diff --git a/vllm/model_executor/models/aria.py b/vllm/model_executor/models/aria.py index dd2ca3e87c3b0..dcad38ef454ea 100644 --- a/vllm/model_executor/models/aria.py +++ b/vllm/model_executor/models/aria.py @@ -471,6 +471,8 @@ def _make_expert_params_mapping( ] ] + # Adapted from LlamaModel.load_weights with the modification of adding the + # expert_params_mapping def load_weights(self, weights: Iterable[Tuple[str, torch.Tensor]]) -> Set[str]: stacked_params_mapping = [