From 882213cca2af7b6a61e46b3a7e665e249efc8396 Mon Sep 17 00:00:00 2001 From: bk-201 Date: Mon, 13 Oct 2025 02:39:01 +0000 Subject: [PATCH] update Signed-off-by: bk-201 --- vllm/lora/models.py | 4 ++-- vllm/lora/worker_manager.py | 4 ++-- vllm/v1/worker/lora_model_runner_mixin.py | 5 ++++- 3 files changed, 8 insertions(+), 5 deletions(-) diff --git a/vllm/lora/models.py b/vllm/lora/models.py index f7b35802f55b8..bfb65bbab9015 100644 --- a/vllm/lora/models.py +++ b/vllm/lora/models.py @@ -336,7 +336,7 @@ class LoRAModelManager: max_num_batched_tokens: int, vocab_size: int, lora_config: LoRAConfig, - model_config: Optional[ModelConfig], + model_config: ModelConfig | None, device: torch.device, ): """Create a LoRAModelManager and adapter for a given model. @@ -709,7 +709,7 @@ class LoRAModelManager: return any([module_name.startswith(prefix) for prefix in prefix_lst]) return False - def _get_mm_punica_wrapper(self, module_name: str) -> Optional[PunicaWrapperBase]: + def _get_mm_punica_wrapper(self, module_name: str) -> PunicaWrapperBase | None: """ Match the corresponding punica_wrapper based on module_name, and return None if lora is not supported for this module. diff --git a/vllm/lora/worker_manager.py b/vllm/lora/worker_manager.py index 49d4c29113760..dc8200fa8d006 100644 --- a/vllm/lora/worker_manager.py +++ b/vllm/lora/worker_manager.py @@ -71,7 +71,7 @@ class WorkerLoRAManager: def create_lora_manager( self, model: torch.nn.Module, - model_config: Optional[ModelConfig] = None, + model_config: ModelConfig | None = None, ) -> Any: lora_manager = create_lora_manager( model, @@ -222,7 +222,7 @@ class LRUCacheWorkerLoRAManager(WorkerLoRAManager): def create_lora_manager( self, model: torch.nn.Module, - model_config: Optional[ModelConfig] = None, + model_config: ModelConfig | None = None, ) -> Any: lora_manager = create_lora_manager( model, diff --git a/vllm/v1/worker/lora_model_runner_mixin.py b/vllm/v1/worker/lora_model_runner_mixin.py index 539f413d34a9b..31094dcbda124 100644 --- a/vllm/v1/worker/lora_model_runner_mixin.py +++ b/vllm/v1/worker/lora_model_runner_mixin.py @@ -131,7 +131,10 @@ class LoRAModelRunnerMixin: @contextmanager def maybe_select_dummy_loras( - self, lora_config: LoRAConfig | None, num_scheduled_tokens: np.ndarray, is_mm_input: bool = False + self, + lora_config: LoRAConfig | None, + num_scheduled_tokens: np.ndarray, + is_mm_input: bool = False, ): if lora_config is None: yield