Skip to content

Added GQA support for CUDNNAttention backend #3984

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

Open
MaciejBalaNV wants to merge 1 commit into
base: main
Choose a base branch
from
+3 −0
Open

Added GQA support for CUDNNAttention backend #3984

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
3 changes: 3 additions & 0 deletions vllm_omni/diffusion/attention/backends/cudnn_attn.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -51,6 +51,7 @@ def __init__(
) -> None:
self.causal = causal
self.softmax_scale = softmax_scale
self.requires_gqa = num_heads != num_kv_heads

def forward_cuda(
self,
Expand Down Expand Up @@ -84,6 +85,7 @@ def forward_cuda(
dropout_p=0.0,
is_causal=self.causal,
scale=self.softmax_scale,
enable_gqa=self.requires_gqa,
)
except RuntimeError as e:
if "No available kernel" not in str(e):
Expand All @@ -100,5 +102,6 @@ def forward_cuda(
dropout_p=0.0,
is_causal=self.causal,
scale=self.softmax_scale,
enable_gqa=self.requires_gqa,
)
return output.permute(0, 2, 1, 3)
Loading