[Misc] Fix Current vLLM config is not set. warnings, assert to avoid issues in the future (#31747)

Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Signed-off-by: Lucas Wilkinson <LucasWilkinson@users.noreply.github.com>
Co-authored-by: Luka Govedič <ProExpertProg@users.noreply.github.com>

tests/kernels/moe/test_cpu_fused_moe.py

@@ -6,7 +6,7 @@ import torch
 
 from tests.kernels.allclose_default import get_default_atol, get_default_rtol
 from vllm._custom_ops import cpu_fused_moe, cpu_prepack_moe_weight
-from vllm.model_executor.layers.activation import SiluAndMul, SwigluOAIAndMul
+from vllm.model_executor.layers.fused_moe.cpu_fused_moe import _CPU_MOE_ACT
 from vllm.platforms import current_platform
 from vllm.utils.torch_utils import set_random_seed
 
@@ -24,11 +24,6 @@ USE_BIAS = [True, False]
 ISA = ["amx", "vec"] if torch._C._cpu._is_amx_tile_supported() else ["vec"]
 DTYPE = [torch.bfloat16]
 
-_CPU_MOE_ACT = {
-    "silu": SiluAndMul(),
-    "swigluoai": SwigluOAIAndMul(),
-}
-
 
 def ref_fused_moe(
     input: torch.Tensor,
@@ -106,6 +101,7 @@ def ref_fused_moe(
 @pytest.mark.parametrize("act", ACT)
 @pytest.mark.parametrize("isa", ISA)
 def test_cpu_fused_moe(
+    default_vllm_config,
     batch_size: int,
     expert_num: int,
     hidden_size: int,

tests/kernels/moe/test_moe.py

@@ -468,7 +468,12 @@ def test_fused_moe_wn16(
 )
 @torch.inference_mode()
 def test_mixtral_moe(
-    dist_init, dtype: torch.dtype, padding: bool, use_rocm_aiter: bool, monkeypatch
+    default_vllm_config,
+    dist_init,
+    dtype: torch.dtype,
+    padding: bool,
+    use_rocm_aiter: bool,
+    monkeypatch,
 ):
     """Make sure our Mixtral MoE implementation agrees with the one from
     huggingface."""