diff --git a/modules/sdnq/loader.py b/modules/sdnq/loader.py
index 6fc34551e..43cf06166 100644
--- a/modules/sdnq/loader.py
+++ b/modules/sdnq/loader.py
@@ -129,7 +129,7 @@ def load_sdnq_model(model_path: str, model_cls: ModelMixin | None = None, file_n
                 model_config.pop("quantization_config", None)
             model = model_cls(**model_config)
 
-        model = sdnq_post_load_quant(model, torch_dtype=dtype, add_skip_keys=False, use_dynamic_quantization=False, **get_quant_args_from_config(quantization_config))
+        model = sdnq_post_load_quant(model, torch_dtype=dtype, pre_quantized=True, **get_quant_args_from_config(quantization_config))
 
     key_mapping = getattr(model, "_checkpoint_conversion_mapping", None)
     files = []
diff --git a/modules/sdnq/quantizer.py b/modules/sdnq/quantizer.py
index a9e9eec55..8509720f5 100644
--- a/modules/sdnq/quantizer.py
+++ b/modules/sdnq/quantizer.py
@@ -469,7 +469,18 @@ def sdnq_post_load_quant(
     quantization_device: torch.device | None = None,
     return_device: torch.device | None = None,
     torch_dtype: torch.dtype | None = None,
+    pre_quantized: bool = False,
 ):
+    if pre_quantized:
+        add_skip_keys = False
+        use_dynamic_quantization = False
+    elif (
+        hasattr(model, "quantization_config")
+        or (hasattr(model, "config") and hasattr(model.config, "quantization_config"))
+        or (hasattr(model, "config") and isinstance(model.config, dict) and "quantization_config" in model.config)
+    ):
+        raise RuntimeError("Quantizing a pre-quantized model is not supported!")
+
     quantization_config = SDNQConfig(
         weights_dtype=weights_dtype,
         quantized_matmul_dtype=quantized_matmul_dtype,
@@ -669,7 +680,7 @@ class SDNQQuantizer(DiffusersQuantizer, HfQuantizer):
             self.quantization_config.add_skip_keys = False
 
             with init_empty_weights():
-                model = sdnq_post_load_quant(model, torch_dtype=self.torch_dtype, add_skip_keys=False, use_dynamic_quantization=False, **get_quant_args_from_config(self.quantization_config))
+                model = sdnq_post_load_quant(model, torch_dtype=self.torch_dtype, pre_quantized=True, **get_quant_args_from_config(self.quantization_config))
 
         if self.quantization_config.add_skip_keys:
             if keep_in_fp32_modules is not None: