fix(modelopt_fp4): skip NVFP4 swiglu-fusion interleave for shared experts with swiglu_limit (#37378)
This commit is contained in:
@@ -787,6 +787,7 @@ class DeepseekV2MoE(nn.Module):
|
||||
ModelOptFp4LinearMethod,
|
||||
)
|
||||
and fc1_n % 128 == 0
|
||||
and self.shared_experts.swiglu_limit is None
|
||||
and not check_cuda_graph_backend(Phase.PREFILL, Backend.TC_PIECEWISE)
|
||||
):
|
||||
self.shared_experts.gate_up_proj._interleave_for_swiglu_fusion = True
|
||||
|
||||
Reference in New Issue
Block a user