diff --git a/vllm/model_executor/model_loader/loader.py b/vllm/model_executor/model_loader/loader.py index 6f9969123d749..764c91f4aa352 100644 --- a/vllm/model_executor/model_loader/loader.py +++ b/vllm/model_executor/model_loader/loader.py @@ -163,7 +163,7 @@ def build_model(model_class: Type[nn.Module], hf_config: PretrainedConfig, return model_class(config=hf_config, cache_config=cache_config, quant_config=quant_config, - pooling_type + pooling_type, **extra_kwargs)