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, 23:30 UTC
[misc] Fix setup.py condition to avoid AMD from being mistaken with CPU (#13022)
Signed-off-by: kevin <kevin@anyscale.com>
Kevin H. Luu committed
Feb 10, 2025 at 18:06 UTC
91e876750eace8e899ab25cd5d93fc365906c07b
1 file changed
+3
-2
setup.py
+3
-2
@@ -48,8 +48,9 @@ elif not (sys.platform.startswith("linux")
48
"so vLLM may not be able to run correctly", sys.platform)
49
VLLM_TARGET_DEVICE = "empty"
50
elif (sys.platform.startswith("linux") and torch.version.cuda is None
51
- and os.getenv("VLLM_TARGET_DEVICE") is None):
52
- # if cuda is not available and VLLM_TARGET_DEVICE is not set,
51
+ and os.getenv("VLLM_TARGET_DEVICE") is None
52
+ and torch.version.hip is None):
53
+ # if cuda or hip is not available and VLLM_TARGET_DEVICE is not set,
54
# fallback to cpu
55
VLLM_TARGET_DEVICE = "cpu"
56