diff --git a/src/transformers/modeling_utils.py b/src/transformers/modeling_utils.py
index 08c760dcbf57..711c0adb3b73 100644
--- a/src/transformers/modeling_utils.py
+++ b/src/transformers/modeling_utils.py
@@ -2368,6 +2368,7 @@ def _initialize_weights(self, module, is_remote_code: bool = False):
         # otherwise
         if (
             is_remote_code
+            and module is not self  # some remote code models do special things in _init_weights for top-most module
             and all(getattr(param, "_is_hf_initialized", False) for param in module.parameters(recurse=False))
             and all(
                 getattr(buffer, "_is_hf_initialized", False)