[Moe Refactor] Make Inplace Flag for FusedMoEModularKernel part of the constructor (#33375)

Signed-off-by: Bill Nell <bnell@redhat.com> Co-authored-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
2026-02-05 13:07:18 -05:00
parent 1ee95841bd
commit a57c8228ff
37 changed files with 132 additions and 109 deletions
--- a/tests/kernels/moe/test_flashinfer_moe.py
+++ b/tests/kernels/moe/test_flashinfer_moe.py
@@ -108,6 +108,7 @@ def test_flashinfer_fp4_moe_no_graph(
        flashinfer_experts = FusedMoEModularKernel(
            MoEPrepareAndFinalizeNoEP(),
            FlashInferExperts(moe_config=moe_config, quant_config=quant_config),
+            inplace=False,
        )

        fi_activation = {"silu_and_mul": "silu", "relu2": "relu2_no_mul"}[activation]