fix cpu test

vllm-project · Aug 4, 2024 · 4a86201 · 4a86201
1 parent 0ba00ab
commit 4a86201
Showing 1 changed file with 3 additions and 3 deletions.
diff --git a/vllm/model_executor/layers/fused_moe/__init__.py b/vllm/model_executor/layers/fused_moe/__init__.py
@@ -1,11 +1,8 @@
-from vllm.model_executor.layers.fused_moe.fused_moe_awq import (
-    fused_experts_awq)
 from vllm.model_executor.layers.fused_moe.layer import (FusedMoE,
                                                         FusedMoEMethodBase)
 from vllm.triton_utils import HAS_TRITON
 
 __all__ = [
-    "fused_experts_awq",
     "FusedMoE",
     "FusedMoEMethodBase",
 ]
@@ -15,9 +12,12 @@
     from vllm.model_executor.layers.fused_moe.fused_moe import (
         fused_experts, fused_moe, fused_topk, get_config_file_name,
         grouped_topk)
+    from vllm.model_executor.layers.fused_moe.fused_moe_awq import (
+        fused_experts_awq)
 
     __all__ += [
         "fused_moe",
+        "fused_experts_awq",
         "fused_experts",
         "fused_topk",
         "get_config_file_name",