[Bugfix] Fix null modules_to_not_convert in FBGEMM Fp8 quantization (#6665)

Author: cli99

vllm/model_executor/layers/quantization/fbgemm_fp8.py

@@ -31,7 +31,7 @@ class FBGEMMFp8Config(QuantizationConfig):
     """Config class for FBGEMM Fp8."""
 
     def __init__(self, ignore_list: List[str], input_scale_ub: float):
-        self.ignore_list = ignore_list
+        self.ignore_list = ignore_list if ignore_list else []
         self.input_scale_ub = input_scale_ub
 
         # For GPUs that lack FP8 hardware support, we can leverage the Marlin