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 14, 23:45 UTC
[Bug] Fix `has_flashinfer_moe` Import Error when it is not installed (#21634)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Wentao Ye committed
Jul 26, 2025 at 10:06 UTC
c215f5c877e0518f328613a98f31bb9cdb32a889
1 file changed
+2
-1
vllm/utils/flashinfer.py
+2
-1
@@ -82,7 +82,8 @@ autotune = _lazy_import_wrapper(
82
@functools.cache
83
def has_flashinfer_moe() -> bool:
84
"""Return ``True`` if FlashInfer MoE module is available."""
85
- return importlib.util.find_spec("flashinfer.fused_moe") is not None
85
+ return has_flashinfer() and importlib.util.find_spec(
86
+ "flashinfer.fused_moe") is not None
87
88
89
@functools.cache