Fix onnx export for quantized MPT models (#1646)

Empty dict raises an exception when exporting quantized MPT models because when `state_dict = {}`, the HF loading pipeline will try to initialize the model by calling `_init_weights` for each module. However, for MPT models, this function is not defined and will therefore raise an exception.

src/sparseml/transformers/utils/model.py

@@ -407,7 +407,7 @@ def _loadable_state_dict(
             f"after SparseML recipes have been applied {model_name_or_path}"
         )
 
-        return {}, True
+        return None, True
 
     @staticmethod
     def _check_tf(model_name_or_path: str):