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
fix: clean up function never use in setup.py (#29061)
Signed-off-by: yihong0618 <zouzou0208@gmail.com>
yihong committed
Nov 22, 2025 at 21:00 UTC
2d4978a57e0addf55cde6113e9615ed064b72fb7
1 file changed
-34
setup.py
-34
@@ -74,18 +74,6 @@ def is_ninja_available() -> bool:
74
return which("ninja") is not None
75
76
77
-def is_url_available(url: str) -> bool:
78
- from urllib.request import urlopen
79
-
80
- status = None
81
- try:
82
- with urlopen(url) as f:
83
- status = f.status
84
- except Exception:
85
- return False
86
- return status == 200
87
-
88
-
77
class CMakeExtension(Extension):
78
def __init__(self, name: str, cmake_lists_dir: str = ".", **kwa) -> None:
79
super().__init__(name, sources=[], py_limited_api=True, **kwa)
@@ -533,28 +521,6 @@ def get_nvcc_cuda_version() -> Version:
521
return nvcc_cuda_version
522
523
536
-def get_gaudi_sw_version():
537
- """
538
- Returns the driver version.
539
- """
540
- # Enable console printing for `hl-smi` check
541
- output = subprocess.run(
542
- "hl-smi",
543
- shell=True,
544
- text=True,
545
- capture_output=True,
546
- env={"ENABLE_CONSOLE": "true"},
547
- )
548
- if output.returncode == 0 and output.stdout:
549
- return (
550
- output.stdout.split("\n")[2]
551
- .replace(" ", "")
552
- .split(":")[1][:-1]
553
- .split("-")[0]
554
- )
555
- return "0.0.0" # when hl-smi is not available
556
-
557
-
524
def get_vllm_version() -> str:
525
# Allow overriding the version. This is useful to build platform-specific
526
# wheels (e.g. CPU, TPU) without modifying the source.