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 13, 08:15 UTC
[Misc]Fix typo in llm_engine.py (#28584)
Signed-off-by: Wei Wei <wwei6@meta.com>
Wei Wei committed
Nov 12, 2025 at 12:59 UTC
478ee511de0776159163877924f7e77ce1f472d0
1 file changed
+4
-4
vllm/v1/engine/llm_engine.py
+4
-4
@@ -281,11 +281,11 @@ class LLMEngine:
281
return []
282
283
# 1) Get EngineCoreOutput from the EngineCore.
284
- with record_function_or_nullcontext("llm_genine step: get_output"):
284
+ with record_function_or_nullcontext("llm_engine step: get_output"):
285
outputs = self.engine_core.get_output()
286
287
# 2) Process EngineCoreOutputs.
288
- with record_function_or_nullcontext("llm_genine step: process_outputs"):
288
+ with record_function_or_nullcontext("llm_engine step: process_outputs"):
289
iteration_stats = IterationStats() if self.log_stats else None
290
processed_outputs = self.output_processor.process_outputs(
291
outputs.outputs,
@@ -295,11 +295,11 @@ class LLMEngine:
295
self.output_processor.update_scheduler_stats(outputs.scheduler_stats)
296
297
# 3) Abort any reqs that finished due to stop strings.
298
- with record_function_or_nullcontext("llm_genine step: abort_requests"):
298
+ with record_function_or_nullcontext("llm_engine step: abort_requests"):
299
self.engine_core.abort_requests(processed_outputs.reqs_to_abort)
300
301
# 4) Record stats
302
- with record_function_or_nullcontext("llm_genine step: record_stats"):
302
+ with record_function_or_nullcontext("llm_engine step: record_stats"):
303
if self.logger_manager is not None and outputs.scheduler_stats is not None:
304
self.logger_manager.record(
305
scheduler_stats=outputs.scheduler_stats,