Address review comments

Signed-off-by: Neta Zmora <[email protected]>

Author: nzmora-nvidia

tensorrt_llm/_torch/auto_deploy/custom_ops/fused_moe/trtllm_moe.py

@@ -364,20 +364,20 @@ def trtllm_quant_nvfp4_moe_fused_fake(
     x: torch.Tensor,
     selected_experts: torch.Tensor,
     routing_weights: torch.Tensor,
-    w1_weight_q: torch.Tensor,
-    w2_weight_q: torch.Tensor,
-    w3_weight_q: torch.Tensor,
-    w1_weight_gs: torch.Tensor,
-    w2_weight_gs: torch.Tensor,
-    w3_weight_gs: torch.Tensor,
-    w1_blockscale: torch.Tensor,
-    w2_blockscale: torch.Tensor,
-    w3_blockscale: torch.Tensor,
+    w1_fp4: torch.Tensor,
+    w2_fp4: torch.Tensor,
+    w3_fp4: torch.Tensor,
+    w1_global_scale: torch.Tensor,
+    w2_global_scale: torch.Tensor,
+    w3_global_scale: torch.Tensor,
+    w1_blockscale_fp8: torch.Tensor,
+    w2_blockscale_fp8: torch.Tensor,
+    w3_blockscale_fp8: torch.Tensor,
     fc1_act_global: torch.Tensor,
     fc2_act_global: torch.Tensor,
-    fc1_global: Optional[torch.Tensor] = None,
-    fc2_global: Optional[torch.Tensor] = None,
-    input_sf: Optional[torch.Tensor] = None,
+    fc1_alpha: Optional[torch.Tensor] = None,
+    fc2_alpha: Optional[torch.Tensor] = None,
+    input_blockscale: Optional[torch.Tensor] = None,
     output_dtype: Optional[torch.dtype] = None,
     mlp_style: str = "gated_mlp",
     act_fn: str = "silu",