diff --git a/vllm/model_executor/layers/quantization/fp8.py b/vllm/model_executor/layers/quantization/fp8.py
index a49744913251..f07be0855492 100644
--- a/vllm/model_executor/layers/quantization/fp8.py
+++ b/vllm/model_executor/layers/quantization/fp8.py
@@ -125,6 +125,10 @@ def from_config(cls, config: dict[str, Any]) -> "Fp8Config":
         ignored_layers = cls.get_from_keys_or(config, ["ignored_layers"], None)
         weight_block_size = cls.get_from_keys_or(config, ["weight_block_size"],
                                                  None)
+        if not ignored_layers:
+            ignored_layers = cls.get_from_keys_or(config,
+                                                  ["modules_to_not_convert"],
+                                                  None)
         return cls(is_checkpoint_fp8_serialized=is_checkpoint_fp8_serialized,
                    activation_scheme=activation_scheme,
                    ignored_layers=ignored_layers,