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 6, 02:00 UTC
[BugFix] Fix order of compile logging (#38012)
Signed-off-by: Richard Zou <zou3519@gmail.com>
Richard Zou committed
Mar 24, 2026 at 14:58 UTC
71a4a2fbd043260a0e693e909fc7d00ec23c9a5f
1 file changed
+4
-2
vllm/compilation/decorators.py
+4
-2
@@ -299,8 +299,10 @@ def _try_load_aot_compiled_fn(
299
# is populated by _verify_source_unchanged.
300
with maybe_use_cudagraph_partition_wrapper(model.vllm_config):
301
loaded_fn._artifacts.compiled_fn.finalize_loading(model.vllm_config)
302
- compilation_counter.num_aot_artifacts_loaded += 1
303
- logger.info("Directly load AOT compilation from path %s", aot_compilation_path)
302
+ compilation_counter.num_aot_artifacts_loaded += 1
303
+ logger.info(
304
+ "Directly load AOT compilation from path %s", aot_compilation_path
305
+ )
306
return loaded_fn
307
except Exception as e:
308
if os.path.exists(aot_compilation_path):