Skip to content

[OpenCOOD] Bypass self-attention for single-CAV inputs #195

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

Merged
Haralishev77 merged 2 commits into from
Sep 1, 2026
Merged

[OpenCOOD] Bypass self-attention for single-CAV inputs #195

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
3fe923d
[OpenCOOD] Bypass self-attention for single-CAV inputs
Haralishev77 Sep 1, 2026
7149fe8
Shorten the comment
Haralishev77 Sep 1, 2026
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
5 changes: 5 additions & 0 deletions OpenCOOD/opencood/models/fuse_modules/self_attn.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -32,6 +32,11 @@ def __init__(self, dim):
self.sqrt_dim = np.sqrt(dim)

def forward(self, query, key, value):
# Single-element attention is an identity operation. Skip backend kernel
# dispatch for this degenerate case (for example, Triton on CUDA).
if query.size(1) == 1 and key.size(1) == 1 and value.size(1) == 1:
return value

score = torch.bmm(query, key.transpose(1, 2)) / self.sqrt_dim
attn = F.softmax(score, -1)
context = torch.bmm(attn, value)
Expand Down
Loading