
from vllm.v1.attention.backends.flash_attn import FlashAttentionBackend
return FlashAttentionBackend()
if backend_name == "TRITON_ATTN_VLLM_V1":
if backend_name == "TRITON_ATTN":
from vllm.v1.attention.backends.triton_attn import (
TritonAttentionBackend)
return TritonAttentionBackend()
@@ -539,7 +539,7 @@ def make_backend(backend_name: str) -> AttentionBackend:
from vllm.v1.attention.backends.flex_attention import (
FlexAttentionBackend)
return FlexAttentionBackend()
if backend_name in ("TORCH_SDPA", "TORCH_SDPA_VLLM_V1"):
if backend_name == "TORCH_SDPA":
from vllm.v1.attention.backends.cpu_attn import TorchSDPABackend
return TorchSDPABackend()
if backend_name == "FLASHINFER":