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 8, 15:15 UTC
Revert #34208 (#34216)
Cyrus Leung committed
Feb 10, 2026 at 15:59 UTC
998e2d91f84e2b30dc40c8543b879d4e412d6f14
2 files changed
+5
-5
vllm/benchmarks/datasets.py
-5
@@ -1314,11 +1314,6 @@ class _ValidateDatasetArgs(argparse.Action):
1314
1315
1316
def add_dataset_parser(parser: FlexibleArgumentParser):
1317
- parser.add_argument(
1318
- "--trust-remote-code",
1319
- action="store_true",
1320
- help="Trust remote code from huggingface",
1321
- )
1317
parser.add_argument("--seed", type=int, default=0)
1318
parser.add_argument(
1319
"--num-prompts",
vllm/benchmarks/serve.py
+5
@@ -1300,6 +1300,11 @@ def add_cli_args(parser: argparse.ArgumentParser):
1300
"bursty requests. A higher burstiness value (burstiness > 1) "
1301
"results in a more uniform arrival of requests.",
1302
)
1303
+ parser.add_argument(
1304
+ "--trust-remote-code",
1305
+ action="store_true",
1306
+ help="Trust remote code from huggingface",
1307
+ )
1308
parser.add_argument(
1309
"--disable-tqdm",
1310
action="store_true",