[Misc] Modularize CLI Argument Parsing in Benchmark Scripts (#19593)

Signed-off-by: reidliu41 <reid201711@gmail.com>
Co-authored-by: reidliu41 <reid201711@gmail.com>
This commit is contained in:
Reid
2025-06-14 16:54:52 +08:00
committed by GitHub
co-authored by reidliu41
parent 06be858828
commit 6fa718a460
7 changed files with 47 additions and 8 deletions
+7 -1
View File
@@ -123,7 +123,7 @@ def main(args: argparse.Namespace):
save_to_pytorch_benchmark_format(args, results)
if __name__ == "__main__":
def create_argument_parser():
parser = FlexibleArgumentParser(
description="Benchmark the latency of processing a single batch of "
"requests till completion."
@@ -171,6 +171,12 @@ if __name__ == "__main__":
# V1 enables prefix caching by default which skews the latency
# numbers. We need to disable prefix caching by default.
parser.set_defaults(enable_prefix_caching=False)
return parser
if __name__ == "__main__":
parser = create_argument_parser()
args = parser.parse_args()
if args.profile and not envs.VLLM_TORCH_PROFILER_DIR:
raise OSError(