[Bugfix] Add swiglu limits to deepgemm fp8 methods (#41986)
Signed-off-by: Yongye Zhu <zyy1102000@gmail.com> Co-authored-by: Claude <noreply@anthropic.com>
This commit is contained in:
@@ -856,6 +856,7 @@ class Fp8MoEMethod(FusedMoEMethodBase):
|
||||
a1_scale=a1_scale,
|
||||
a2_scale=a2_scale,
|
||||
block_shape=self.weight_block_size,
|
||||
swiglu_limit=getattr(layer, "swiglu_limit", None),
|
||||
)
|
||||
|
||||
# Inject biases into the quant config if the model has them
|
||||
|
||||
Reference in New Issue
Block a user