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, 14:45 UTC
Fix erroneous "model doesn't support compile" warning (#16486)
Signed-off-by: rzou <zou3519@gmail.com>
Richard Zou committed
Apr 11, 2025 at 12:24 UTC
70de35a8816e224663aede45b7f54eef250a5cfe
1 file changed
+4
-1
vllm/config.py
+4
-1
@@ -3880,7 +3880,9 @@ def set_current_vllm_config(vllm_config: VllmConfig, check_compile=False):
3880
try:
3881
_current_vllm_config = vllm_config
3882
yield
3883
- finally:
3883
+ except Exception:
3884
+ raise
3885
+ else:
3886
logger.debug("enabled custom ops: %s",
3887
vllm_config.compilation_config.enabled_custom_ops)
3888
logger.debug("disabled custom ops: %s",
@@ -3898,6 +3900,7 @@ def set_current_vllm_config(vllm_config: VllmConfig, check_compile=False):
3900
" does not support it. Please open an issue on GitHub"
3901
" if you want it to be supported.",
3902
vllm_config.model_config.model)
3903
+ finally:
3904
_current_vllm_config = old_vllm_config
3905
3906