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, 05:30 UTC
[Benchmark] Finish documented v0.11.0 deprecation of --endpoint-type (#26007)
Signed-off-by: Nathan Scott <nathans@redhat.com>
Nathan Scott committed
Oct 1, 2025 at 22:41 UTC
f9e714813ae17641d5bc8d3d3c992e20c195792f
1 file changed
-24
vllm/benchmarks/serve.py
-24
@@ -52,21 +52,6 @@ TERM_PLOTLIB_AVAILABLE = ((importlib.util.find_spec("termplotlib") is not None)
52
and (shutil.which("gnuplot") is not None))
53
54
55
-# TODO: Remove this in v0.11.0
56
-class DeprecatedEndpointTypeAction(argparse.Action):
57
- """Argparse action for the deprecated --endpoint-type flag.
58
- """
59
-
60
- def __call__(self, _, namespace, values, option_string=None):
61
- warnings.warn(
62
- "'--endpoint-type' is deprecated and will be removed in v0.11.0. "
63
- "Please use '--backend' instead or remove this argument if you "
64
- "have already set it.",
65
- stacklevel=1,
66
- )
67
- setattr(namespace, self.dest, values)
68
-
69
-
55
class TaskType(Enum):
56
GENERATION = "generation"
57
EMBEDDING = "embedding"
@@ -883,15 +868,6 @@ def add_cli_args(parser: argparse.ArgumentParser):
868
choices=list(ASYNC_REQUEST_FUNCS.keys()),
869
help="The type of backend or endpoint to use for the benchmark."
870
)
886
- parser.add_argument(
887
- "--endpoint-type",
888
- type=str,
889
- default=None,
890
- choices=list(ASYNC_REQUEST_FUNCS.keys()),
891
- action=DeprecatedEndpointTypeAction,
892
- help="'--endpoint-type' is deprecated and will be removed in v0.11.0. "
893
- "Please use '--backend' instead.",
894
- )
871
parser.add_argument(
872
"--base-url",
873
type=str,