A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 15, 04:45 UTC
raise exception for pin_lora (#19809)
Signed-off-by: Andy Xie <andy.xning@gmail.com>
Ning Xie committed
Jun 19, 2025 at 13:57 UTC
c7b370c603daf287e27d852ac36a7347536b01c6
1 file changed
+2
-3
vllm/worker/worker_base.py
+2
-3
@@ -202,8 +202,7 @@ class LoRANotSupportedWorkerBase(WorkerBase):
202
raise ValueError(f"{type(self)} does not support LoRA")
203
204
def pin_lora(self, lora_id: int) -> bool:
205
- return ValueError(
206
- f"{type(self)} does not support LoRA") # type: ignore
205
+ raise ValueError(f"{type(self)} does not support LoRA")
206
207
def list_loras(self) -> Set[int]:
208
raise ValueError(f"{type(self)} does not support LoRA")
@@ -398,7 +397,7 @@ class LocalOrDistributedWorkerBase(WorkerBase):
397
398
model_input, worker_input, kwargs = inputs
399
num_steps = worker_input.num_steps
401
- if (execute_model_req is not None and execute_model_req.spec_step_idx):
400
+ if execute_model_req is not None and execute_model_req.spec_step_idx:
401
kwargs["spec_step_idx"] = execute_model_req.spec_step_idx
402
403
self.execute_worker(worker_input)