We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
1 parent 9729886 commit e843e17Copy full SHA for e843e17
1 file changed
python/sglang/srt/layers/attention/linear/kernels/gdn_flashinfer.py
@@ -199,7 +199,6 @@ def extend(
199
alpha_fi = torch.exp(g[0].to(torch.float32))
200
beta_fi = beta[0].to(torch.float32)
201
202
-
203
if self.is_sm100plus:
204
# SM100+: slot 0 is reserved as dummy/scratch (never assigned to real
205
# sequences), so clamp(-1 → 0).
0 commit comments