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 12, 17:15 UTC
[Bugfix]Fix a conditional to not check zero value (#28754)
Signed-off-by: Yanan Cao <gmagogsfm@gmail.com>
Yanan Cao committed
Nov 21, 2025 at 19:59 UTC
933f67ecd81231ebfa5e2434d3ae3819b6c28068
1 file changed
+2
-1
vllm/compilation/caching.py
+2
-1
@@ -116,7 +116,8 @@ class VllmSerializableFunction(SerializableCallable):
116
the AOT compiled path.
117
"""
118
compile_inputs = [
119
- inp or example_inputs[i] for i, inp in enumerate(fn.example_inputs)
119
+ inp if inp is not None else example_inputs[i]
120
+ for i, inp in enumerate(fn.example_inputs)
121
]
122
with tracing(TracingContext(fake_mode)):
123
fn.optimized_call = vllm_backend(