[Bugfix] Fix missing per_act_token parameter in compressed_tensors_moe (#20509)

Signed-off-by: Lu Fang <fanglu@fb.com>

Author: luccafong

vllm/model_executor/layers/fused_moe/cutlass_moe.py

@@ -322,7 +322,7 @@ def cutlass_moe_fp8(
     topk_ids: torch.Tensor,
     w1_scale: torch.Tensor,
     w2_scale: torch.Tensor,
-    per_act_token: bool,
+    per_act_token: Optional[bool] = None,
     activation: str = "silu",
     a1_scale: Optional[torch.Tensor] = None,
     a2_scale: Optional[torch.Tensor] = None,
@@ -366,6 +366,9 @@ def cutlass_moe_fp8(
     Returns:
     - torch.Tensor: The fp16 output tensor after applying the MoE layer.
     """
+    if per_act_token is None:
+        per_act_token = a1_scale.numel() != 1 if a1_scale is not None else (
+            a2_scale.numel() != 1 if a2_scale is not None else False)
     per_out_ch = w1_scale.numel() != w1_q.size(0)
 
     num_experts = global_num_experts if global_num_experts != -1 else w1_q.size(