Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Misc] Optimize flash attention backend log #4368

Merged
merged 2 commits into from
Apr 25, 2024
Merged

[Misc] Optimize flash attention backend log #4368

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
e626315
optimize
esmeetu Apr 25, 2024
ee233b2
format
esmeetu Apr 25, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 5 additions & 5 deletions vllm/attention/selector.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -25,7 +25,7 @@ class _Backend(enum.Enum):
def get_attn_backend(dtype: torch.dtype) -> Type[AttentionBackend]:
backend = _which_attn_to_use(dtype)
if backend == _Backend.FLASH_ATTN:
logger.info("Using FlashAttention backend.")
logger.info("Using FlashAttention-2 backend.")
from vllm.attention.backends.flash_attn import ( # noqa: F401
FlashAttentionBackend)
return FlashAttentionBackend
Expand Down Expand Up @@ -62,21 +62,21 @@ def _which_attn_to_use(dtype: torch.dtype) -> _Backend:
# NVIDIA GPUs.
if torch.cuda.get_device_capability()[0] < 8:
# Volta and Turing NVIDIA GPUs.
logger.info("Cannot use FlashAttention backend for Volta and Turing "
logger.info("Cannot use FlashAttention-2 backend for Volta and Turing "
"GPUs.")
return _Backend.XFORMERS

if dtype not in (torch.float16, torch.bfloat16):
logger.info("Cannot use FlashAttention backend for dtype other than "
logger.info("Cannot use FlashAttention-2 backend for dtype other than "
"torch.float16 or torch.bfloat16.")
return _Backend.XFORMERS

try:
import flash_attn # noqa: F401
except ImportError:
logger.info(
"Cannot use FlashAttention backend because the flash_attn package "
"is not found. Please install it for better performance.")
"Cannot use FlashAttention-2 backend because the flash_attn "
"package is not found. Please install it for better performance.")
return _Backend.XFORMERS

backend_by_env_var = os.getenv(VLLM_ATTENTION_BACKEND)
Expand Down
Loading