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 17, 03:00 UTC
[BugFix] Use different mechanism to get vllm version in `is_cpu()` (#3804)
Nick Hill committed
Apr 2, 2024 at 23:06 UTC
c9b506dad43c1439fd9aa5c1dd358d03967a7790
1 file changed
+5
-3
vllm/utils.py
+5
-3
@@ -119,9 +119,11 @@ def is_hip() -> bool:
119
120
@lru_cache(maxsize=None)
121
def is_cpu() -> bool:
122
- from importlib.metadata import version
123
- is_cpu_flag = "cpu" in version("vllm")
124
- return is_cpu_flag
122
+ from importlib.metadata import PackageNotFoundError, version
123
+ try:
124
+ return "cpu" in version("vllm")
125
+ except PackageNotFoundError:
126
+ return False
127
128
129
@lru_cache(maxsize=None)